mirror of
https://github.com/safijari/openpuck.git
synced 2026-11-04 06:34:45 +00:00
[GH-ISSUE #72] Intermittent Disconnects #36
Labels
No labels
pull-request
wontfix
No milestone
No project
No assignees
1 participant
Notifications
Due date
No due date set.
Dependencies
No dependencies set
Reference
safijari/openpuck#36
Loading…
Reference in a new issue
No description provided.
Delete branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Originally created by @slychoober1 on GitHub (Jun 20, 2026).
Original GitHub issue: https://github.com/safijari/openpuck/issues/72
Im having frequent disconnects/reconnects when the controller is just sitting idle on my desk or when playing games. The dongle and controller are within 3-4 feet of another with no obstructions.
I do not get this same behavior with the Official Puck. I thought it was a faulty board, but tested this out with a brand new Pro Micro.
Let me know what I can provide to help troubleshoot!
@safijari commented on GitHub (Jun 20, 2026):
Thank you for reporting. I've experienced this once myself but thought it was a fluke. I will test more and investigate.
@safijari commented on GitHub (Jun 23, 2026):
Can you see if things are any better with the latest code?
@MansoorTB commented on GitHub (Jun 23, 2026):
Same issue persists. Reflashed with all commands after putting board into dfu mode. Controller disconnects and reconnects immediately
@PS1 commented on GitHub (Jun 24, 2026):
I just flashed the dongle 5min ago I just got now and I confirm that I have those disconnections.
@stickman-dev commented on GitHub (Jun 25, 2026):
Same issue here. Regular disconnect/reconnects with Steam input. Seems stable for me in Xbox mode with no disconnects so far over about 2 hours of use. Not sure if that's just coincidence but I'll keep testing.
@Domali commented on GitHub (Jun 25, 2026):
I can get mine to pretty reliably disconnect/restart by rapidly moving the touchpad around. Oddly enough this only seems to happen when steam is running. I'm using the Teyleten Robot Pro Micro NRF52840 and when the controller disconnects the little red LED comes on(maybe the whole microcontroller is restarting?). I'll update if I have time to troubleshoot and figure anything out.
@joshbirnholz commented on GitHub (Jun 25, 2026):
Just got my board and flashed it, I'm also seeing pretty common disconnects, to the point where it's basically unusable.
@Domali commented on GitHub (Jun 25, 2026):
I also wanted to add that the official puck seems to make the mouse movement much smoother - I wonder if the board isn't keeping up with commands from the controller which is causing this issue.
@safijari commented on GitHub (Jun 25, 2026):
Pretty wide gamut of issues. Would like more detail on frequency.
I see disconnects some times but it takes almost an hour for them to show up. I don't see lots of inputs creating disconnects though. Will investigate more.
I wonder how much difference there is between different boards. These are all nicenano copycats afterall.
@safijari commented on GitHub (Jun 25, 2026):
@Domali steam sends back haptics that's probably why.
@safijari commented on GitHub (Jun 25, 2026):
Can you check the command rate in the WebUSB panel?
@joshbirnholz commented on GitHub (Jun 25, 2026):
Mine was disconnecting every minute or two. It even disconnected while sitting on my desk while writing my last comment, with no obstruction to the board. It connected again after a few seconds on its own. I did see RF Link go down on the web USB page when it was down.
It was bad enough that I checked GitHub issues to see if anyone else was having the issue, but actually since writing my last comment a few minutes ago, it hasn't disconnected since then.
Currently seeing this status:
@Domali commented on GitHub (Jun 25, 2026):
The delivered/new reports is around 250-265 sometimes spiking to 280. It does look like its a watchdog+soft reboot. When its just sitting there doing nothing its around 237/s.
@MansoorTB commented on GitHub (Jun 25, 2026):
I was going to submit this same finding. Have been testing back and forth between xbox mode and Steam mode. Xbox mode has never disconnected. But...I also thought it may just be a coincidence
@joshbirnholz commented on GitHub (Jun 25, 2026):
Captured a video of some stats while it was dropping out:
https://github.com/user-attachments/assets/8a7fc521-9b23-41a4-8e7d-2bb8ec9cf513
I unplugged it and plugged it back in and now it seems stable again.
