gz_test troubleshooting notes

So I did the "sudo ./gz_test".  The gz_test seems to run, but I don't know which header pins to need to check.  Then looking at the board, I found very small letters saying FB1, FB2, etc.

I used a scope to check the FBn pins but disappointedly found all pins are dead zero volts.

Then I remember that the program should have used the RPi's clock pin somewhere.  I have not yet played with the clock and PWM pins yet.  So I need to check the header pin map again.

I read it and found that the core thing taks the clock input RPi GPIO4.  Now I have a problem.  Because I only extended the JTAG signals from the RPi chassis to the CPLD board.  I now need to open the chassis and get the GPIO 4 signal out.  It is annoying because it was the first time in months that I closed the chassis a couple of hours ago, and I need to open it again!

I thought about using the 1K Hz signal from my cheap multimeter or the socpe, but then remember that they should be 5V logic and so it is not a good idea, though I remember that this CPLD is 5V logic tolerant.

I think I better call it a day.


.END


gz_test waveform - Guzunty Pi

https://github.com/Guzunty/Pi/wiki/gz_test-waveform

No need for a waveform for this one.

Use this core to test your newly built Guzunty Pi SB. Although very easy to build (we have not seen a failed board so far), it is good to be able to test that it all works before using it with a project.

This core takes a clock input from the Raspberry Pi GPIO4. It then divides this clock down into different frequencies dividing by powers of 2. We use 8 different powers of 2 and output each on a number of output pins. This produces a square wave output on every pin of the CPLD (except of the clock input of course). There are four groups of 8 pins. Bit 0 of each group divides the input clock by 2^13, up to bit 7 which divides the clock by 2^20. There are a total of 34 IO pins on the CPLD. We test 32 of the pins using the 4 groups of 8 pins, 1 pin is dedicated to the clock input, so we would see no output at all if there were something wrong with this pin. That leaves just one pin remaining, which we test by passing a divide by 2^20 signal to it.

These divisions are selected, together with the input clock frequency, so that the device may be tested with minimal equipment. The slower signals can be seen on a multimeter. The faster ones may be tested using an light emitting diode with series resistor with a value around 90 ohm. If you have an oscilloscope, so much the better. You can check each output has a nice clean square signal output.

.END