Neve 5104 - PSU Requirements for Running Two Channels?

[smilan]

Hi,
I have two Neve 5104 channels that were previously racked with a DIY PSU. Both channels are working, but the PSU gets extremely hot and sometimes there is a hum noise in both channels together (which makes me think this problem comes from the PSU).
Anyhow I don't have the schematics for the 5104 and I would like to build a new PSU.
Does anyone have the PSU schematics / have any experience with racking 5104s?
How much current draws each channel for each one of the supply voltages:
48v
25v +/-
16v +/-
15v +
?

 

[Bo Deadly]

Measure it. Find some resistor of a known value in the supply, look at the voltage drop and calculate I = V / R.

Or insert a resistor that is small enough to limit the drop to a 100mV or so and again calculate.

For higher precision, make your measurement resistor by paralleling a bunch of carefully measured resistors. Then insert that and again calculate. Derate 100% and those are your current requirements. But if you end up using SMPS (and you should IMO) derate only 20% (maybe less if you're confident that there won't be a large variation in draw) because SMPS work better under load.

 

[smilan]

I can measure it, but since the current PSU getting extremely hot I'm afraid that something there drawing too much current. Does it make sense or I can measure the current drawing with this PSU without any concerns?

 

[Bo Deadly]

I can measure it, but since the current PSU getting extremely hot I'm afraid that something there drawing too much current. Does it make sense or I can measure the current drawing with this PSU without any concerns?

That could be. But the channels are not going to draw more current because of a faulty supply. If you think there might be a problem with the channels then fix them and then maybe your supply is fine. If the channels are having a problem then obviously you would need to fix that first. It could be the caps are shot or there's a bias issue or something else.

But take your measurements anyway because either way you'll need to see if any change has an effect on current.

And measure each channel separately. If the channels are identical and yet there is a difference in current consumption, that is a strong indication that there is something wrong with one or both and maybe the supply is fine.

What has been done to the channels? Who racked them?

 

[smilan]

OK, I'll take the measurements and report back.
I received the channels from a friend who asked me for helping him to make the channels operating well. He complained that sometimes there's a stromg hum noise in both channels (at the same time).
I have the channels here for a couple of days and so far the only thing I've noticed is that the PSU tetting really hot.
I've lso tried to run the PSU with a variac and found the the phantom voltage reaches to +48V at 148V (the mains voltage here is 230V), so I'm gussing that the regulators working too hard and dissipating a lot of heat that in some point making problems?
I checked the frequency responce of the channels and it seems flat, so I can assuming that the coupling caps are still in good shape.
I don't know who racked those channels.

 

[smilan]

Those are the measurements I took.
There is a difference between the two channel at the + and - 24 rails.
Now how can I tell which one is wrong?

Voltage Rail	Both channels	Channel 1	Channel 2
-15.13	0.113	0.056	0.053
+14.98	0.482	0.243	0.241
-15.22	0.511	0.26	0.245
-24.56	0.065	0.023	0.043
+24.51	0.062	0.022	0.041
+48		0.003	0.003	* measurement taken with a Rode nt1

 

[Bo Deadly]

The difference in +-24 is interesting. An extra 40mA from +24 to -24 is almost 2W. But I don't see it being an obvious issue. The 0.5A between +15 and -15 is 15W. That seems a little high for a couple of channel strips.

How is ventilation? Post pics of the channels and the enclosure. Are there some kind of louvers or vents?

Do the channels exhibit hum when not in the enclosure?

Do the channels get hot?

Do you have schematics?

 

[smilan]

The difference in +-24 is interesting. An extra 40mA from +24 to -24 is almost 2W. But I don't see it being an obvious issue. The 0.5A between +15 and -15 is 15W. That seems a little high for a couple of channel strips.

There are two -15V regulators in this PSU (two different LM7915).
The one on the third row on the table drawing almost the same current (+/- 0.25A per channel) as the +15 rail while the -15V rail (on the first row) drawing 0.056A per channel.

How is ventilation? Post pics of the channels and the enclosure. Are there some kind of louvers or vents?

None of the above...

Do the channels exhibit hum when not in the enclosure?

The enclosure is made of wood and chipboard, I don't think it makes any affect on the noise.

Do the channels get hot?

Not really

Do you have schematics?

No, I wish I had...

