Noisy Sunn Coliseum 300

[gbono]

I brought this amp back from being "air locked" but it seems to be very noisy. Not the 60/120 Hz kind of noise but more "white" or 1/f noise associated with semiconductors or resistors. All the active devices were replaced on the pre-amp board along with adding sockets. Power amp board also has new active devices as well. I haven't really been able to isolate the noise from either preamp or power amp since the power amp out is located before the CMOS limiter section and the preamp out "sounds" okay - hard to hear over the fan noise. Were these old (80's) Sunn amps just noisy?

[Enzo]

C1 on the power amp is its input, looks like pin 2 of the conector.  Don;t see why you can't scope there or use a signal traqcer there to listen.

Fan noise?  SO turn the fan off for a few seconds.  Or again, use a scope or signal tracer.


I am not sure quite what you mean about the CMOS limiter.   Whatever is before or after that point, it will be useful to know if the noise is present there.


And as always, what controls have ANY effect on the noise and which have NO effect?

[gbono]

I checked the noise levels with a scope at C1 and the reading was 1-2 mV of random noise. I looked at the test point 9 (base of Q4) and the noise voltage was still around 2 mV. The noise signal jumps up to 10-11 mV at test point 12 (13 dB!) I get the same readings on test points on the other rail. I have replaced ALL the active devices on this board so I'm stumped - is this just the going noise level for this design?

[J M Fahey]

Dear gbono.
The gain of that power amp is 120k/3k9=30 X. (or around 30dB)
So if you have 1-2 mV noise on C1 (the input) you *will* have 30X that (30-60 mV) at the output terminal.
I think the noise comes from an earlier point, so Enzo´s question remains: what control(s) , if any, affect the noise?

I only knew the "old", classic Coliseum 300 version; this one seems to be a later one where they "put all seasonings" at the same time: graphic EQ, compressor, parametric EQ, contour switch, even a DC servo in the power amp.
With so many add ons, sometimes it´s difficult to keep noise away.
Again: what controls affect noise?
Maybe you can find a setting that is not annoying.
Good luck.

[gbono]

Thanks JM,

   How did you get the differential amp gain?

   I had a "duh" moment after writing and posting the reply that I was just seeing the noise voltage multiplied by the comp stage(s). 

   The controls are not too effective at changing the noise level - the tone controls are set at the zero point (i.e. no boost or attenuation). There is some slight change in noise level with gain and master controls. I did socket all the ICs (maybe that's an issue?). I didn't chase the signal through the preamp board looking for a change in noise level but I did a "shotgun" replacement test of all the op amps/voltage amp with no change in output noise level. 

   One thing I have to say about this design - it can easily/reliably drive a 2 ohm load. Very smooth (it soft clips like a tube amp with tube rectifier) but a bit noisy at idle 

[Enzo]

The point of the controls settings was not that it would reduce the noise.  WHat we want to know is which controls do affect the noise and which do not.  COntrols that affect the noise in some way are FATER the source of the noise.  COntrols that have zero effect on the noise are either before its source or not in the signal path.

[Loudthud]

The CMOS stage is what gives the amp the soft clipping. Note how the -48V is used to derive the supply for the CMOS so the CMOS clipping level tracks the power amp sag.

What happends to the noise level when a cable is inserted into the Power Amp Input jack? That disconnects the preamp but it is before the CMOS stage so if the CMOS stage is the source of noise, the noise will still be high.

[gbono]

I cannot find any combination settings that has an effect on the noise level.

I connected a phone plug into the power amp input jack and the noise is still there. Assume it's the CMOS stage - I replaced the IC without a change in noise level. Any other suggestions?

Thanks.

[J M Fahey]

Hi gbono.
Now that we are narrowing it, yes, the two gain stages between the so_called power amp input jack (which really is not; the true input is, as Enzo noted, pin 2 connected into C1 on the actual power amp board) and the actual one are a regular Op Amp, IC ¿12A? at TP13 , with gain around 3X (1M/680K ) driving a CMOS clipper, IC ¿15? , 1/6 of a CD ¿4049? (quite poor unreadable scan), with gain around 3X (100k/¿39k??).
Have you changed the Op Amp? Can´t imagine a CD4049 being noisy there ... although it´s not impossible either.
As a side note, considering the Power amp sensitivity will usually be around 1V, your measuring 1mV there means a S/N ratio of 60dB, excellent.
Maybe that hiss is heard in a quiet room or at the shop; very much doubt it can be heard at all on stage.
It´s a 300W amp, meant for heavy metal players anyway.
Remember they were famous for using 18" speakers in folded horn cabinets, not much treble response anyway.
If you want to hear one of the "classic" Coliseum Bass amps, here you have a 1976 recording of Argentine Jazz Rock band "Alas".
They were using a Rickenbacker Bass.
Listen from 0:55 on, it´s clearly heard.
http://youtu.be/kiM_laQjFlM
Notice the bass sound is slightly "broken", these SS amps were used on the edge of clipping, to add some presence; the treble-less cabinets helped for the end sound not being too buzzy.

[gbono]

Thanks JM,

Yup that's a pretty bad scan of the schematics - if someone has something better it would be appreciated.

Thanks for pointing out the DC servo amp on the output board.

I'll look at that section of the preamp board again but I know those are all new opamps on there now. Funny that they only used one hex inverter off the MC14069? How does that clipper circuit work?

  Enjoyed the track - made me think of Flora Purim and Airto - back in the day   :dbtu: 

[Loudthud]

The CMOS clipper stage just softly pre-clips the waveform before it's sent to the power amp. The power amp does not have enough gain to clip the signal any harder. If the power supplys dip because of currrent draw or low line voltage, the supply to the CMOS also dips so the soft clipping is maintained. The trimpot R101 allows a slight adjustment to the CMOS supplys.