ARII R3 burning out under load

Bluxgore

New member

Donor 2024
Joined
Oct 31, 2024
Messages
7
Reaction score
1
Bench testing our ARII without load shows all good voltages. Under load from the game board or a 2ohm, 20 watt resistor is rapidly heating up R3; a 33ohm 1/4 watt resistor. It burned all the way out once and then fried Q1, the LM305 voltage regulator. After replacing both, we can watch R3 rapidly heat up with a thermal camera again. We're unsure where to go from here. R29 and R30 are the commonly listed symptoms of issues, but I haven't seen anything on R3.

Thanks in advance
 
Last edited:
When you bench test with a dummy load, are you jumpering R29 and R30? You have to.

That alone won't be your problem, since you say it's happening with a game board as well. I would check the solder and traces around Q2. You can get cracked traces and not see them.

Also, is there a silpad and nylon screw on Q2? Both are required. Q2 must be isolated from the heatsink.
 
are you jumpering R29 and R30
We've got 1+2, and 3+5 jumped on R7.
Also, is there a silpad and nylon screw on Q2?
Q2 has also been replaced with a new silpad and the old nylon screw.

We've replaced almost every part of the regulator at this point with an (almost) full rebuild kit. It's clear the board has failed through R30 in the past, so perhaps there could still be a burnt out part we skipped.

I would check the solder and traces around Q2
Will do! Thanks. We replaced Q2 at the same time we replaced R3 because we had it on hand, so the component itself I don't think is the problem but the traces could still be.

Any tips for checking traces across the board as a whole other than one-by-one continuity testing?
 
You just have to tone out each point-to-point trace. When you have a problem like this, replacing more parts is not the solution.

Where did you get your replacement parts?
 
All the resistors on the board tested at their expected values except for two. R30 because of temporary intentional bench testing modification for sense. It tested fine off the board. R7 was around 3k instead of 7.5k on the board, but it appears there are other connected traces and it also tested fine off the board.

We continuity checked every trace on the back of the board and didn't find any broken.

We are at a total loss of how to further debug the board at this point. If it could still be a short, what methods can be used to find it? Is there anything else we can check?
 
Last edited:
Are any areas of the PCB carbonized (burned) due to previous resistor failures? If so, the carbonized areas are conductive and can change the effective resistance between pads and traces - and present a current path where none should exist. If R29 and R30 have overheated in the past (or any other components), it is common to see both sides of the PCB carbonized in the immediate area. Parasitic resistance from the carbonized epoxy could be a factor here.
 
Did you clean your edge connectors on the game board? The tin becomes resistive, which drives the AR to push more voltage. That 1/4 watt resistor is intended to fail before the resistance overheats the edge connector.
 
Back
Top Bottom