8 Channel, 3U NEVE 1073 completed

[remsouille]

Do you have fuses on everything while you go?

Yup, it might be oversized though... what would be a good value for 6 channels with 3 acdc's at 230v?

 

[remsouille]

Okay, this thing is starting to drive me crazy. I spend everyday trying to figure out what's going on, but it seems I'm way over my head or I'm missing something really stupid and obvious. So I'll try to write down all the facts so you guys can have a clearer view of what's going on.

Symptoms: Four of the six channels exhibit noise and/or hum at random positions (I previously stated that they only occured in the 70-75db region, but I was wrong). Random means: one position will be quiet then, and noisy as hell 5 minutes later. The four channels affect each other, which means for example if channel 4 is on a quiet position, dialing channel 1 may induce some noise/hum in channel 4 output, even with channel 1 not being powered. Also, you can clearly hear pops when switching gain on other channels (even un-powered channels)....

My wiring (everything I'll write down here has been tested and double checked):
-Input xlr goes to input transformer with shielded wire (just like in the wiring guide). XLR,transformer, and grayhill switch are grounded to chassis.
-Trim (fader) pot goes to board with shielded wire, pot grounded to chassis.
-Board output goes to output transformer with shielded wire. Output transformer is grounded to chassis.
-Output transformer goes to phase switch with shielded wire and then goes back to output XLR with shielded wire (the two cables shields are connected together). Phase switch and output XLR are grounded to chassis.
-Every two boards are star-powered by one ACDC. ACDC's are bolted/grounded to chassis.
-+48v is disconnected/ruled out.

Something I think is strange (but I might be wrong) is that the resistance between one acdc's 0v to ground varies depending on how many acdc's are plugged. If only one is plugged, I read 10ohm, fine, when I plug the second, I read around 6ohm on both, and with the three on, it goes down to 4,6ohm. The acdc's are star-powered from an external 2x22v secondary through a 4-pin xlr grounded to chassis.

Everything points to a grounding/shielding issue, but I really can't see which! Go on, fellas, throw me something, before I pop my brains out!!

Thank's in advance

 

[sr1200]

random noise might be a cooked polystyrene cap no?

 

[madriaanse]

Some things to consider:

Sounds like all the 0V's are somehow, somewhere connected together between your different ACDCs? You don't want to do that.
XLR pin 1 is connected to 0V, but not chassis, right? Is output XLR pin 1 disconnected?

I'd suggest working on one channel at a time and get them straightened out one by one. Leave the other channels/PSU's powered off.

M.

 

[remsouille]

Sounds like all the 0V's are somehow, somewhere connected together between your different ACDCs? You don't want to do that.

You are absolutely right. The 0v rails ARE connected to each other, even with the preamps unplugged from the acdc's.
When I disconnect one ACDC from the transformer, resistance from 0v to ground reads 10ohm and continuity between the other 0v is broken, so I must have wired the acdc's to the transformer wrong, though I don't quite understand how to do it otherwise. Here is how it is now (sorry for the lazy-ass drawing)

Sill, I tried a pair of channels with one acdc and the two others disconnected, but unfortunately, the audio problems didn't go away...
EDIT:.......... I need diodes? ..........right?

 

[madriaanse]

That actually looks ok, but since the 0V's are connected together, disconnect 2 of the 3 ACDC's from earth/chassis.

The audio issue you're describing sounds like you may have input/ouput transformers too close together. Try experimenting with transformer placement. Magnetic lines of flux can be a real showstopper. Start with one unit (not a pair) and see if you can get it worked out (see FAQ at beginning of thread etc.).

M.

 

[remsouille]

That actually looks ok, but since the 0V's are connected together, disconnect 2 of the 3 ACDC's from earth/chassis.

Ok, will do!!

The audio issue you're describing sounds like you may have input/ouput transformers too close together. Try experimenting with transformer placement.

Just tried that, without success. As you advised I'm now experimenting with only one unit. All other acdcs  and preamps are unplugged. I posted a video on youtube, so you see what I'm experimenting:
http://youtu.be/zm_3NrSC5gM

 

[spase]

did you tried to replace those wires?

 

[remsouille]

did you tried to replace those wires?

Yup, these were just cut for the tests. Unless you mean using another brand/type of wire? Cause these are as good as anything else!

 

[dmp]

Seems like you have coupling / oscillating through the power supply rail.
Have you checked and tested your psu and made sure it is correct?  Do you have a known good psu to try, like a DC supply?.
Do you have the correct caps in the power supply decoupling positions? i.e. c4,c5,c7,c8?

