#✅ - MKE Mini -failing to open mask Walsh3D mk38

120 messages · Page 1 of 1 (latest)

primal path
#

I recently got the mke mini and it cant seem to open the mask of the mk 38 walsh3d ..seems to struggle/only partially move, but will close fine (but does seem to exert a lot of force on a close).

If manually opened and set to close it closes pretty forcefully so it almst feels like its doing the open/close in reverse.

Tried swapping to different motors, tried swappin motor 1+2 on board...no luck.
Tried swapping and redoing all the helper arms bit they move smoothly, and easlit lift if I am just pushing the faceplate lightly up to position.

Slighlty going crazy over it

Seemed to work fine on bech under no load, powered via a 5v bank through penelope board.

ancient cairnBOT
#

Hey @primal path, thank you for contributing to this forum.
Please provide as much information (images, video, description) as you can. This will increase the chance that someone can help you with your issue.
Use the Close button when your issue is resolved.

primal path
#

Ill send pics of the wiring +assembly later today when I can

dark locust
#

Can you post a short video?

It sounds like you might have your servo cables reversed. By design the servos open at a fast speed to provide enough "oomph" but closes slowly so the faceplate doesn't slam down and potentially damage the paint.

Do the LED eyes turn on when the faceplate is closed? If not but they turn on when the faceplate opens then it is very likely your servo cables are reversed.

Other issues that might cause the faceplate to not open properly:

  • Servo arms or helper arms binding because the screws are too tight or misaligned
  • Faceplate is too heavy. Typical print settings are 2 or 3 walls with 5% infill
  • Power to the board, some USB cables are not capable of handling the amount of power required
primal path
#

I swapped motor cables around. One config was very wrong, failed to open and close at all.
Current config has the home and led lights on at correct spot.
Faceplate is 3 walls, 10%infill, doesnt seem like much weight diff.
I adjusted servo arms several times, reprinted a set just in case. Sanded down the fittings for smooth rotation along the mount points as well. (Not the geared fittings)
Power....that might be something I can try. Using a mini usb to usb cable that I salvaged from an old wireless controller. Figured it should be a power only, but might be possible its limiting.

#

Ill get a vid up later when Im off work

oak iron
primal path
#

5v 2.4A

oak iron
#

Do you have a wall plug with a USB output? 2amps or higher? if the issue still persists then battery bank isn't the issue. I have a kitchen plug with USB outputs at 5v 4amp which is good for testing

#

but power cable is crucial. I have an old playstation controller charging cable

primal path
#

I can check....wondering if my cable is a problem. Its an old wireless xbox.controller charging cable...but wondering its also data transmitting. Going to try finding another one to try as well

#

Ill have some pics+ vids in a little while

oak iron
#

Also confirm your servos are true 180 degrees. Manually spin then to confirm a “stop” point in either direction

primal path
#

Tried a new cable, pretty much the same

#

Did try a 5v 3A supply as well

#

But didnotoce that the servos do the fun wierd bit where they zero out in different directions first

primal path
primal path
#

both types giving the same issue, so not sure its the case, esp since it does seem to rotate to the correct spot

oak iron
primal path
#

😓

oak iron
#

For your servo arms, did you use the file with built in gears or the slotted arm that utilizes the horn included with your servo

primal path
#

Had to use the file with built gears..walsh didnt have the horn variant

oak iron
#

Ah ok. Personally I prefer to use the slotted arm as the gears from the factory appear to he better quality then what my printer could achieve

#

Is this MK39?

primal path
#

I know, wish he had the variant

#

Yup

#

Its def part of the issue

#

with each failure its eating those gear slots

oak iron
#

That was my first helmet and I used the slotted arms. Or maybe I drilled out the geared arm and glued in the horn. Ill check this evening

primal path
#

Would be awesome ifi t existed...wasnt in the file list I got

#

Might try my other motors (different brand..but seem to go slightly past 180)..they didnt really work prior, and seem to be slightly out of sync but maybe with new wires and such...dunno

#

otherwise guess I can try that source on ebay I saw floating around in other discussions

primal path
#

I think its the stupid servo arms lol

#

reprinted a set 1% bigger to see if it would more cleanly fit over the motor gear...and now magically works (well mostly, still a little clunky but its opening for once)

#

Def..if you have the file or know where I can find one for this helm that uses the horns vs the printed gear that woudl prob make this so much better

#

Doesnt stay up when worn though...guessing my alignment is off

#

Stays up if helmet is tilted back, but when in a normal angle, servos dont keep it held up or dont make it all the way up

dark locust
#

By design, the servos shut off when they reach their destination for safety. The servos can heat up and also drain the battery.

Several people here have solved the problem in various ways:

  • Adjust the tension of the helper arms
  • Attach a piece of elastic between the faceplate and the dome
  • Attach a magnet

Search the discord, I'm sure there's other methods ppl have used. I prefer the right amount of tension.

primal path
#

Its still not fully making its way open in that position, so weight/alignment I think may still be off. But I can add that in to the troubleshooting

primal path
#

Hopefully last question: Could you elaborate on the elastic idea? I searched the discord a bit, but only reference to a working one was a samurai-iron man model that wasnt up anymore...
Wouldnt the elastic put further strain against the servo + helper arms? Where would it even attach to in order to not bind/change angle/or get caught in the face when wearing?

oak iron
primal path
#

Not a bad plan...I may eventually just rebuild the model entirely and make ones that have the horn indent for future use, but for now I got it to the stage that it can open while laying flat by upscaling the servo arms by 1% and the helper arms by 2%. Had to sand out the holes a but on the helper arms as well as they were slightly catching.
So it opens pretty consistently, but now on the seemingly common issue of keeping it open when its vertical/worn.

