Sensorless - Printer halting when adjusting SGTHRS

[glass ice]

Hello
I'm building a v2.4 from a fysetc kit since last year!
Just git'ed my configs and I'm proud of it https://github.com/Apocallur/Voron2.4

This month I've returned to the project and I'm trying to set up sensorless x and y.
I'm following the Voron and klipper tutorials
https://docs.vorondesign.com/tuning/sensorless.html
and
https://www.klipper3d.org/TMC_Drivers.html#sensorless-homing

But stuck in the same part.
When I start decreasing SGTHRS value (starting with 255, then decreasing 1 or 2 per test) the carriage moves like 10mm and stops.
I'm getting these currents on my G28 x0 (Run Current: 0.70A Hold Current: 0.70A) OR (Run Current: 0.80A Hold Current: 0.80A)
This keeps happening while I decrease StallGuard Treshhold, always the ~10mm. But sometimes when I'ts on ~240, other times when is ~200, the printer halts and i got this error
TMC 'stepper_x' reports error: GSTAT: 00000004 uv_cp=1(Undervoltage!)

I'm working with tmc2209
I assume that the problem is that the voltage of my TMC's are quite low, and because of that the printer head is moving too little every test. Is this a thing?
Also, I'm scared to burn my TMC's by changing the voltage to a unsafe value. Any tips (or tutorials) on how to proceed?
Any help is appreciatted 🙃

[robust dawn]

what sgrths value are you on? voron's guide suggests moving in 50s until you a maximum value

[frozen tree]

first comment, working through the printer.cfg

homing speed is very low- it should be equal to or higher than the rotation distance

[robust dawn]

log? 😂

[frozen tree]

(log would be good i'm being naughty preseuming the git cfg is up to date)

diag pins appear correct for typical x/y motor positions on a manta m8p v2

you have autotune and appear to have the sg4_thrs values in the correct spots there which is where you should be adjusting them

[glass ice]

By log, its just what is showing on my console page?

[frozen tree]

so things to request/confirm/change:

klippy log to make sure cfg matches what your printer is actually getting

photo evidence that your diag jumpers are in place

(i rechecked homing speed for x/y and it's at 55 so ignore me on that front)

[glass ice]

I've reached 220 once (printer halted)
then started again from 255 to 240 when the printer halted again

[frozen tree]

oo it's a spider i was wrong

[frozen tree]

which spider version does it say on the end

[glass ice]

V3.0 H8

[robust dawn]

sounds like you're going the wrong way to me. I'd start at like 125 and go up until it doesn't move really after the retract

[frozen tree]

@surreal chasm sorry to disturb you, but if you're about, can you remind me of the terminology fysetc use for the spider 3.0 H8 on their wiki pages? just trying to clarify diag pins for this user

[glass ice]

Srry, Misremembered the name
V3 H7 (not H8

Followed this pinout

[frozen tree]

that would make more sense

(sorry james in advance i was hallucinating a revision)

so really we want the endstop pins that relate to the two motor drivers that are being used for X and Y (a/b)

which are actually written on the board

presuming you're using motors m0 and m1

[glass ice]

exactly

[frozen tree]

nope. error. those aren't the pins. i always fall for this.

[robust dawn]

X- PB14
Y-PB13

[frozen tree]

yes

which is both bad and good in that it matches your configuration already so doesn't change where we are

can we see a video of current homing

[glass ice]

just a sec

Tap is installed

[frozen tree]

have you still got autotune active?

driver_sgthrs is showing as 255 for x and y in that log but the autotune stuff from your github seems to have gone

with 255 being the active value the video behaviour is expected

[glass ice]

Where do I check if I have autotune? I dont remember seeing anything like that

[frozen tree]

don't worry about it. it's not active per your log so don't worry about it for now- it won't be a factor

so basically you just need to do the iterative process from the sensorless guide

[glass ice]

The same way I was doing? 255-- until it homes correctly?
Or should I try any variant?

[frozen tree]

255-240-220-200-180-160 or so would be my normal approach til you get close

hovering over the e-stop

then when you get close use finer adjustments

for context on my similar machine with 2209 sensorless i have around 165 on both x and y- they can be different values though

[glass ice]

understood
okay! gonna try it!

Thanks for all your imputs bomtarnes!

[frozen tree]

sorry if it's been muddly at times

[glass ice]

No problem, any light is a lighthouse to those on the pitchblack