#off-topic-tech
1 messages · Page 98 of 1
Based G502 user
welcome to the cool kids club
Put all of the weights in it for extra cool points
g403 superiority
No
It's my mouse for work and gaming
Invalid opinion
It suits me best out of the three mice I've tried
🧱
I rubbed through the rubber of my g502 and it just fell off.... 😦
Good thing they're cheap
How well does Qi charging work for you all with Logitechs? I know it's not officially supported, but it apparently works?
the thumb rubber is gone now 😦
if you like cables yes
Darmoshark M3, VXE R1 Pro
I'm wearing through the top buttons on my g403
I find it hilarious how so many people think cables actually make an impact
They're starting to get very polished
also, Pulsar X2 is super discounted rn
Try having a proper desk setup
Wired peripherals are a must on desktop for me
my G502 had some of the worst cable drag I ever experienced on a mouse
well... they do for me. At least only for the mouse. It's annoying.
Cable drag definitely exists but if you route your cables right it's negligible
I often forget my mouse even has a cable
(I also use it with all the extra weights so its possible the cable just isn't heavy enough for me to notice)
same with my mouse - mainly cuz it's super soft and flexible
"this is my own personal opinion and experience so it must apply to everyone"
that's very likely it - I'm using ultralight now
heavy as shit, buttons give in in no time, mediocre sensor, mediocre software
Yeah I run 2560DPI on it so low sense + heavy it ends up working out
Use the HERO or HERO SE
Buttons are way better on that model
(I have owned 4 G502's)
I broke the scroll on the first one from gamer rage, buttons gave out on the second one, then got a hero, then got another hero and gave the old one to my work desk
I owned one, it failed, I noticed how impossible it is to service, decided to vote with my wallet and take my business elsewhere
got fucked by Asus software, super happy with my Darmoshark M3 now
deathadder v2 pro was way better than the g502
I'll either go with a 502X for my new mouse or go crazy with a madcatz one
I think I bought my G502 in 2015? It's worked perfectly ever since
I feel that the mouse wheel is maybe a bit gunked up or starting to give out, but that's the only thing
this is their only good looking wireless one and I want it
Rip they no longer get produced.
Or did they resurface
just like a mad cat
All this g502 slander makes me sad
Nvm, they seem to be alive still
However retailers are "sus" for delivery in sweden.
there are like 2 here. Amazon and another one (but I already bought from them once)
I used to have a RAT 7 back in the days or whatever
yeah those were always sick
Same, 2 here (but different retailerse), CDON and delenordic.
CDON is pure crap, and delenordic I have no idea about.

Why did my mouse come with a USB extender?
so you can extend the USB
6ft / 2m cable it included isn't enough? For who? Honestly curious.
couch
Huh. Okay, for people who stubbornly want wired over distance? That's crazy to me, but I guess enough people like it they threw it in the box.
also free USB extender
for that one fucking raspi you have powering your 3d printer but you only have 1m micro usb and no extension cords left 😦
I've been on the fence before about ordering 3m / 10ft USB-C cables before, but I've never encountered a scenario where I'd need that. I'll throw it in my possibles bag regardless.
you never need it until you need it
What you're saying is exactly the sort of weird scenario that I'd run into somewhere too. It hasn't happened yet. But the space cost on being prepared for it is small.
it's that once in a lifetime issue and you'll love yourself for keeping that one usb extender you got for free with your mouse
#CableHoarding #Grindset
I use a toiletry hanger for mine. I found bags marketed for actual cable storage aren't as good. I like the zipper pockets on the toiletry bag instead.
Basically something like this.
https://www.amazon.com/Premium-Hanging-Toiletry-Travel-Bag/dp/B01LLLI8IW/
Mine has more mesh pockets though, and only one big plastic one at the bottom.
and you roast my mouse 🤣
Your mouse isn't as beautiful
And adjustable
The madcatz ones are contrary to their looks actually extremely comfortable, since you can adjust width and length plus it comes with different thumb/pinky rests you can swap out
just like pirate software once said i need a mouse not a perfected transformer
the only thing im really interested in swapping would be my headset because it has run its time and its starting to pop and crackle every once in a while
but i have a feeling its cable wear because its very random when it happens its not at all consistent
Sounds like cable wear yeah
502X feeling great in my hand, so that's settled.
nice to see youtube ended up adding a X to close the AD block detected popup letting you continue using it
Is Brilliant a legit operation? Someone seems to have signed up with my Email. So I went to the trouble of resetting the password, logging in, and permanently deleting the account.
From what I’ve heard it’s basically paid khan academy
Well if someone was paying for it, they made the mistake of using an active email belonging to someone else to sign up for it and do coursework, and now it's deleted. Why would someone do that?
Because some people are just idiots with emails.
I know a person who has somewhat common name, and gets monthly(?) registrations with his email from other people.
One or two others with the same name who keep using his email address.
I mean with name.surname@gmail.com type of address.
Is it even legal to have such small power consumption?
It doesn't even consider that enough to lower remaining time from max 30h displayed
i hit the jackpot
i found a 64GB USB stick just laying around in some of my old stuff
might come in handy
Proceeds to get Rubberducky'd by past self
its not even partitioned so ods are i either dident use it or i reset it expecting to use it for something like a recovery
"Why did the monitors of my secondary computer go black?"
Oh, I hadn't used it for 5 hours so it turned them off 🤣
why not just set it to nerver at that point
cause some hours off are still hours
Mainly for automatic lock of the Windows account if I have been away for that long and didn't remember to manually lock.
fair enough
i allways set it to nerver because i cant stand coming back from something and your device having turned off every time
Anyone here wanna reccomend a server hoster for a server when 1.0 comes out ?
i've used shockbytes had a few issues there, any other "better" ones?
I host my minecraft servers usually myself, also I thought we were way beyond 1.0 at this point
huh? :P 1.0 is what comes out in a week no ?
Factorio 1.0 was released 2020-08-14?
Minecraft 1.0 was released 2011-11-18?
im talking about satisfactory
what's that?
And for anything Satisfactory, the off-topic channels are the wrong place?
Like the channel topic says?
i just assumed serverhosting was a tech question im sorry
And the place that might have something to say is probably #dedicated-servers
also a serverhoster can host any game no ? i was asking in general about a reccomendation for a serverhoster.
thanks
Yourself
Grab a laptop with at least 4 physical cores, 8gb of ram, put it next to your router, and host it yourself
You will get better value out of doing that than paying monthly for a provider
rather pay monthly for my electricity bill (/hj)
Is there a way to allow high refresh on battery saver on Android
You are the ultimate computer guy out of everyone here
I think I pick up information pretty well, but you put me to shame
Those two things don't really go together
No, thats the point of battery saver
Is there a way to go 250kph in my Porsche while saving fuel?
if you're getting towed by a cybertruck
That can do 250kph while towing?
Oh right, the gas pedal slides off and locks in place, forgot about that
Was referencing their teaser video
It'll do at least 200 with a porche on a trailer though
If the frame doesn't break off 
Yeah 💀
There was a video of a dude towing a f150 in a not so gentle way and the entire back frame of the cybertruck just broke off
Cybertruck is such a silly car elon truly knows his fanbase
I know 💀 that's my dilemma
What do you need hrr for?
60hz is way too jarring but I like using this app to manage battery
/codeberg.org/s1m/savertuner
60hz is standard refresh for all phones
You have to go out of your way to enable higher
Is it jarring because you're used to having it set higher?
modern AMOLED from what I know is 120Hz by default
at least on android
I stand corrected then maybe
21% in 8.5h at 2W
this thing can be recharger just once a day
Depends on the phone
Mine is 90hz default
I see no need for it
Refresh rate isn't the biggest contributer to motion clarity, and 90hz is plenty for smoothness
I like 120, but I'd take 90hz maximum with really good dynamic refresh over a high maximum. Not sure how phones compare in this regard these days.
more frames more better.
