Counerpoint SA100/200 power amp scheme

[aparatusonitus]

There is a neat trick to get around this that I used in a very low noise preamp - two toroids stacked one on top of the other wired in series so that their stray fields cancel. It took some time explaining what I wanted to the local toroid winding guys, but they got it eventually. I used the flat type core, and they potted the two into one of their standard cans for me. Worked perfectly!

Hi ilimzn,

Could you provide me with all necessary informations needed to order such a toroid compound, and from whome, thank you.

 

[ilimzn]

Hi ilimzn,

Could you provide me with all necessary informations needed to order such a toroid compound, and from whome, thank you.

Any toroid winder can do it for you, but you actually wind two equal transformers, each of half power, each carrying a full primary and half of each secondary. For instance, if it is 230V promary and 24V secondary, you split that into two transformers, having one 230V primary and one 12V secondary each. You stack the two toroids one on top of the other, keeping in mind that the mechanical orientation MUST be the same. You connect the primaries in ANTIPARRALEL and half secondaries in ANTISERIES, to form a compound 230 to 24V transformer. A flat form toroid core is used for each transfomer in order to get the core center lines as close as possible. This way the magnetic fields are oposite, and so are the stray fields, in order that they mutually cancel.
This is the basic idea and all that I can tell you as the technique is patent pending, but it is not so difficult to deduce if you have a basic understanding of transformers.

 

[anatech]

Hi ilimzn,
I'll stick to a properly made EI core. Reducing the idle current in the output stage and going to BJT's will drop the core temperature a fair amount (the original transformers become reliable now - no joke!). I don't think those transformers are dual secondary either, just two leads but I've never taken one apart to confirm this. You would know better.

Nice trick with your toroid design though, using the fields for cancellation. Might be useful in a preamp design, would it work for short EI core types?

Hi Brian,
Your oscillation has got to be in the final stage, or more likely in the output stage. Those models have normally been burnt out already and I don't get a chance to see them. The thought of one of these having response up that high is a little foreign to me.

-Chris

 

[ilimzn]

Hi ilimzn,
I'll stick to a properly made EI core. Reducing the idle current in the output stage and going to BJT's will drop the core temperature a fair amount (the original transformers become reliable now - no joke!). I don't think those transformers are dual secondary either, just two leads but I've never taken one apart to confirm this. You would know better.

Yes, two leads, connected to the same secondary winding. You can see that in the pictures I posted towards the begining of the thread. This incidentallly makes for an additional ground loop since the actual ground wire from the transformer is split in two, then recombined in the amp.
Still, the core is simply underdimensioned. It's a 200VA transformer AT MOST, and this would, with some clever design, perhaps suffice for 60-80WPC reliably - but you and I both know clever is not really in the picture. Ohm's and Joule's laws simply don't work 'by ear'.

Nice trick with your toroid design though, using the fields for cancellation. Might be useful in a preamp design, would it work for short EI core types?

There is some cancelation with EI core types but not enough. Using a flat core toroid brings the cores very close together, while with other types the distance increases quite a lot due to bobbins etc, usually to the point where the distance is on the same order as that of the distance between the transformers and other circuitry. Still, it is possible to place two EI or C-core transformers next to each other and use a common shorting winding (this shorts out all differential flux).
I came up with this idea for a preamp for a 24 bit measurement ADC. It should work even beter in a power amp, especially hig bias, where surge currents due to rectifier conduction angle momentairly saturate the core(s). It's not a cheap solution, but it worked so well I couldn't measure any stray field from the prototype at all (using an inductive probe).

 

[aparatusonitus]

Any toroid winder can do it for you, but you actually wind two equal transformers, each of half power, each carrying a full primary and half of each secondary. For instance, if it is 230V promary and 24V secondary, you split that into two transformers, having one 230V primary and one 12V secondary each. You stack the two toroids one on top of the other, keeping in mind that the mechanical orientation MUST be the same. You connect the primaries in ANTIPARRALEL and half secondaries in ANTISERIES, to form a compound 230 to 24V transformer. A flat form toroid core is used for each transfomer in order to get the core center lines as close as possible. This way the magnetic fields are oposite, and so are the stray fields, in order that they mutually cancel.
This is the basic idea and all that I can tell you as the technique is patent pending, but it is not so difficult to deduce if you have a basic understanding of transformers.

