Q68 Serial devices

[Peter]

Also, I am in touch with BlackBox tech support and forwarded them a question regarding buffering. They did respond to me initially which I was surprised since the product is 15 years old.

Very surprising indeed.

I'm not typing fast enough to overload the Q68 at 50 characters per second as I'm typing probably more around 10 characters per second.

There can be keycode sequences though, when you press special keys. For example "Delete" sends $E0 $71 when you press ist, and $E0 $F0 $71 when you releasse it.
"Pause" is even worse, it sends $E1 $14 $77 $E1 $F0 $14 $E0 $77

You could, for example, check if pressing the "Pause" key brings the Q68 into a problematic state. With a normal PS/2 keyboard, nothing happens, and you can continue to type normally.

If it actually is caused by lack of buffer in the converter, I would still have ideas how to solve it. But I'd need a device here, and it might be a little "dirty" software-wise.

[bwinkel67]

I would lend you one of mine but I think shipping is prohibitive (you are in Europe, correct). Of course if it doesn't even work on other PS/2 connectors at school maybe I can stick it in a bottle and toss it in the ocean to get to you for cheap How does China do its shipping again since I can buy things for a quarter including shipping?

There is one other person on this thread that is waiting to test it on their Q60...maybe the two of you live a bit closer though even regional mail might be costly. Then again I want to get into FPGA coding though have no experience...

[Dave]

I had similar behavior to this with my Q68 very originally. It turned out to be a tin whisker on the PS/2 pin. Juuust enough to cause power glitches. It's very unlikely to ever happen twice, but it would be good to rule it out.

Another possibility is that the power supply isn't quite providing enough current to run the external 5V devices. They might almost be crashing but not quite.

[bwinkel67]

What if I plug a powered hub into the USB-to-PS/2 device and give it 5 volts? Don't want to damage the Q68.

[bixio60]

..........

There is one other person on this thread that is waiting to test it on their Q60...maybe the two of you live a bit closer though even regional mail might be costly. Then again I want to get into FPGA coding though have no experience...

It will be here Monday(Tomorrow), I will test and report the results. Considering I bought 2 I could borrow one to Peter for further investigation.

Let you know

Fabrizio

[bwinkel67]

Tried yet another keyboard (one of the newer Dell models) and typed one key at a time, pausing a second each time. After 10 to 12 key presses it starts misbehaving. I only typed the alphabet keys. So not sure how buffering works but that seems odd. Just seems like the communication between the Q68's PS/2 port and the keyboard, through the device, is not working. I also plugged it in directly to the port (without the splitter cable) and that had no difference.

I also tried adding an unpowered hub and it stopped working even quicker. I will be eagerly waiting to see how the Q60 does.

[Peter]

Tried yet another keyboard (one of the newer Dell models) and typed one key at a time, pausing a second each time. After 10 to 12 key presses it starts misbehaving.

That would not point to a buffering issue.
Does it never misbehave for the first few keys?
Is the number of working keypresses before misbehaviour reproducable?
What if you press "Pause"? Does it then start to misbehave earlier?

I also tried adding an unpowered hub and it stopped working even quicker.

Is that reproducable? This would point to power supply or the USB side rather than the PS/2 side! What is the max. current of your (hopefully regulated!) 5V supply?

I will be eagerly waiting to see how the Q60 does.

It would also be good to try with a different Q68 and power supply. Not that we draw false conclusions.

Another idea that comes to my mind ist the pullup resistors. The Q68 implements only "weak" pullups inside the FPGA around 20 kOhm and relies on the keyboard side to have it's own pullups. But an issue with missing pullups inside the converter would not explain why the first few keys work.

[Dave]

It does sound like the power supply. You plug the keyboard in and it delivers the extra current, but then after a few seconds it hits the thermal limit and starts current limiting?

[bwinkel67]

Peter and Dave,

My power supply is 5v at 2.4 amps (I had emailed Derek with the specs to make sure the extra amps were ok). It's a modern power brick so I'm guessing it is regulated...how can you tell? So the big question here, can I put a powered USB hub on the thing (I have one)? I don't want to damage the Q68 but if the risk is minimal I will try.

With regard to key presses, it's not exactly reproducible in that it sometimes takes 10 and sometimes takes 12 but until you get there they all work. If I type fast for the first few key presses I usually lose a key (i.e. I type Hello and it gives me Helo).

The unpowered hub is reproducible in that the keyboard right away behaves differently.

BTW, what is "Pause" as I'm having trouble identifying that on my Dell keyboards.

[Peter]

The unpowered hub is reproducible in that the keyboard right away behaves differently.

This sort of contradicts thoughts about PS/2 side buffering and pullups. Can you test a different power supply?

BTW, what is "Pause" as I'm having trouble identifying that on my Dell keyboards.

Maybe "Break"?

So the big question here, can I put a powered USB hub on the thing (I have one)?

Damage risk for the Q68 should be low. Most risk is for the power supplies.