Motion clarity is what matters for mobile, retaining text readability while scrolling
I guess factory does multithread some things.. starting to wish I had a more powerful cpu now
Satis is pretty heavily parralellized
Until autosave
which is a lie, because it just happens to be spread out in that screen shot..
most of the time it's 98% on one thread, and moderate usage on the rest.
it does try, I'm just trying to keep it from hitting the swap file which i think is more of a proxmox issue than anything else.
Do you not have a way to disable swap?
Or do you not want to/are unable to allocate enough ram for that?
I could. I dropped it from 4 to 1 gb swap and upped the ram from 18 to 20.
all my devices are more than 60
120 at the minimum
peak Internet
( I downloaded these 2.5GB in less than 30 seconds )
when Satisfactory 1.0 launches it will probably be ready to open in less than 5 minutes
now it's on 8GB
epic games launcher spotted
This filament is so beautiful
But I can't wait to try this 50% copper one
You can apparently polish or oxidise it
Strawberry flavour.
Something for sure.
When will you guys learn there is a limit to what can be parallellized? If you have 4 tasks, and one of those take 250% longer one thread will obviously be working while the others are not working, hence it looks the other threads are idle. (Extreme example but can be applied anywhere)
I mean sure, but I think htop shows a better percentage of per thread use versus something like top that gives an overview of what's doing what. as far as you guys, I'm well aware that games are primarily st workloads :p
raspberry
Yes
I will print a pi case out of it now
Still have some wood filament for the base so it will look like a pie
Made an idiot rage-quit: #1280631115396419605 message
You really seem to have little idea how driver level API interception used by various antivirus software etc. works...
RenderDoc seems to be doing the same for various graphics API calls.
Which is basically the only way you CAN do such capture without per program support coded in.
Yes, and that is separate from the issue.
That can be used to control RenderDoc from program being debugged.
But this has nothing to do with that, just the constanly loaded driver.
You would get the same behavior if you made manual copy of the driver, uninstalled RenderDoc, and then reinstalled the driver.
And if you manually uninstalled or disabled the driver, with RenderDoc installed, the issue would go away.
But RenderDoc wouldn't work.
Don't mix Windows base graphics APIs and the RenderDocs own control API.
This is the RenderDoc hooking the first type causing the crash, not the game calling the second ones.
When the game calls the first ones, but not the second ones.
Something in the graphics API calls the game makes is causing the RenderDoc driver to barf.
Like the graphics API calls caused old GPU drivers to barf.
U8 needed new drivers from both AMD and Nvidia to work properly without constant crashes.
So there is high likelihood that UE5 is outputting malformed, non-standards-compliant, graphics API calls, but everyone else has adapted their code to work, instead of making Epic fix the issue.
Which is why I said to make RenderDoc issue, so they can look into if the bug is on their side, or UE5/U8 side, and contact CSS/Epic to get that fixed.
Because the game is NOT calling RenderDoc, RenderDoc is intercepting calls to Windows the game makes.
(The person removed the post he made to this channel)
(And to the thread too)
Did they remove all messages in this server?
New user, don't remember the name anymore.
"That was uncalled for" or something similar after the link containing post, and started again about how Satisfactory shouldn't be calling the APIs for RenderDoc control in release builds in here, trying to link the documentation for that API.
Etc.
Not a fan of renderdoc ngl. Because it loves to hook into stuff, easiest fix would have been just to uninstall renderdoc.
That was the complaint, that he shouldn't have to uninstall renderdoc to run the game, and that the game should stop calling renderdoc in release builds.
When it is that global hooking via special driver that is ongoing at all times when renderdoc is installed that is causing a crash.
When that driver does something to the games output that crashes the graphics stack later.
That doesn't happen without the driver being loaded.
So RenderDoc issue, not game or system issue.
Could in theory be caused by nonstandard output by the game, that the RenderDoc driver isn't able to handle.
But the system without RenderDoc is able to handle the output.
And games that actually do output fully standards compliant DX12 are very few.
But as the GPU drivers aren't standards compliant either, it only matters that the game and driver can work together most of the time.
When the driver initialises it will hook into every entry point into the API such that when application uses the API it passes through the driver wrapper. In the case of D3D11 this is the D3D11CreateDevice and CreateDXGIFactory functions.
Not understanding that the "API" in this quote from the How RenderDoc works document doesn't mean the RenderDoc APIs, but the Windows Graphics APIs.
Tdd good or bad? What do you think?
Tdd as in test driven dev?
Depends.
Solo projects, nah.
Bigger complex products, sure. But know your limits. E.g 100% coverage is probably overkill.
I found personal projects the only place were i can have tdd, haha
And it works (especially since i started writing new entities with protocol -firsr aproach)
It's a tool you should have in your dev toolkit
Whether or not it's applicable depends on the project and situation at hand
You wouldn't use a hammer to tighten a screw
I like TDD, but it really depends on the team and the project
How do you deal with bugs (before release and after) without TDD?
Logs + debugger and generic knowledge to look based on debugger / what logs say.
There are a shocking number of developers that have never written a unit test
I don't know if you could ever work in a professional environment that way, but I'm sure there are plenty of freelancers that do
I've seen professional codebases that didn't have any unit tests
I can understand why because unit tests aren't shiny and therefore worthless to management, but still
Pushed for unit tests (and test-first devving) anyway because there was a fuckload of fancy mathematics involved
which is ironic because fixing an issue later on is far more costly than putting in tests to stop those from happening
didn't like Spotify or smth say "we want to make mistakes faster than anyone else"?
And I was almost an idiot, because there was space, just by moving the assembler block one block left without touching anything else...
Fortunately I had thought to make a backup before starting.
Wrong channel...
Correct!
just write perfect code and there's no need for testing or expensive fixes.
and if your devs can't do that, fire your programmers and hire new ones
Yeah, "we have integration tests so there's no need for unit tests"
I wish windows had a lspci command.
Because when you have a function that adds up two numbers, you input 2 and 2 and return 5 you can keep it very consistent with end to end integration tests, but not correct at all
And when your test input datasets are large by necessity it's gonna be very hard to just spot incorrectness
the best part will be starting over. the worst part will be starting over
12900KF for $210 on Amazon 😶🌫️
im surprised 12th gen hasent gone up in price any after all the chaos
because its the only "safe" generation for LGA 1700
And 8 P cores on LGA 1700 is what people are really after anyway.
"Brand New Ford Pinto! only $450!"
somehow i worked in 2 teams where goal was in automation tests for other ... stuff.
and despite being just a fellow programmer, you get assigned to QA and become hated just for having that QA title by everyone else 🙂
I can agree that sometimes it's faster to make prototype, let into the wild for a bit and then decide how to change/evolve/fix it with new knowledge.
But i never seen a manager who would be like "ok, this new feature demo was a success/failure, let's now make it in a proper way and cover with tests"
also, as recent case of crowdstrike shown - even decent quality tests can give false sense of reliability
one of my games' dev teams doesn't have an analogue to the live servers, so what they come up with is basically in beta when it hits live because nobody has actually tested it against what customers actually use
it's hilarious what can make it thru
and on the other hand, when TDD, as Primeagen said, you have to make interface beforehand, but that means that you already write tests for the interface that might not be very usable in practice when writing features/other code.
So this is kind of a place for me to sometimes tradeoff into integrating new code in small usecase, confirm its usable, and cover in tests for wider scope of usecases.
the opposite of exhastive testing...
you'll love their official excuse
in an mmo where they could merge all the servers because the population is so low that it couldn't cause lag, they don't want to make a live-analogue test server "because if they went through all that trouble, we might as well be putting it live for more customers"
how did we got convinced that it is impossible?
I mean, ofcourse, larger the project the furthest are we from practical 100% coverage (and even more so in terms of actual edge cases, loads etc)
But, i'v seen it being used as a reason to not write quality tests too many times
w t f
welcome to the world of WotC property
that's like "if we test shampoo on cats, we might as well take director's daughter as test subject"
oh there's a reason i don't pay into it 🤣
this is a company that forgot to pay for its SSL keys. twice. TO THE DAY.
recently they had an issue where they added in an "expansion" along with the ability to teleport to its main public area like you can with other expansions/locations.
but they may have forgotten that that means that you have to either make the UI panel longer, or add in a scroll bar, so people can, ya know, use it.
so after a week they undid it, then spent another 6 weeks screwing around until they made a decision. and their decision was.......longer panel. only now for some reason you can manually highlight the NPC's names & their text boxes, which screws up every NPC dialogue panel in the fucking game
and these people have the balls to ask up to $130 for an expansion
as history shows, most amounts of money go from idiots to idiots people who make poor decisions to other people who make poor decisions, haha
why
Can't even see the fuckin screen

Anyone have any thoughts?
