Revisiting some "old" ideas from 1970's - IPS, OPS

Index of the built and tested designs:

1) CDC-VFA-CCS V2 - fast VFA front-end, utilising a complementary differential cascade.
... Test with NS-OPS - Bode plot, spectrums, square wave response, clipping.

2) AmpliWire BJT version - cool simple front-end with current-driven VAS stage.
... My build and test results with NS-OPS.

3) AmpliWire Tubie version - 6DJ8 or ECC88 tube at the input + current-driven VAS stage.
... Tubie, tested by Terry with ECC88 - boards and scope screen photos.
... good Tubie layout, designed by Idefixes.
... and its single-layer etching-friendly version with PDF files for transfer.
... AmpliWire resistors value charts for different rails voltage, covering both BJT and Tubie versions.

4) VERTICAL CFA IPS - a fully symmetric top quality CFA front-end, utilising two complementary differential circuits.
... VERTICAL CFA IPS + Slewmaster OPS - test by Thimios with the newest Slewmaster OPS, designed by Ostripper.
... Good grounding illustration by Thimios, showing that the speaker ground has to be connected to PSU clean ground, rather than the amplifier board's ground.

5) NS-OPS - excellent non-switching OPS with building instructions and initial setup procedure.
... VERTICAL CFA IPS + NS-OPS. Top performance. Very low distortion throughout the bandwidth.
... Cool compact layout option, amended by Idefixes for NLW3281/1302 output devices.
... Photo of jwilhelm's build.
... Photo of my build.
... DIY-friendly build on a single-layer PCB - member filenet has designed a combined VERTICAL + NS-OPS layout.
... TubSuMo IPS + NS-OPS test - another excellent combination with 12AU7 tube at the input.
... and TubSuMo amendments, required for +/-50V rails operation. New R27-R31 values are good for any version.
... NEW! Modular NS-OPS based series of modules - snap-on IPS boards on top of the flagship "carrier" OPS board. Test.

6) VERTICAL VFA IPS - new, tested! Rival of the CFA version. Excellent symmetric VFA.
... Test with NS-OPS being outside the global NFB loop. Cool :cool:
... VERTICAL + NS-OPS - reference build by Jeff Wilhelm. Equipped with "21-st century" control/protection board.

7) VirtualZero-X - a new series of front-end designs. This is a CFA version. Not prototyped yet.
... VirtualZero-X2 - latest "current-drive VAS" VFA front-end. No Miller capacitors. Fast and linear. Not prototyped yet.
... VirtualZero-X4 - ultra-symmetric "four wheel drive" front-end. Tested! Updated for running cooler.
... Current-drive TIS based amplifier variation by Paul (mcd99).
... DiagoHex+ - low distortion front-end with single LTP, symmetric VAS drive and OPS from VHex+. NEW!
... Parallel - also low distortion "current drive" based amplifier. Very clean square wave response. NEW!

8) TRIBUTE-3000 (In memoriam of and inspired by the good old Sansui BA-3000)
... TRIBUTE-3000 Driver board layout
... TRIBUTE-3000 OPS board layout
... TRIBUTE-3000 Spectrums (live prototype)

9) Aureaux high-quality no-global-loop headphone amp design

Some great layouts, designed by members still4given, jwilhelm, marsupialx and Idefixes are available throughout the thread.
Additional details on some of these projects are available at my website VIRTUAL ZERO distortion AUDIO.
___________________________________________________________________________________________________
Here goes initial post:

Hi All,

Came across an interesting article in AES library, published in 1978 by Sansui engineers, describing the vertical differential cascade as a good way of driving VAS - higher slew rate, lower TIM, good stability (see the small picture attached).

So, I thought - why not giving it a try, using the modern active devices.
The whole thing is called Double Complementary Differential Circuit - I call it DC2.

