High Frequency Oscillation Related to Presence Control - Carlsbro 60 T.C

[Grantorino]

Hello EL34 World.

I am working to restore this vintage amplifier - a Carlsbro 60 T.C.

Following the schematic I obtained from this site, I have fully recapped, replaced the input jacks, installed a new NFB resistor, new screen resistors and installed 1ohm cathode resistors for bias measurement. All the voltages are good and the bias is set around 35mA per tube.

The amp passes signal on both channel 1 and 2 - this amp has two inputs (assumed hi/lo) per channel. The clean channel 1 works great at volume. The higher gain channel 2 produces a squeal/high-frequency oscillation when the presence control is set anywhere past midway.

This amplifier does have a 'limiter' control and also a small trim pot on the board, secondary to the bias trim pot. The Limiter control can tame this squeal and allow the presence to be turned up fully. I understand that it limits the high frequencies entering the phase inverter. 

I understand the trim pot to control negative feedback stability, affecting the feel and distortion of the amplifier. It's set to roughly 40% from the fully CCW currently, which I found to be most stable.

Suspecting grounding issues, I tightened the presence pot to the chassis and the threaded shaft broke. I am going to replace the presence pot. Temporarily, I fixed a ground wire from the volume pot to the chassis which made no difference.

Switching V1/V2 made no difference.

The amplifier had this squealing issue before any restoration effort was done, whilst also sounding terrible on both channels, with two jacks not working and it was also red-plating and emitting smoke. There are some slight differences in stated values on the schematic vs parts fitted - for example, schematic calls for a 3kpf cap, I fitted a 3.3nf.

I'd appreciate advice from anyone with experience of similar issues.

Thank you.

[AlNewman]

What they're calling a presence control on the schematic isn't what is normally considered a presence control.  The limiter control would be closer, as it is tied to the NFB.

It does make sense that the 'presence' could add oscillation on channel 2, because it's attenuated at channel 1 by the 100k resistor in series.

If the limiter is a trim pot inside the chassis, and you've found it can make all the demons go away, then maybe somebody at some time twiddled the trim pot, and only the limiter needs to be adjusted to make the amp work as intended.

[Grantorino]

Thanks for the reply Al — just to clarify a few points so we’re on the same page:

• The “Limiter” on this amp is not an internal trim pot. It is a front-panel pot, factory-installed.

• Turning up the limiter audibly reduces treble and reduces overall level — basically a variable HF clamp into the phase inverter.

• Separate from this, there is an internal trim pot on the PCB, but this adjusts negative feedback stability / damping, not the limiter function.

• Adjusting that internal trim pot helped, but it does not fully remove the channel-2 squeal. Channel 1 can be maxed out cleanly, but channel 2 still oscillates when the presence control is above halfway.

• Presence at zero = stable. Presence past noon = oscillation. Limiter pot can tame it, but obviously that reduces the usable range.

I appreciate the insight. I’m now looking more closely at the 100k feeding the PI, as you mentioned — that makes sense, because channel 1 is attenuated by the 100k mixer resistor, but channel 2 is “hotter” going into the PI and more likely to trigger oscillation.

If anyone has experience with the Carlsbro 60TC PI design or common oscillation points, I’d appreciate your insight.

[pdf64]

Thanks for the reply Al — just to clarify a few points so we’re on the same page:

• The “Limiter” on this amp is not an internal trim pot. It is a front-panel pot, factory-installed.

• Turning up the limiter audibly reduces treble and reduces overall level — basically a variable HF clamp into the phase inverter.

The schematic doesn't show which way is 'up'; I assume that the pot's CW track end is 'up', connected to ground?
I also assume that rotating the wiper such that the wiper (and LTP grids) were at ground would mute or at least massively reduce the signal?

• Separate from this, there is an internal trim pot on the PCB, but this adjusts negative feedback stability / damping, not the limiter function.

• Adjusting that internal trim pot helped, but it does not fully remove the channel-2 squeal. Channel 1 can be maxed out cleanly, but channel 2 still oscillates when the presence control is above halfway.

The trimmer on the V3 (labelled V2 on schematic) direct coupled cathode follower DCCF just looks to adjust the overall gain / level / volume.
It doesn't look to affect any negative feedback circuit.

• Presence at zero = stable. Presence past noon = oscillation.

The presence control is a bit weird / unusual, as it includes a negative feedback loop around the V3 DCCF.
It's not obvious to me which track end is 'zero', please could you advise?
With no negative feedback loop, it would be end connected to the grid, but with feedback, the grid becomes a low impedance virtual earth.

... I’m now looking more closely at the 100k feeding the PI, as you mentioned — that makes sense, because channel 1 is attenuated by the 100k mixer resistor, but channel 2 is “hotter” going into the PI and more likely to trigger oscillation.

The phase splitter is V4 long tail pair, but you seem to be referring to the 100k at the V3 DCCF input grid.

If anyone has experience with the Carlsbro 60TC PI design or common oscillation points, I’d appreciate your insight.

Overall, the schematic seems to have been designed by someone wanting to try out some different ideas.
As those ideas don't seem to have been used in any subsequent designs, it might be inferred they weren't well received.

