das iii typos at higher speeds

[lam47]

Hey. Finally got my dasIII sorted so enjoying the blues very much.
Does anyone else find the enter key to be a bit stiff? Its not the switch as its the same blue as the rest. I have looked at the metal bar etc and all looks fine.
It just seems stiff and mushy.
It is a UK one so it could be different if you have the US type.

[sargon]

Does anyone have a Deck Legend that they can test for this problem?

[BigWopHH]

Hey. Finally got my dasIII sorted so enjoying the blues very much.
Does anyone else find the enter key to be a bit stiff? Its not the switch as its the same blue as the rest. I have looked at the metal bar etc and all looks fine.
It just seems stiff and mushy.
It is a UK one so it could be different if you have the US type.

When you say, "Hey. Finally got my dasIII sorted so enjoying the blues very much.", does that mean you found a resolution to the problem?

[lam47]

No sorry it does not.
I had another issue.
There is no fix for the rollover issue.
They are aware of the problem though and said there will be a fix at some point.
Not sure how they plan on doing it though.

[Deckguy Adam]

Does anyone have a Deck Legend that they can test for this problem?

Hi, this is Adam from Deck Keyboards.  Per sargon a post has been started in our forums concerning this issue.  I thought it may benefit others if it was posted here as well.

For anyone else wondering if the Deck Legend has this problem the short answer is this: no problems have been noticed thus far from our testing and customer experiences but it may be possible if you type faster than 1 character every 4 or 8 milliseconds.

Here is the thread in our forums:

http://deckkeyboards.com/forum/viewtopic.php?f=2&t=548

From that thread as provided by our lead engineer with respect to USB:

USB HID keyboards are strange; there’s no first-in-first-out queing of the data.  Instead, there is a “key-down array”  Note the “Report order is arbitrary and does not reflect order of events” in the “Comments” column on page 73 of the USB HID specification:

http://www.usb.org/developers/devclass_docs/HID1_11.pdf

Code: [Select]

B down 01000000 04 05 1B Report order is
                         arbitrary and does
                         not reflect order of
                         events
USB polls our keyboard about once every 8 ms.  So there is about an 8 ms window in which the ordering will be arbitrary...Then the next limiting factor is the scanning frequency of the keyboard [and we are not planning to change this for our current products].

With respect to PS2:

No, PS/2 has a first-in-first-out buffer (assuming the keys are pressed on a different scan, but faster than the USB polling rate).

The next limiting factor for PS/2 is the scan rate.  If you truly press two keys simultaneously (stop time, press two keys, restart time), whatever key the scanning routine found pressed first would be reported first.  

The same thing could happen during a single key scan period – even if the keys are pressed in a specific sequence, they could be reported out-of-sequence.

So what I think you’re getting at is would he be happier with a PS/2 keyboard?  I typically scan at a 4ms period and have about 25ms debounce time (the debounce time is irrelavent of the reported key sequence).  So they’d only be getting about 4 ms of extra resolution (8 ms usb polling period – 4ms scan rate).

As you can see our products work just about as fast as possible.  We could scan more but it really doesn't provide a huge noticeable difference.  I would bet that other companies either don't scan as much or have something else going on with their firmware...or maybe you all just type ridiculously fast and our products aren't able to cope with the mad typing skills!

[itlnstln]

This would seem to explain the odd NKRO results with USB (only) 'boards such as the ABS and Das.  The Das also suffers from a slow scan rate.

[itlnstln]

BTW, Deckguy_Adam, thanks for coming to GeekHack.    I hope we get to hear more from you and other companies as well.  Your consideration is greatly appreciated.  Thanks, again.

[alpslover]

thanks adam.  it pretty much confirms what we've suspected about the das iii (and other usb keyboards that suffer from this problem) - the keyboard isn't scanning the keys fast enough.

funny how metadot couldn't give us this kind of response.

[Biggs]

Would using a PS/2 to USB converter on a PS/2 version negate its advantages?

[itlnstln]

To a certain extent, yes, but I doubt you would see the same issues that some of these native USB 'boards have (unless you have a crappy PS2 -> USB adapter).  I would imagine that your PS2 NKRO would just be reduced to 6 keys.

[cmr]

i'm still not really convinced it isn't scanning the keys fast enough. i make exactly the same kinds of typos on my M10 as on my das iii. i wonder though whether people were making typos because of the relatively low resistance of the blue cherry switches. i'd be interested to hear from anyone who notices more transposition errors on a das iii compared to other blue cherry boards.

[BigWopHH]

I have a Cherry G80, M10, and DAS III, and definitely had more typos on the DAS.

[itlnstln]

i'm still not really convinced it isn't scanning the keys fast enough. i make exactly the same kinds of typos on my M10 as on my das iii. i wonder though whether people were making typos because of the relatively low resistance of the blue cherry switches. i'd be interested to hear from anyone who notices more transposition errors on a das iii compared to other blue cherry boards.