 

[smilan]

I checked the supply voltages on the regulators and the voltage on the 24V rails seems to be too high, around +/- 42V at the inputs of the 7824 /7924 which is 2V over the max input voltage.
This can explain the overheating of the PSU, isn't it?
So according to the measurements I've took I'm thinking about order this PSU from JLM

 

[Walrus]

It looks like the two 12-0-12 transformers are wired in series with their centre taps isolated.. Hence you have 24-0-24 AC. There is also another transformer (i think) which feeds them. So possibly the 12v transformers may be 110v or something, and the other stepdown? transformer doesn’t have the right volts for them, hence why you have 42v DC before the regs. No wonder the box gets hot.
If the bias on the output card transistors is set too high on one of the channels, that would explain the difference in consumption on the 24 V readings.
The first -15 rail in your table is for the logic, so will be lower.

That JLM supply looks a little overkill!
I’d go with a +-24V SMPS and then regulate down for the 15 Volt rails.

 

[smilan]

It looks like the two 12-0-12 transformers are wired in series with their center taps isolated.. Hence you have 24-0-24 AC.

This is correct in theory, what is really happens is that those couple of 12-0-12 producing much higher output voltage (33ACV that rectified into 42DCV).

There is also another transformer (i think) which feeds them. So possibly the 12v transformers may be 110v or something, and the other stepdown?

No, the black transformer is 230V to 15-0-15.
Those voltages feeding the two 7915 and the 7815 for the 15V rails.

transformer doesn’t have the right volts for them, hence why you have 42v DC before the regs. No wonder the box gets hot.

Yes, that make sense

If the bias on the output card transistors is set too high on one of the channels, that would explain the difference in consumption on the 24 V readings.

Well I wish I had schematic so I could set the right bias.

The first -15 rail in your table is for the logic, so will be lower.
That JLM supply looks a little overkill!
I’d go with a +-24V SMPS and then regulate down for the 15 Volt rails.

I don't have too much experience with SMPS, can you recommends on good ones for this application?

Maybe I'll just replace the two 12-0-12 transformers with transformers with the right output voltage and see if it will fix those issues?

 

[Walrus]

Just had a quick look, and there doesn't seem to be any sensible +-24V SMPS's. So as you suggest, the best thing would be to replace the
two 12V trannies with one 24-0-24V.
Attached is the schematic for the 10882 board which has the Insert driver on, and has a trimmer to set the bias for the output transistors.
Unfortunately I don't have a factory line up sheet, I think when I racked some I just tweaked it so that clipping was symetrical.
There may be another output driver somewhere on board. Do you those strips have two toroidal output transformers?
The schematics I have belong to a desk that had all the output transistors off board in a card cage under the desk.

 

[trampo]

The original power supply was a Kingshill supplying +/-25V , +/-16V and -15V. If memory serves me correctly the +/-25 Volts is only for the eq module and the output cards with the pumpkin transformers on. You could disconnect the eq module and see if that helps.

 

[smilan]

Just had a quick look, and there doesn't seem to be any sensible +-24V SMPS's. So as you suggest, the best thing would be to replace the
two 12V trannies with one 24-0-24V.

I'll do that, I was looking at mouser and found those 24-0-24 transformers:

https://www.mouser.co.il/ProductDetail/Hammond-Manufacturing/266L48?qs=vcFyY%2BOcAyhaKWkHhqWN4A==
https://www.mouser.co.il/ProductDet...ormer/DP-241-8-48?qs=FESYatJ8odIRtNMctr0ZGA==
Does the Hammond having any advantages over the Signal?

Attached is the schematic for the 10882 board which has the Insert driver on, and has a trimmer to set the bias for the output transistors.
Unfortunately I don't have a factory line up sheet, I think when I racked some I just tweaked it so that clipping was symetrical.

Thanks, I can see this trimmer, this is a good idea

There may be another output driver somewhere on board. Do you those strips have two toroidal output transformers?
The schematics I have belong to a desk that had all the output transistors off board in a card cage under the desk.

The only transformer on the channels I have is a TF10015 input transformer.
There's an empty place for a transformer on the 10882 board.

The original power supply was a Kingshill supplying +/-25V , +/-16V and -15V. If memory serves me correctly the +/-25 Volts is only for the eq module and the output cards with the pumpkin transformers on. You could disconnect the eq module and see if that helps.

I'll try that.

 

[Walrus]

The Bel transformer should be enough, i don’t think you’ll need 4A for the 24v rails.

 

[smilan]

The Bel transformer should be enough, i don’t think you’ll need 4A for the 24v rails.

Can I go with this transformer?
https://www.mouser.co.il/ProductDetail/Hammond-Manufacturing/186D48?qs=mHL17HKVUJDWkq9%2BhHMIJg==

 

[Walrus]

Can I go with this transformer?
https://www.mouser.co.il/ProductDetail/Hammond-Manufacturing/186D48?qs=mHL17HKVUJDWkq9%2BhHMIJg==

Yes, that should be OK.