Thank you very much for the input ilimnz, aspecially as the idea is patent pending. Brilliant staff, as allways from you...

One more thing, would you mined if I buzz you from time to time to take a quick peep in "my" analog deseign, just to make sure that I'm heading in right direction?

 

[anatech]

Hi ilimzn,

- but you and I both know clever is not really in the picture. Ohm's and Joule's laws simply don't work 'by ear'.

LOL!
Now I'm cleaning my coffee of my monitor and keyboard! Of course I more than agree with you, also on :

It's a 200VA transformer AT MOST, and this would, with some clever design, perhaps suffice for 60-80WPC reliably

Never mind the low quality and variable reliability. You sure know how to get a laugh. First thing in the morning too!

There is some cancelation with EI core types but not enough.

I was thinking that, but you never know what another guy can figure out. I'm not so much bothered by radiated fields because you can "box" the power supply sections. I wonder how much electrostatic stuff comes off large supply caps? Never measured that.

Great idea though. I still would like to see an air gap and I like an EI core for many other reasons (with electrostatic shield).

-Chris

 

[ilimzn]

Thank you very much for the input ilimnz, aspecially as the idea is patent pending. Brilliant staff, as allways from you...

One more thing, would you mined if I buzz you from time to time to take a quick peep in "my" analog deseign, just to make sure that I'm heading in right direction?

No problem.. just remember that I don't and never claimed to know everything

 

[ilimzn]

Hi ilimzn,
LOL!
Now I'm cleaning my coffee of my monitor and keyboard! Of course I more than agree with you, also on :

Thanks... maybe i sound harsh, but there it is. I mean, OK, you really can 'voice' amplifiers, and we know that the sound you get is a rather ephemeral thing to predict and qualify, but for goodness sake, at least first get the basics right.

I still would like to see an air gap and I like an EI core for many other reasons (with electrostatic shield).
-Chris

I quite like C cores for that reason, you get the power density of toroids, and a gap. Still, I would love a gapped toroid, and believe it or not, they do exist - at least that's what my local winder tells me they can get. Basically, it's the same trick as a C-core, except it's not would oval like a C-core starts, but as a toroid, then cut - just like a C core.

 

[bscally]

Hi ilimzn,
Hi Brian,
Your oscillation has got to be in the final stage, or more likely in the output stage. Those models have normally been burnt out already and I don't get a chance to see them. The thought of one of these having response up that high is a little foreign to me.

-Chris

This is a plot of the feed point to the FETs. It appears that the oscillation is caused by the path delay from the input triode to the feedback triode - psudo LTP thing.

If the circuit is redesigned to avoid the coupling cap you don't see the oscillation.

But as mentioned there are so many other issues....


Brian

 

[jlw]

SA12 vs SA100

Seems like the tube stage of the SA12 is a lot simpler and easier to handle than the SA100. I never liked the SRPP-like follower stage in the SA100/220 anyway,

I just finished repairing one and noted the difference, and made a few adjustments that resulted in pretty positive improvement.

anode coupling between gain stages 1 and 2 got me to section them off using a 560ohm and bypass with a 30uF for stage 1

stage 2 plate resistor now gets 50k vs 100k and plate increases to 115V

Running the follower a lot hotter now with 18k against 110V along with a 24pF for stability

Reduced coupling cap to 0.1uF hopefully eliminates chance of motorboating
Grid to ground was 1M anyway so I figure smaller doesnt matter.

Most importantly reduced the stock output stage bias from 480mA to 200mA (was surprised to see the amp actually was still working when it landed on my bench)

I cant comment on the SA100 yet cos while Ive fixed a few I never tried to mod the tube portion of the circuit like I did to this SA12.