Simple Symetrical Amplifier

[Magna]

Hi jay,

What would be the "ideal" bias? About 17ma current flows across the resistor with the current setup(+/-75v)....Is it too high?

Shaan,

Thanks for the kind words. I've chosen keltron caps for most locations except for the main decoupling caps. In my experience these caps are much better than the regular samwa types, known for leakage at its early life span. Of course these caps cost atleast twice more than samwa here in the local market.

Everything worked to my expectation, except the dc offset, which goes up / down (+ / - 50mv).

Once i get this sorted out, planning to test the amp in a PA system to explore further on this amp. will post my results soon...

[Jay]

Originally Posted by Magna

Hi jay,

What would be the "ideal" bias? About 17ma current flows across the resistor with the current setup(+/-75v)....Is it too high?

Hi, I just noticed your circuit is different than the one used in BIGBT HP. If I'm not mistaken your approach has been used by Shaan. The 17mA you mentioned is flowing to the regulator. The input itself is biased less than that. Around 8mA I guess.

I think there is nothing wrong with "hot" regulator (as long as there is sufficient heatsinking). I don't know about the "optimum" bias for the input. I have used around 10-12mA.

ADD: May be Shaan has a comment regarding this regulator (CCS). I think he didn't use it at the end.

[Lazy Cat]

Originally Posted by Magna

Hi all,

After a month's work, finally got my version of SSA singing...

Is your sch exact from post #3550 or it is modified?

Regards L.C.

[Magna]

Hi jay,

Thanks for your suggestions. yes, the circuit is based on shaan's version, except the CCS which i intend to keep to improve PSSR. I'll try to modify the bias to 10 - 12 ma and see if it affects the sound. I'd like to keep the sound as it is....

Hi LC,

Its the same circuit as shown in the post 3550....

[Lazy Cat]

Originally Posted by Magna

Hi jay,

Thanks for your suggestions. yes, the circuit is based on shaan's version, except the CCS which i intend to keep to improve PSSR. I'll try to modify the bias to 10 - 12 ma and see if it affects the sound. I'd like to keep the sound as it is....

As you can see it is also Lazy Cat's version from year before.

If you like I can comment your sch later today. Otherwise very professional SSA you have made.

[shaan]

Originally Posted by Magna

What would be the "ideal" bias? About 17ma current flows across the resistor with the current setup(+/-75v)....Is it too high?

IMO, the amount of current needed to flow through the input pair for a VAS current of 10mA at +-65V in your case would be "ideal". It could be as low as 1mA into the input trannies. Measure the VAS bias and if it's about 10-15mA then all's well.

Shaan,
Thanks for the kind words. I've chosen keltron caps for most locations except for the main decoupling caps. In my experience these caps are much better than the regular samwa types, known for leakage at its early life span. Of course these caps cost atleast twice more than samwa here in the local market.

Ya, I discarded those samwha caps a couple of years ago and now use Keltron for all my DIY jobs. Double the price no doubt, but solid performance.

Everything worked to my expectation, except the dc offset, which goes up / down (+ / - 50mv).

It should not if the input pair is thermally coupled with a cable tie(zip tie) and a drop of heatsink paste. In the picture it looks like the input pair and the cascodes are well placed(face to face) but the center two transistors(possibly the input pair) aren't tied together. I think this is causing you the offset problem.

Also the schematic shows no 10K from the conjoined bases of the input pair to ground. This could also cause offset problems due to lack of base current at the inputs.

Once i get this sorted out, planning to test the amp in a PA system to explore further on this amp. will post my results soon...

Lookin' forward.

Originally Posted by Jay

The 17mA you mentioned is flowing to the regulator. The input itself is biased less than that. Around 8mA I guess.

Yes. Most of it flows to ground through the 100ohm resistors. My setup has 3.5mA through the input BJTs, but total 12mA from the +-15V zener supply.

May be Shaan has a comment regarding this regulator (CCS). I think he didn't use it at the end.

It sounded bad. It didn't have the collectors of the voltage sense transistors sinked/sourced to/from ground, but floating. This definitely is an example of bad design resulting into bad sound. I will connect it again soon with corrected design. Hope to have good results.

Originally Posted by Magna

shaan's version

It's based on LC's SSA as the VAS has cascodes. My PS is +-28VDC, so cascodes there is a no-no. But, maybe in future, with +-56V... We'll see.

[Magna]

Hi LC,

Thanks LC. Its my pleasure, to have your comments on the scheme....i'm sure that will certainly improve the performance of the amp further.

Hi Shaan,

I'll take your views on bias....i'll try to couple all the input trannies with heatsink compound....perhaps that could solve the DC drift problem....i do have the 10k input bias resistor in place with this amp

How do i reference the CCS to ground with this design....any idea?

Will that have positive influence on the overall sound quality of the amp...?
I'm already very happy with the sound....

Do you think a simple resistor bias is better than the CCS, especially the PSRR and of course sound quality as well...what's your experience...? probably i'll also experiement with that....

[shaan]

Originally Posted by Magna

Hi Shaan,
How do i reference the CCS to ground with this design....any idea?

Your CCSs are already sourced/sunk to/from ground. Mine was not.

Will that have positive influence on the overall sound quality of the amp...?

I can't say for sure, but I must first try it myself before deciding over its influence on sonics. The floating nature of the CCS I built is my suspect for the fatiguing sound.

Do you think a simple resistor bias is better than the CCS, especially the PSRR and of course sound quality as well...what's your experience...? probably i'll also experiement with that....

Tough question.

It's easy to understand that CCS will improve the precision of the feedback network at HF by an order of magnitude than that with a resistor. Probably the PSRR will also be higher with CCS. but,

BUT

what if your brother or friend likes the "precise" sound but you don't???

If you like the sound and it doesn't give you fatigue then no matter whether its a CCS or a resistor, it's perfect.

Well, my SSA's FB network is fed through two 1K2 resistors and this amp sounds more precise to my ears than any other amp I have heard, right up to full volume. It has stunned more than 100 people with its sound(me included) to-date and is keeping up the shock-job for new people almost every week.

No hum no noise no no instability no fatigue so far.

My SSA is made for music, and its music says it's true.

That said, there is no guarantee that you will like the resistor-fed SSA sound more than the CCA-fed SSA. Here's the devil, subjective quality, totally private property.

[Magna]

Hi Shaan,

Lots of ideas for further experiementation on CCS....Thanks. Will let you know about my results without CCS...As you've pointed i may / may not like the sound....

[shaan]

My local DIY friend Bhaskar has printed six SSA boards. Yesterday I was on a visit to his home and had some unique experiences.

First, he's a genuine music-lover; has over 3000 pieces of gramophone records and most of them are pure gem recordings. Many dating to 65's! I had my first experience of listening to analog recording and I was completely blown away. The realistic feeling of the music just froze me and now I know why people say that once you get into analog, there's no coming back.

Second, he is a very kind person. He provided me with two of the six SSA boards made and now I have my own designed peeceebee at hand.

The peceebees look very well etched without any errors.

Now I'm on the way to build a pair of SSAs in the pro way LOL.

Thanks for the good wishes all.