Blue Alps board only recognizing every other key

[zooksman]

I recently found a Datatech SPK-100, which is a relatively unknown blue alps board. It uses a standard 5-pin DIN connector, and the soarer's converter picks it up as XT, which makes sense since there is no AT/XT switch. It does have lock lights though (maybe controlled by the board itself?). Anyway, the keyboard came somewhat dirty, so I cleaned all the keycaps and shell and it looks good as new. The switches feel great, which would make one think that the board would work. However, I'm encountering some weird issues.

When I plugged the board into PS/2 (through an AT adapter), the lock lights and lock keys functioned, but my PC wasn't recognizing any keypresses. So I pulled out my Soarer's and tried that instead. I got the same behavior. In order to troubleshoot it, I booted up the hid_listen program linked in the Soarer's docs. It picks up keypresses as read codes (eg. rA6), but they have no HID output, and the computer sees nothing. However, every OTHER key in any given row just throws read errors (R05 R06). Some keys on the right of the board output tons of those error codes per press, and one key outputs five different read codes. I cleaned up the connector pins and dusted out the switchplate to no avail. This thing just seems to be broken in a very specific way.

I know it's a longshot, but has anyone seen this kind of weird behavior before? Maybe one of the traces on the PCB is dead? I really have no idea how to diagnose this, and I would really like for this beautiful board to work . If anyone has any advice as to how I could go about fixing this, I would greatly appreciate it. Otherwise, I'll just have to sell it as a broken board...

[Polecat]

I recently found a Datatech SPK-100, which is a relatively unknown blue alps board. It uses a standard 5-pin DIN connector, and the soarer's converter picks it up as XT, which makes sense since there is no AT/XT switch. It does have lock lights though (maybe controlled by the board itself?). Anyway, the keyboard came somewhat dirty, so I cleaned all the keycaps and shell and it looks good as new. The switches feel great, which would make one think that the board would work. However, I'm encountering some weird issues.

When I plugged the board into PS/2 (through an AT adapter), the lock lights and lock keys functioned, but my PC wasn't recognizing any keypresses. So I pulled out my Soarer's and tried that instead. I got the same behavior. In order to troubleshoot it, I booted up the hid_listen program linked in the Soarer's docs. It picks up keypresses as read codes (eg. rA6), but they have no HID output, and the computer sees nothing. However, every OTHER key in any given row just throws read errors (R05 R06). Some keys on the right of the board output tons of those error codes per press, and one key outputs five different read codes. I cleaned up the connector pins and dusted out the switchplate to no avail. This thing just seems to be broken in a very specific way.

I know it's a longshot, but has anyone seen this kind of weird behavior before? Maybe one of the traces on the PCB is dead? I really have no idea how to diagnose this, and I would really like for this beautiful board to work . If anyone has any advice as to how I could go about fixing this, I would greatly appreciate it. Otherwise, I'll just have to sell it as a broken board...

Not sure this is any help, but here goes. This DT post:

keyboards-f2/is-this-old-datatech-spk-1 ... t8772.html

and its photo link:

https://imgur.com/a/gyg3P

show an SPK-100. The SysRq on the PrintScreen key and the extended layout make it by definition an AT-compatible board. If there's no switch it's probably auto-detect.

I have an SPK-101 here, same layout, but white Alps and not as pretty of a case. The SPK-101 has a knockout on the back for an XT-AT switch, but it's not punched out on mine, and there's no switch behind it on the inside, nor any jumpers on the bottom of the PC board where the switch would be. Mine works normally on a PS/2 connection with a passive adapter. I don't have an XT setup available at the moment to check for auto-detect.

Datatech/DTK was a huge supplier of IBM clone boards and parts, and their stuff was if anything *more* compatible than IBM. By that I mean they sold motherboards and cards made to be used by hobbyists building clone computers, so they had to work with everybody else's hardware. For instance, if you had a motherboard that wouldn't recognize a particular card the first thing was to drop in a DTK (or Phoenix) BIOS chip. They made good stuff.

I guess one question is what happens when you plug an auto-detect keyboard into an auto-detect controller? But regardless it should work correctly on a PS/2 motherboard.

First thing I'd try would be to replace any electrolytic capacitors on the PC board. Those can dry out over time, and cause strange problems like you're experiencing. Electrolytic caps have an aluminum case, usually with a colored plastic sleeve covering it, with microfarad (uf) and voltage ratings. They're polarized, with + or - marked with a stripe on the case, very important to install the new ones the same way.

