das iii typos at higher speeds

[bhtooefr]

Here's iOne's European site, FWIW: http://www.ione-europe.com/component/page,shop.product_details/flypage,shop.flypage/product_id,1642/category_id,10/manufacturer_id,0/option,com_virtuemart/Itemid,42/

[wellington1869]

Well, that's iOne's US site...

I always forget theres a whole non-US world out there
For that matter, non-NYC

[itlnstln]

Yeah, us country bumpkins down here in the South need our own special layout with a few less keys to match our teeth.  Yeeeee-haw!

[lam47]

So even on the Euro site they only have a US layout.

You know if I can fix the c I will keep it. Im in love even with its flaws.

Can I open the switch from the top?

[bhtooefr]

Well, I've been looking around, and that looks like it's just the pic, here's a German layout: http://lauwe.de/index.cgi?session=90aea7ae49944957927f505e6ec47334708b71286854dde3bbacb4a207289c401&LANG=DE&id=1592&menu=1&WG_ID=1499_1500&do=details

[wellington1869]

Can I open the switch from the top?

not on the das, AFAIK.

Im in love even with its flaws.

now thats love

[lam47]

Well, I've been looking around, and that looks like it's just the pic, here's a German layout: http://lauwe.de/index.cgi?session=90aea7ae49944957927f505e6ec47334708b71286854dde3bbacb4a207289c401&LANG=DE&id=1592&menu=1&WG_ID=1499_1500&do=details

they must do then as thats the same as UK!

[bhtooefr]

I'd contact iOne Europe, and see what they can do for you, to get a Scorpius M10 in UK layout...

[LlamaZorz]

I think I am a decently fast typer, but there is no way I can write code like "printf("");" at those speeds to seen here.  My lord, when do you think about what you are coding.

[wellington1869]

they must do then as thats the same as UK!

lam you can buy blank MX keycaps (to put on the m10 or anything else):

http://geekhack.org/showpost.php?p=12168&postcount=85

http://mykeyboard.co.uk/keycaps/

[lam47]

Thats interesting. Im going to have a look at getting a UK M10

not great opinions on amazon.

http://www.amazon.com/Ione-Scorpius-Mechanical-Keyswitch-Keyboard/dp/B000UC1W3C

[rush340]

Are you fast typers really noticing a significant amount of typos?  Even of I concentrate on typing "the" as fast as I can while still being sure I typed the letters in order, it never comes out wrong.  The problem only seems to occur if I hit all of the keys at the same time.

[sprintf32768]

Would anyone like speculate on what sorts of USB HID implementation errors might have caused this problem?  

I know almost nothing about USB HID programming.  I just skimmed the HID 1.1 Device Class Definition.

I imagine that the keyboard is sampling the keyswitches at 60Hz, and sticking all the depressed keys into one HID data report packet.  The keyboard does not buffer keypresses into separate reports (pg. 62 "Keyboards may buffer events that would have otherwise resulted in multiple event in a single report.")

   So, instead of buffering keypresses into separate reports so that their sequence is preserved, the keypresses all go into one report.  The USB host sees the report packet and sees that 'i' and 'n' were simultaneously depressed.  And, the USB host decides to produce "ni" (pg 62: "The order of keycodes in array fields has no significance. Order determination is done by the host software comparing the contents of the previous report to the current report.")

So I have a couple questions for the USB experts in the crowd.  Can you tell a keyboard device to increase its idle rate, thereby increasing the rate of data packets?  It looks like it is theoretically possible to set the idle interval to 4ms (from the HID spec).  Wouldn't you be surprised, though, if you could just tell a USB keyboard to increase its polling rate?  I would be.  I figure that USB manufacturers set an idle rate that is convenient for them and the USB host can not simply demand more packets/second.

  Anyway, if anyone happens to know if idle rate is a way of tweaking the performance of a keyboard, that would be pretty useful!  A brief google search turned up nothing of particular interest except for some linux kernel guy, trying to set the idle interval to zero, and failing.  But ymmv.