That you have HW problem in your computer? What about it?
I think you need to run some memtest
Disable XMP as first step and rerun the Windows Memory Diagnostic.
I can't find out what the problem..
That first pic is the lite version inbuilt into Windows.
So if the error continues without XMP, then the RAM will need replacing.
But first need to test it isn't just that the RAM is being run at too high settings.
Bios?
Yes
could the intel micro code problem be that or was that only vram?
Yes. Set XMP/DOCP to not be on in BIOS, so that all RAM settings are at auto.
Or just clear CMOS.
Wouldn't cause issues in the Windows Memory Tester in any likelyhood.
The VRAM thingie is that the GPU driver data gets corrupt input because of unstable CPU.
But the tester shouldn't run the CPU at speeds where the issues can happen.
But the first likelyhood is that the RAM you have just cannot run at the marketing spec (XMP/EXPO) in your computer.
Those settings are guaranteed for the RAM in golden sample system, not in buyers system.
I prefer a memory test that shows info beyond "something's wrong, talk to your oem uwu"
The more in-depth results are in Event Viewer, but not needed at first.
Yes, MemTest86 is much better.
But also needs more knowledge and effort to use.
Yeah, I suppose, I just don't like windows' memtest because it's so opaque and therefore there's no real further steps you can take
So just disabling XMP and then rerunning Windows Memory Tester gives good baseline to differentiate between physically faulty RAM, and if the current detection was just unstable XMP.
Fair
Apparently xmp was never on..
So almost certainly bad RAM stick that would need to be replaced / warranty RMAd.
Yeah, my guess'd also be either bad ram sticks or 4 ram sticks
I've got 2 with 8 on each..
4 sticks on all auto should just work in most cases, even today.
Yeah, probably, but I've had an issue where 4 sticks tested good in pairs, but failed all put together
That was with xmp on however
XMP usually will not work.
With DDR5, with 4 sticks both Intel and AMD will run the RAM at 3600MT/s on auto.
💸
dead memory
No idea if the multicore test in Memtest86+ is usable at this point.
It generated failures from just bad code for long time.
MemTest86 and MemTest86+ are two completely different programs.
But probably dead RAM, but not sure with SMP testing with 86+.
I’m becoming an IT specialist in the military 🥳
Training starts in 2 months after my drivers course where I get a free C-category license
Hey, that's pretty sick
Cheers
Testmem5 with absolut preset is a good test
Base course is ending tomorrow, the main pain is over hopefully
For RAM OC stability testing.
This is diagnosing if there are HW level RAM faults.
Which seem likely, but the MemTest86+ SMP mode was broken for 10+ years in meantime and I haven't checked if the new versions are actually functional or not.
I’d imagine if it faults, it’s also unstable
Because MemTest86 has free version again, and 86+ was in limbo for long time.
No, just code bugs in the multi-core mode.
I think we can discount software bugs causing 100% false positives here, since Windows' built-in memery tester also saw issues
i love a random instant bluescreens with event viewer empty after
i tested stability but i guess i dident test it enough
hmmm, anyone willing to do a quick and dirty mic check with me ? apparently I'm not hearing other participants, despite sound working in general
these any good? https://www.amazon.com/dp/B084CJLNM4/ref=twister_B08KXCTVVJ looking for a raspi boot drive..
Built to capture up to 15,000 hours(1) of footage for your home security cameras or dash cam, the SanDisk MAX ENDURANCE microSD card is engineered not only for continuous recording and re-recording, but also for continuous peace of mind. Designed to last(1), this microSD card can withstand a vari...
so who's excited to see starliner not be able to undock from the station tomorrow
Er, I don't know that's the right attitude. SpaceX having credible competitors would be good.
I totally get the satisfaction of watching bad corporate culture eat itself though.
real
me
i'd be nicer about it but they don't deserve it
dont be suprised if its hatch flies off
and if i were SpaceX i'd 100% send them a bill for the rescue
i would too
they shouldnt be the one to pay for their loss
its not spacex's fault for them not making a good spacecraft
this starliner wasn't even "the final one", the next one would also need to have its own cert flight, which means this entire mission has been meaningless besides taking crew & a bit of cargo
if i were nasa i'd be seriously looking at suing boeing for fraud, lying about the readiness of their craft
That's not at all what's happening, since nasa's paying them anyways
They're not sending a special mission to "rescue" them either, they're transitioning them to part of the crew 9 rotation
Which was a launch that nasa had already purchased and scheduled
boeings stock is cosplaying its doors
Look at intel stock
ik
that means they're bumping a mission
a crew & training as well
if they hadn't launched at all, that bump would not have happened
Just two people from a four person mission
But yes, it still costs nasa
still 2 less people
However, it's not like boeing hasn't suffered enough
i don't think they have
They're several billion in the hole on the starliner program since they're on a fixed price contract
They've spent about 3x what they're ever going to get paid, especially since it looks like nasa might not even purchase more than the inital three launches
who in a position of responsibility & significant financial power within boeing has been fired explicitly for their incompetence over this. if the answer is nobody, then they haven't learned anything
actually, i take that back
they've learned that they can get away with it again
and i'd send congressional investigators to dig thru every email at boeing & their suppliers, because somebody knew they weren't ready
i'll take a cover-up of incompetence over them being stupid in this
but i also don't want nasa to cancel their contract, i want boeing to back out of it
i want boeing to be the one to admit that they can't do it
all their shit-talking about SpaceX being some kind of underdog, only to get outcompeted by them at every turn. i want to see them admit that they lost the competition
or, they could invest in doing something right
I just wish instead of the cost plus contracts and everything to do with SLS, all the money instead went to literally anything else.
forget sls
that stuffs never gonna happen
i don't mind it going to space R&D, i just don't like it being used on garbage
with all the launch tower costs and devoplement?
Right, I'm saying SLS should have been forgotten long ago.
if the launch tower was going to be something guaranteed to be used for public & private stuff for the next 50 years, i could understand the cost
but right now it's like 4 missions
use unfair human rights 🗿
wut
Lawmakers boast about SLS providing jobs in all 50 States, but I would just as soon have them spend that money on paying those people to play DDO all day for Cuyar's amusement.
real
Or maybe have them start their own Star Citizen or EVE group or something. Basically, the whole SLS industry just forms some giant faction in some MMO.
nah nah
maybe KSP2
it can host jobs in all 50 states
a functional one
just dont pay em!
infinite money glitch dropped!
turn the launch tower an island in the ocean
since theres no minniumummumius wage there
i wonder how many DCFCs for parking lots that'd fund
apartments, offices, restaurants, hotels
Nuclear power! Modern nuclear power!
with my luck someone in that crowd would know exactly how to solve certain issues with the game
and they still wouldn't lift a finger to actually do it
Or they would just give you a choice of playing the legacy server with no updates forever, or the new one that works.
forget dcfc parking lots, we got concrete
fun fact that i got out of one of the devs: they don't keep old versions of the game
once an update is out, they basically delete everything previous
why is even touching topic of cache (without even actually impelmenting cache itself, let alone the invalidation) can be so hard?
Does it make sense to replace cache fully operating by "we take object, make a hash by its properties, save it's data as a bunch of files in a folder named by the hash" with SQLite3 in a form of "we take object, make a hash by its properties, store metadata in in sqlite and the binary data in a file/folder named by the hash" ? (note: single object can take up to 30MB with a decent size metdata - it's a http request-response cache)
Expect gobs of Intel rumor type stories in the next months as people take early retirements.