@Veri7ion commented on GitHub (Jun 26, 2026):
I have the same problem and created issue https://github.com/safijari/openpuck/issues/27
I thought it was related to my controller adapter so I closed the issue, but it might not be after all.
AYWHP Pro Micro NRF52840 here.
@stickman-dev commented on GitHub (Jun 27, 2026):
Had a little sesh last night using the x-input mode but it did eventually become unstable as well. Disconnects start happening much quicker in steam input mode though.
@safijari commented on GitHub (Jun 27, 2026):
Please try https://github.com/safijari/openpuck/pull/89
Kids have been playing Mario Kart with 2 steam controllers and one ReversePuck'd Steam Deck for a half hour with no issues.
Note that the compile command has changed.
Edit: 10 minutes later they had a disconnect but I'm unsure if that was caused by something else. Continuing to "test" as they play another round 😂
@Domali commented on GitHub (Jun 28, 2026):
After fighting with flashing issues(my windows machine doesn't like flashing - had to use MacOS to get it to work) I was able to get the stability-fix branch installed. I'm still able to crash the board pretty quickly by spamming usage of the right track pad.
@safijari commented on GitHub (Jun 28, 2026):
Thank you for testing. How hard are you spamming? I've gone ham on the trackpads of two contollers together for a while but couldn't get it to crash now. Does this happen if Steam is closed?
Edit: I just did another test on your branch just going to town on both trackpads for a solid minute without issue.
@joshbirnholz commented on GitHub (Jun 28, 2026):
I'm still getting frequent disconnects as well. For me, it happens even when just using the trackpads normally to control my mouse and navigate my desktop. I tried playing some Hades II, and it disconnected often -- the game was unplayable. At one point, the game even through up a "We couldn't detect your controller" dialog telling me to try disabling Steam Input. I wasn't using the trackpads at all and I myself haven't really noticed trackpad usage to be linked to when the disconnects happen.
At least it seems to be catching it correctly, the config page shows:
It also just disconnected while sitting on my desk untouched, multiple times while I was typing this comment.
I've seen it disconnect both with and without Steam open while in Steam controller mode. I played some Star Fox in Switch Pro mode while connected to my Switch 2, and didn't see any disconnects the entire time. I also tried playing more Hades II on PC using the PS5 mode and had no disconnects either.
@safijari commented on GitHub (Jun 28, 2026):
This is so bizarre. I'm on the latest commit in the PR and just played in Steam Controller mode for a solid 2 hours with no stability issues.
So,
Please leave a link to the board you bought.
Instead of -DCFG_TUD_TASK_QUEUE_SZ=64 try with 128
@joshbirnholz commented on GitHub (Jun 28, 2026):
This is the board I bought: https://www.amazon.com/dp/B0DCZJKYL1?ref=ppx_yo2ov_dt_b_fed_asin_title&th=1
Will try with 128 and report back!
@safijari commented on GitHub (Jun 28, 2026):
Ok.
I'm ordering both AITRIP and Teletyn pro micros to see if I can repro.
@joshbirnholz commented on GitHub (Jun 28, 2026):
Tried with 128... still getting disconnects. It does seem more infrequent, though it's hard to say, since the timing of the disconnects I was seeing was never really consistent.
EDIT: Tried some more in PS5 mode as well, it doesn't seem to actually disconnect, but while playing Hades my character will sometimes keep running in a direction after I let go of the analog stick.
@stickman-dev commented on GitHub (Jun 28, 2026):
I think you could be right about right about board differences. After my disconnects I get last reset reason as "watchdog + soft reboot". It's happening whether I am using the controller or not. I've tried #89 with no noticeable difference. Sometimes it will take 30 minutes to show itself, sometimes much quicker. Once it starts happening it seems to continue happening with a higher frequency. Instability takes longer to show in x-input mode.