   Oh, and here is the HID spec that I downloaded, if anyone is interested.
http://www.usb.org/developers/devclass_docs/HID1_11.pdf

[lam47]

I know that USB ports can have polling rates upped to 500hz.
Hold on.
http://www.overclock.net/computer-peripherals/173255-cs-s-mouse-optimization-guide.html
halfway down the first post, overclock your usb ports.
Would this work?


On a side note I dismantled a das3 yesterday (wont go into why)

The plastic is very cheap, 2 of the nubs that hold the pcb up were snapped.
The shiny top panel is actually black transparent plastic.

The PCB had many badly soldered points and 3 repaired traces!

This keyboard is shockingly badly made for the price.

[bhtooefr]

Right, but that's the rate at which the keyboard is polled, not the rate at which the keyboard is sampling.

Also, raising the USB polling rate increases CPU usage.

[alpslover]

I imagine that the keyboard is sampling the keyswitches at 60Hz, and sticking all the depressed keys into one HID data report packet.  The keyboard does not buffer keypresses into separate reports (pg. 62 "Keyboards may buffer events that would have otherwise resulted in multiple event in a single report.")
   So, instead of buffering keypresses into separate reports so that their sequence is preserved, the keypresses all go into one report.  The USB host sees the report packet and sees that 'i' and 'n' were simultaneously depressed.  And, the USB host decides to produce "ni" (pg 62: "The order of keycodes in array fields has no significance. Order determination is done by the host software comparing the contents of the previous report to the current report.")

yes, if the host receives one report in which 'i' and 'n' are shown as depressed, and then in some subsequent report, 'i' and 'n' are shown as released, then the sequence of the keys is indeterminate since the host interprets that sequence of reports as simultaneous key depressions.  it has no choice but to assign some arbitrary sequence, for argument's sake 'ni'.

the problem is, what if you had in reality pressed 'i' before 'n', but so quickly that the keyboard misinterpreted this sequence as simultaneous because it wasn't scanning the keys quickly enough?  it then sends the above sequence of reports (with incorrect data) and you get 'ni' instead of 'in'.  in this case, the problem lies with the keyboard, not the host.

from my observations, i'd say 60hz is a bit on the high side.  i don't think i'm hitting keys within 16.67ms of each other, but then again i've no way of timing it either.  perhaps 30hz is more realistic.

i've tried setting the usb polling rate to 1000hz, it doesn't make a difference.

[sprintf32768]

As it happens I was able to time my bad das-3 keystrokes fairly well, using a microphone and a sound-editor to measure the click-gaps.  It is very easy to hit two keys within 16ms of eachother, especially for commonly typed words like printf.  The DAS-3 seemed to be able to sequence them properly when I hit them at a > 16ms interval, but not if the gap was < 16ms.  

  I arranged my index and middle fingers over the 'n' and 'i' keys like a "little hopping man" and adjusted the angle of my hand so that the keys were getting pressed more/less simultaneously.  I started with the 'i' first, and kept adjusting my hand until the keyboard produced 'ni'.  At that point I stopped recording and measured the gap in the sound-editor (I used GoldWave, fwiw).

  This was a simple way to determine the practical sampling rate of this keyboard, but you should be skeptical.  I believe that my mind can detect which keys I've hit first!  But if you think I am crazy or cannot perceive my own fingertips accurately, we may need to develop a key-depressing apparatus that we all can trust.  DAS claimed I was actually hitting 'n' first in the 'ni' cases.  I claim I was hitting 'i' first.  In any case I chose to return the keyboards for a refund rather than make a key-pressing machine whose key-order perception would be indisputable; as I say, DAS did not really take this problem very seriously.

[sprintf32768]

the problem is, what if you had in reality pressed 'i' before 'n', but so quickly that the keyboard misinterpreted this sequence as simultaneous because it wasn't scanning the keys quickly enough?  it then sends the above sequence of reports (with incorrect data) and you get 'ni' instead of 'in'.  in this case, the problem lies with the keyboard, not the host.
...
i've tried setting the usb polling rate to 1000hz, it doesn't make a difference.

  Absolutely, the keyboard can only scan so quickly!  We can't just expect a keyboard to scan at 1KHz because we demand it.  So yes, it will just send the same report N times and then finally get around to sampling again.  

