Blown 22KΩ 5W resister on CS4 Board

[jjvornov]

I assembled my BeePre and got everything checked out after corrected a few misteps. But I continued to have a problem with one channel making noise after a few minutes and then cutting out. The LEDs on the B side of the channel lit on the bench when I tested it, but once the channel went out, they no longer lit. I removed both boards on that side and resoldered every connection without fixing the problem. I decided to try swapping the boards between the two channels to see if the problem followed the CS4 or PSU board.

When I powered up the unit after swapping just the CS4 board, the board in the bad channel gave a puff of smoke and it looks like the 22KΩ 5W has a blackened area and reads open with my meter. But after taking the resistor out it reads okay, it's the PN2907A that has a bubble and smoked out. So the board from the good channel is now blown and I'm no closer to figuring out where the problem is.

I might just need to finally send one of builds in for repair, but I'm open to suggestions about how to fix this. When it was working, the sound was great.



James

[jjvornov]

I replaced the transistor on the board that smoked and reinstalled it.

Going back to square 1, I started checking resistances again to see if I shorted something when trying to repair.

I'm getting only 19K Ohms to ground at 6, while I get 141K Ohms on the other, good channel. It seems that the resistance comes from the RC network on the PSU board, but all components are in place and I get the same 19K across the PSU board from -Reg to +Reg.

UPDATE: Further resistance checks show the R8 143K Ohm resistor checks out as 143K on the good side, but only 19K on the bad side. Identical resistors, so is it possible that a faulty resistor is the problem?

[Paul Birkeland]

You can heat up one solder joint on the 143K resistor and lift half of it off the board, then measure it out of the circuit. 

The power dissipated by the 22K/5W resistor is controlled by the available B+ voltage, which is measured at IA and a few other terminals. 

Blowing 2N2907's can occur from solder bridges on the PC board, miswires, backwards or incorrect transistors, etc. 

It would've been helpful to have a full measurement of DC voltages when the channel dropped, as that would've told us a lot about what exactly the failure mode might be.  Without that piece of information, it will be hard to progress on debugging the build.

-PB

[jjvornov]

Paul-

Thanks for the help.

So I pulled it out of circuit and does indeed read 143K. I put it back in and it reads 20K. What puzzled me is that the resistor on the good side read 143K and I saw the correct resistance at terminal 11, not at 6.

The voltage at 6 was 7.6V.  the other side was 146.  Another odd behavior was when I touched the meter probe to Y+, the LEDs went back on. The other odd behavior was when I powered down the preamp, the sound through that channel would come back for a second and then fade.  I thought it was mechanical and a bad solder joint, but I'be over the joints many times and don't see anything that looks less than solid.

I'm going to recheck that PSU board through the whole instructions as I must have miswired something. Good thing the Foreplay still sounds good and is happy to take its place back in the system for as long as it takes.

UPDATE: I checked parts identity on the power board and matched everything against the working side. I rechecked the reisistances and Everything checks except 6, C7 and C9 all of which read 19k, while the good side all of them read 140K.



James

[jjvornov]

I went ahead and powered up the preamp to get some voltages. On the good side, the LEDs all lit and voltages look good.

On the bad side I'm seeing:

1   5
2    140
3   0
4   9.77
5   0
6   146
10   150

22   192
34   12.5
35    6.2

[Paul Birkeland]

You should swap 300B's just to be sure the voltage behavior doesn't follow a tube. 

Just to be clear, you have swapped C4S boards, and now the board that was on the troubled channel remains working on the other side now?

I would double check your jumper connections that connect the board into the circuit.  Based on your voltages, your 300B is heating, but not drawing any current, or not being allowed to do so. 

Also double check that you don't have the board 180 degrees off, which will cause some unusual and annoying problems!

-PB

[jjvornov]

Paul-
I switched tubes with no change.

And you're right, I switched the CS4 boards and the one from the bad channel now works,  at least by voltage check and LEDs lighting.

Board orientation is correct and the wiring is correct. No LEDs lit at all.

Is it possible I fried some other components on the board other than the PN2907A on the one side. I'm tempted to swap back to test the board or maybe replace the other PN2907A. Does the low  resistance at 6 bother you? It's the only thing I find different between the two sides.

James

[Paul Birkeland]

You're getting the correct voltage at 6, which is the most important characteristic. 

On the offending channel, you can run the preamp with just the 300B in that channel and no 6BQ5.  This will cause the voltage at terminal 6 to rise, but we would be interested in seeing if the voltage at OB settles down with no 6BQ5.

-PB

[jjvornov]

So this morning on powering it up and changing nothing, I have 180V at 6 and 2. Kreg is 1.9V, So I'm not getting good voltages out of the PSU board it seems to me. Do you think that's the intermittent problem I had originally, now back?

With the 6BQ5 the 6 rises to 214, but 2 and 0b also move to 214.

Two problems here?

James

[Paul Birkeland]

Yeah, you may have 2 problems here.  With the 6BQ5 out, the regulator is disabled, and the 300B should draw OB down properly. 

Now you have 1.9V at Kreg, which also suggests that the regulator is not working properly (which can be a symptom of an issue with the "A" side of the big C4S board).

Can you post photos of the offending side of your build, including the tops and bottoms of the power supply board, as well as the wiring under the C4S board?

-PB

[jjvornov]

Here are some photos of the bad channel. The C4S board was working in the other channel, but I messed up on the switch and it gave a puff of smoke. I see a dimple in the B side 2N2907 so I wonder if its bad too. I only replaced the A side since it was clearly damaged.

The PSU board looks identical to the working side, but the resistances don't match, so I'm convinced something odd is going on there. I've noticed that when I measure the resistance at 6 to ground it starts off high (0L), goes down to below 8 KOhm and it eventually settles at 19 KOhm. The good side settles right at 140KOhm.

James

[Paul Birkeland]

On the "B" side of the C4S board bottom surface, where the MJE5731A mounts, it looks like some of the copper trace is gone, which would cause some issues.  (BTW - the 300B not drawing any current will throw off your Kreg voltages up on the power supply PCB)

You can restore this connection by using a bare piece of wire to attach your current solder joint to the empty pad on the same trace.  I've made a little picture to assist you.

-PB

[Doc B.]

Blowing that image up, it looks also like there might be a lot of flux and not enough solder on the pin associated with that trace.

[jjvornov]

Tried resoldering and the jumper- no change in the voltages. Actually, it checked out by continuity testing with the multimeter as well. It's something I've become habitual about- checking continuity across the build as I go along.

Any other ideas?My only other idea was to switch the good boards onto the bad side one by one and see if the wiring outside of the boards was okay. Since I have one side working.

James

[jjvornov]

Just to close the loop for the board for now, I'm going to send this in for repair. I could rebuild both boards, but thinking about the time and the original intermittant nature of the problem that I couldn't find, I've decided that it's best to get the experts to look at it.

James