If I were the one typing, I would say you are correct; however, there are several tests earlier in the thread that would suggest otherwise.

[alpslover]

i'm still not really convinced it isn't scanning the keys fast enough.

the rollover test results are evidence.  the fact that the das iii outputs the same patterns very consistently when you press several keys simultaneously (which, in reality, are rarely truly simultaneous), points to an issue with the rate at which keys are scanned.  keyboards that don't suffer from this problem output more random results.

i make exactly the same kinds of typos on my M10 as on my das iii.

that just means that in your case, something else is causing you to make more typos.  it may very well be the lightness of the switches that you need to get used to.  but once you do, you may find yourself making typos on the das that you don't with the m10, in which case it would probably be due to the slower key scanning rate.

[cheater1034]

I got rid of the das because of the typos which made the board worth about as much as a pile of sticks to me, but this m10 i'm on has no problem, although it doesn't have a 6-key rollover like they advertise , but that doesn't really bother me i'm just happy it's not giving me any problems like the das.

[cmr]

typed on my m10 by pressing the four keys simultaneously:

fdsa
fdsa
fdsa
fdsa
fdsa
fdsa
fdsa
fdsa
fdsa
fdsa
fdsa
fdsa
fdsa
fdsa
fdsa
fdsa

it fails the same test as the das iii...

fdsa
fdsa
fdsa
fdsa
fdsa
fdsa
fdsa

does anyone know what the sampling clocks of these keyboards actually are?

[bhtooefr]

sadf
dfsa
sdaf
sdfa
sdaf
sdaf
sdaf
sdfa
sdfa
sdfa
sdfa
sdfa
sdfa
sadf
sdfa
sadf
sdaf
fsda
fsad
sadf

On my Scorpius M10, connected to a ThinkPad X61 Tablet (via USB,) running Windows XP Tablet SP3.

[ozar]

cmr... here's my M10, but I'm connecting with PS/2 if that makes any difference:

fdsa
sdaf
dfsa
sdfa
dfsa
dfsa
dfsa
dfsa
fdsa
fdsa
dfsa
dfsa
dfsa
dfsa
fdsa
fdsa
dfsa
fdsa
dfsa
fdsa
dfsa
fdsa
fdsa
fdsa
fdsa
dfsa
fdsa
dfsa

[cheater1034]

dafs
afds
adsf
adsf
dasf
adsf
dafs
adfs
fdas
dfas
dfas
dafs
adfs
adfs
adsf
fdas
dfsa
dafs
dafs
dafs
afds
dafs

On my scorpius m10, hooked up as regular usb on the ole' pc

on the das keyboard i had it definitely didn't behave correctly, the M10 does (although it doesn't have 6-key rollover like advertised)

[bhtooefr]

Where was it advertised?

[wellington1869]

Where was it advertised?

scorpius has a youtube video where it sounds like they claim n key rollover. I dont know that they use the phrase but he 'demonstrates' it by whacking at lots of keys at the same time and triumphantly looks at the camera.

[bhtooefr]

http://www.youtube.com/watch?v=ZXnGH6IMOfI

Hmm, they do claim nkro.

You know what I find funny? The guy, who supposedly works for Qtronix (read: Ione) has a crappy Dell rubber dome board on his desk.

[itlnstln]

http://www.youtube.com/watch?v=ZXnGH6IMOfI\
You know what I find funny? The guy, who supposedly works for Qtronix (read: Ione) has a crappy Dell rubber dome board on his desk.

Strong.

[MamiyaOtaru]

Does anyone have a Deck Legend that they can test for this problem?

Adam was already here, and seems to have laid concerns to rest, but here goes.  Deck Legend PS2

afsd
safd
dsaf
sadf
dfsa
fdsa
adfs
dsfa
sfda
afds
fdsa
fdsa
sfda
sdfa

[Hamps]

Yesterday, May 13th
Daniel, the co-founder of Das Keyboard did the honorable thing and admitted to the key transposition error the Das Keyboards have. http://www.daskeyboard.com/blog/?p=82

This was brought to our attention last November. (Most likely by you guys) Our testing has shown that this is in fact a ”limitation” of the first version of our keyboards.  The detection time of the current version was designed as 100ms which means that if you are a very, very fast typist, approximately 120 words per minute then, you might encounter this limitation.  It’s something we are redesigning for the next version of Das Keyboard.