   This is all very interesting.  Right now, I am at home using the super-lame Dell keyboard that came with this computer.  It does not reorder keys on me.  It has some n-key rollover issues, but it keeps the keys in the right order!  I do a little programming on it, which is not much fun, but I am not going crazy like with the DAS-3.

   And surely this crappy keyboard is the cheapest-ass keyboard ever.  It is free.  So how can they get this *right* when DAS does it wrong?

   Is it that DAS needed fancy firmware to support N-key rollover, but it accidentally broke key-sequencing?

[alpslover]

 This was a simple way to determine the practical sampling rate of this keyboard, but you should be skeptical.  I believe that my mind can detect which keys I've hit first!  But if you think I am crazy or cannot perceive my own fingertips accurately, we may need to develop a key-depressing apparatus that we all can trust.  DAS claimed I was actually hitting 'n' first in the 'ni' cases.  I claim I was hitting 'i' first.  In any case I chose to return the keyboards for a refund rather than make a key-pressing machine whose key-order perception would be indisputable; as I say, DAS did not really take this problem very seriously.

nice work.  i hadn't thought about recording the keystrokes and using audio software to determine the timing.  i still have my das, i may have a go at it to confirm your results.

[bhtooefr]

Is it that DAS needed fancy firmware to support N-key rollover, but it accidentally broke key-sequencing?

That's my guess. The scan delay was probably intentional, so that they could simulate N-key rollover without quite the necessary hardware (are there diodes installed?) or so that they could simulate it over USB without drivers. (And, they didn't even simulate n-key rollover, they simulated 12-key rollover.)

[alpslover]

This is all very interesting.  Right now, I am at home using the super-lame Dell keyboard that came with this computer.  It does not reorder keys on me.  It has some n-key rollover issues, but it keeps the keys in the right order!  I do a little programming on it, which is not much fun, but I am not going crazy like with the DAS-3.

   And surely this crappy keyboard is the cheapest-ass keyboard ever.  It is free.  So how can they get this *right* when DAS does it wrong?

   Is it that DAS needed fancy firmware to support N-key rollover, but it accidentally broke key-sequencing?

i've also tested three el cheapo usb keyboards - 2 (different model) dells and an hp.  one dell and the hp exhibited the same problem as the das, but the other dell did not.  none of them have n-key rollover support, so while i don't believe that the das's n-key rollover is the cause of the problem, it does make it easier to test for it.

[Chloe]

Does anyone have an idea of what WPM would trigger these typos? I'm thinking of getting a DAS for when I'm in the mood for something clicky.

[wellington1869]

Does anyone have an idea of what WPM would trigger these typos? I'm thinking of getting a DAS for when I'm in the mood for something clicky.

Someone else was asking that before... I think the consensus was that its not really a WPM issue, because even slow typers will occasionally type a short sequence of letters quickly (and run the risk of the problem potentially showing up) even if the overall WPM is pretty slow.

That said, I think the people who first noticed the problem were people who typed around 100 wpm to begin with.

[bhtooefr]

Just get a Scorpius M10 if you want Cherry blues, avoid that issue altogether. They're sold in Europe, too, after all.

[Chloe]

I want blank keycaps as well. I'm not sure if the switches in the b18 are clicky otherwise I'd probably get one of those.

[safinace]

I just got my Das Keyboard today, and I wanted to test this out. If I press two buttons so close to each other that I hear _ONE_ click ... then yes, what's said in this forum is true. I cannot duplicate the key error, whenever I hear two clicks - no matter how close together they are, and how hard I try.

This is absolutely a non issue for _ME_. My typing is ~90wpm. I hope this is helpful for others. I also hope that for those people, which this is an issue, find a resolution.

[wellington1869]

I just got my Das Keyboard today, and I wanted to test this out. If I press two buttons so close to each other that I hear _ONE_ click ... then yes, what's said in this forum is true. I cannot duplicate the key error, whenever I hear two clicks - no matter how close together they are, and how hard I try.

This is absolutely a non issue for _ME_. My typing is ~90wpm. I hope this is helpful for others. I also hope that for those people, which this is an issue, find a resolution.

