The symptoms since a firmware upgrade (and downgrades, re-upgrades…):
- strange characters on the display for SEC/DIV
- system freezes
- +/- buttons are ignored (but still work physically)
- the ADC chip doesn’t get a CLK signal during a freeze
My theory:
The ATmega64-16AU is overclocked to 20MHz, which is a good reason for failure. But all the other KITs seem to work, so maybe I have a CPU that is a) really faulty or b) doesn’t tolerate 20MHz or c) both.
Plan:
- Get a replacement CPU (ordered)
- Get a chipquik set (ordered)
- Scope the supply voltage lines during a crash for anything odd
- Disconnect the lines to the ADC and see if it still crashes
- If so, inject external CLK to the ADC
- Scope the data lines and see if it works during a freeze
- If so, ADC is OK. Replace the CPU and hope it fixes the problem
- If the problem persists: salvage parts and TRASH the rest
A somewhat updated plan:
- Get a replacement CPU (ordered)
- Get a chipquik set (ordered)
- Scope the supply voltage lines during a crash for anything odd
- Disable the ATmega64 by setting its reset pin to LOW
- Use a Boarduino to supply a clock signal to the ADC (e.g. Timer1 interrupt)
- Read data and transmit to my PC and see how long it runs without a crash
By keeping the ATmega64 in reset-state permanently, it should be disabled and I’ll see if the ADC does funny things on its own. If this setup runs for say an hour without issues, it should be quite sane to assume that the cpu is the cause of all evil here.
→ Update 1: 23.07.2009:
I’ve received a free replacement DIY KIT from Jyetech today. After assembling it, it was at first acting a bit strange, but that somehow resolved itself (at least last time I checked). This time I powered it with a regulated power supply at 9V DC. The new one draws a current of 240mA (-080 firmware), the older KIT just 190mA (-070 firmware). The 5V voltage regulator just gets a bit warm. Sometimes button presses seem to be ignored, maybe I need to adjust the way I press them. There is no hardware debouncing, it seems to be done in software. This time the front panel was cut to size properly and the LCD fit snugly without modifications. I will do some stability tests later. I guess I will compare my big power supply vs. connecting it to my PC with a USB/TTL adapter. The test will basically consist of me trying to make the scope crash or freeze by continuously working with it, switching timebase and modes. First I will compare the new KIT with the latest -080 firmware against the old KIT with the -070 firmware. Then I’ll upgrade the old KIT to -080 and test it again.
That’s all for now.
Any comments from jyetech scope users and further ideas for testing and hopefully fixing it are highly welcomed.