[zooksman]

I recently found a Datatech SPK-100, which is a relatively unknown blue alps board. It uses a standard 5-pin DIN connector, and the soarer's converter picks it up as XT, which makes sense since there is no AT/XT switch. It does have lock lights though (maybe controlled by the board itself?). Anyway, the keyboard came somewhat dirty, so I cleaned all the keycaps and shell and it looks good as new. The switches feel great, which would make one think that the board would work. However, I'm encountering some weird issues.

When I plugged the board into PS/2 (through an AT adapter), the lock lights and lock keys functioned, but my PC wasn't recognizing any keypresses. So I pulled out my Soarer's and tried that instead. I got the same behavior. In order to troubleshoot it, I booted up the hid_listen program linked in the Soarer's docs. It picks up keypresses as read codes (eg. rA6), but they have no HID output, and the computer sees nothing. However, every OTHER key in any given row just throws read errors (R05 R06). Some keys on the right of the board output tons of those error codes per press, and one key outputs five different read codes. I cleaned up the connector pins and dusted out the switchplate to no avail. This thing just seems to be broken in a very specific way.

I know it's a longshot, but has anyone seen this kind of weird behavior before? Maybe one of the traces on the PCB is dead? I really have no idea how to diagnose this, and I would really like for this beautiful board to work . If anyone has any advice as to how I could go about fixing this, I would greatly appreciate it. Otherwise, I'll just have to sell it as a broken board...

Not sure this is any help, but here goes. This DT post:

keyboards-f2/is-this-old-datatech-spk-1 ... t8772.html

and its photo link:

https://imgur.com/a/gyg3P

show an SPK-100. The SysRq on the PrintScreen key and the extended layout make it by definition an AT-compatible board. If there's no switch it's probably auto-detect.

I have an SPK-101 here, same layout, but white Alps and not as pretty of a case. The SPK-101 has a knockout on the back for an XT-AT switch, but it's not punched out on mine, and there's no switch behind it on the inside, nor any jumpers on the bottom of the PC board where the switch would be. Mine works normally on a PS/2 connection with a passive adapter. I don't have an XT setup available at the moment to check for auto-detect.

Datatech/DTK was a huge supplier of IBM clone boards and parts, and their stuff was if anything *more* compatible than IBM. By that I mean they sold motherboards and cards made to be used by hobbyists building clone computers, so they had to work with everybody else's hardware. For instance, if you had a motherboard that wouldn't recognize a particular card the first thing was to drop in a DTK (or Phoenix) BIOS chip. They made good stuff.

I guess one question is what happens when you plug an auto-detect keyboard into an auto-detect controller? But regardless it should work correctly on a PS/2 motherboard.

First thing I'd try would be to replace any electrolytic capacitors on the PC board. Those can dry out over time, and cause strange problems like you're experiencing. Electrolytic caps have an aluminum case, usually with a colored plastic sleeve covering it, with microfarad (uf) and voltage ratings. They're polarized, with + or - marked with a stripe on the case, very important to install the new ones the same way.

Your talk about it being an AT board for sure made me question the integrity of the Soarer's converter. Sure enough, I plugged my Model F AT into the converter and it got all sorts of wacky errors no matter what mode I forced, AT or XT. So I guess that converter is dead. :/

Regardless, this keyboard still does not work when plugged into the passive AT to PS2 adapter, which I know works because I use it daily in conjunction with my Model F AT. With the SPK-100, all I get are lock lights.

So is there any chance of this still being an AT vs. XT problem? Could it be that the keyboard is somehow auto-detecting XT despite being plugged into AT via PS2? While it wouldn't make much sense, it would explain why the PC isn't detecting anything. Even though the soarer's converter is definitely broken, it was picking up unique signals from the keyboard in XT mode versus nothing in AT mode. My only guess at this point is that the keyboard is somehow stuck in XT mode. Do you think that's plausible enough to purchase a new soarer's converter to test it? I probably need it anyway...

[Polecat]

Your talk about it being an AT board for sure made me question the integrity of the Soarer's converter. Sure enough, I plugged my Model F AT into the converter and it got all sorts of wacky errors no matter what mode I forced, AT or XT. So I guess that converter is dead. :/

Regardless, this keyboard still does not work when plugged into the passive AT to PS2 adapter, which I know works because I use it daily in conjunction with my Model F AT. With the SPK-100, all I get are lock lights.

