Keyboard - every keystroke is double

Get help here getting your replica 1 going

Keyboard - every keystroke is double

Postby boolleon » Oct Wed 15, 2008 6:41 am

I have a question about the Replica 1 TE revision 1. Soldered it myself wich was great and all seems to be working fine. The thing is when I connect several keyboards to the PS/2 connection all keystrokes ar double qqwweerraassdd This happens with Microsoft/Compaq/IBM keyboards. Do I need specific PS/2 keyboards?

Added is the answer I just got from Vince Briel.

shouldn't need any specific PS/2 keyboard. Are all of your keyboards you are trying giving you this action? If so, check solder connections on 6821 and PS/2 connector.

I will check this and give an update.

A friend of mine has the exact same problem with his self soldered board.
boolleon
 
Posts: 3
Joined: Oct Wed 15, 2008 6:35 am

Postby vbriel » Oct Fri 17, 2008 4:38 pm

I think I have an answer for you. Another user sent their board to me with almost the same problem. I have been working on this for a couple of nights and finally got it to work correctly. It wasn't bad chips, bad resistors or even a bug in the firmware, it was the solder.

I'm still testing this board but here's what I found. The user built theirs and either added flux or it was built in the solder. AVOID these types of solder because the flux is causing issues with the clocked keyboard strobe. I don't have the exact answer why, maybe the board I worked on just needed cleaning up but the type of solder you use is important.

I personally use RadioShack .032 8.0oz Rosin-Core solder cat # 64-009 that is low temp and doesn't spred flux all over.

Remember I will fix your board for free (return shipping is all I ask) but I would like you to try to fix it on your own before sending me your board. Also, don't just send in your board, email me first so I know what is going on.

So, check your solder job, clean the solder side of your board and use good solder. Let me know what you find.

Vince
User avatar
vbriel
Site Admin
 
Posts: 1184
Joined: Jul Tue 19, 2005 12:10 pm
Location: Ohio

Postby vbriel » Oct Sun 19, 2008 8:24 am

Upon futher testing I found that if I put a 3.3K resistor on the strobe signal to ground (pull down resistor) it appears to fix the issue as well. Funny that some boards have this issue while others do not.

Easy fix, grab a resistor 3K-10K and put it on pin 10 and 11 of the 74LS245. That is across the bottom 2 pins on the chip and this should fix your problem.

Try this and post back here.

Vince
User avatar
vbriel
Site Admin
 
Posts: 1184
Joined: Jul Tue 19, 2005 12:10 pm
Location: Ohio

Double character output serial port or keyboard input

Postby sbenoit » Oct Mon 20, 2008 7:24 am

After a few weeks of trouble free running this started happening, both keyboard and serial port:
<---snip --->\

FFF000000000RRR


F000: A9
KRUSADER 1.3 BY KEN WESSEN
? NNN

000
<--- end snip -->
Notice that the system vectors to the Krusader? It also will allow BASIC Programming:

<--- snip -->EEE000000000RRR


E000: 4C
>>111000 PPPRRRIIINNNTTT """IIIMMM BBBRRROOOKKKEEENNN"""


>222000 NNN

*** SYNTAX ERR
>>222000 NNN

*** SYNTAX ERR
>>222000 EEENNNDDD

>>LLLIIISSSTTT

10 PRINT "IM BROKEN"
20 END

>RRRUUUNNN

IM BROKEN

>>
<--- end snip --->

[/quote]
sbenoit
 
Posts: 5
Joined: Oct Sun 19, 2008 8:50 pm

Postby sbenoit » Oct Mon 20, 2008 4:03 pm

update:

I placed the 10k resistor on the 74ls245. No change.
Cleaned board and touched all solder points. No Change.

Burned up a new EEPROM replaced 2764. No Change


Measurements:
@ R8 R20 junct.
My O-scope shows a -480 volt spikes with keyboard plugged in.

the baseline voltage is 88mv
How much noise can the U8 deal with before it thinks it triggered?

Removed ATX supply

Place system on a 9 vdc radio shack supply.
Noise less but present.
Problem of double key went away, for now.

Will keep you posted .

@ R8 R20 junct.
My Oscope shows a -480 volt spikes with keyboard plugged in.

the baseline voltage is 88mv
How much noise can the U8 deal with before it thinks it triggered?
sbenoit
 
Posts: 5
Joined: Oct Sun 19, 2008 8:50 pm

Postby sbenoit » Oct Mon 20, 2008 4:04 pm

Correction:

-480 mV dc spikes


Sorry not -480vdc

What a goof!!!
sbenoit
 
Posts: 5
Joined: Oct Sun 19, 2008 8:50 pm

Postby vbriel » Oct Mon 20, 2008 4:36 pm

480V would be something. Not sure on the prop what kind of spikes it can take but 480 shouldn't register. You said the problem is gone with the wall supply but it is there with ATX. Any noise on the 5V or 3.3V when ATX is used?

This looks like a firmware issue. Do you have a serial cable, you can try and reprogram the EEPROM onboard using the parallax software. There's a copy on the CD that comes with the TE.