A retiring employee just told me a few stories about why he really likes Pat, and why other people do. There's a lot of hope in the company that Pat survives the missteps that have happened of late.
it's just kinda weird that he's been at intel for both of its MAJOR screwups
Probably, but what about Samsung Pro Endurance in price comparison?
As that has been the spec endurance leader and also has great random read and write performance, even when it doesn't have official A1 or A2 rating.
Fixed price with milestones.
Boeing isn't getting paid anything for the flight unless NASA accepts that it completed all requirements for the certification flight and the contract can move to actual crew flights.
Ah, that was next, but that line seemed weird.
$2 more for 64GB card
https://www.amazon.com/gp/product/B09W9XYQCQ/ref=ox_sc_act_title_1?smid=ATVPDKIKX0DER&th=1 Don't really need the space.. but if it has better wear leveling then that's a win
Samsung PRO Endurance New Portable SSD for monitoring devices is built to last up to 140K hours of recording (256GB model).** This solid state drive stands up against magnets¹, X-rays², water³, drops⁴ and more. Record and play in FHD and 4K. The PRO Endurance is engineered to deliver long lasting...
I'm not going to run anything other than octopi.. and a jmri server, both of which are servers of sorts that'll be doing mostly log writes.
Those Samsung Pro Endurance cards fullfill the A1 spec in testing and have been giving better performance in testing than many officially A2 cards.
octopi does have writes when you upload files. but you get the idea.. not dashcam/camera recording use.
I'm just after a long service life. I've already dealt with a failing boot drive on one of my servers... it's been awhile, but if you remember the proxmox memory address errors I had posted.
Yeah, I know.
I got Pro Endurance 64GBs in my Pis.
But trying to do current comparison on the write endurance ratings.
that whole month long ordeal was a because of a failing boot drive lol
Ah, 15k hours on SanDisk 32GB, with Full-HD video without spec on bitrate.
And 35k hours on Samsung 64G, with Full-HD video with 3.25MB/s / 26Mbit/s constant writes.
17.5k for the Samsung 32GB.
so 30k on the sandisk ( no spec ) and 35k on the samsung, which I'm not sure I believe in their testing/rating.. so about equal with the sandisk being more expensive/GB
Yeah, 30k hours on the 64GB Sandisk.
But as they don't specify the bitrate, it can be much lower.
10Mbit/s is the normal max for full-hd recording on many of the cameras the cards are meant for etc.
So not specifying the actual write speed gives lot of leeway to overstate the endurance in hours.
So I trust the Samsung rating much more than the SanDisk rating.
In 2019 the SanDisk packaging has specified that Full-HD is that same bitrate.
but that samsung card hasn't been around for 35k hours so their testing is just writing one cell to death and calculating a rating on it, if even that. That's kind of nitpicking and it becomes a game of chance
Higher rate writing works the same.
And it can be somewhat calculated by knowing the flash and the controller.
And same for the SanDisk?
that's what I'm curious about, and there's little to no idea on either of them as far as that that I could find.
I guess I'll find out if the pi still runs when it's taken out of service in however many years. I'm leaning more towards the samsung because at least in theory larger card = less utlization = longer life.
That Samsung Pro Endurance line has been in existence for long time already, and they raised the endurance ratings between 2018 and 2022.
My 64GB old Pro Endurance card in my dashcam since 2018 is still kicking fine, no idea about the total amount of writes.
And they test well for App-type rating too, even when they don't have one, so no issues with Pi usage from bad random read and write performance caused stalls.
The old SanDisk High Endurance cards have much lower endurance ratings and that Max Endurance lineup that almost matches the Samsung Pro Endurances in that is pretty new.
samdung it is then.. just don't buy their tv's
No idea about the random read and write performance for SanDisk Max Endurance.
Samsung FLASH products are solid if there aren't firmware bugs that have caused issues with some SSDs until fixed.
toms hardware said they were pretty good in pi's.. the cheapest silicon power card was the fastest.
Yeah, great until the log writes kill the card.
how long has it been in service for?
actually toms says the sandisk is the fastest
Since like April?
Would probably happen between 1-2 years of service, possibly in as short as 6 months of runtime.
with samsung is better for endurance.
Oh no then I'll go bankrupt 
I think my current octopi server has been in service since.. 2018 or so.. I'm upgrading it from buster to bullseye
No SanDisk Max Endurance in that testing.
Samsung Pro Endurance is and is giving respectable performance like I said, not chart topping but good enough.
I want to play the long game >.>
Yeah, depends if you have backups or have to reinstall etc.
If you just make image of the finished card, and then just deploy that on the next one, not much harm.
it's been on and off useage, but it has a samsung 32G card.. the tip is orange, it might be an endurance one
I'm gonna set it up on a USB stick anyways soon.
But I rather paid extra to not need to do that from time to time.
I don't mind paying more.. I'd like the storage to last the service life of the machine.
I just got my Pi4 for my printer yesterday so I can actually boot from USB
Before I used a pi2 
I ran containers from pimox off of a usb.. it was awful
Thank god my printer doesn't do that
it was usb 2.0 but oh boy IOwait just stayed pegged at 100%
sometimes you've just got to try things.
For some reason Toms Hardware sorted this chart by random write speed, but for random reads the Samsung Pro Endurance is second:
I'm running my unifi server on a pi5 from a USB stick and its running pretty good
After Samsung EVO Plus.
you should look into getting a n100 or something similar.. something like this https://www.amazon.com/ASRock-N100M-Micro-ATX-Motherboards plus 16G of ddr5 spanks the pi5 silly with what you can run on it.
Yeah but it also needs more power
not really.
15W or so.
Intel® Processor N100 (6M Cache, up to 3.40 GHz) quick reference guide including specifications, features, pricing, compatibility, design documentation, ordering codes, spec codes and more.
6W
mine draws about 20W total as reported by the ups.
that's not really at idle either.
mostly idle but still doing stuff
And I was meaning that the power draw of any Atom box is going to be significantly higher than a Pi:s, even at idle.
Might get closer at full-load with the newer Pis, or not.
not to get sidetracked from the pi talk, from what I understand the pi5 is quite powerful just as far as running services goes I would pick the n100 over it.
Depends on what is needed.
that's true, mines running a terraria server, pi hole, wireguard server, home assistant, serveral docker containers, wikijs, and nginx lol
That N100 gives more performance if needed, but if that is needed or not depends.
For significantly larger space requirements, and higher power usage.
So goes into performance needs vs. space etc.
and much more flexibility. depends on needs/wants for sure. I wouldn't use a n100 for octoprint xD
so from what i've been reading all this time, intel 12th gen is basically unaffected by any of the bullshit, right
Yes.
just a sanity check 🤣
You could use the VID request capping to set that to 1.50-1.55V if you wanted to be safe even on 12th gen.
i'm getting antsy about getting a family member a built system to replace an aging laptop
might donate my 5600X platform & upgrade
12900KF is $250 rn 🤣
IA VR Voltage limit on Asus and Gigabyte
MSI has decided to not expose that. No idea about ASRock.
As in that is Intel provided functionality, that MSI just has declined to explose to user.
Even after this shitstorm hit.
Alder Lake is a good buy at the right price, 12900KF for $210 the other day was great.
Intel cancelled their 20A node and products that were supposed to be produced on that will just be done by TSMC instead.
And concentrating on 18A instead.
Why not get a 5800x3D?
Because there is another person that needs basic computer now.
So moving his computer to that person and getting new gaming grade computer to replace it has plus sides vs. buying just that basic grade new computer.
Ah, misread that
X3D for factory games tho
also why did nobody tell me that 7600X3D was a thing?
Hey guys! That imagine format, is it pronounced jif or gif?
Just came and is very limited run.