 

[remsouille]

Seems like you have coupling / oscillating through the power supply rail.
Have you checked and tested your psu and made sure it is correct?  Do you have a known good psu to try, like a DC supply?.
Do you have the correct caps in the power supply decoupling positions? i.e. c4,c5,c7,c8?

The pres are powered by three ACDC's, delivering 24v and 48v as they should. Unfortunately, I don't have any other psu to run some tests, but I'm gonna more testing on the acdc's!

Thank y'all for chiming in and helping me not to shoot my brains out !
EDIT: I plugged one of the defective units into the acdc usually powering the 2 working units, it didn't make no difference. I compared the acdcs side by side, checked polarity and everything, they look just fine. I did the same between one working ez1290 and one non-working, checked every damn component, still nothing. The only difference is that the working are populated with bc184' and the non-working with mpsa18' (oriented at 180°).

 

[madriaanse]

Interesting. Thanks for the video; that's helpful. It's a bit hard to see, but it looks like in the video you are able to make the buzz/oscillation go away entirely by moving the power cable. Is that correct? It also looks like you're getting it at some of the lower gain settings. If that's the case, I'd take a serious hard look at the resistors surrounding the gain switch.

Can you compare transistor voltages on the good channels vs the bad channels and report back? Pls see first post on what the values should be.

Also, are you able to look at the PSU rails with a scope?

M.

 

[remsouille]

Can you compare transistor voltages on the good channels vs the bad channels and report back? Pls see first post on what the values should be.

Also, are you able to look at the PSU rails with a scope?

M.

Did that today, voltages are correct on both good and bad channels... Won't be able to scope anything till next week, BUT,
I've been doing a whole lot of testing and when I unplugged one good board and plugged it to the in/out transformers of a bad channel, it worked perfectly, SO: the problem definitely comes from the boards, which makes sense, since, as I said, the two good boards were not assembled at the same time. I must have made one single mistake when I assembled the 4 bad boards together.
BUT, I just doubled-(heck, tripled)check every-damn-thing, and still don't get it. I'll see if I have enough bc184 left to populate one bad board, who knows, maybe I came across a bad batch of mpsa18... Any other ideas?

 

[remsouille]

It's a bit hard to see, but it looks like in the video you are able to make the buzz/oscillation go away entirely by moving the power cable. Is that correct?

It is indeed correct, my good sir.

 

[madriaanse]

-All channels only hum on the 55, 60 and 65dB positions when plugged alone (I'll check again, but it seems to get randomy when several boards are plugged).
- I unplugged one good board and plugged it to the in/out transformers of a bad channel, it worked perfectly, SO: the problem definitely comes from the boards, which makes sense, since, as I said, the two good boards were not assembled at the same time. I must have made one single mistake when I assembled the 4 bad boards together.
BUT, I just doubled-(heck, tripled)check every-damn-thing, and still don't get it. I'll see if I have enough bc184 left to populate one bad board, who knows, maybe I came across a bad batch of mpsa18... Any other ideas?

Yes, 55dB is when the second pre-stage kicks in so you could potentially narrow your search there. On the PCB you will see two duplicate gain stages. The gain stage closest to the gain switch is the second pre-stage.  Can you clarify that the 55-65dB settings oscillate, but 70, 75dB work without problems?

I assume that by "all channels" you mean the four boards that are not working, right?

Can you post a closeup of the resistors surrounding the gain switch, and maybe a closeup of the second prestage?

M.

 

[remsouille]

Sorry Martin, I just edited my last post. Actually I found out that on another board, oscillation also occurs in the 20>40 range. It still is randomy in the sense that going back and forth on the gain switch, sometimes a position will oscillate and when you get back to it 2 seconds after it would not. But the 40/45/50/70/75 definitely wont oscillate.
I figured it would be the second stage, so I checked it a thousand times... Anyway here are the requested pics.

Just so we rule it out, here are the transistors I used http://fr.mouser.com/Search/ProductDetail.aspx?R=MPSA18RLRMGvirtualkey58410000virtualkey863-MPSA18RLRMG

 

[dmp]

You might verify that the tant caps are the right way. If their polarity is wrong all kinds of weird things happen.

 

[remsouille]

You might verify that the tant caps are the right way. If their polarity is wrong all kinds of weird things happen.

I did. They are.

 

[remsouille]

I swapped mpsa18' for bc184' on one board...works like a charm... 

 

[desol]

You must have oriented them incorrectly...or something.

I think the legs are backwards to the bc184 - reversed. You can't put them in the same way...

Best to use the bc184 > 500 hfe.

Less possibility for complications.