Unicomp SSK USB issues

User avatar
lhutton

23 Apr 2021, 04:49

Anyone else here with a Mini M having it die randomly or the Q stop working? I saw a couple of threads on Reddit about this and I've noticed my board has started doing it too. At first I didn't think much of it because I've had some issues before with various USB adapters and this board in the past but my other two Unicomp Ms that are USB native always seemed to work OK with it. My UltraClassic would sometimes freak out on boot if I used the ultra l337 g@m3r high polling rate USB plug on the back of the board but if I moved it off that port it would be fine. The 2020 M had no problems whatsoever with the turbo g@m3r l0lz USB port. I thought it might be that at first but the Mini M doest it on any port on the back of the machine. I'm currently testing by running the Mini M through a USB hub to see if that clears the issue up.

I'm running an ASRock X399 board and Fedora Linux for reference. I sent Unicomp and email tonight and am awaiting response to see. My guess is there's going to be a firmware update for these in the not to distant future at least I hope that's the only fix needed.

Dikkus

23 Apr 2021, 09:20

Can't say I've had the issue of Q dying, but if I wake up from suspend, then the keyboard needs to be replugged as the entire thing is asleep.

EDIT: Suppose I should mention my mobo and OS
X570 Aorus Elite
Arch Linux and Windows (both exhibit same issue)

User avatar
Muirium
µ

23 Apr 2021, 10:30

The weird shit they're pulling on the controller / USB port is the Unicomp SSK's red flag to me. If you could buy it with a regular single controller (2KRO like an original SSK is fine by me, if done right) and a USB C or even micro port, then I'd have one of these already. But everything I've heard about the multi-controller hack puts me right off.

User avatar
lhutton

23 Apr 2021, 15:30

I tried it on my work PC too running Debian and same issue. Seeing USB device enumeration errors in the dmesg on both machines. I've noticed that when it screws up it can sometimes take the whole USB bus with it on my main home machine. I've got two Dell monitors with built in USB 3 hubs and those stop working after the SSK crashes.

User avatar
lhutton

23 Apr 2021, 16:25

Muirium wrote:
23 Apr 2021, 10:30
The weird shit they're pulling on the controller / USB port is the Unicomp SSK's red flag to me. If you could buy it with a regular single controller (2KRO like an original SSK is fine by me, if done right) and a USB C or even micro port, then I'd have one of these already. But everything I've heard about the multi-controller hack puts me right off.
Can you elaborate on this? I haven't heard anything about the controllers other than they were a difficulty in the manufacturing process and caused some of the slowdown.

User avatar
Muirium
µ

23 Apr 2021, 23:55

Chyros covered it on his review. Unicomp hacked up the matrix to approximate marketing essential NKRO from what is fundamentally 2KRO membrane hardware. As I recall they also wanted to claim very high polling rate. They wound up needing two controllers to pull all this off, and, being Unicomp, they made a bit of an arse out of it.

Look at that USB cable socket, for instance. Why is it USB A? They’re breaking the standard and put the *host* end on the peripheral. Sure, USB C has rewired the way we think about these things lately, but use USB C then! And why the awkward grip? Did they think any of this through?

Welcome to your sleepy, USB port disabling typing experience. :roll:

User avatar
sharktastica

24 Apr 2021, 00:28

lhutton wrote:
23 Apr 2021, 04:49
Anyone else here with a Mini M having it die randomly or the Q stop working?
Hasn't happened to me personally, but a few on my subreddit and Discord have came across it. Unicomp has got back to some of those, believing it to be firmware issue and they're working on it.
Muirium wrote:
23 Apr 2021, 23:55
Chyros covered it on his review. Unicomp hacked up the matrix to approximate marketing essential NKRO from what is fundamentally 2KRO membrane hardware. As I recall they also wanted to claim very high polling rate. They wound up needing two controllers to pull all this off, and, being Unicomp, they made a bit of an arse out of it.
Basically, they increase the matrix so they could strategically 'place' keys in a way they're unlikely to block with each other. The new matrix is 12x16 for just 8x keys vs 8x16 that 1xx key Ms use. At the same time, they've written new firmware for supporting the numpad layer and better suppressing phantom key generation. The latter of which is probably at fault here.

By "two controllers", do you mean physically or an 'enumeration to the operating system' technique? This is what the only PCB in the Mini M looks like:
protomini_controller.jpg
protomini_controller.jpg (760.18 KiB) Viewed 4574 times
The comments about the USB I can agree with. The standards part makes sense of course, but I like the pressure-locking solution it has versus what seems to be manual-screwing solutions for Type-C at present.

User avatar
lhutton

24 Apr 2021, 01:36

sharktastica wrote:
24 Apr 2021, 00:28

Hasn't happened to me personally, but a few on my subreddit and Discord have came across it. Unicomp has got back to some of those, believing it to be firmware issue and they're working on it.
I found a few threads on Reddit too. Don at Unicomp got back to me today and said they're working on a fix and he will alert me when they have one. Luckily for me Unicomp is literally the next state over so shipping it back won't be a problem if they need to replace the controller or reflash it in the shop. Heck, I could probably drive it there if I wanted. I do wonder if they'll come up with some other solution for international people.

I'm surprised this bug isn't more prevalent TBH. Most computers use one of a handful of USB controllers and it seems to hit such a broad swath of hosts that everyone would have seen it by now.

