Killing CPUs

[olivier.jan]

Over the past few days I’ve faced strange behaviour from a H5 and from a traditional STF, both using originals 68000 in DIP format.
I finally identified the issue this morning, testing both CPUs in a known working Amiga 500 and they are both totally dead.
I was wondering what could have killed these 2 CPUs, it started with random crashes until it didn’t boot anymore.
I haven’t been using original PSUs but bench power supply on these motherboard, with input voltage at 5.13 V. This gives an average 4.6v at the CPU VCC but I saw some peaks at 5.28V due to ripples. Are these above the tolerance of the 68000, should I lower the voltage of the bench PSU to 5.00V?
I have one 68000 left, I’d prefer not to kill it.

[exxos]

Voltage doesn't sound right.. 2 layers are fully used for VCC and GND.. So if you have 5.13 input.. you would have exactly the same across the board.

[olivier.jan]

I checked again with power supply at 5.13V:

• H5 #1 - PLCC CPU : 4.75V on the 68k socket.
• H5 #2 - DIP CPU: 4.70V at the CPU.
• STF with 2k2 resistors on the bus - 4.60 at the CPU.

Now that you mentioned it, I changed the resistors on the bus of the STF last week...

[exxos]

I checked again with power supply at 5.13V:

• H5 #1 - PLCC CPU : 4.75V on the 68k socket.
• H5 #2 - DIP CPU: 4.70V at the CPU.
• STF with 2k2 resistors on the bus - 4.60 at the CPU.

Now that you mentioned it, I changed the resistors on the bus of the STF last week...

It sounds like you have a bad PSU connector ?

[PaulJ]

If you had a 5.13 VNC average but seeing ripple up to 5.28 with ripple that’s an awful lot of ripple. Also you have a significant variance of dc voltage. Do you have any idea what current the current draw is for the board? You may have something again to a near short some where. Have you tried a different supply? Is anything getting really hot?

[olivier.jan]

Ok, just tested with a Meanwell RD-35, and the voltage is clean at 5.06V on each machine.
Concerning the current draw, using the bench PSU it shows on average 1.5A depending ont he machines.
No parts gets hot, and the machines worked properly for a while before they died.

[PaulJ]

As I recall approximately 1.5 amps is normal for a board only. Hopefully you won’t have any additional cpu losses.

[olivier.jan]

So I finally found out what the problem is: the cable between PSU and the Atari. I guess with an Electricity or Electronics background I would have avoided such a stupid mistake.

Here is a picture of the cable, who can spot the issue ?

IMG_5906.jpeg (106.67 KiB) Viewed 2970 times

[olivier.jan]

So I finally made a new cable using "standard" 1.5mm2 copper wire instead of the AWG 25(0.1624 mm2) and the voltage drop is now more acceptable with a steady 5.06V at the CPU when the PSU is set at 5.13V.
The issue was clearly coming from the cable section I was using. I checked using the standard formula : (P x2LxI)/S with my 1m copper cable using AWG 25 and, spot on, this gives 0.4V-0.5V voltage drop depending on the Current draw(1.5-1.7).

Now with a 1.5mm2, the voltage drop is divided by 10 at 0.04V-0.05V.
I shouldn't be killing too many more CPUs hopefully.

[exxos]

I'm not sure I follow how you got for a bad mouse cable to blowing up CPUs ?