I tested main with OPK_LOG=1. I reproduced it with the controller physically next to the board and without me giving any input to the controller. Debug dump shows RECONNECT (block+reinit armed) / RF LINK UP, followed by heavy repeated rid=79 traffic. Copy of part of the dump is here if useful to you: https://pastebin.com/P9dt3i6Y
I'm in the UK and my options locally for NRF52840 boards seems a little bit more limited. Here's my board: https://www.amazon.co.uk/dp/B0G4L4P4B6?ref=ppx_yo2ov_dt_b_fed_asin_title
Board enumerates itself in DFU mode as "VID:PID 239A:00B3, USB Serial Device (COM6)"
I'm keen to try the specific board you're using if could possibly provide a link. Will be nice if the solution is as simple as just getting a better board. Really appreciate your work on this. It will be a fantastic solution to the Steam Controller's short comings.
@MansoorTB commented on GitHub (Jun 28, 2026):
Getting disconnects on the latest commit with both compile edits.
This was the board variant i bought from amazon..
Brand DORHEA (https://www.amazon.com/dp/B0GSZ4HCH4?ref=ppx_yo2ov_dt_b_fed_asin_title)
Included Components nRF52840
Number of Items 1
UPC 687117794147
@safijari commented on GitHub (Jun 28, 2026):
Latest commit on main or on the stability fix PR?
@MansoorTB commented on GitHub (Jun 28, 2026):
Both actually. Main and Stability Fix PR #89 (
github.com/safijari/openpuck@fc1258dd2d).I also noticed that the trackpads dont seem as responsive / smooth / as much detail in movement with the pull request #89
@safijari commented on GitHub (Jun 28, 2026):
I had bought this board
https://a.co/d/0i4qAxpB
Which it seems is also from DORHEA so that's more perplexing.
I've ordered two other distributor's boards that also physically look different and will try to reproduce these issues.
@MansoorTB commented on GitHub (Jun 28, 2026):
I have been reflashing over 2 different boards out of the 4 that came in my order. My process has been to jump ground and reset putting the board into usb mode and then running the commands.
Doing this way, the controller never loses pairing. I think i should probably nuke the board into a blank slate but i'm not sure how to do that.
@Oneseat1 commented on GitHub (Jun 28, 2026):
I am having this issue. It's like every 10 min or so. I will say I had a similar issue with my oem puck right when I first got it I'm going to plug it into an internal 2.0 USB and see if there is a difference and switch back and forth between my oem puck and my open puck try to stress it
@Oneseat1 commented on GitHub (Jun 29, 2026):
Ok I know this is odd but I have it connected to my 19/20 pin USB 3gen 1 connect via a USB c to usb c wire and it seems to be more stable than ever before. Not sure what that means but I'll keep stressing it on my bazzite machine
on't know what that means but
@safijari commented on GitHub (Jun 30, 2026):
Alright I got a new branch for y'all, please test this
https://github.com/safijari/openpuck/pull/94
I have the alternate boards arrive today as well and will test two controllers with each of them tomorrow. 🤞
@joshbirnholz commented on GitHub (Jun 30, 2026):
Tried building from that branch and I'm getting disconnects just like before. Actually, the connection seems less stable, since now it will sometimes stay "connected" but drop inputs much more than it did before.
@stickman-dev commented on GitHub (Jun 30, 2026):
Just tested #94 on my board too. It's the most problematic version I've tried so far. It starts disconnecting within 10 minutes and then more frequently after that. Plus there's a new thing where the controller haptics does random little pulses and occasionally longer pulses which I stop by pressing the steam button.
I've got a Supermini board coming to see if that improves things over my current unbranded board. Judging by the photo, it has a different crystal on the board so might make a difference..? Worth a punt anyway.
@safijari commented on GitHub (Jun 30, 2026):
https://github.com/safijari/openpuck/discussions/26
The buzzes are a known prior issue.
@stickman-dev commented on GitHub (Jun 30, 2026):
I've experienced the occasional long buzzes with previous versions, but this is new for me. #94 is giving multiple short split second buzzes, pretty much from the moment the controller is connects. Doesn't seem to be related to whether I'm touching the controller or not.
@xamindar commented on GitHub (Jun 30, 2026):
I just set this up last night and noticed these disconnects. Shame, it's a cool idea. I'll keep messing around with it. I got 3 boards from amazon to try.