So is there any chance of this still being an AT vs. XT problem? Could it be that the keyboard is somehow auto-detecting XT despite being plugged into AT via PS2? While it wouldn't make much sense, it would explain why the PC isn't detecting anything. Even though the soarer's converter is definitely broken, it was picking up unique signals from the keyboard in XT mode versus nothing in AT mode. My only guess at this point is that the keyboard is somehow stuck in XT mode. Do you think that's plausible enough to purchase a new soarer's converter to test it? I probably need it anyway...

I've never used Soarer's converter, perhaps someone else can help with that. Hopefully the SPK-100 didn't fry it, from bad caps or otherwise.

Have you tried powering up the computer with the keyboard connected, rather than hot plugging it? An auto-detect keyboard, assuming this is one, might be picky about that. The Model F AT board is AT-only, so it might not care about that.

Having just lock lights is a symptom of a keyboard being in the wrong mode. Have you opened up the case to see if there's a hidden switch or jumper to set the mode? The photos in the earlier post show the SPK-101 having a blank spot (with QC sticker) where the switch would be, just like mine, and if the SPK-100 uses the same PC board the switch or jumpers would be directly under that blank.

[zooksman]

Your talk about it being an AT board for sure made me question the integrity of the Soarer's converter. Sure enough, I plugged my Model F AT into the converter and it got all sorts of wacky errors no matter what mode I forced, AT or XT. So I guess that converter is dead. :/

Regardless, this keyboard still does not work when plugged into the passive AT to PS2 adapter, which I know works because I use it daily in conjunction with my Model F AT. With the SPK-100, all I get are lock lights.

So is there any chance of this still being an AT vs. XT problem? Could it be that the keyboard is somehow auto-detecting XT despite being plugged into AT via PS2? While it wouldn't make much sense, it would explain why the PC isn't detecting anything. Even though the soarer's converter is definitely broken, it was picking up unique signals from the keyboard in XT mode versus nothing in AT mode. My only guess at this point is that the keyboard is somehow stuck in XT mode. Do you think that's plausible enough to purchase a new soarer's converter to test it? I probably need it anyway...

I've never used Soarer's converter, perhaps someone else can help with that. Hopefully the SPK-100 didn't fry it, from bad caps or otherwise.

Have you tried powering up the computer with the keyboard connected, rather than hot plugging it? An auto-detect keyboard, assuming this is one, might be picky about that. The Model F AT board is AT-only, so it might not care about that.

Having just lock lights is a symptom of a keyboard being in the wrong mode. Have you opened up the case to see if there's a hidden switch or jumper to set the mode? The photos in the earlier post show the SPK-101 having a blank spot (with QC sticker) where the switch would be, just like mine, and if the SPK-100 uses the same PC board the switch or jumpers would be directly under that blank.

I never hotplug them anyway. No PS/2 keyboards or mice work on my system unless they’re plugged in at boot.

I have scoured the internals for some kind of hidden switch for jumper. I don’t know what QC sticker you-re referring to, but all that’s inside the case is a PCB covered in solder joints and a controller. On the controller there is nothing that looks like a jumper either. There is one component labeled “JP1” but it doesn’t really look like a jumper to me and I don’t want to fry anything. I’ll attach photos of the PCB and controller in case anyone sees something I’m not seeing. https://imgur.com/a/3bmk3

[Polecat]

I never hotplug them anyway. No PS/2 keyboards or mice work on my system unless they’re plugged in at boot.

I have scoured the internals for some kind of hidden switch for jumper. I don’t know what QC sticker you-re referring to, but all that’s inside the case is a PCB covered in solder joints and a controller. On the controller there is nothing that looks like a jumper either. There is one component labeled “JP1” but it doesn’t really look like a jumper to me and I don’t want to fry anything. I’ll attach photos of the PCB and controller in case anyone sees something I’m not seeing. https://imgur.com/a/3bmk3

The JP1 jumper (hardwired) appears to be connected to the same IC pins as the switch pads on my SPK-101. Pin 1 to one side, and pins 7 and 20 (and others?) on the other. There's a chance that snipping that jumper may force it into one mode or the other, but that's only a maybe. Again my SPK-101 doesn't have anything connected there, just unused pads where the switch would have gone. And my 101 does come up in AT mode (when plugged into a PS/2 motherboard using a passive adapter), but after seeing your photos the PC board is quite different between the 100 and 101 models.