I pronounce my imagine format as a "think"
Thought was that it would be Microcenter only again, but now it is MindFactory only for Germany, no idea if other countries have access.
mindfactory mentioned
I might need to switch to prusa slicer instead of cura. It got a lot of nice features but also seems to be very cluttered
I'm slowly trying to get my girlfriend to not rely on Google
Its so hard you don't even understand
Oh I've gone through the same thing so many times
She keeps complaining about targeted advertising from Gboard, but won't switch
I think I'm gonna start with password manager and different Keyboard
FUTO keyboard is so good, minus gifs but eh who needs them
I'm fighting the passwords in the notes app thing still
sometimes id rather have a good product than a good company behind it
And that's what is wrong with society today
ltt did some degoogling videos.. part 2 got taken down lol
googledebunkers
I guess I'm stuck on 5.10
I upgraded from buster to bullseye and it completely broke octoprint.. had to reinstall it
Raspibian and/or Raspberry Pi OS are not made to be dist-upgradable.
well, it's linux under the hood.. so it's possible lol
Even when the base Debian is.
But the Raspberry Pi folk don't bother with supporting it with their custom packages.
I think the dist-upgrade wiped out a bunch of python packages, and erased some of the venv stuff octoprint was setup in
Remember that the current version is Bookworm (Debian 12).
I'd like to move the kernel to 6.1
So even that Bullseye is major version out of date and from year old.
Still supported, but not for long.
Yeah, but if I can't move the kernel to 6.1 then I'm not going to bother upgrading it to 12.
Linux raspberrypi 5.10.103+ #1529 Tue Mar 8 12:19:18 GMT 2022 armv6l GNU/Linux
Raspberry Pi 12 is on 6.6 IIRC.
Or was it 6.9?
I guess 2 years old is okay..
Need to manually change the kernel version tree package.
Because Pi uses fixed major version ones by default.
Yeah. I had to do it with ubuntu for an intel gpu.. I've slept since then
While base Debian just specifies kernel arch by default.
But the base answer from Raspberry Pi developers on this kind of stuff on their forum was that dist-upgrade isn't supported, just reinstall each time...
Which is idiotic IMHO.
it is. makes me with I had just used debian arm or something for a distro rather than raspbian
And they have multiple cases of that kind of package name changes between major versions that don't get auto-changed.
Yeah, I used Debian Raspberry Pi Bookworm images, and then added couple of specific packages from Raspberry Pi repos as addons on the Pi4, and just base Debian Raspberry Pi for the Pi1.
But basically all the Pi config utilities etc. are missing from those, as they are Raspberry Pi OS only.
Most of the stuff is possible to do other ways too, but no-one talks about those and just tells to use Raspi-Config etc.
raspi-config makes it easy.. and pi being somewhat unique I can see how that would be an issue if you were trying to enable pcie on the cm4.
well, I guess this system isn't as old as I thought sudo tune2fs -l /dev/mmcblk0p2 tune2fs 1.46.2 (28-Feb-2021) Filesystem volume name: rootfs Last mounted on: / Filesystem UUID: 87b585d1-84c3-486a-8f3d-77cf16f84f30 Filesystem magic number: 0xEF53 Filesystem revision #: 1 (dynamic) Filesystem features: has_journal ext_attr resize_inode dir_index filetype needs_recovery extent flex_bg sparse_super large_file dir_nlink extra_isize Filesystem flags: unsigned_directory_hash Default mount options: user_xattr acl Filesystem state: clean Errors behavior: Continue Filesystem OS type: Linux Inode count: 1918752 Block count: 7748608 Reserved block count: 318497 Free blocks: 5915918 Free inodes: 1766113 First block: 0 Block size: 4096 Fragment size: 4096 Reserved GDT blocks: 219 Blocks per group: 32768 Fragments per group: 32768 Inodes per group: 8096 Inode blocks per group: 506 Flex block group size: 16 Filesystem created: Fri May 7 10:06:09 2021 Last mount time: Fri Sep 6 11:23:44 2024 Last write time: Sat Aug 31 00:55:59 2024 Mount count: 64 Maximum mount count: -1 Last checked: Fri May 7 10:06:09 2021 Check interval: 0 (<none>) Lifetime writes: 31 GB
2021
Yeah, but raspi-config etc. are not in all the OS options.
And is just thing that makes things easier, so what the specific options do in actuality would be nice to be documented somewhere so the same changes could easily be manually done.
Because it isn't doing anything special, just UI to select things.
it's kind of the case of making shit harder to use by making it easier to use ig
But basically the Raspberry Pi folk don't care about anyone not using their own distro.
need something specific and unique to a pi? use our software or get fked.
Which does some idiotic stuff to the base Debian.
So I went with the base Debian Arm with ready made images.
But for the Pi4 in the end I had to install the Raspberry Pi 4 kernel etc. to get video decode hardware acceleration.
Which isn't supported by upstream kernels and so not supported by Debian.
some of that might have to do with restrictions imposed by qualcomm rather than raspberry
Basically.
Raspberry Pi OS folk have to keep patchset, some of which is sent upstream but not accepted.
But it was annoying, because they have split stuff differently in packages compared to base Debian kernel & firmware stuff, same package names but some specific files were in different package.
Which made it impossible to use the base Debian files in combination with the Pi Kernel.
SUPPORT OUR REPORTING: Buy a coaster pack, solder mat, or more at https://store.gamersnexus.net/
In this report, we close out the story on what's going on at the tech company that sold over $250,000,000 of products in its time in business. The company has been collapsing from within for years now, silently, but grew louder this year as contrac...
Here we go
So for others: About EKWB
wow, is this the first time GN's broken an hour?
I don't think so
Hopefully the audio doesn't sound sped up. It's already long and in need of breaks for most people. Go all in and give good delivery on the audio. Though I'm not going to watch it regardless.
Its too long for me to watch
Mostly I'm just not interested in the subject either way.
So apparently I inadvertently have Tuesday off. In addition to Wednesday, Thursday, and Friday. Fun times.
Hopefully it's PTO
i wanna hear all the dirt 🤣
Nope, but meh.
hopefully they keep track of the players to see where the violators escape to
I got plenty of dirt out of Intel employees last night.
Honestly, it makes me inclined to believe pretty much anything people like MLID say about the internal history of Intel.
What's the tldr on that
Honestly, lots of things. And tell people will pretty much talk openly about anything that's 5+ years old. The Beast Lake video from MLID has plenty of kernels of truth to it, though no one I talked to really knew details. Mostly in that conversation I tried to impress upon them that gamers definitely still need more CPU performance. Especially gamers who play games like Satisfactory, or anything else where the amount of pathfinding, physics, and other calculations is basically unlimited. Unfortunately, gamers are a fart when it comes to Intel's revenue.
Employees would definitely go to war for Pat if it feels like he's in trouble.
10nm was actually one of the things that someone dished on last night. Cannon lake went nowhere because they tried to use the 14nm tools to make 10nm with Cannon lake. The China only Cannon lake part was basically only released as a preventative measure against lawsuits.
Krzanich was hated as a CEO, and a lot of the internal politicking problems you hear about that era are most likely true. Bob Swan wasn't necessarily hated, but he was a finance guy who didn't know shit about technology and was unfit to run the company. One of the people I talked to was head of an engineering team, and another in fabs. Both of their faces immediately lit up when I brought up Murthy. Apparently he was one of the worst (to them).
As for why any one of them would talk? Intel employees don't operate like military intelligence it turns out. They're normal people who will bitch about their jobs and bosses like anyone else. Last night I just decided to dig into them a little.