@Doze-Zoze commented on GitHub (Jun 30, 2026):
I have noticed disconnects myself when trying to use Switch mode on a switch consistently within 15 seconds of connection. If I use the Hori mode instead, it seems to work flawlessly with no disconnects so far, although my playtime has not been extensive with it.
@safijari commented on GitHub (Jul 1, 2026):
Alright alright, lemme cross all of my fingers here first.
https://github.com/safijari/openpuck/pull/94
I have yet to have a disconnect/crash/hang on the latest commit for PR 94. I had two controllers connected to an AITRIP board (which in general was much more unstable than my DORHEA one) and they stayed stable for 40 minutes before I ended the test (there's a stability test now in the webusb panel that will keep time and if the board crashes it'll keep a log of the reasons until refreshed). If testing this, please ensure to build with
make buildto ensure parity between your build params and mine.Some additional niceties here are a better (I think) laid out webusb panel, the ability to export and import configs (which can allow two pucks to share pairing info for the same pairing slot), and general multi controller fixes.
@Veri7ion commented on GitHub (Jul 1, 2026):
I tried #94 but unfortunately still experience the disconnects.
used
make build:Tested in Steam mode.
On some of the resets the haptics also got very loud and stuck.
openpuck-hanglog.csv
The logs at the very bottom of the config page:
The board I use:
https://www.amazon.nl/dp/B0FKH4M63S
@intentionwindsurfer commented on GitHub (Jul 1, 2026):
It seems there are some differences in boards. I went down this rabbit hole: https://github.com/joric/nrfmicro/wiki/Alternatives
@safijari commented on GitHub (Jul 1, 2026):
Thank you for the link. I'll go through it and make a list of recommendations based on it and my other testing.
@stickman-dev commented on GitHub (Jul 1, 2026):
Same issues with the latest #94 on both my boards. Clearly too many boards out there of varying quality. I've got an actual nice!nano v2 board coming to me from overseas so hopefully that'll eliminate any random board issues I'm having. I'll try testing again once I've the new board.
@xamindar commented on GitHub (Jul 1, 2026):
So it looks like some brands are just crap? I assume I should just send these back instead of hoping for a software resolution? And yes, a recommended board would be highly appreciated.
That said, any idea if the Seeed Studio XIAO nRF52840 would work? That one fits a size contraint I want to use.
@safijari commented on GitHub (Jul 1, 2026):
I have put a warning on the readme for now. I recommend folks that are having disconnect issues even on the final commit from here https://github.com/safijari/openpuck/pull/94 please send back your boards if still possible. That commit has been tested to be rock solid (for at least a half hour, with two controllers connected in steam mode) on DORHEA brand and AITRIP brand but NOT on Teletyn brand.
I am going to merge this today and then work on trying to improve Teletyn.
As for Xiao, my order got cancelled. I need to order again.
@xamindar commented on GitHub (Jul 1, 2026):
My boards are the Teyleten variants. I'll go ahead and return them to Amazon. Probably many bought these because they are on there and marked as "Amazon Choice" lol.... One of the reviews on Amazon does list all the things they found wrong with them. Basically avoid.
They are so unstable it makes them completely unusable. Even worse when using the reversepuck thingy on steam deck. In that case they disconnect constantly and eventually stop working all together.
This is an awesome project and thank you for all the effort you are putting into it, but I wouldn't bother wasting time on these defective variants.
@intentionwindsurfer commented on GitHub (Jul 2, 2026):
Just confirming that the Teyleten boards I bought from Amazon are also disconnecting with the newest build. I was playing in xbox 360 mode and about a half hour in, the controller disconnected and reconnected to Steam mode. The disconnects became frequent thereafter. I'm going to return them to Amazon.
I took a photo of the board to help ID the chipset.
@mnul commented on GitHub (Jul 2, 2026):
How can I know what variant I have? I do have a lot of disconnects, yet mine is from a different "brand". (from Amazon Germany). Looks exactly like the photo above.
