Simple Symetrical Amplifier

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi All.

Salas Solution with a CCS in the Driver stage is a small masterpiece.... First I was sceptical but the drivers BD139 and BD140 will be able to handle all the current flowing into the basis.
He has choosen a CCS of about 2mA, I would say why not set it bit higher, like 10mA, to get a higher Ft value...

The bias circuit will have nearly optimum conditions this way, and there is no thermal runaway in the VAS stage.

PS: i am sorry, that i am not telling you guys more...

- Sonny

Yes, very clever idea from Salas indeed. :cool:

P.S. Hehehe one day you will ... :D
 
diyAudio Chief Moderator
Joined 2002
Paid Member
He has choosen a CCS of about 2mA, I would say why not set it bit higher, like 10mA, to get a higher Ft value...

The bias circuit will have nearly optimum conditions this way, and there is no thermal runaway in the VAS stage.

PS: i am sorry, that i am not telling you guys more...

- Sonny

Had mentioned that option once. Its just testing phase, and since with Sanyo FBET 300MHz Ft drivers and 5mA went wild once, thought to start conservatively this time.:forbiddn:

With the CCS the sqw rise time went up significantly also, and the faster TO-92s showed periodic oscillation. Fortune had it that the PHILIPS brand BD pair kept behaving in all areas, with the open possibility of running much more Iq in a future experiment when all other parameters will be fixed. The CCS is good for PSRR there also.
 
Beautifully made speakers

For these speakers it is possible to use only stereo amp (one pair of binding posts per speaker only), for the other speakers I will need 4 amps per channel. :rolleyes:

The beauties You have surely sounds excellent?, especially with big subw. behind :D, a 4.way system?
For the sub You also use the SSA High Perf. same as for main speakers?
Vinyl playback on Yours system is perhaps sounding outstanding?

- For really comfortable driving 2.5 OHm speakers, how many BIGBT modules would be necessary? :scratch:

Kind regards,
 
Yes, very clever idea from Salas indeed.

But, the CCS in the VAS means that this is no longer a symmetrical VAS. It controls the current through one side of the VAS well, but leaves the other side more room to move with temperature. It's asymmetric, a different amplifier.
 
But, the CCS in the VAS means that this is no longer a symmetrical VAS. It controls the current through one side of the VAS well, but leaves the other side more room to move with temperature. It's asymmetric, a different amplifier.

I agree completely that's why I didn't put this fet in my schematics. Salas find it usefull so be it, but for me this is not necessary. Today I did some testing with SSA BIGBT HP front-end plus mosfet drivers and I was stunned. Measurements will follow ... :eek:
 
SSA BIGBT HP front-end initial power-up, stability check and some measurements.

Prior to full power testing with complete circuit (BIGBT power outputs) I checked what can I expect from SSA High Performance front-end plus power mosfet drivers. SSA input bias set to 2,5 mA, VAS bias 15 mA, drivers bias 50 mA.

Offset trimmed to zero and it was inside +/-20 mV all the measurement time with no instability issue. Bias of the drivers was completely thermally stable.

Than AC measurements took place and here are the results:
- output impedance 38 mohms, tested at 1 A/4,7 ohm load
- 1 ohm stable (drivers as output devices) inside SOAR
- bandwidth -3 dB unmeasurable with my current equipment
- input/output phase delay 12 ns at 2 MHz (period 500 ns), that is equivalent of 8,6° phase shift

I was satisfied with initial testing, so we have a go for further development. :)
 

Attachments

  • SSA1.JPG
    SSA1.JPG
    108.6 KB · Views: 563
  • SSA7.JPG
    SSA7.JPG
    113.4 KB · Views: 185
  • SSA6.JPG
    SSA6.JPG
    106.6 KB · Views: 169
  • SSA5.JPG
    SSA5.JPG
    113.6 KB · Views: 495
  • SSA4.JPG
    SSA4.JPG
    112.9 KB · Views: 520
  • SSA3.JPG
    SSA3.JPG
    105.6 KB · Views: 530
  • SSA2.JPG
    SSA2.JPG
    101.4 KB · Views: 540
For TO-126 at that power level (2mA or up to 10mA across 30V) don't expect much thermal wandering. Was moving around much more without. Its a solution for when some versions won't behave.

Yes it very much depends on topology. SSA front-end was for me first time tested in high performance mode (input cascodes + VAS cascodes).

I found out that this circuit is even faster as before (without cascodes). Phase shift in a SSA basic was 24° and now is 8,6° at 2 MHz, without any instability issue. :eek:

So based on this ground I can do a lot with this circuit in the future. :wiz:
 
diyAudio Chief Moderator
Joined 2002
Paid Member
If the SSA was driving mosfet like my CFB with EXICON it would also be assymetric by a few volt. because of difference in the N and P channel.

So the few volt accross the JFET does not matter if the total voltage swing is +/-35V...

It will only be noticed by clipping....

There are ~0.6V pinch off Jfets to use.;) My latfets ain't matching also, just picked out off the tube.
 
diyAudio Chief Moderator
Joined 2002
Paid Member
Yes it very much depends on topology. SSA front-end was for me first time tested in high performance mode (input cascodes + VAS cascodes).

I found out that this circuit is even faster as before (without cascodes). Phase shift in a SSA basic was 24° and now is 8,6° at 2 MHz, without any instability issue. :eek:

So based on this ground I can do a lot with this circuit in the future. :wiz:

Its a very worthy bridge front end from the moment you get to know its traits. Such stuff usually lives on miniscule silicone area, trimmed and thermally homogenous current feedback op amp world. Was a refreshing idea you brought it back to DIY discrete world power amps Andrej. Hats off for that.:trapper:
 
How much low do you trim the front BJTs current feeding resistors so they run 2.5mA Iq? ~?

Well it goes like this:
- 15 V supply voltage on zeners
- 2,5 mA input BJT collector current
- 1,220 V SSA bridge DC voltage, that means 0,61 V on 93,7 ohm (100 ohm parallel to 1,5k) is another 6,5 mA
- so all together is 9 mA through current supply resistor, voltage on it is 14,39 V (15 V - 0,61 V) so the resistor-trimmer value is 1.597 ohm
- 15 V zener bias current is 4 mA
- VAS current 12 mA at start, stabilized at 15 mA later
- driver bias current fixed to 50 mA (iron fist Vbe multiplier)

Images:
- green LED side with input BJT's (below middle blue cap, BJT's glued together) plus cascodes
- VAS side plus cascodes
- complete amp
 

Attachments

  • SSA8.jpg
    SSA8.jpg
    78.8 KB · Views: 546
  • SSA9.jpg
    SSA9.jpg
    99.7 KB · Views: 476
  • SSA10.jpg
    SSA10.jpg
    110.1 KB · Views: 473
Last edited:
diyAudio Chief Moderator
Joined 2002
Paid Member
Yes, around the 8-9mA mark the stronger side seemed good in mine too that I CCSed the feeds.

Your Vbe mult keeps bias well. In latfet CFP with drop over resistor for bias, the Vas warm up tips the bridge much more readily in my experience for output stage Iq, thus offset.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.