Tried some magnets, but the not great mounting locations to get a solid face to face contact. - Tried at dome but contact point was small, and on the sides with the helper arms would either slip off or drag the arms out and collide.

oak iron
primal path
#

Nope, swapped to what *should * be 180s, but they do go a bit beyond 180 before a hard stop

#

seems to work save the general keeping up issue

primal path
#

Though just got a "tower pro" set that was recommended by others in several channels....and they rotate 360, are analog and have plastic gears..so not sure if the seller is scamming or I am missing something here lol

oak iron
primal path
primal path
#

Caved...going to try some adafruit ones.
Current ones that had "worked" started to only really work partially, and I think their size in relation to the arms were causing some binding...on top of some potential torque issues.

Tried the knockoffs for fun...those were absolutely terrible 🤣

#

Any idea on how many amps the mini is supposed to allow to the motors? I tested using a bench supply out of curiosity as well and only saw like a 1.5 amp draw on motor use...which seems crazy low

dark locust
#

You'll need at least 2 amps.

primal path
#

I know, it didnt seem to draw that though

#

so Im wondering the board is limiting or its just crappy motors lol

dark locust
#

It will draw more amps when the faceplate is attached and trying to lift it. The board is designed to allow up to 4 amps (but we don't recommend it)

primal path
#

The faceplate was in place

#

This was when I was running it fully assembled, but ran the power lines through a benchtop power supply instead that was set to 5v

#

it read a 1.5 amp draw during the operation (and failed to fully lift)

dark locust
#

Maybe @nocturne sage can do a better job of explaining the current draw on the board?

#

Are you powering the board through the power input or are you using the USB port on the ESP32-C3 module?

primal path
#

Bench to power input

#

Havent touched the usb

#

for bench test I skipped penelope

dark locust
#

Are you using the power cable that came with the kit?

primal path
#

yup

#

fed the red/black lines to a powered breadboard

#

then to the power input slot on the mini

dark locust
#

Ok, that all checks out. When you measured the current draw was that at the servo?

primal path
#

it was overall at supply readout

#

I could try probing the servos....but not sure how to get accurate if Im only tapping one line/servo at a time

#

when I gave the signal to open under load it flashed around 1.5 amps as the motors tried to lift, then fell to zero as they shut off and motors closed/mask collapsed

dark locust
#

Do the servos work outside of being mounted in the helmet?

primal path
#

They work, but draw .5 amp under no load

#

well "work" Im still not sure my motors arent just crap

#

No load run. Still seems low even for no load. Under load it hit 1.5.

#

and these are the motors that have had the at least partial working success

dark locust
#

Yeah... IDK, what do your servos look like?

primal path
#

Bit messy now, since Ive been changing em around to test others

#

they are at minimal digital 180 (though...might be 220) servos

#

but not name brand towers

#

the other two off brands ones Ive tried are far worse. (Both 360s, one an analog with some plastic servos)

#

could check the draw on those thoug

#

the digital 360s draw .3 under zero load...

#

so possibly explains why they also failed insanely hard..on top of being 360s

#

but hence my curiosity if the mini is limiting current draw in some way. I cant check the code apparently on a mini so Im just guessing at that, but feels like the motors just dont give the torque they should

dark locust
#

Yeah, knock-offs just don't seem to work.
Say you use 2 MG90S servos.

  • Peak current (worst case, both stall): ~2 A.
  • Supply should handle at least 2–2.5 A @ 5V.
primal path
#

Well I orderd the ada fruit mg90d s that people are saying may work..if those fail, not sure what to do next

dark locust
#

Out of curiosity... what did you set the amps to on your desktop PSU? I know on mine I can adjust it.

primal path
#

Its not set? Its set to 5v supply, the amps display reads out the draw at any time

#

might need to tweak that under a consant load if so

#

though think I had the dial set to max allowed

#

was a touch under, adjusted dial to max, .6amp under no load on "open"

dark locust
#

To adjust my amps on my PSU, I have to power off my PSU, connect ground & power cables to each other, power on the PSU and then adjust.

primal path
#

yea, it was allowing 10 amps if needed

#

okay...because I could, I rewrote an nano from a older project with the motor code and rewired that to a bread board

#

crappy motors still draw about .3-.4 under no load...sooooooo it may just straigt up be crappy AF motors that dont actually draw the correct current

dark locust
#

At least you've exhausted all the other possibilities.

primal path
#

welp...will see with the new motors hopefully

#

either that or I overwite the nano code's speed and see just how far I can push the crappy motors till they burn 🙃

primal path
#

Ignore the sad led eye falling out

dark locust
primal path
#

Only took 4 tries...guess imjust going to stick with ada fruit. Amazon/ebay motors just way too uncertain

dark locust
#

ya... sometimes it's hard to find good servos. we're always open to ideas on other servos but so far the MG90s are the best for helmets.

primal path
#

Part of me would want to run a drive cable from helm through a channel under the skull capand have the face plate be aligned via a track on the sides instead. Would however mean the motor and such would have to run to a suit/back panel though

dark locust
#

That's ok for a full suit. The majority of makers just make the helmet. A lot of full suit ppl like the option of being able to take the helmet off but still have the suit operational.

#

Bottom line... there's no single way to build a helmet or suit and everyone likes to customize in their own way. Hopefully we've built a board and code that facilitates the infinite possibilities.

primal path
#

Yup! Appreciate the help + sticking with the crazy amount of back and forth I was having lol

ancient cairnBOT