0
Please be positive and constructive with your questions and comments.

LED Issue?

by EvilXyzzy on Thu Jun 02, 2005 4:46 am

FInally escaped from work long enough to really start playing with the x0x and I'm seeing a weird LED issue.

In Pattern Edit mode the Accent and Slide LED's never seem to light. On powerup they light up with every other LED, but at least in pattern edit they dont seem to. Specifically if I switch to pattern edit, then hit Next for step edit mode, then type a note, hit next, type another, then hit accent, I dont get any indication, however if I hit Run, the notes are properly accented and slid.

Perhaps I do not yet grok its operation yet but I figured I'd ask.
EvilXyzzy
 
Posts: 19
Joined: Sat May 21, 2005 8:21 pm

by adafruit on Thu Jun 02, 2005 2:31 pm

hmm. that doesnt sound right. in pattern run-edit mode, do they light up? Does accent light up in Keyboard mode?

adafruit
 
Posts: 12151
Joined: Thu Apr 06, 2006 4:21 pm
Location: nyc

by EvilXyzzy on Thu Jun 02, 2005 3:56 pm

Accent doesnt light in keyboard mode.

If I run a pattern in run-edit mode, also no light. But I know the microcontroller can actually Light the LED because it gets lit at powerup.

I was hoping I'd made a mistake during assembly like soldering the LEDs in backwards, but no such luck.

Other LEDs off the same latches do light such as the DONE light.
EvilXyzzy
 
Posts: 19
Joined: Sat May 21, 2005 8:21 pm

by adafruit on Thu Jun 02, 2005 4:09 pm

just cause it 'lights at powerup' dont mean much, since the microcontroller hasnt actually done anything at that point. the synth is in a "unknown state" while it waits for the subsystems to stabilize.

Im not sure what to tell you. You could try swapping out those LEDs or checking that the voltage is correct or changing the latch chip. in hindsight,maybe i should have made a "test mode" that goes thru and lights each LED, performs a scale, etc. hmm.

adafruit
 
Posts: 12151
Joined: Thu Apr 06, 2006 4:21 pm
Location: nyc

by jcoplin on Thu Jun 02, 2005 4:52 pm

Try checking the voltage at pin 1 on IC8. It should go low when the accent led is supposed to fire. It normally is around -5v when the LED is off, it goes to close to 0 volts when the LED is active. If that is behaving correctly, it is either the LED, solder bridge, or a cold solder in R205. If the voltage isn't swinging it could be a bad chip or something downstream. If you socketed the chips, swap it for one of the other ones and see if that fixes the problem.

James
******************
If anyone asks of my whereabouts
simply tell them I've gone out the window
for a spot of tea and am not
expected back anytime soon.
******************
jcoplin
 
Posts: 18
Joined: Sat May 28, 2005 11:18 pm
Location: Minneapolis

by EvilXyzzy on Thu Jun 02, 2005 4:53 pm

And here I assumed you had included a "turn all LED's on at powerup" test...

I cant see how it'd be 3 bad LED's. Even if the cpu comes up in an unknown state, the latch can only output +5, tristate, or 0 ( I dont even know if those latches are tristate ) so the only way the LED could light incorrectly is if the ground they're on is floating or something. Maybe I do have a couple fried latches or something...
EvilXyzzy
 
Posts: 19
Joined: Sat May 21, 2005 8:21 pm

by EvilXyzzy on Sun Jun 05, 2005 5:14 am

HAHA!

So, I replaced the 595's in question. No change. I then pulled my hair out and drank heavily for a while. Then I came up with an idea. I pulled the CPU and Flash memory. Powered the unit up. Connected pins 8 and 32 on the CPU socket, applied +5 to pin 6, and then tapped +5 on the 8/32 pair, and all the LEDs light up one by one. Attach ground to pin 6 and I can turn them off one by one the same way. So I'm satisfied that they work and now I'm glaring at the CPU. Next stop is nabbing the firmware off Sourceforge and reloading the CPU using my handy-dandy STK500. Maybe I'll get real motivated and try a full build from source, but I'm almost out of good booze...
EvilXyzzy
 
Posts: 19
Joined: Sat May 21, 2005 8:21 pm

by EvilXyzzy on Sun Jun 05, 2005 5:48 am

Installed Winavr. Did a checkout of the latest source, built it, zorched it into the Mega162 and.. PROBLEM SOLVED!

Yay!

Now to put it back toegether and take it to bed with me.
EvilXyzzy
 
Posts: 19
Joined: Sat May 21, 2005 8:21 pm

by adafruit on Sun Jun 05, 2005 5:12 pm

bizaare! i guess we'll start doing more thorough verifications of the burned micros from now on...

adafruit
 
Posts: 12151
Joined: Thu Apr 06, 2006 4:21 pm
Location: nyc

by EvilXyzzy on Mon Jun 06, 2005 5:55 am

Other things it was doing before I updated the flash.
In Pattern Edit, if you hit r/s to play the pattern and rotated the tempo knob, it acted very slow ( large number of turns ) and would occasionally stop the pattern, dumping whatever was in scratch-pad at the time. This would also happen occasionally while it was just playing a pattern without being touched. I chalked it up to perhaps some crosstalk between traces or something and hadnt gotten around to hunting for a cause but happily it went away. I mention it in case there was a known firmware rev that had these quirks that the cpu might have been burned with.
EvilXyzzy
 
Posts: 19
Joined: Sat May 21, 2005 8:21 pm

Please be positive and constructive with your questions and comments.