RME Fireface 800 PSU Question

[smilan]

Hi, I have here a RME Fireface 800 with a faulty PSU.
There are no voltages at any of the PSU outputs.
I would like to test the unit with an external PSU before ordering a replacement from RME.
The problem is that I don't know what voltages I should supply and where each voltage should be connected to the main board?
There are two connectors connecting the PSU to the main board.
A two wires (white and red) and a four wires (red, black, yellow and blue).
The black and the white wires connected to chassis ground.
Does someone knows what voltage should be supplied to the other wires?

 

[Khron]

You could remove the power supply, and see if there are any markings on the board, near the roots of the wires. And/or trace them back to whatever electrolytics they may be connected to, for at least a hint of what voltage is supposed to be there.

I'd expect a +/- 6-18v to power the analog stuff, a 5v rail for the converters and/or digital stuff, a 48-50v rail for the phantom power, and that's about it.

Does the front panel power switch connect the mains? Or is it just providing a signal to the mainboard?

 

[smilan]

You could remove the power supply, and see if there are any markings on the board, near the roots of the wires. And/or trace them back to whatever electrolytics they may be connected to, for at least a hint of what voltage is supposed to be there.

I'd expect a +/- 6-18v to power the analog stuff, a 5v rail for the converters and/or digital stuff, a 48-50v rail for the phantom power, and that's about it.

I've checked that, there are no makings on the board.
According to the capacitors values the two wires connector is the phantom (the red wire connected to a 63V cap)
The red wire on the four wires connector should be the 5V rail (it connected to a16V cap)
The black wire is 0V
The yellow wire is the +18/15V (connected to a positive terminal of a 35V cap)
The blue wire is the -18/15V (connected to a negative terminal of a 35V cap).
There's a chance the the supposed to be 5V rail is actually lower than 5V and by connecting it to 5V I'll destroy something?

Does the front panel power switch connect the mains? Or is it just providing a signal to the mainboard?

The power switch is connected to the mains.
I've measures 230VAC at the input of the PSU, so the switch is not the problem.

 

[Khron]

Most digital things that work at 5v, there's a good chance they'll work a bit lower, sometimes "even" down to 3v or something.

 

[smilan]

Most digital things that work at 5v, there's a good chance they'll work a bit lower, sometimes "even" down to 3v or something.

So it will be safer to start with 3V and +/- 15V and rise it to 5V and +/- 18V if it won't work?

 

[Khron]

15 to 18 volts won't make any difference whatsoever.

I don't suppose RME were generous enough to label any sort of voltage test points on the mainboard, have they?

But 3.3v would be safe(r) to start with, for the digital supply rail.

 

[smilan]

I don't suppose RME were generous enough to label any sort of voltage test points on the mainboard, have they?

No, no voltages marks on the main PCB...

 

[smilan]

Thank you Khron, It seems to work with an external PSU.
I'll order a replacement one from RME.

 

[Khron]

I'd be quite curious what exactly kicked the bucket in the original power supply, though. It seems to be relatively common(?) for it to fail...

 

[beatnik]

the electrolytic capacitors are often bad with these... i fixed several of them and it was always the capacitors. it might be more economically viable to recap the unit with high quality long life capacitors than purchasing a replacement psu from rme, with the same low quality capacitors. you'll end up having the same problem a few years down the line

 

[Khron]

At least the ones on the secondary do look to be United ChemiCon ones, for the most part, which should be a good thing. But if the thing won't even start up, especially since no "skidmarks" are visible, i'd start suspecting the two little capacitors near the UC3842 controller, on the primary side.

 

[tinn]

as written in the manual, the unit needs some ventilation space.
Otherwise it gets too hot. The capacitors in the original PS are good
but life at 60°C+ with high ripple current is limited.
Very often the two 2200uF/16V go bad , sometimes the 470uF as well.
The replacement PS from RME is not expensive though (~100€).

 

[Khron]

Well, can't quite tell what brands / series of caps the stock unit uses, but especially if they're not the Japanese brands, i'd call bs on whatever lifetimes the datasheets might quote. But i'm just cynical like that.

I'd rather trust Japanese electrolytics that i've scavenged from various boards, than brand new ones from, say, CapXon or Elite or Aishi or any from an endless list of cheapos / nonames...

Even the lowliest low-esr caps from reputable brands are spec'd for a minimum of 2000h @ 105C @ full ripple current. Lifetime roughly doubles for every 10C below rated temp, so call it a minimum of 32000h @ 65C, assuming full ripple current, so worst case. That works out to about 3.6 years of 24h full-load.

 

[tinn]

Measured Voltages RME FF800 Power supply output :
+ 5,8 V red (Linear post regulators )
GND black
GND white (48V)
+ 15.4 V yellow
- 15.4 V blue
+48 V red

 

[smilan]

Measured Voltages RME FF800 Power supply output :
+ 5,8 V red (Linear post regulators )
GND black
GND white (48V)
+ 15.4 V yellow
- 15.4 V blue
+48 V red

Thanks, RME customer support was kind enough to reply me.

This is what they wrote to me:

Hope this helps:


Red 5,6V/3A

Black GND

Yellow +15V/0.6A

Blue -15V/0.6A



White GND

Red 48V/130mA

 

[Khron]

Measured Voltages RME FF800 Power supply output :
+ 5,8 V red (Linear post regulators )
GND black
GND white (48V)
+ 15.4 V yellow
- 15.4 V blue
+48 V red

But wait, since all those voltages are within spec... Is the power supply actually ok, after all? Or were these measurements taken with the power supply disconnected from the mainboard? Or.....???

 

[smilan]

But wait, since all those voltages are within spec... Is the power supply actually ok, after all? Or were these measurements taken with the power supply disconnected from the mainboard? Or.....???

I think that Tinn's PSU is working well while my PSU is dead...

 

[Khron]

Oh, right, sorry - didn't notice who wrote what... Can you tell it's friday afternoon, at the end of a sleep-deprived work-week? <sigh...>

 

[fragletrollet]

I brought my friends UFX back from the dead (symptoms were pretty wild with all the digital stuff going haywire, screen in las vegas mode and terrible noise on outputs) by replacing the electrolytics in the SMPS. Well known problem, either crappy caps or faulty PS design to begin with? RME didn't save any money on that choice...

 

[Khron]

I want dead expensive stuff too!...