Thats probably accurate, and I too type about 90wpm. But for me there were two issues that kept me from being able to live with the problem. One, quite honestly, was the principle of the thing (as others mentioned, knowing there is an issue makes me indignant enough to move on). The other is that, even at 90wpm, I'm certain that I will sometimes type a short sequence of keys that close together. And if it results in a typo, it will be impossible to tell if its the board or me. And once again I'd be indignant. I suppose its my own reaction (if I didnt care about the occasional typo then sure, I could live with it too, as a practical matter).  I guess this is something everyone has to decide for themselves. Of course if DAS fixed the problem it'd be a non issue for everyone...

[lam47]

I just got my Das Keyboard today, and I wanted to test this out. If I press two buttons so close to each other that I hear _ONE_ click ... then yes, what's said in this forum is true. I cannot duplicate the key error, whenever I hear two clicks - no matter how close together they are, and how hard I try.

This is absolutely a non issue for _ME_. My typing is ~90wpm. I hope this is helpful for others. I also hope that for those people, which this is an issue, find a resolution.

do you think you could do a little test.
press asdf all at the same time then move your hand and press enter.
Now do it again 10 - 20 times.
Post the results.
It will let us see if the issue has indeed been fixed.

[wellington1869]

do you think you could do a little test.
press asdf all at the same time then move your hand and press enter.
Now do it again 10 - 20 times.
Post the results.
It will let us see if the issue has indeed been fixed.

I dont think he was suggesting it was fixed, was he? I thought he was saying that it looked like something he could "live with"...

[lam47]

Just re read. My mistake. I thought his was not showing any problems.

[safinace]

Here's what happens when I press asdf at the same time (x20) in Emacs

asdf
asdf
asdf
asdf
asdf
asdf
asdf
asdf
asdf
asdf
asdf
asdf
asdf
asdf
asdf
asdf
asdf
asdf
asdf
asdf

@wellington1869 - It _is_ a principle thing.

Far From Left Field: I deleted all my hotmail contacts (moved them over using an Emacs macro) and abandoned my account after they rolled the new UI interface which had compatibility issues with Ubuntu. Not because I didn't know how to fake my agent string and fix the issue in the 15 seconds it takes to type about:config and replace a string, but because of principle.

If you are putting up with the flaw, you are selling yourself out.

If there is open hardware / firmware, you have the power to fix it yourself. In a case like this, that power is already stripped from you. So the only power you have left, is not to put up with the flaw.

As, I mentioned before, I verified the problem; I just don't run into it during typing.

I'm rooting that the company addresses and fixes this.

[wellington1869]

that was very philosophical.

[lam47]

Yeah nice response mate.
Thanks for doing the test.

[safinace]

****. I'm considering getting a second Das Keyboard. I love this thing.

[rush340]

I just got my Das Keyboard today, and I wanted to test this out. If I press two buttons so close to each other that I hear _ONE_ click ... then yes, what's said in this forum is true. I cannot duplicate the key error, whenever I hear two clicks - no matter how close together they are, and how hard I try.

This is absolutely a non issue for _ME_. My typing is ~90wpm. I hope this is helpful for others. I also hope that for those people, which this is an issue, find a resolution.

Exactly.

I don't see how this could be an issue for anyone.  Even if I concentrate on typing two letters in an order that would be switched, I cannot replicate the error.  I only see it happen when I pressed them at the same time and can only hear one click.  If you're typing that fast, it's very likely that you made the typo (by hitting them at the same time).

If your typing two letters that fast, there's no way to tell whether you're hitting them in order, in which case it may as well be a typo so you can learn from it and improve your technique/control.  Then again who knows, maybe some of you have insane finger senses.  I'm probably going to get some defensive responses of people who think they can tell.

[wellington1869]

Exactly.

I don't see how this could be an issue for anyone.  Even if I concentrate on typing two letters in an order that would be switched, I cannot replicate the error.  I only see it happen when I pressed them at the same time and can only hear one click.  If you're typing that fast, it's very likely that you made the typo (by hitting them at the same time).