Amazon-Link
@safijari commented on GitHub (Jul 2, 2026):
Teyleten folks can you please see if the disconnects exist on this commit?
https://github.com/safijari/openpuck/commits/main/
@MaximusVX commented on GitHub (Jul 2, 2026):
Hi, which commit? The latest commit I'm seeing is a typo fix
@safijari commented on GitHub (Jul 3, 2026):
Coming up on 30 minutes with my least stable Teyleten board (disconnects within 10 minutes typically, can disconnect in under 1 by going ham with the trackpads which I did a few times during this 30 minute period) with the latest commit on #102 !
Going to test that build on an AITRIP board and if I don't get any hangs I'll merge it.
Edit: AITRIP board also hit a half hour. I'm feeling good about this one.
@safijari commented on GitHub (Jul 3, 2026):
Merged into main. Please test and confirm.
@skullbone commented on GitHub (Jul 3, 2026):
Hi, I have the teyleten boards and so far so good. I got 32 minutes, but I had to go do something, so I ended it. But now I'm currently 25 minutes into a new test, and I'm going to see if I can get an 60+
Update: It has cleared 75 minutes
@safijari commented on GitHub (Jul 3, 2026):
@safijari commented on GitHub (Jul 3, 2026):
I've got some other independent reports of stability with the latest build so I'm going to close this issue for now. If Someone runs into a stability issue with a board please open a new issue and leave info on which build and which board.
@MansoorTB commented on GitHub (Jul 4, 2026):
Tested the board for 2 hours straight with 1 disconnect after about 70 minutes. Disconnect and reconnect was instantaneous.
Played on steam input mode. About 1-2 feet away from dongle. Using the same DORHEA boards I posted earlier.
Thank you folks!
@safijari commented on GitHub (Jul 4, 2026):
It would be good to know if this was a hang or a different interruption. The WebUSB panel lists the last reboot reason.
Was this on 0.9 1 or 0 9 6?
@Domali commented on GitHub (Jul 4, 2026):
Thanks a lot for the updates! Just wanted to stop by here and say I can no longer crash the board by spamming the trackpad 🕺 I'll use it later today for a longer gaming session and see if it crashes at all. The mouse on windows still feels not quite like the OG steam puck(as smooth) but its not too bad.
I was trying to go through the PR to see if I could understand the fix but there are soooo many changes so I was just curious what the issue ended up being? Also, was curious if you had a tip jar so I could help support some of the work. I really appreciate it!
@safijari commented on GitHub (Jul 4, 2026):
It was some deadlock creating problems. The PR also addressed a few other long standing bugs like the buzzing that could happen some times and introduced a fair bit of diagnostic tooling so there's a bunch in there.
I'll look at the mouse smoothness, I think there's a bit of polling rate loss that's probably due to the code just waiting around some times when it doesn't need to which might be creating this problem.
As for tips, just donate some money in the name of the project to a charity of your choosing.
@xamindar commented on GitHub (Jul 16, 2026):
I'm still seeing disconnects with this on the latest version of openpuck which makes it unusable. I can't be the only one, but it happens at least every half hour and in some games (Palworld for example), it never regains controller functionality back once this happens until the entire game is closed and re-opened.
Are we sure this disconnect issue has been resolved?
@safijari commented on GitHub (Jul 16, 2026):
I have personally not experienced it on any of the boards that I have. Longest session I've done was about 90 minutes
@xamindar commented on GitHub (Jul 17, 2026):
Well, in any case, they disconnect on the three Teyleten boards I bought so I will just send them back to Amazon. I bought another brand and those do not disconnect at all in my testing.
When the Teyleten disconnect, most times it's short enough steam doesn't even notify that it happened. But some of the games I test it in will lose controller and not recover unless I completely close the game (steam input still works).
@safijari commented on GitHub (Jul 17, 2026):
I see. I will leave a warning about Teyleten boards. Thank you.
@nintendogamer15 commented on GitHub (Jul 18, 2026):
Not sure if this is relevant or helpful, but I was getting some random missed inputs/disconnects with my teyleten board and after plugging it into a usb 2 port instead of usb 3 those issues seemingly just went away. I'm using the current release 0.9.31.