Too bad this didn't come a couple days sooner because I just purchased Das Ultimate on May 10th. I purchased it entirely aware of the defect, but I was hoping that it wouldn't show up in regular typing. I type at 120 WPM though and I experienced it nearly every time on certain words. Ironically, in my refund request I linked to the very article they referenced in their blog post. I'm glad Das at least admitted to it officially though; because of that I'll consider purchasing their "next version"

[wellington1869]

Yesterday, May 13th
Daniel, the co-founder of Das Keyboard did the honorable thing and admitted to the key transposition error the Das Keyboards have. http://www.daskeyboard.com/blog/?p=82



Too bad this didn't come a couple days sooner because I just purchased Das Ultimate on May 10th. I purchased it entirely aware of the defect, but I was hoping that it wouldn't show up in regular typing. I type at 120 WPM though and I experienced it nearly every time on certain words. Ironically, in my refund request I linked to the very article they referenced in their blog post. I'm glad Das at least admitted to it officially though; because of that I'll consider purchasing their "next version"

interesting. the thing is tho, 120 wpm isnt uncommon. Its fairly common among professionals who work at keyboards all day. I could hit it in bursts i'm sure. (my average is about 90-100 wpm). All you'd have to do is hit that speed in a burst and you get nailed by the "limitation".

[Hamps]

Scratch that last post. I have that problem all the time. That is, as soon as Majestouch ships those Blue Cherry Filcos.

I lold

Yeah wellington, I couldn't believe how often the error occured. I honestly believed I wouldn't experience it at all unless purposefully looking for it. Because I mean c'mon it's a Das-- its sole purpose is to be a high quality keyboard for enthusiasts right? But at 120 wpm it happened every. single. time. Not to mention the input lag during my bursts above 120. I would literally see the letters come out all at once as an entire transposed word instead of letter by letter. It was really odd.

[wellington1869]

Because I mean c'mon it's a Das-- its sole purpose is to be a high quality keyboard for enthusiasts right?. .

ya, I agree. Just on principle I wouldnt get one. Its supposed to be for professionals and it fails there. Even if it is a gorgeous looking board.

[Manyak]

ya, I agree. Just on principle I wouldnt get one. Its supposed to be for professionals and it fails there. Even if it is a gorgeous looking board.

Agreed. It's really a shame. But if they do fix it in their next revision like they claim they will, I'll probably pick one of those up.

[ozar]

Anyone here happen to know if the Das Keyboard folks will pay the shipping on having the defective boards shipped back to them for a refund?  I'm guessing they won't but if anyone else got the shipping paid, I'll try the same.

[ozar]

On second thought... I'm only about an hour and twenty minutes from Austin, so I could just deliver it to them the next time I'm there.  I may or may not buy the next version of their board, depending on the price, what it looks like, and what switches it comes with.

[wellington1869]

Shipping is cheap (if you're in the US) so it's not a big deal.

When I RMA'd before they wouldn't pay for shipping.

I wouldn't push it anyway.  It is actually a pretty classy move for something that most people won't really see.

I'm impressed with Hamps having visible input lags.  I may have had them with the first Apple II I touched but that was long ago....

maybe i'm one of these spoiled americans, but personally I think classy would have been to pay shipping both ways - it was their defect after all

[sprintf32768]

Wow.  I am delighted that they are finally admitting this problem happens.  I almost can't believe that blog post is real.  It might be a very intricate troll for the geekhack forums.

  I wonder what changed their mind?  I RMA'd two DAS-3s back to metadot, back in July 2008 with the key-transposition.  And basically I was told that I couldn't correctly perceive the order in which my fingers hit keys.  This disparaging opinion about the sanity of Das-3 complainers came up a few times at geekhack as well.

  I had thought that perhaps the only way to defend our sanity was to build a keyboard-testing apparatus -- I won't describe it for you, but it was going to be a real contraption -- and this apparatus would demonstrate (or fail to demonstrate) the key-transposition bug.  But I knew that if I made this keyboard-testing machine, then I would be demonstrating that I was *not* sane.  I would be proving to the world that I am a freakish keyboard eccentric, railing against hardware bugs that only I can see, muttering and soldering in my basement.  Even if the bugs are real, I am still crazy.

   And so, I don't really feel like I've won.  I have this DAS-2, which is really unsexy.  But a few weeks ago I got one of Majestouch's NKRO Brown-Cherry 104s, and that thing has changed my life.  I am getting used to the Browns, the other guys in the office think the lack of resistance is crazy unusable -- they're Blue guys.  

   But on those Brown cherries I am having a great day, every day.  That keyboard is a dream.  I am making typos like crazy, but every single one of them is mine.  

   Suck it, DAS.

[ozar]

maybe i'm one of these spoiled americans, but personally I think classy would have been to pay shipping both ways - it was their defect after all

I generally feel the same when it's a defective product, or not as described.

[wellington1869]

 I am making typos like crazy, but every single one of them is mine.  