Oh military talks. Just gotta do it in a bar
Real DJ moment, testing my own speakers.
im having a great laugh at this OCCT run
just pure proof that no stability testing was done in its rawest form
Checks out, that's where I got Intel people talking.
il deal with your false advertising, software that dosent even support your own products, oboard speakers and microphone that are permenantly disabled... and so on but shipping unstested faulty hardware is where i draw the line
just throw a satellite dish on the roof, it'll be fine
(context)
jesus I hate how extreme the difference of temp is on nozzle sizes. 260°C on 0.6mm prints like trash, 0.4 and it's smooth like butter
now I need to learn how to compile my own firmware so I can go over 260 ;-;
what's better is using an air line that's 13mm & not 1/2"
i had one go flying off an air compressor
because 13mm is NOT 12.7
What heater do you have in your hotend?
stock heater of the v2 neo
can someday manufacturer stop saying they do not sell at msrp
What print speeds are you pushing?
it's getting so much on my nerves
hotend itself is the spider v3
currently fairly slow actually
What is fairly slow? 100 mm/s?
printing carbon fiber petg rn
currently at 35mm/s
Oh weird, you shouldn't need to push much higher than 240 for that
Still
my bed alone is at 90 otherwise it doesn't want to stick to my cf-optic bed (which is fairly difficult to print on anyways)
That's not too surprising
it's also a hardened steel nozzle so I need more heat since it doesn't transfer as well as brass does
Hello everyone! Can you share with me, your graphic prefference or gama balance? I play on the laptop rtx 3060, 16gb, ryzen 7 5600. My graphic pref is low. But after 1-2 hours in game, my eyes are leaking. Instead in Whicher 3 or Titanfall 2 everything ok even after 6 hour playing. Who have the same or near problem?
as I said in the other channel I play on ultra with similar hardware but on PC and get fine frames, so you can probably handle at least medium
Thanks)
SHUT UP OCTOPRINT
The clippy of 3d printing
Yeah.. I've got that feature on mine too since I had to reinstall it... what a bunch of bullshit
Someone needs to fork em. dumbass achievements but no fucking dark mode.
What's the use case for wanting to switch scroll wheel modes on the G502 on the fly? So far that's my only complaint about the mouse, is that's basically a dead button for me, as it can't be reprogrammed. I knew that going in, so meh.
windows 11:
oh look they are talking about their dark mode now i have to run!
goddamn i thought i lost my monitor for a minute
cleaned off the screen for the first time in way too long, and it wouldn't turn back on
i had moved it around to get the right angle vs the overhead light to see all the little dirty spots
turns out i had moved it just enough to yank the power cord out
scared the crap outta me
Or just find the location in the firmware binary file where that limit is and hex edit it.
Just another reason to use klipper instead 😄
ngl i wouldn't mind if logitech licensed out their infiniscroll wheel lock thingy to other mouse makers
that was the one thing on a logitech mouse that i never had any issue with
there is but i like the freewheel feeling when i want it
it's like getting used to something like the lights vs wipers on different sticks in japanese cars. it just makes sense & getting into another car definitely feels like it's missing something important
Standard anywhere not US again?
basically
Never seen a car in Europe that didn't have separate stalks for those...
also a 3rd stick for cruise control just felt right
Including Fords etc.
ford still has to comply with euro standards
Yeah, but meant that nominally US manufacturers having dual-stalk controls outside of US.
true
So they could very trivially use that part in US production too.
but my info is also with older cars, from what i can tell there's no rules anymore with a lot of it
depends
Pretty good, but on the G-series mice the buttons can get broken quite quickly
sometimes logitech uses decent quality switches
So basically mostly good but often crap switches.
So if you are able to do switch replacements, no problems.
If you cannot, expect to need to constantly replace the mice.
As in change once to a long-life switch after the originals start to crap out.
BAsically mod a 3 year old mice to a never dieing mice
still have to change the switches in my G501
Less than 1 year in cases.
i just switched over to redragon, cuz if the switches are gonna last a year, i might as well get something with everything on it for half the price
Cases of people having time to get multiple RMA replacements in warranty period for switches going bad.
That's warranty terrotory
and logitech likes to put that stupid grippy shit in places where it'll fall apart or get super slimy
Those can fortunately be removed in few hours when starting to go bad to get just normal plastic surface.
Some scraping and then final isopropanol cleanup for the plastic parts when disassembled.
Razer has always used shitty cheap switches for the side buttons, even when they did use Omron for the mains.
Often exact same form factor, but chepest they could find for the sides, mid-life Omrons for the mains.
And later went to shit switches for the mains too, or lowest rated Omrons.
i wish they were more easily disassemblable
just to clean them out
perfect mouse:
-logitech scroll wheel
-steelseries switches
-redragon price
i consider longevity to be near the top when it comes to peripherals
the worst thing that mouse makers do is stick the case screws under the pads
I know that Keychron M6 ticks the Logitech scroll wheel
49€ idk if that counts as redragon price
my M913 loves to get buildup on the edges of the side buttons
think i paid $35usd for it
still rocking an old microsoft mouse, yet to have the side buttons fail
i got addicted to the Logitech G602/604 6x side buttons, & now having a bunch of side buttons is just too nice to miss out on
yea up to 12 now too 🤣
idk why i was worried about accidentally clicking them
isn't it a popout window that you can maximize?
when I open history, there's a button at the bottom "manage history" which does exactly that
(well, it's "Chronik verwalten" in German, so I just eyeballed the translation)
crap. now i want a mouse with touchpad left/right clicks 🤣
oooooooh. how about not a scroll wheel in the middle, but a touchpad there that will let you scroll, but also has some left/right as a joystick
it got eaten by a xenomorph
I'm at 91 at the moment, Its lower than normal for me
search? taking away adblock?
oh wait, you're talking about Chrome
yeah, using Chrome is just big skill issue
Starliner is back, but:
- one of the 12 RCS on starliner (not sm) didn't fire at all
- sounds to me like the navigation system is tripple redundant, string 3 (SIGI 3) failed of after the plasma was left, string 2 had also "hickups"
And the two planned Boeing representatives skipped the after landing press conference at last moment.
And to the yesterdays high write endurance SD card conversation, that dashcam of mine is recording at about 36Mbit/s total from the dual cameras when the car is on.
As that 64GB card can only hold about 4 hours of video.
My scroll wheel tilts to scroll horizontally
So yeah, one capsule RCS thruster not working at all is confirmed and that the navigation system fucked up during the plasma blackout and some parts had to manually be reset after that ended have been confirmed...
So in how many ways can Boeing keep making the situation worse?
And if you only have 3 in voting array, if one is fucked you just lost all redundancy already.
As you cannot determine anymore between the two if they differ.
IIRC SpaceX uses four dual-units.
Which autoreset the unit if the two subunits in same unit disagree.
So four units voting (so one can be lost and still have redundancy), and then unit-level sanity checking and autoresetting etc. if the unit itself thinks it is fucked somehow, which they can do as if one unit is resetting, they still have three units voting.
more whistleblowers disappearing
We're at two already, right?
https://support.discord.com/hc/en-us/articles/22045487931799-Members-List-Recent-Activity-FAQ this is the most annoying thing ive seen let me turn it off
no one wants to move to better things
discord has the popular name brand at the moment, and haven't really "done anything" to ruin the name yet. once that kind of thing happens, we'll start to see movement
What alternatives are there?
skype was one of the big ones before discord, but its name kinda got ruined when microsoft bought them out
and i remember ventrilo as the "gamer one"
but i wouldn't be surprised if a completely new one, or one that's around but is little, starts to get more popular when discord starts to become unpalatable
i suspect that it'll be difficult to break into this area now tho, since now people will be demanding more things from the start like encryption & bot prevention
Skype is EOL, replaced by Teams
Which doesn't have the social network public server aspect of Discord
TS kinda has that but its UI is horribly clunky and the chatting capabilities are less than bare minimum
theres guilded but roblox owns that
sooo...
I think I found out how to improve cf-petg printing quality dramatically
it's one of those "water expands when it gets colder" bullshit things. I had to go faster
Oh very interesting
my guess is, you need to constantly push an amount of material through the nozzle otherwise you get those annoying mini-clogs I kept getting
gotta go fast
That's what I get for using chopped fibers instead of ground
And for that SpaceX Falcon 9 landing failure, the FAA grounding lasted grand total of 15 days 😆
Hadn't noticed before.
idk why it got grounded in the first place, it didn't happen during ascent
Without FAA knowing the cause, they had to go in the direction of safety.
Until SpaceX proved that the issue wasn't something that could endanger the public (like for example engine underperformance could).
i think sending nasa a letter that just says
SpaceX: 345 landings
Everyone else, ever: 0
should be enough
FAA, not NASA
a barge in the ocean is a weird fuckin jurisdiction for the FAA
And no, it isn't.
Because the problem wasn't losing booster on landing.
It was if the cause was something that could affect going up too.