One more thing - an OPS topology, used in a couple of Nakamichi power amps (420, 620) - I consider it great. I have simplified it a bit, still performing very good. In order to have a thermal feedback, Nakamichi places all transistors on the same main heatsink. I think, having the first pair together with the drivers on a separate one could be a good option as well - needs some testing. I call this arrangement Double Diamond Follower - D2F.

Simulation shows rather lower THD and narrower crossower region, comparing to conventional EF2/EF3 topology. No bias spreader required - the left side of a diamond keeps the bias of the right side regulated properly.

Attached is an experimental full power amp schematic (jpg and pdf), combining both cool topologies, along with stability analysis for the whole thing. I have also simulated the D2F OPS with my CF-FET V2.0 front-end - showing an outstanding performance.

Most likely, I'm going to test the OPS first. The whole thing may be built on a single PCB or as two separate modules (servo should be located on the IPS board). Number of output devices may be increased, although, with 2-3 pairs of MT200 Sankens it's already pretty much ok ;)

I did not have time to come up with any PCBs as yet.

So, what do you think? Comments/suggestions are welcome.
Brother Pete, if you'd like to simulate/check/amend/improve something - just go ahead ;)

Cheers,
Val
 

Attachments

  • Complimentary Diff Circuit.JPG
    Complimentary Diff Circuit.JPG
    25.5 KB · Views: 10,380
  • @DC2-Sch-01.JPG
    @DC2-Sch-01.JPG
    168.4 KB · Views: 10,454
  • @DC2-Sch-01.pdf
    64.9 KB · Views: 756
  • DC2 Pilot - AC Analysis.JPG
    DC2 Pilot - AC Analysis.JPG
    88.6 KB · Views: 10,175
Last edited:
Interesting OPS. Much like the plain diamond buffer, it affords you more buffer stages with the voltage drop of merely an EF2. I don't see why XO region would be lower than for an EF2 though - I suppose being able to run the drivers fairly hot would have something to do with that.

But MJE340/350 at only ~2 mA? Won't that be a tad slow?
 
these (and other) designs were popular during the "slew rate" wars some years ago.
some designs were so hung up on speed they forgot about things like overload & rail-sticking, common mode conduction in the output stage, latch-up, etc.

be careful or have a stock of spare devices ...
:)

mlloyd1
 
Interesting OPS. Much like the plain diamond buffer, it affords you more buffer stages with the voltage drop of merely an EF2. I don't see why XO region would be lower than for an EF2 though - I suppose being able to run the drivers fairly hot would have something to do with that.

But MJE340/350 at only ~2 mA? Won't that be a tad slow?

Actually, after closer digging into EF2 and D2F, comparing performance with different loads / frequencies at the same quiescent currents - the crossover region difference is barely noticeable, but THD is very slightly lower with the latter one. Well, certain advantage is expected in terms of the bias regulation accuracy, I also like low voltage drop and ability to have the bias fully independent from the VAS current.

Simulation does not show any slow-down, comparing to EF2/3...
 
these (and other) designs were popular during the "slew rate" wars some years ago.
some designs were so hung up on speed they forgot about things like overload & rail-sticking, common mode conduction in the output stage, latch-up, etc.

be careful or have a stock of spare devices ...
:)

mlloyd1

Well, CF-FET V2.0 front-end is even faster, but you-re right, this is already close to the OPS speed capabilities. Clipping looks good though - controlled by IPS. D11, D12 prevent sticking in overload conditions.

Anyway, thanks for "heads up" ;)
 
Searching for "purity" :)

Tried to simplify the front-end, sticking to the same principle and preserving the quality.

These two options - having single end VAS with CCS (1) and bootstrapped R (2) loads - utilize just one Complementary Differential Circuit (CDC) and use BJT transistors only, nothing difficult to find.
Choose the one you like :)

Both offer harmonics at -100db level in audio bandwidth, low TIM and rather high speed. Distributed compensation is pretty light. Local feedbacks keep the circuit rather linear even open-loop - around 0.1% THD @ 20KHz, 20V RMS, 80db gain.