Vince
User avatar
vbriel
Site Admin
 
Posts: 1184
Joined: Jul Tue 19, 2005 12:10 pm
Location: Ohio

Postby sbenoit » Oct Fri 31, 2008 2:01 pm

Vince,

Good news:
1. after re-soldering and re-seating u9 the replica1 TE has been running for two weeks without issue. also spread the pins on U9 to give a "tight" fit in the socket. As a precaution I did reprogram the eeprom.

2. The external computer power supply connected to the ATX seems to be defective. It's a dumpster dive find. way to much noise

Summary:
Check and recheck solder work and make sure you have good pin to socket connection for U8 and U9... oh and don't use a dumpster dive power supply

Thanks Vince
sbenoit
 
Posts: 5
Joined: Oct Sun 19, 2008 8:50 pm

Postby boolleon » Nov Sat 01, 2008 4:42 am

Soldered a 10k resistor as Vince recommended, but didn't work here as well. (tried a 3.3k resistor as well)

Will try the solution offered by sbenoit and will report back.
boolleon
 
Posts: 3
Joined: Oct Wed 15, 2008 6:35 am

Postby vbriel » Nov Sat 01, 2008 7:24 am

I am concerned about a trend growing here but lets see what happens. Bummer about the dumpster power supply, probably why it was there anyways.

The propeller is a 3.3V device that I have attached to a TTL (5V) circuit. The rules for doing this with little hardware is grey but I thought I had it resolved.

Besides, the 24LC256 EEPROM, Propeller chip, 6821, also check the 74LS245 because that acts as a buffer between the PS/2 keyboard and the propeller.

Sending TTL level signals into the proepller isn't a problem as the designer says "It can handle it".

Try different things for results, keyboard, power supply, etc. I just fixed one that required the resistor, so I may make this a revision change next board run. I'm running tests now to see what else I can try or do.

Funny thing is all the boards I've built don't have this issue, never have. I've wondered if the type of solder is causing this issue. I recomend .032 dia. 8 oz or thinner 60/40 Rosin-core solder. I get mine from Radio Shack because I always seem to run out on a weekend.

Vince
User avatar
vbriel
Site Admin
 
Posts: 1184
Joined: Jul Tue 19, 2005 12:10 pm
Location: Ohio

Ringing?

Postby mfortuna » Nov Mon 03, 2008 6:57 am

The 74LS245 has a stong output stage. The edge rate may be enough to cause ringing which in turn may cause the strobe signal to be sampled multiple times. A noisy power supply could add to the ringing.

You could implement a digital filter in the Propeller chip. When you see a change in the strobe signal, wait some amount of time before you sample it again. Maybe 50 msec.

Another experiment could be to lift pin 11 on the LS245 and put a 33 or 49 ohm resistor in series between the pin and pad.

Mike
mfortuna
 
Posts: 49
Joined: Dec Wed 12, 2007 2:30 pm

Postby boolleon » Nov Mon 10, 2008 11:34 am

I have tried a 9v dc power supply. Still the same problem. Also tried different keyboards. The best result I got was from a microsoft ps2 keyboard. Every 3rd/4th keystroke is double.
boolleon
 
Posts: 3
Joined: Oct Wed 15, 2008 6:35 am

Postby mfortuna » Jan Mon 26, 2009 7:28 am

Was there any resolution to this issue? Just curious.

Mike
mfortuna
 
Posts: 49
Joined: Dec Wed 12, 2007 2:30 pm

Postby vbriel » Jan Tue 27, 2009 7:53 pm

I ran into this again today and this time the resistor did not fix it. I tried fudging the code and nothing fixed it. I replaced the 6821 and presto...

This can happen, Jameco carries them as well as Unicorn electronics.

http://www.jameco.com

http://www.unicornelectronics.com/IC/6800.html

Vince
User avatar
vbriel
Site Admin
 
Posts: 1184
Joined: Jul Tue 19, 2005 12:10 pm
Location: Ohio

Postby Ckelly » Mar Sun 01, 2009 5:39 pm

I have a new Replica 1 TE, and I ran into a similar issue. But in my case the current keystroke is not immediately doubled, but rather the previous key is displayed (again) followed by the current key.

So for example if you type 'ABC' the following happens:

After 1st key, 'A', is pressed:

A

Then after 2nd key, 'B', is pressed:

AAB

Then after 3rd key, 'C', is pressed:

AABBC

So my best guess for what is happening was that the STROBE line is somehow going high to the 6821 just before it gets the new key data, causing the 6502 to read the old key data first, followed immediately by the new key data.

Vince suggested putting a resistor between the STROBE and GND, and putting a 3.3k resistor there did resolve the issue.

==> But I'm really struggling to understand what exactly the problem was, and how this solution works?

I'm guessing it has something to do with either:

(a) the relative timings of the /OE (pin 19), STROBE (pin 11), and key data (pins 12-18) on the 74LS245,

OR

(b) possible "open/floating" ("high impedance") of the STROBE line (even though I measure it to be ~0v) before being explicitly set high by the Propeller chip on a keystroke.

If anyone knows or even has any guesses of exactly what could be happening, I would love to hear.

Thanks,
Christiaan.
Ckelly
 
Posts: 9
Joined: Feb Sat 21, 2009 7:34 pm

Next

Return to Troubleshooting

Who is online

Users browsing this forum: No registered users and 2 guests

cron