   Suck it, DAS.

lol, thats how I feel. I'm proud of my typos - so long as they're mine

[watduzhkstand4]

I think all DAS keyboards are ripoffs

[ssb]

I hate that when happens!
Since neither me nor my wife are fast typists (< 90 WPM) I didn't notice this issue. Having three Das III is quite costly to send them back. :smile:

[itlnstln]

I think all DAS keyboards are ripoffs

The Das II was a good keyboard.  That one was Cherry-made (basically, a G80-3000).  Switching to Costar caused the problems.  For as many good keyboards Costar makes, they also make some that aren't so great.  Unfortunately, the Das III was one of them.  The overall quality is good, but I think they have some issues in their software engineering department.

[wellington1869]

ripster your board collection reads like the wishlist of most geekhackers I just noticed. The only thing missing there is a white alps.

[itlnstln]

I don't know the Costar/Metadot specifics but most of the time the OEM manufacturer is at the tail end of the design process. In other words, I have a hard time blaming Costar for NKey Rollover issues. They are given a design. They build it. Metadot should be checking to see if the design/manufacturing handoff is seamless.
 
I even blame Metadot for case/plastic issues. Their Industrial Engineer needs to live at that plant until they get 100% QC. If Metadot even HAS a IE.
 
Metadot's roots are in software design. Beware of software companies building hardware (insert Microsoft example here....).
 
LOL - I'm getting myself so worked up I may just send the Das back for the principle of the thing.

Some of those things are Metadot's fault.  Scan rate problems, IMO, are Costar's.  It's a keyboard.  They make many different keyboards.  They make some really good keyboards.  For them to make an error like this, though, is careless.  Now, other than people like us, something like a scan rate issue will slip through QA.  When most people buy a keyboard, they assume that if they press keys, those keystrokes are going to appear on the screen.  They might test each key to make sure it works, but they might not test to make sure it outputs correctly at 150 WPM.  They probably don't even have anyone on their staff that can type that fast.  The problem, now, is Metadot is left holding the bag since they approved the design and went to market with it.  They were unaware of the problem until some ridiculously fast typers here found it and reported it.  I don't know what kind of contract they have with Costar, but it sure seems like they got into something where Costar doesn't have to fix defects found after QA.  That is both their problems.

[majestouch]

Like I said earlier I don't know the Costar/Metadot specifics.  All I am saying is Costar may just be a manufacturing outfit.  They are given the shell design, the electronics list of parts, the PCB layout and told to go build it.  

If Costar HAD engineering design expertise I don't know why they would screw up the Das Nkey and get the Filco's right.  I take that back, I know lots of engineers.....

We need a spy in Costar to get the true story.  It's probably somewhere in between.

:spy:

[itlnstln]

Mmm... Hawaii.

[alpslover]

Not to mention the input lag during my bursts above 120. I would literally see the letters come out all at once as an entire transposed word instead of letter by letter. It was really odd.

i've seen this happen on other cheap usb keyboards as well, although it's not as immediately noticeable because i find it hard to get a fast typing rhythm going on a cheap keyboard.  i have never seen this same input lag on any ps/2 keyboard i've used.

i'm pleasantly surprised that they've acknowledged this as a problem and are offering a full refund.

[sprintf32768]

i'm pleasantly surprised that they've acknowledged this as a problem and are offering a full refund.

   I'm pleasantly surprised also, but they are hardly to be applauded.  Was it barneyb.com's two-pencils-taped-together trick, posted the day before, that finally got DAS to admit the failure?  Or was it a wild impulse?  Who can say.  But they were hyping this premium keyboard for months, and they knew it couldn't even scan in-order.  Any claims they make about the Das-4 will be false until proven true.  

  There are quality keyboard manufacturers who we can trust.  What is Das bringing to the table that we can't get elsewhere?

[itlnstln]

There are quality keyboard manufacturers who we can trust. What is Das bringing to the table that we can't get elsewhere?

An easy-to-find source of Cherry blues.  Too bad the logic sucks and it's overpriced (IMO).  Otherwise, there are not a whole lot of places to get Cherry blues.

[itlnstln]

I'll stop now; I couldn't stop wine-ing last night, though.  Ugh.  Cherry blues are one of the only switches I haven't tried.  Cherry blues, reds and Topres are the only major switches I haven't tried.  There are some Mitsumis, Acers and few other less popular switches I haven't tried, either, but I don't care as much about those.

[o2dazone]

What board did you try reds from?

[itlnstln]

Cherry blues, reds and Topres are the only major switches I haven't tried.

They're one of the switches I haven't tried yet.  I am not too interested, though, because they are linear, and I prefer tactile switches.  I would like to try some Topres, but I want to find someone that has them, and check their 'board out first.

[o2dazone]

my reading comprehension sucks, thanks for the clarification

[itlnstln]

my reading comprehension sucks, thanks for the clarification

No problem, my writing sucks today, so we're even.