Hey guys! I'm back and armed with an all new iTead Studio Logic Sniffer and my trusty 1960's Eico Oscilloscope!

you may remember that sign I built back in November/December that has huge signal issues, they were at the time really hard to disgnose without the proper tools. However even now, with the tools, I find myself needing help to figure out what exactly is wrong.

I have discovered via the oscilloscope numerous things:

1. there are 5-6v spikes for only a fraction of a 70hz wave ( I think anyhow, based on how I know how to use this old analog device) These causes particular columns of LEDs to remain dimly lit at all times. why? how?
2. The computer PSU I am using has ridiculous amount of noise in the power lines.

I have also discovered some oddities with my logic analyzer hooked up DIRECTLY the the arduino's output pins designated for CLOCK, DATA OUT, LATCH (Strobe) and Output Enable:

1. The arduino outputs seems to "bounce" much like a bounce of a button looks like & ONLY when powered by the computer PSU, However when powered by my laptop instead USB port, the signal is perfect.
2. The LED driving MIC5891 chips do not appear to receive those "bounces" from the arduino, the code runs the same on the sign whether it is plugged into the PSU or my laptop - with the exception of one thing: timing. The MIC5981s farthest from the clk/data lines will literally speed up (as if they ARE receiving those bounces)
3. There are single 5v spikes about 40nS wide every 10-20 seconds in every line AT the arduino's output when plugged into the PSU.

Next I was planning to check the data/clk/latch/OE lines at the first chip 12" away from the arduino, and again at the next chip, to compare logic graphs.

The oscilloscope is where I feel I need improvement, I really don't know what i should be checking for to determine if it really is my PSU that is making a lot of these issues.

I will get some graphs and pictures up soon as I get out of school. If you have suggestions on testing, I would gladly take them.

 

Well, I managed to match the impedances of the data clock latch and reset lines on the "F" side of the sign with pull-up resistors at the end of the lines for near perfect output!

But no matter what values I set the resistors on the "P" output control lines, I can only get the top half or the bottom half to work correctly, there is a finely divisible area where tuning of the clock line either hits or misses and only one half or the other will display perfect data, I can't seem to set my 15 turn 10kohm potentiometer to a good value.

I was reading the MIC5891 datasheet like a whore puts lipstick on. And after a while I noticed These chips work with TTL 12v signals as well (I shouldve known...)

This gave me the idea of setting VDD (logic input) to 12v, and controlling all the chips with higher voltage logic, therefore bypassing all the 5 and 6v spikes that show up on the oscilloscope!

Anyone else think this is a good idea?