Attached AC analysis is done with a simple EF3 OPS attached. I keep VAS rather warm (15mA), so MOSFETS (Lateral or Hex-FET) can be driven directly from it with no problem.

If somebody wants to "practice" PCB development - you are welcome ;)
Can be SMD or TH or a combination of both.

Cheers,
Valery

P.S. This circuit can be also used as a high-end pre-amp line gain stage (rails can be lowered for that purpose, like +/- 24V, with couple of resistor values amended), or an excellent amp for the headphones - with optional addition of a single push-pull EF (MJE340/350) - at the same +/-24V rails.
 

Attachments

  • 02-CDC-Boot-Sch.pdf
    42.2 KB · Views: 227
  • 01-CDC-CCS-Sch.pdf
    42.8 KB · Views: 224
  • CDC-VFA AC Analysis.JPG
    CDC-VFA AC Analysis.JPG
    80.2 KB · Views: 1,471
  • @02-CDC-Boot-Sch.JPG
    @02-CDC-Boot-Sch.JPG
    111.3 KB · Views: 1,576
  • @01-CDC-CCS-Sch.JPG
    @01-CDC-CCS-Sch.JPG
    114.9 KB · Views: 1,622
Tried to simplify the front-end, sticking to the same principle and preserving the quality.

These two options - having single end VAS with CCS (1) and bootstrapped R (2) loads - utilize just one Complementary Differential Circuit (CDC) and use BJT transistors only, nothing difficult to find.
Choose the one you like :)

Both offer harmonics at -100db level in audio bandwidth, low TIM and rather high speed. Distributed compensation is pretty light. Local feedbacks keep the circuit rather linear even open-loop - around 0.1% THD @ 20KHz, 20V RMS, 80db gain.

Attached AC analysis is done with a simple EF3 OPS attached. I keep VAS rather warm (15mA), so MOSFETS (Lateral or Hex-FET) can be driven directly from it with no problem.

If somebody wants to "practice" PCB development - you are welcome ;)
Can be SMD or TH or a combination of both.

Cheers,
Valery

P.S. This circuit can be also used as a high-end pre-amp line gain stage (rails can be lowered for that purpose, like +/- 24V, with couple of resistor values amended), or an excellent amp for the headphones - with optional addition of a single push-pull EF (MJE340/350) - at the same +/-24V rails.

You have 5 comp. caps ? Lead /lag , miller , and a shunt ! How did you arrive at these ?

BTW , a very interesting "blameless" -mutant. I like the decoupled low voltage input and overload clamp.

OS
 
Funny that we have to look back to the 70's to find really good amps. I'd say that is because stereos "mattered" back then. Having a great stereo was a status symbol and folks were willing to pay good money to impress their friends. Today they need to stuff 6 channels into a small box that will fit on a shelf under their flat screen.

How I remember going to the audio store and listening to Emerson, Lake and Palmer through a huge set of speakers and monster amps. Bass you could feel and mids and highs that seemed to come from every corner of the room. The store was filled with young men dreaming of owning that perfect "system". Those were the days! Great memories.
 
  • Like
Reactions: 1 user
At least 2 or 3 Hi-fi shops in every city, I remember them days too. Nowadays it is pods or something that young people want and a smart phone that everyone carries in there hand, not pocket. I saw a wreck on the interstate a few years back where the SUV rolled a few times with the lady thrown out the window during some flips and guess what was still in her hand when she landed? There is still a market for hi end 2 channel but the market is a lot smaller.
 
  • Like
Reactions: 1 user
Watch sometime who is driving that "dream" hot rod. He'll have a few gray hairs. He finally got one. Whenever I listen my big JBL's through the Slewmaster or one of Valery's creations I think, "If my friends could see me know".

My younger brother just brought me a ZZ Top vinyl the other day. I'm going to have to dust off the SL-1200-MK2 and give it a spin. That is going to bring back some memories.