• Hello, Guest! Welcome back, and be sure to check out this post for more info about the recent service interruption and migration.

LC TDK power supply starts when it feels like it

Arisotura

Active member
I ordered some supplies so I can redo my soldering on my Mac LC motherboard the right way. In the meantime I figured I'd focus on the power supply, which is the typical TDK one.

Of course, it had crapoed capacitors, so I replaced all the output-side capacitors and cleaned up the mess. The power supply now provides correct voltages, but... only when it feels like starting.

Generally, if I mess with it for a while, it will happily start, but if it's been left alone for several hours, it won't start, just doing nothing at all, no clicks, nothing.

I thoroughly cleaned the board, desoldering some components to clean up some more capacitor gunk, reflowing all solder joints, but the problem persists.

Something interesting I found out while probing around with my voltmeter: the moment I probed between the microcontroller's OVP pin and its ground, that kickstarted the PSU. As if something about the overvoltage protection circuit was bad.

In the meantime I desoldered the controller, cleaned it and put it back. We'll see.

 

techknight

Well-known member
I was gonna say this board needs thoroughly cleaned alongside all the caps being replaced. 

Then if that fails, replace opto-isolator, shunt regulator, and the SMPS controller if necessary. 

 
Last edited by a moderator:

Arisotura

Active member
Lil' status update on this, after some more rework, it's finally working reliably. I have gotten some replacement parts if it ever shits itself again, tho.

 

Arisotura

Active member
Well, I spoke too soon. The damn thing is being flaky again. I replaced the microcontroller and the optoisolator, to no avail.

I could try going for the shunt regulator but I doubt that is the issue - when the PSU is in non-working state, there is no current at all in the output side. However, in that situation, the microcontroller is getting 13-14V instead of the normal 12V, which I think is what trips the overvoltage protection.

Mains current goes through R2, R3 and R4 before powering the microcontroller.

R2 is a black thing that looks like a tantalum capacitor, but the board says R2 with a resistor symbol, so... well. Says 16D-9 on it. It measures at around 16 ohm.

R3 and R4 are both 150k ohm resistors. I measured them, they give 149k and 151k respectively. I will try measuring them again after they cool down, to make sure they aren't drifting or anything weird.

I had already measured a bunch of other resistors on the input side and they measured good.

 

techknight

Well-known member
Check the film and ceramic capacitors. They arnt known for going bad, but they can. 

Also make sure the PCB has been washed and squeeky clean from all electrolyte or you will continue to have problems. 

 

Arisotura

Active member
Well, I did a few hunts for leftover cap gunk. Basically a visual inspection, but also heating up solder joints -- any bubbling through the liquid solder is a telltale sign of cap gunk.

I thought I had thoroughly cleaned the damn thing, guess I wasn't thorough enough, because I did find quite some leftovers. For each affected component, I desoldered it, cleaned it and the PCB, and put it back.

I'm going to see if it works better now, but not right now -- waiting on some new components, due to some idiocy on my part. If that doesn't do the trick, I will try replacing the shunt regulator, and checking resistors and capacitors and whatnot.

 

Arisotura

Active member
Lil' update on this. I replaced the few resistors I'd blown, still waiting for a replacement MOSFET (original one went bad too). I took a 2SK2611 MOSFET from another power supply, which is electrically compatible. The power supply now seems to work reliably, however the 2SK2611 is bigger than the original 2SK1024, so I can't quite put the thing back together. I'm waiting for a proper replacement (still).

 

Arisotura

Active member
Well, finally got the proper 2SK1024, and put the power supply back together. It seems to work really nicely now. So I guess in the end all is good... we'll see how this goes.

 
Top