FAA has jurisdiction for any US company doing space stuff.
also makes it pretty clear that it wasn't one of the engines that was the problem if they cleared it up that fast
the slo mo that scott manley showed of that right rear leg made it pretty clear that was the problem
One satellite company tried to go around that and launched on Indian rocket without FAA license for the satellites and got hit pretty hard for that.
US based company.
lul?
TBH I don't mind people who like spacex for what they do, since what they do is AFAIK pretty decent work in making rocket launches cheaper, IIRC it's the only Musk company that hasn't been touched much by his sheer pants-on-head idiocy
In a lot of ways it has, but in a lot of ways it's served spacex well
He's also had a really really good team of competent people around him at spacex
now if he would just make a tiny tesla that i could afford
i would totally be in the cult of musk then
Ohboy alpine is very different from debian.. now how the hell do I make a proxy server with it lol
you would probably have the 3rd gear behind monthly pay wall
For sure
It used to be there's an app for that.. now there's a subscription for that
And of course, companies will claim that customers definitely wanted that subscription.
delete that, don't give them any ideas
There are no Teslas with 3rd gear anyways?
And not sure if any of the current production even has 2 gears.
All the later ones are fixed reduction gearing "direct" drive.
Some of the competing EVs use 2-ratio gearboxes to get better performance at very high speeds.
I'm fairly certain there are at least two gears somewhere in that car.
EVs may not need a big transmission block but I'm fairly certain even they can't run entirely without gears
The first tesla's had 2 gears the one's today has 0 gears with direct drive motors and tranmission.
Not a single gear in the entire car?
Are you taking about hte piece of metal Cog/Gear or a transmission?
Single fixed gear ratio reduction gear per motor.
First one
So then putting the third gear behind paywall would be same as putting "the car is able to move" behind paywall.
Vs. limiting max speed via putting transmissions higher "gears" behind paywall.
⚙️
Yeah it will no doubt have a cog somewhere.
One of cybertrucks drive unit reduction gears.
Front one that is shared for both front wheels, so also contains differential
See? More than two gears
lol 🤣
i was more thinking of the weird magnetic phase stuff that tesla figured out
AFAIK Tesla (the car company) hasn't done anything revolutionary in electric motors, inverters or reduction gears.
Just incremental improvements in combining existing stuff into more integrated wholes.
So unless you are referring to the person, no idea what you are referring to.
Those are some ball bearings
(Insert inappropriate joke about musk fans riding his balls)
45kW electric drive of my desing 😄 Will have 3 "virtual" gears from 400V frequency converters. 70:1 reduction in the hub
Why not just normal continous control everyone does?
If DC fed and not AC fed.
As variable frequency from DC to AC conversion is just programming compared to fixed.
Baldur you're literally holding back innovation and the betterment of society with comments like that
Three fixed frequency outputs is worse than variable frequency covering the whole range?
As you get that variable frequency output "free" with any non-mechanical DC-to-AC conversion process.
It's AC and the gears are more like ranges to allow precision with analog remote controller
So wall powered and not battery powered?
But yeah, so you are just doing virtual "gearing".
CAT C9 250kW diesel gen
Please tell me your analog controller looks like this
Like current CVT gearboxes etc. in cars do.
When drivers used to old gearboxes didn't like the continously changing ratios.
Yeah something like that. Driving is secondary function in this tracked mobile jaw crusher.
Currently at lunch but I'll get a pic
isn't the current "big issue" with CVTs that they can't handle high torque?
No variable transmission can do as much as gear based at same size.
No matter if older type automatics with belts etc, or the current more advanced CVTs etc.
Which is why computer controlled manuals are the high-end, often with dual-clutches etc.
But the current CVTs could allow for drive-by-wire where the pedal just sets the speed, and engine RPM is set to most efficient for the wanted load.
But that was too weird for most drivers, so that was dropped very quickly and replaced with just virtual gears.
Yeah, that's how most modern farm tractor CVT work
yea i tried a gen1 CVT from Jeep & it was GARBARGE 🤣
But like Fendt Vario still has 2 ranges you select stationary, for field and road.
i wonder what hooking a CVT up to an electric drive would feel like
Extra two speed manual in there for more optimized ranges probably.
Like most "low gear" vs. "high gear" selectors
Not much point.
As electrical motor already provides all the same functionality except at very high speeds where the max torque starts to drop.
Which is the reason to some EVs having that low/high-gear two speed gearbox to give more very high speed performance.
wouldn't running the electric drive at a lower RPM reduce heat?
No, actually it is the high RPM that needs less electricity.
ah, using the syn drive
Max current is at stationary, at max torque.
How much the high RPM effects things depends on the motor type (induction, permanent magnet etc.) but the end effect is same.
So adding dual-speed gearbox, you could reset the curves to second point at higher starting speed.
But for normal car users that really isn't necessary.
But the high-performance EVs do that to get better performance above 100km/h and to be able to do 200+km/h.
That percentage is battery charge when testing.
SOC = State of Charge.
So getting same torque from 0 to 60-70km/h.
Looks like the remotes are not with the machine. Understandable as only thing missing before test drive to do is the radios and fuel lines. You could lift the diesel generator of and run the machine from "wall" or battery pack and inverter. But price and energy density is just not there yet to be practical.
i'd also assume that you'd still need the motor's hp to produce enough power to overcome air resistance
Still 300 HP at 150km/h with 45% battery, 375 with 90+% one.
With 190-245 Nm of torque.
i did recently subject my Bolt EUV to a "worst case scenario" drive 🤣 it actually did ok
4 people, some cargo, full charge, blasting the A/C the entire time, highway speeds with minimal braking
and it needed to do the A/C cooldown 3x
But that is the point why some EVs have those dual-range gearboxes, to reset the curves at some suitable point after the dropoff at first gearing starts.
So 70-100km/h if that Model 3 for example.
If Tesla was to add such thing to make it get higher accelerations and speeds.
it said the trip efficiency was 2.8m/kwh, on a total driven distance of ~76 miles it lost ~86 miles of range
crap efficiency, but it was definitely the worst i've put it through so far
So with gear change at 100 km/h, you would then get another straight section until about 150-200km/h.
(not going to see if the length would be the same or double etc. so just estimation).
Which is why Porsche EVs etc. have them and normal ones don't.
it'll be interesting to see what the next-gen stuff looks like
i'm hoping that'll be motors in each wheel
Already done by multiple higher end EVs at top trims?
Not really worth it?
As you end up with way more expensive drivetrain.
Because two half-size motor-driveunits will cost almost double one larger one.
i feel like a lot of it is going to eventually get cheap enough, like how even a baseline car today has stuff that luxury cars didn't have 20 years ago
And only benefit is torque vectoring.
In better way than just automatic brake done one.
i wonder when electric motor braking is going to be allowed to be the only brakes in a car
Cannot be.
As that doesn't work when battery is full or speed is slow etc.
yea it'd have to be dumped to heat in some way then
in which case rotors would still be fine
good point
But back the the previous, as the current brake applying antislip is good enough for normal use to limits of the tires.
So having dual-motor torque vectoring instead isn't really needed outside of track use etc.
And two motors will always cost more than one, even when they would be identical (usually mirrors and not identical) to get efficiencies of scale.
We also had some talk with Dana representatives to do custom hubs with integrated electric motors and reduction drive, cost just too high for this prototype. So now its just built off-the shelf parts. Custom parts inside the hubs to remove the stock hydraulic drive and mount the bevel gearbox.
When the input materials are the same, but you then need more processing steps.
As in one big one using all the materials vs. two smaller ones each using half.
unless they came up with a way to make the internal rotors equally efficient in both directions, then they could just rotate the same motor 180° in the chassis for left/right
More about the casings.
Most already use the same internals.
But with mirrored casings.
Just casing and gear teeth orientations reversed.
Otherwise you have one motor before the drive shaft, one after.
Or one of the drive train units has to be upside down compared to the other.
that sounds very volkswagen 🤣
ok, stupid question time
for the electric motors that drive 2 wheels
how do they do "differential stuff"
Integrated into that reduction gearbox.
This specifically was such dual unit.
The farthest gear is the output one with differential showing inside the "hub".