I'd suggest just selling it to me, but if another keyboard showed up right now I'd be sleeping with it, in my van.

[zooksman]

I think I’m going to try another soarer’s converter first to see if it even works in XT mode. If that doesn’t work then I’ll have to assume the board is dead. I have one for a 240-degree din plug that I can probably wire up to a regular one.

[zooksman]

Update: I've tested it a bunch more. It seems the converter is NOT the problem, since every keyboard besides this one and my Model F AT works flawlessly with it. (I don't know why the F AT doesn't work with it. It works fine with a passive converter.)

In case you're not familiar with the Soarer's Converter, it's basically an all-in-one keyboard USB converter that supports three protocols: XT, AT, and terminal (122-key). You can change its firmware to force it to one of those protocols for testing. There is a program called hid_listen that reads the raw output of the converter including error codes for diagnosis. When I plug the SPK-100 in, the converter autodetects XT, and the output is a mix of R05 and R06 error codes and read codes depending on the key pressed. If I force the converter into AT or terminal mode, the output is simply "R06 wFE" on every keystroke. So I pulled out some XT and AT keyboards to check those results against them.

When I tried to force an AT board into XT mode on the converter, the output was eerily similar to the one the SPK-100 gives, being a mix of R05 and R06 with read codes. That gave me the idea that the SPK-100 really is in AT mode. However, changing the converter to AT mode simply outputs "R06 wFE", which doesn't make sense if the board really is AT.

Secondly, I tried to force a different XT board into AT mode on the converter. The output was exactly "R06 wFE" on every keystroke, which is exactly what the SPK-100 outputs when forced into AT mode on the converter.

These two conflicting behaviors have led me to believe that the SPK-100 is autodetecting whichever protocol is opposite of the one set on the converter. This is consistent with the lock lights functioning but nothing else working when plugged into a passive AT-PS2 adapter; it is stuck in XT mode. So why on earth would it be doing this? I'm almost positive that it is an autodetect board; it has a SysRq key like you said, but the date sticker on the back says 1988, so it is very unlikely that it had no XT support at all. Maybe someone else could weigh in on this: why would it be autodetecting the wrong protocol, and is there any way to "trick" it? I could cut that JP1 jumper but I'm not trying to fry anything.

[Polecat]

Update: I've tested it a bunch more. It seems the converter is NOT the problem, since every keyboard besides this one and my Model F AT works flawlessly with it. (I don't know why the F AT doesn't work with it. It works fine with a passive converter.)

In case you're not familiar with the Soarer's Converter, it's basically an all-in-one keyboard USB converter that supports three protocols: XT, AT, and terminal (122-key). You can change its firmware to force it to one of those protocols for testing. There is a program called hid_listen that reads the raw output of the converter including error codes for diagnosis. When I plug the SPK-100 in, the converter autodetects XT, and the output is a mix of R05 and R06 error codes and read codes depending on the key pressed. If I force the converter into AT or terminal mode, the output is simply "R06 wFE" on every keystroke. So I pulled out some XT and AT keyboards to check those results against them.

When I tried to force an AT board into XT mode on the converter, the output was eerily similar to the one the SPK-100 gives, being a mix of R05 and R06 with read codes. That gave me the idea that the SPK-100 really is in AT mode. However, changing the converter to AT mode simply outputs "R06 wFE", which doesn't make sense if the board really is AT.

Secondly, I tried to force a different XT board into AT mode on the converter. The output was exactly "R06 wFE" on every keystroke, which is exactly what the SPK-100 outputs when forced into AT mode on the converter.

These two conflicting behaviors have led me to believe that the SPK-100 is autodetecting whichever protocol is opposite of the one set on the converter. This is consistent with the lock lights functioning but nothing else working when plugged into a passive AT-PS2 adapter; it is stuck in XT mode. So why on earth would it be doing this? I'm almost positive that it is an autodetect board; it has a SysRq key like you said, but the date sticker on the back says 1988, so it is very unlikely that it had no XT support at all. Maybe someone else could weigh in on this: why would it be autodetecting the wrong protocol, and is there any way to "trick" it? I could cut that JP1 jumper but I'm not trying to fry anything.

