Negative feedback problems on ECL82 SE amp.

[colinB]

Hi,

I recently finished an ECL82 SE amplifier based on the following schematic that I found on a Korean website (see attachment). I copied the amplifier circuit from this schematic but used a different power supply that I had available from a previous project (Gary Kaufmanns DC coupled 6EM7 SE) with suitable RC combinations to reduce B+ voltage.

Initially, all the testing (dc voltages) and limited listening was carried out with no negative feedback applied but there was a noticeable background ‘hiss’ audible from the listening position during quiet sections of the music. Last night I decided to apply negative feedback to reduce gain and distortion. I started with the recommended 4k7 & 220pf combination as detailed on the schematic for the feedback loop. Unfortunately, this resulted in a seriously loud ‘squeal’ from both loudspeakers suggesting positive feedback.

So my question is what is the easiest way to remedy this problem?

At the moment I’m using a pair of Hammond 125GSE output transformers connected to FE167E speakers (8ohm) as follows:

Primary: Brown cable > B+, Blue cable > plate/anode of ECL82 tube.
Secondary: Yellow cable > +ve terminal of speaker, Black cable > -ve terminal of speaker (+ attached to ground).

So in this case do I simply reverse secondary connections so Black > +ve terminal and Yellow > -ve or am I seriously missing something else altogether?

Thanks
Colin.

[Tom Bavis]

Yes, reverse either the primary or secondary leads, whichever is easier - but of course, not both....

[Shoog]

If you have the secondary of the transformer grounded as in the schematic, you will also need to change the ground connection over, otherwise you will be feeding back zero signal.

Shoog

[Miles Prower]

Originally posted by colinB
Hi,

I started with the recommended 4k7 & 220pf combination as detailed on the schematic for the feedback loop. Unfortunately, this resulted in a seriously loud ‘squeal’ from both loudspeakers suggesting positive feedback.

Yup, Mr. Murphy done wired it up for positive feedback.

So my question is what is the easiest way to remedy this problem?

At the moment I’m using a pair of Hammond 125GSE output transformers connected to FE167E speakers (8ohm) as follows:

Primary: Brown cable > B+, Blue cable > plate/anode of ECL82 tube.
Secondary: Yellow cable > +ve terminal of speaker, Black cable > -ve terminal of speaker (+ attached to ground).

So in this case do I simply reverse secondary connections so Black > +ve terminal and Yellow > -ve or am I seriously missing something else altogether?

What I'd do is to connect the blue primary to Vpp and the brown primary to the plate of the ECL82. That should reverse the phasing and get you negative feedback.

[colinB]

Hi,

Thanks for the suggestions guys. Initially I was going to change the brown & blue primary connections over but I discovered I had cut the blue leads too short so swopping would have been difficult.

So, I was forced to switch the secondary connections resulting in negative feedback.

I probably just need to spend some time playing with the R & C values in the feedback loop to suit my output trafos but I will use 4k7/200pf as a starting point as per the schematic.

BTW is the best way to optimise R & C to use a signal generator to produce a 10Hz square wave signal into the amp and then tweak the R & C to give the best looking square wave on an oscilloscope at the output?

Thanks for help.

Colin.

[gingertube]

Colin,
The square wave test (at 10kHz NOT 10Hz) is the way to tweak the value of the cap. Run into a resistive dummy load at say 1/2 maximum power output, then tweak the capacitor. Note: make absolutely sure you are not saturating the output - else what you see on the CRO will be meaningless for this set up. Generally as you increase the cap value you will see less ringing, if you go to far you will see undershoot. Usually a single half cycle of overshoot with no ringing is about right. If the output tranny is not ideal you may wish to do this at 5kHz or even 2kHz.
Cheers,
Ian

[richwalters]

On a new amp build, another simple way of checking if the o/p tranny is the correct way round is disconnect global nfb loop and to pass a low level signal though amp into a speaker. Temporarily swap the suggested value of the global nfb resistor for a value 10X or higher in value i.e if 5K select 50K. If volume increases then either sec or primary windings need reversing.

[colinB]

Hi Gingertube,

Thanks for the tips regarding the optimisation of the value of C. I've only recently acquired an oscilloscope so I guess what you make sound simple in your post will probably turn out to be at least a few days work for a novice like me

Ok I have just a couple of further (dumb) questions if anybody can assist:

1 The value of 4k7 for feedback resistor suggested on the schematic. Is this there a rule of thumb for the value of this resistor? Should I stick to this or experiment with higher values (e.g increase to 10k or 15k?) prior to tweaking the feedback cap value?

2. Having got the required 'best' feedback resistor I presume that I use maybe a 1V pk to pk 10kHz (or 5kHz) from the signal generator into input. Use a calibrated scope probe (set at 10x?) on output (with 8ohm dummy resistors) in combination with a variable capacitor connected across the feedback resistor Tweak variable capacitor to minimise ringing. This seems ok but on the scope should I be set as AC coupled or something else? I'm not really sure about the scope settings here.

3. Briefly used the 4k7/220pF set-up tonight but there was still quite a bit of background hiss audible during quiet parts in the music. Will playing with the feedback rectify this or is the hiss likely to be from elsewhere (power supply noise?)

Rich, thanks for the tip on how to quickly sort out if the o/p tranny is connected the correct way round. Will use this if I ever get in this fix again - which is highly likely for me!

Thanks

Colin.

[Shoog]

In my experience the ECL82 is not a particularly hissy valve, I have used it in two headphone amps and hiss was never an issue. Feedback is unlikely to cure a serious hiss problem because of the random nature of hiss generation. How sensitive are your speakers.

What is more likely is that you have some unresolved stability issue. Some photos of your build might be useful.

Shoog

[colinB]

Hi Shoog,

Just a quick reply to your questions in your post

How sensitive are your speakers.

I use Fostex Fe167E in the recommended enclosure provided by Fostex with the drivers. I think the efficiency is around 94db.

Some photos of your build might be useful.

I've attached a photo of the underside of the amp. It's not pretty and I used what components I had available to hand.
Note this build is put together in a temporary chassis that has been used previously for a 6EM7 amp and before that a EL34SE amp. In fact the components & batteries for the 6EM7 amp are still attached to the 8 pin sockets at the top of the photo - ignore this!

I usually use a MP3 player or Walkman CD into a 5687 pre-amp into the ECL82 power amp. Hiss is audible at ~5ft from speakers when everything is connected together but no music, increasing the volume makes it worse.

I've been playing with the feedback loop tonight will post some photos of the CRO traces shortly.

Cheers
Colin.