The splines on the left one are for the electric motor, then some reduction to second shaft, then more reduction to the outside of the third, then that internal differential and output to top and bottom.
Same thing is done in almost all front-wheel drive cars, with differential being part of the gearbox final output stage.
Today, not in far history.
ah so it's more like a transaxle
Yes, just without multiple gear ratios.
got it
for some reason i was thinking they'd want to stick a half shaft on either end of the motor
You need that reduction gearing to get any usable max speed.
makes sense to do it that way tho, it's really just changing what's ahead of the gearbox
But as the electric motors generate so much torgue and have so high usable operating range, you only need the fixed ratio one.
Trading max torque for max speed.
that's still just weird to me, not needing to have a bunch of gear ratios
Because ICE engines are shit in generating power in wide range. Electric engines aren't.
Which is why diesel-electric trains etc. have been thing for so long.
Using fixed speed diesel engine to generate electricity, to drive electric motors.
i'm kinda surprised that Subaru hasn't been going gangbusters with electric cars considering their heavy industry stuff
But when you compare ICE engine torque curve to that electric engine one from before...
You don't get much anything before you have high enough RPM, and then after that it start to drop again rapidly.
And not many normal ICEs can get to 11k RPM at all without blowing up.
Where as that Tesla 3 electric motor got fixed torque from 0 to 5000 RPM.
You can tune those ICE curves to be wider, especially with Turbo, but still significant limits.
ngl i haven't even tried out my bolt's sport mode yet
EV vs ICE in theory.
And the EV maker can tune where the specific drivetrain moves from that constant torque to constant power.
By making max torque less, but moving the switchover point to higher speed.
And you don't need more torque than your tires can use 😁
Which is the limiting factor for almost any EVs on off-the-line acceleration.
Or at least easily could be, if the maker didn't tune the reduction ratio to be even lower.
To move that switch point to right and down.
just stick some of those ridiculously wide tires on it, np
But that is the reason the higher performance trims have the dual-back motors, to give better acceleration with better traction control.
And as the higher performance single motor would start to get too large.
And then the front motor(s) are usually optimized for higher speed cruising.
And the back motor(s) are disabled at cruise to save electricity.
i hope that big fat chargers & inverters start becoming standard in cars, fast charging is more important to me than super long range
Limited by what the battery can take usually.
right, but as far as taking up volume in the vehicle
Just sizing the charger to what the battery can take.
Not artificially limiting the charge rate.
i'd rather have 200 miles of range that can charge in 10 minutes instead of 500 miles that takes 4 hours
Unfortunately the possible charge rate only goes up with battery capacity if there isn't some big breakthrough.
Limit is charge current per battery cell.
And that max depends mostly on the cell capacity.
Somewhat on the form factor etc. causing different heat transfer characteristics.
So the 400V to 800V change for DC charge is way to make the current lower for same power.
But only halves the current in total, and needs double the cells in series, so for same capacity battery, half the cells in parallel, making the cell current still be the same.
But giving some benefits from the other parts of the per cell current limit.
When using exact same cells.
But by doubling the voltage, the power electronics and charge cables can get smaller for same current and/or power.
As we had been reaching point where 400V charging at max current needed insane sized cables.
Tesla was using water cooled ones with smaller conductors, almost everyone else was just using giant, very heavy and inflexible, copper cables.
For the charger side I mean, car side doesn't have water cooling even for Tesla, just thick enough cables.
Not completely sure about the following numbers for the C ratings available right now, but general idea is the same:
Currently it is very common for Li-Ion cells to be able to take 1C charge rate (1A per 1Ah of capacity, so in theory 1 hour to charge).
But that max rate cannot be reached nearer to full the cell is (which is why often car charge speed is quoted as to 80% full).
You can get higher charge rate cells, I would expect that 3C charge rate is commonly available (20 min to full in theory).
But the higher rate means more internal space needed for heat transfer etc., making the volumetric efficiency worse.
So for same size cell you get less Ah, so the actual max charge rate is less than 3x the 1C cell and you also lost max capacity for same volume or weight. And the per cell price also went up.
I don't think anyone has 5C charge rate cells in volume production.
And like previously said, that is only the speed when the charging voltage doesn't yet have to be limited to not permanently damage the cells from overvoltage, which limits the last 20% or so of charging to slowing charge rates.
Which can easily make it so that 0-80% takes as long as 80-100% takes.
Max discharge speeds are significantly higher.
So EV:s are using as high charge/discharge rate cells before the costs balloon and the volumetric efficiency goes to trash, and then sizing the car side charger to what the battery as whole can take, except some early models that had undersized car side electronics for the batteries.
But larger battery makes you charge faster for same range, until the extra battery weight starts to add up too much in power usage for same range from the extra weight causing more losses.
And just charging to that sweet-spot and not full makes the charging much faster for per range charging speed.
So 20 min fast charges for enough drive time/range for next human necessity (for normal long distance travel) pit-stop is already a thing.
So if the fast chargers are ubiquituous enough, you could already drive until you need to stop for toilet break, stop at suitable location, plug in, go to toiled and possibly grab coffee etc., get back to car, unplug, drive onwards until needing to repeat.
How about building power lines above the streets?
And to guarantee that the car will be connected for long enough, we could maybe add something like rails or so into the ground where the wheels go into.
And while we're at it, just make it way bigger
Oh shit I've invented the train again...
And trolley busses first.
Been thought as solution for long range electric trucks, but probably will not be taking off because of the infrastructure costs and the limits it would cause to oversized item hauling convoys.
Bruh just use a freight train at that point smh
It's a failure, there were way too many problems on the two test tracks and the third never was built
also can you imagine the cost to build the infrastructure needed across an entire continent? not going to happen
Point would have been to have those on the long distance highways, and the trucks would have switched to batteries for the lower speed roads.
But the costs to built the infra would never be acceptable today to governments.
Use trains for long distance, trucks only need to go short distances. Problem solved
Rebuilding the track infra would be even more costly than those overhead lines...
Thing is: moving the freight from train to truck takes time
That, and space.
Also needing more dual-track to be able to support enough freight trains.
ship → train → truck
ship → truck
You save one moving process
And often currently you cannot even do ship -> train, but need to do ship -> truck -> train.
maybe even two if you would've gone
storage → truck → train → truck → storage
Cuz not every piece of cargo is shipped via ship
Also needing an actually functioning rail system
-# LOOKING AT YOU DEUTSCHE BAHN!
Trains are the answer
Last mile truck. Otherwise trains lol
Needs a train network that can handle it
Also there's a distance/accessibility cutoff at which transferring to train at all isn't worth it
You can even do shipping containers so the whole cargo is picked up from the railcar and loaded on the truck
And that needs loading stations
Intermodal is a good solution most of the time
Which need place
Just less effecient from the rail side
But you need a functional rail system
I thought they could do it like a passenger train, stations scattered around where the train stops and containers are swapped and the train moves on to the next city or something like that
for freight trains to be worth it and still be operable in a workable train network, they need to be fairly large
which means separate stations from the passenger system
which means even more space required
bottom text lol
If this filament wasn't so fucking insanely good I would never ever print with it
it's horrible to print
guys, so my pc alr has a 500gb ssd. i decided to upgrade and bought an extra 1tb m.2 ssd
now, since basically my windows and my primary boot is on my 500gb ssd
idk what to do
i want everything to be on my 1tb ssd and use that 500gig as an extra (the 1tb is a 4.0 and the 500gb is 3.0)
you wanna do it how you want to or do you want a recommendation?
Any tech savy guy can help me block internet from Epic games so I can play in offline mode tomorrow?
Or just copy the game to another directory? No DRM.
Wait, the only reason Satisfactory open Epic launcher is because it's in the epics games file directory?
Might need to create shortcut with "-NoSteamClient" launch argument to the main exe.
And possibly change the contents of the "launcher_id" file "steam"
Seems that just deleting that file in the copy works without any shortcuts.
So make a copy of the game, then delete that launcher_id from the copy.
I just renamed the launcher_id and it work, thanks
- OS and most software are perfectly fine on 3.0
- just use the new drive as secondary for games n shit
- what are the models?