You've done a superb job of troubleshooting the problem, and I agree with the conclusion that the autodetect is working backwards. I don't really know what an autodetect keyboard looks for, other than possibly a pulse on the reset line from an AT/PS/2 computer. If the converter doesn't generate a reset in AT mode, or if it does in XT mode that could explain what's happening, but that's clutching at straws on my part.

Snipping the jumper almost certainly won't hurt anything, and it's easy to put back the way it was, but chances are it won't help either. The programming on the controller chip is almost certainly different between hard-switched and auto-detect versions. The DTK keyboard was one that was sold separately, not with a computer, so I agree that it's probably auto-detect since it doesn't have an XT/AT switch. I just tried to fire up my old Televideo TPC-II to try my 101 on an XT, but it apparently sat too long, and I'm not getting any video.

[orihalcon]

I had a batch of Lynk branded model M's which were brand new in box that I slowly sold off about a year or two ago that wouldn't work with Soarer's Converter 1.12 which gave the same R06 pretty much on EVERY key that was pressed. Solution was go back to version 1.03 which apparently handles XT differently than 1.12. I think Soarer made the change so that it would support the Leading Edge DC-2014 which uses a somewhat non-standard XT protocol apparently and 1.12 was supposed to work for both IBM and the Leading Edge

See here for the full story:
https://geekhack.org/index.php?topic=62168.0

If you want an AT/XT Soarer's Converter with the 1.03 firmware on it to give it a shot, PM me!

[zooksman]

Well, I went ahead and bit the bullet, cutting the jumper. And to my surprise, it just completely fixed the problem! The converter now autodetects the board as AT, and everything works properly! It works with a passive converter too. However, the board acts like any other AT board when I force it into XT, with most keys not functioning as intended. Perhaps I was wrong about it being an autodetect board in the first place, and that jumper was just a hardwired AT/XT switch. Or it could be that the XT on the board is the same variant as the Lynk Model Ms that Orihalcon mentioned. Since I think this thing was supposed to be a model M clone/competitor, it would make sense. Perhaps cutting the jumper somehow reversed the autodetect back to how it should be, or perhaps it was just a placeholder for an AT/XT switch they never included.

I must thank you for all your help Polecat. I never would have cut that jumper unless you said that bit about it being connected to the same points as the switch on your SPK-101. Since this board doesn't have a wiki entry yet I'll probably add one with this information. And Orihalcon, I'm a big fan of your converters In fact, I'm sorry for ever assuming that the converter was the problem! They are very well made and work great (except, randomly, with my Model F AT).

[Polecat]

Well, I went ahead and bit the bullet, cutting the jumper. And to my surprise, it just completely fixed the problem! The converter now autodetects the board as AT, and everything works properly! It works with a passive converter too. However, the board acts like any other AT board when I force it into XT, with most keys not functioning as intended. Perhaps I was wrong about it being an autodetect board in the first place, and that jumper was just a hardwired AT/XT switch. Or it could be that the XT on the board is the same variant as the Lynk Model Ms that Orihalcon mentioned. Since I think this thing was supposed to be a model M clone/competitor, it would make sense. Perhaps cutting the jumper somehow reversed the autodetect back to how it should be, or perhaps it was just a placeholder for an AT/XT switch they never included.

I must thank you for all your help Polecat. I never would have cut that jumper unless you said that bit about it being connected to the same points as the switch on your SPK-101. Since this board doesn't have a wiki entry yet I'll probably add one with this information. And Orihalcon, I'm a big fan of your converters In fact, I'm sorry for ever assuming that the converter was the problem! They are very well made and work great (except, randomly, with my Model F AT).

Excellent, glad it worked! Definitely sounds like the 100 came hardwired as (pseudo) XT. My 101 may well be AT-only then with no jumper or switch installed. Connecting a switch across the jumper connection would probably allow for XT-AT in both cases then. Good stuff. Once I have a working XT again I can play with my 101 a bit more. It was common for the "extra" keys on an extended/101 layout not to work correctly on early XT machines unless you upgraded the BIOS chip. A few keyboards came with a switch having two XT positions, regular and extended, which was supposed to help with that. Strictly speaking there was no such thing as "IBM compatible" if it wasn't actually built by IBM.