If your typing two letters that fast, there's no way to tell whether you're hitting them in order, in which case it may as well be a typo so you can learn from it and improve your technique/control.  Then again who knows, maybe some of you have insane finger senses.  I'm probably going to get some defensive responses of people who think they can tell.

"who knows" is exactly the problem for me. I'd rather have a keyboard where, in fact, I know

[lam47]

You can see its an issue from posts on the forum.
I have seen members who's typing is normally perfect write teh twice in one post whilst using the Das, and not once with another board.
I think its in this thread actually.

[alpslover]

Exactly.

I don't see how this could be an issue for anyone.  Even if I concentrate on typing two letters in an order that would be switched, I cannot replicate the error.  I only see it happen when I pressed them at the same time and can only hear one click.  If you're typing that fast, it's very likely that you made the typo (by hitting them at the same time).

not when these particular typo combinations are occurring with a significantly greater frequency on this particular keyboard versus other keyboards for those of us experiencing this problem.

If your typing two letters that fast, there's no way to tell whether you're hitting them in order, in which case it may as well be a typo so you can learn from it and improve your technique/control.  Then again who knows, maybe some of you have insane finger senses.  I'm probably going to get some defensive responses of people who think they can tell.

yes, you will get defensive responses, because you've presumed that because you can't type very quick key sequences and know that they're in the right order, no one else can either.

due to its nature, this issue isn't going to affect everyone.

[safinace]

Hmm, not related to the key transposition issue: I expected this keyboard to be "Made in Germany", but reading the fine print reveals: "Designed in Austin, Texas", "Made in Taiwan". That's deceptive.

[bhtooefr]

Not that deceptive, they never said that it was made in Germany, they said the keyswitches were German-engineered. (Although the pseudo-German name...)

[itlnstln]

Weren't the Das IIs made in Germany (since they were pretty much off-the-shelf Cherry 'boards)?  Or were they made in the Czech Rep. like some of Cherry's other stuff?

[lam47]

 I'm probably going to get some defensive responses of people who think they can tell.

Are you trying to be insulting?

[Therac-25]

So, are the Filco N-key rollover keyboards affected by this bug as well?  Given that both the Das III and the Filco boards are made by Costar...

I'm probably going to break down and order one of the Filcos sometime this month, but I think I'll steer clear of the N-key rollover ones just incase.

[bhtooefr]

I believe the Filco N-key rollover boards are PS/2, and don't have to resort to tricks like the Das uses to get more than 6-key rollover on USB.

[Therac-25]

I believe the Filco N-key rollover boards are PS/2, and don't have to resort to tricks like the Das uses to get more than 6-key rollover on USB.

This is good.

I'll probably order one in a bit.  I want to see how the blue Cherrys[1] are on the M10 before deciding on which tenkeyless to order.

[1] grammar disclaimer -- i operate under the assumption that pluralization of proper nouns follow different rules.

Edit: Fixed.... :-P

[bhtooefr]

Proper nouns are also normally capitalized, though.

[Chloe]

[1] grammar disclaimer -- i operate under the assumption that pluralization of proper nouns follow different rules.

I wondered about this myself. In the past I have used Cherrys but it doesn't look right. If I were writing an article I would probably avoid using Cherrys altogether. I prefer to use "blue stem Cherry MX switches" for sake of clarity, but lately I've been using "blue Cherries" for convenience.

[Therac-25]

I wondered about this myself. In the past I have used Cherrys but it doesn't look right. If I were writing an article I would probably avoid using Cherrys altogether. I prefer to use "blue stem Cherry MX switches" for sake of clarity, but lately I've been using "blue Cherries" for convenience.

Again, being a Leafs fan, it doesn't bother me to see a brand name pluralized oddly.  It's a brand, not a normal word.

[lowpoly]

Cherry is a surname (Walter L. Cherry).

[Therac-25]

Cherry is a surname (Walter L. Cherry).

Guess that settles it then.  

Of course, we aren't referring to multiple family members here, but I can't imagine why the rules for brands would be different than the rules for names.  

I know that the brand police at my company would throw a fit if the company name got changed in pluralization, so I figure I'll respect the brand managers at Cherry and not mangle the name in pluralization.