Generally, oscillation is typically due to poor schematic implementation, how the schematic has been made into a real world thing.
I've no idea how that been done, so lots of good photos would help us to help you.
They should show the overall layout inside and outside the chassis, and then a bit closer up, showing the wiring from a couple of viewing angles, of the left, middle and right hand ends of the inside of the chassis.
I suggest you upload them to a hosting site eg imgbb, and provide links.

What controls affect the oscillation, eg with presence at 0, is the amp stable even at any setting of the treble / response / volume controls?

And with the presence set high, does the oscillation stop at particular settings of the above controls? Or is the oscillation solely determined by the presence setting?

[Grantorino]

Thank you for the detailed response pdf64.

To provide some clarifications you asked for:

With the presence control at zero, the amp is completely stable on both channels at any setting of volume / treble / response / limiter. The oscillation is introduced when turning the presence control clockwise anywhere past midway. If the presence is set midway or below, the amp sounds good on both channels, at any EQ setting and at any volume.

If I lift one leg of the 3n3 cap between the presence wiper and the limiter pot, the oscillation disappears completely. Presence has no audible effect, which is expected as it's removed from circuit. So oscillation is dependent on the presence network. Before the recap, the oscillation was present.

Limiter control - Yes, turning the limiter pot clockwise (toward ground) kills highs and level. This also suppresses the oscillation and will allow the presence control to be turned further CW, but the range of presence is reduced.

The internal trim pot does change the feel/gain, but at no setting does it eliminate the oscillation if the Presence is above halfway. When I first started restoring the amp, this trim pot was set to around 60-70% CCW (from the rear of the chassis). Right now, I think the amp sounds best (feel, articulation) with this trim around 30-40%.

Therefore, only the presence control directly introduces the onset of oscillation.

Before digging deeper, I’m going to replace the presence pot (the threaded shaft snapped) and reassemble the chassis. With the chassis out of the cab, placing a metal panel over it noticeably reduces the oscillation, so I’d like to test it with the cab’s original shielding in place.

If the amp ultimately needs the presence kept below halfway to remain stable, that’s acceptable — but I’ll come back with photos of the wiring and layout once the chassis is refitted, in case any lead dress improvements would help.

Thanks again for your help — I am beginning to suspect the presence circuit in this design might be inherently touchy.

[pdf64]

...
If I lift one leg of the 3n3 cap between the presence wiper and the limiter pot,  ...

I don't understand what you're describing, the presence control is related to the V3 circuit, the schematic shows that 3nF cap between the presence wiper and ground, whereas the limiter control is part of V4 circuit.

... With the presence control at zero, the amp is completely stable on both channels at any setting of volume / treble / response / limiter. The oscillation is introduced when turning the presence control clockwise anywhere past midway. If the presence is set midway or below, the amp sounds good on both channels, at any EQ setting and at any volume. ...

Does it oscilate if volumes are turned to CCW minimum?

[Merlin]

If you disconnect the 68k global feedback resistor, does it stop oscillating?
I suspect it is the global feedback loop that is going unstable depending on the source impedance of the cathode follower (since the presence control will vary that source impedance). Disconnecting the 68k resistor will confirm. If the answer is yes it stops oscillating, I suspect we can stabilise the amp with a speed-up cap. Also, can you confirm there is no capacitor connected between the anodes of the phase inverter, which might not be shown on the schematic?

I don't understand what you're describing, the presence control is related to the V3 circuit, the schematic shows that 3nF cap between the presence wiper and ground, whereas the limiter control is part of V4 circuit.

That cap is soldered to the tabs of the pots, between the presence wiper and the limiter pot which happens to be the nearest grounded point, that's all.

[Grantorino]

Merlin - the answer is yes, removing the 68k global feedback resistor has completely eliminated the HF oscillation. I can confirm that I replaced an old paper-in-oil 200pf cap connecting to the anodes of the PI (V4) with a 200pf mica when restoring the amp. Is the speed-up cap mod you mentioned applied directly to the 68k NFB resistor?

I'm very grateful for your advice - I should have sought help from this forum sooner.

[Merlin]

Merlin - the answer is yes, removing the 68k global feedback resistor has completely eliminated the HF oscillation. I can confirm that I replaced an old paper-in-oil 200pf cap connecting to the anodes of the PI (V4) with a 200pf mica when restoring the amp. Is the speed-up cap mod you mentioned applied directly to the 68k NFB resistor?

That explains a lot, someone has added an anode-to-anode cap like Fender/Marshall used, which often makes a feedback loop go unstable. I suggest removing that 200pF cap entirely. Put a 33pF or 47pF cap in parallel with the 68k feedback resistor (that's a speed-up cap). That value is just my estimate, feel free to try different values until the amp remains stable at any setting of the limit and presence controls.

[danhei]

Why is it called a speed up cap?

[Merlin]

Why is it called a speed up cap?

It's a casual name, but the cap allows lagging frequencies to regain some phase shift as they pass through the feedback loop, so they 'catch up' with the input again. A similar concept is used in high-frequency transistor switching circuits to help the transistors switch on/off faster