At any rate my attempt to mitigate it with a USB hub didn't not succeed. I've also noticed that sometimes when the keyboard dies it takes out all the USB ports on the machine. My USB hub quits working (even though the keyboard was plugged straight into the computer when this happened) as does my Logitech trackball. Doesn't happen every time the keyboard goes kaput but enough to notice.
sharktastica wrote:
24 Apr 2021, 00:28
Basically, they increase the matrix so they could strategically 'place' keys in a way they're unlikely to block with each other. The new matrix is 12x16 for just 8x keys vs 8x16 that 1xx key Ms use. At the same time, they've written new firmware for supporting the numpad layer and better suppressing phantom key generation. The latter of which is probably at fault here.

By "two controllers", do you mean physically or an 'enumeration to the operating system' technique? This is what the only PCB in the Mini M looks like:
protomini_controller.jpg

The comments about the USB I can agree with. The standards part makes sense of course, but I like the pressure-locking solution it has versus what seems to be manual-screwing solutions for Type-C at present.
This is what I remember, the enhanced matrix but I don't recall anything about dual controllers anywhere. While I'm personally fine with the rollover situation as it existed in the Unicomp and Classic Ms gamers are a large market that pays a lot for their gear. I know some pretty big time esports/gaming folks and they won't touch a either a Unicomp or a Classic M because "bRuh iT's onLy 2KRo" even thought I can demo it and show there's no meaningful collisions. Unicomp would probably be foolish to R&D a new product and _not_ court gamers to some degree which is what I'm guessing the increased rollover is for.

I am seeing USB enumeration errors in my personal machine running Fedora and my work machine running Debian from it now that I'm looking for them. Unicomp has done right by me in the past so hopefully they'll come through with a fix soon. I don't mean them any ill will over this I was just looking for data, theories about what it was happening and maybe some mitigations strategies people had come up with.

As to the USB-A vs USB-C thing it doesn't really bother me. Lots of industrial and research gear uses an A port as a device port in my physics lab experience. I thought two the reasons they went for A was the locking mechanism and durability. C is rated for more insertion cycles but A is more study when getting yanked around as it were.

User avatar
Muirium
µ

24 Apr 2021, 11:00

sharktastica wrote:
24 Apr 2021, 00:28
By "two controllers", do you mean physically or an 'enumeration to the operating system' technique?
Your pic answers it nicely: clearly the latter! So hopefully it's something they can fix in firmware. Extra points if the firmware update can be performed by end users.

User avatar
lhutton

24 Apr 2021, 20:55

I swapped out for my old UltraClassic and the issues with my desktop subsided. The Mini M was not only crashing after sleep but was really messing with the USB bus on this machine. I thought my motherboard was dying at first but it seems to just really not like the Mini M's controller. It would seemingly sometimes cause all the USB hubs to drop off the bus and not allow the machine to sleep. Had to hard reboot to get it all to come back. I'll move over to the 2020 M later and see if that one is well behaved as I remember it.

My work machine just had the dead Mini M on wake from sleep issue and it didn't totally nuke the USB ports on the machine like my home workstation. I'm going to try the Mini M on my Dell WD19TB laptop dock and see if it's better behaved there.

micmil

25 Apr 2021, 05:38

You know what Unicomp is? Unicomp is pre-training Santa's Little Helper. Sure you love 'em and want them to do well but then they just stand there and let the frisbee doink them in the face.

I'm wondering if their new firmware has some sort of frippery they had to perform in order to keep it compatible with older hardware that some of their business customers may still be using. If that's the case I could vaguely understand why it may trip out on modern systems but... also, if that's the case... *doink*

User avatar
Muirium
µ

25 Apr 2021, 09:05

You guys want rollover? We got rollover.

Image

micmil

25 Apr 2021, 21:39

But also Unicomp is the one throwing the frisbee so it's more like
UNISANTA.gif
UNISANTA.gif (729.82 KiB) Viewed 4320 times

User avatar
wcass

24 Jul 2021, 01:30

I posted this in another thread. Sharing here too. Biggest risk of blocking/ghosting on columns C and H. Hold down any two numbers and it will block two keys on the QWERT row.
M Mini matrix.png
M Mini matrix.png (24 KiB) Viewed 3706 times

User avatar
wobbled

25 Jul 2021, 00:51

I've finally experienced the strange issue where the Q key dies. Disconnecting it from the keyboard end (A royal pain in the gooch due to that recessed tab) did fix the issue.
Now I'm having more of a hardware related issue with mine where the spacebar decides to unmount itself if I press down and release my finger quickly so I probably need to replace the spacebar which is wonderful. The fact a used k120 or a Dell with mashed potato for rubber domes from ebay that has suffered from the hands of a greasy gamer and pornography addict would outlast something from the descendants of IBM's keyboard division is a complete joke.

bonk

23 Aug 2021, 04:20

I opened a support ticket with Unicomp and they sent a prepaid shipping label to return my Mini M. I got in back in about 9 days and the Q key issue is history. According to them it needed a firmware update.

User avatar
raoulduke-esq

23 Aug 2021, 05:09

Replacing the controller is VERY worth it.

tiltowait

24 Aug 2021, 21:13

raoulduke-esq wrote:
23 Aug 2021, 05:09
Replacing the controller is VERY worth it.
If I were to buy one, this is probably the route I would go. I replaced the controller in my 1987 SSK, and it's like a whole new keyboard with all the customization I can do. I wish Unicomp had just gone with QMK from the get-go, but I guess I can sort of understand why they didn't. (Nah, scratch that. I really can't.)

Post Reply

Return to “Keyboards”