Another sorted Stereo 20

[Jazid]

When I was doing the same to my TL12+ amps I fiddled with the gain and felt it was better after, but reverted before sale for authenticity.
What made the biggest difference on the 'scope was tweaking the feedback components. Tidied up the 10kHz square wave quite nicely! No idea about sonic change though 🤣

 

[rabski]

When I was doing the same to my TL12+ amps I fiddled with the gain and felt it was better after, but reverted before sale for authenticity.
What made the biggest difference on the 'scope was tweaking the feedback components. Tidied up the 10kHz square wave quite nicely! No idea about sonic change though 🤣

The sensible way to attack it is to do both, as I'm sure you already know. Changing the NFB is the simplest way to drop the gain, but the RC network needs to be bang on to keep phase coherence. I used 1% silver mica caps and closely matched resistors but kept the original values. Testing shows a nice square wave at 1kHz and 10kHz, so I'm happy with that. Changing the feedback resistors needs the capacitor values to be changed as well, of course. The easy way to do it is take a stab at the equation, then put decent caps in and a pair of trimpots to get it spot on.

Once I've sorted the third one, that'll leave two that are cosmetically dire. I won't feel so precious about keeping those stock, so I have a couple of thoughts. The really mad one is to parallel them to run as monoblocks. That won't double the power, but it will give more current and more output before distortion kicks in (if I've worked it out right). It also means setting each pair of OPTs to 16 ohms to match an 8 ohm load. That has a major benefit, because they're wound ideally for 16 ohms so the whole ouptput stage should be a lot happier.

Or I'll just respray both, rebuild them as standard and start moving some stuff on. I need the space. As much as anything, to finish all the rest. Anyway, that's just the Stereo 20 collection. I haven't even started unboxing any of the rest yet

 

[Jazid]

Solution? Well. Durr! You simply need more amps in pieces 🤷‍♂️

Luckily I can help you out, my dear wife will even thank you, just whistle 😘

 

[rabski]

You simply need more amps in pieces 🤷‍♂️

I respectfully wish to disagree.

My wife and my garage also wish to disagree. My workshop refuses to comment.

 

[Lurch]

I respectfully wish to disagree.

My wife and my garage also wish to disagree. My workshop refuses to comment.

Let me know when your done, as I have a nice little 50s portable Pye valve radio I'd like to get checked over/fettled.

 

[Tarzan]

Learning question; The LEAK amps are much sought after and seem to be highly respected- so how do they sound LEAKERS?

 

[audio_PHIL_e]

Learning question; The LEAK amps are much sought after and seem to be highly respected- so how do they sound LEAKERS?

• Excellent (Stereo-20 owner).
• "Bloody brilliant" - Les Wong formerly of Walrus Systems in Marble Arch

 

[rabski]

Learning question; The LEAK amps are much sought after and seem to be highly respected- so how do they sound LEAKERS?

Having just finished the third one and with two still in bits, I suppose I'm probably a bit biased...

My answer would nevertheless be 'way better than anyone would expect'. The original circuit is excellent, but using tight tolerance, low noise components pays additional dividends.

 

[Amormusic]

Learning question; The LEAK amps are much sought after and seem to be highly respected- so how do they sound LEAKERS?

I love mine.

Partnered with appropriate speakers. Just beautiful and organic and natural and... right.

Mine is a keeper.

 

[audio_PHIL_e]

I love mine.

Partnered with appropriate speakers. Just beautiful and organic and natural and... right.

Mine is a keeper.

Do I take it that yours is a "Valve 1961 fully rebuilt beauty"? My "

appropriate speakers

" are a pair of 15Ohm 2-way Sandwich 600 speakers. Yours? I feed my Stereo-20 from a passive pre via a set of 15dB Rothwell attenuators. Works quite nicely.

 

[Nagraboy]

Having just finished the third one and with two still in bits, I suppose I'm probably a bit biased...

My answer would nevertheless be 'way better than anyone would expect'. The original circuit is excellent, but using tight tolerance, low noise components pays additional dividends.

Richard, do you think there’s much value in using a different input valve with some resistors to reduce the gain to a more sane level?

Having said that, I had good results with a stock Leak via a unity gain Croft 25 preamp.

 

[Amormusic]

Do I take it that yours is a "Valve 1961 fully rebuilt beauty"? My "

" are a pair of 15Ohm 2-way Sandwich 600 speakers. Yours? I feed my Stereo-20 from a passive pre via a set of 15dB Rothwell attenuators. Works quite nicely.

Yep. I actually stuck a photo up today on from your listening chair thread.

 

[audio_PHIL_e]

Yep. I actually stuck a photo up today on from your listening chair thread.

Found it

 

[rabski]

Richard, do you think there’s much value in using a different input valve with some resistors to reduce the gain to a more sane level?

Having said that, I had good results with a stock Leak via a unity gain Croft 25 preamp.

Definitely 'no' to the first bit Greg.

The gain is set primarily by the feedback circuit. If you change the input valve to a lower gain, not only is it not running in the right operating point, but all that happens is the feedback compensates for it, so you end up with almost the same gain, increased distortion and reduced linearity.

There are three ways to drop the gain. The simplest is to use a voltage divider on the input, either built in or by adding attenuators. Works fine, but doesn't sound the best. The hardest is to change the entire input stage. The best is to change the input stage loading values (plate and cathode values and cathode bypass) and most importantly, change all the values in the negative feedback circuit.

The last one is what I'm going to do with the final two rebuilds, as I'll probably keep them. It's time consuming though, as it needs a signal generator and 'scope to get it spot on and avoid adding instability.

 

[DomT]

Having just finished the third one and with two still in bits, I suppose I'm probably a bit biased...

My answer would nevertheless be 'way better than anyone would expect'. The original circuit is excellent, but using tight tolerance, low noise components pays additional dividends.

I have never heard one. What would be its equivalent performance level today? Would it be on par with say an Audio Note OTO?

 

[rabski]

I have never heard one. What would be its equivalent performance level today? Would it be on par with say an Audio Note OTO?

Haven't had an Oto in the system here, so although I've heard one, I can't give a direct comparison Dom. I'd tentatively say that assuming a decent set of valves, tight tolerance components and decent quality coupling capacitors, then it would be comparable though different. There are also quite a few different versions of the Oto, and the higher spec ones have 'serious' output transformers, so I would expect them to be a step up.

The last two 20s that I'll probably keep are going to get a bit more heavyweight components, including (ironically) Audio Note coupling capacitors and some slight changes to layout, etc. That should be interesting. The final step may be trying to run them paralleled as monoblocks. Fun awaits...

 

[George 47]

There are also quite a few different versions of the Oto, and the higher spec ones have 'serious' output transformers, so I would expect them to be a step up.

A step-up transformer, then

 

[hifinutterreturns]

Haven't had an Oto in the system here, so although I've heard one, I can't give a direct comparison Dom. I'd tentatively say that assuming a decent set of valves, tight tolerance components and decent quality coupling capacitors, then it would be comparable though different. There are also quite a few different versions of the Oto, and the higher spec ones have 'serious' output transformers, so I would expect them to be a step up.

The last two 20s that I'll probably keep are going to get a bit more heavyweight components, including (ironically) Audio Note coupling capacitors and some slight changes to layout, etc. That should be interesting. The final step may be trying to run them paralleled as monoblocks. Fun awaits...

Done a cracking job there with the stereo 20's. I had thought about mono blocks is it a big job do you think 👍

 

[rabski]

Done a cracking job there with the stereo 20's. I had thought about mono blocks is it a big job do you think 👍

That depends on workshop test results basically. Bridging would techically be best, but very complex, as it needs a complete redesign of the circuit to change the phase. A bridge too far, basically

Paralleling should be more straightforward. Just simply run L and R channels in parallel and then use the 16 ohm taps to run 8 ohm speakers. But... unless the two channels are all but identical, one will be trying to drive the other, rather than summing. It will almost certainly need a balancing resistor on each output transformer, though that will drop efficiency and also upset the load a little. Just an ohm or so may well be enough. I'm hoping that using closely matched R and C values throughout will make it feasible. Dynaco did it with the ST70, and that didn't use anywhere near the component tolerances I'm using.

The only way to find out is getting it on the bench and playing around. I'll probably replace a few resistors with trimpots to get things absolutely matched. It can wait. These two are going to be shot blasted and powder coated first, because they've not had a pleasant life. I'm also going full-on sacrilege mode and nailing a large choke to the top of each. Leak did it with many models, but not on the ST20, presumably simply for cost. A choke in that PSU design makes an awful lot of sense.