An audio amplifier with 7Mhz ULGF, with no frequency compensation (other than that provided by the stray/parasitic elements, like the input capacitance of the output stage) which was for 15 years in production and was successfully sold to customers?
Please, please, do humanity a big favour and tell the brand and the model number of this unbelievable amplifier. Have you measured these amplifiers and confirmed your simulation results?
Are you blind, glasses would be a good investment.
It uses what you call brutal compensation.
Personally I havent, my employer has, his design. Im a few years younger. Review appeared in a japanese publication. British company that also bought the design doesnt do advertising in audio publications.
Build it and learn something instead.
It uses what you call brutal compensation.
Personally I havent, my employer has, his design. Im a few years younger. Review appeared in a japanese publication. British company that also bought the design doesnt do advertising in audio publications.
Build it and learn something instead.
Yes, my sight is no longer what it used to be. I'll use a magnifier to better analyze the partial schematics you posted. Is that the cap in series with a resistor towards the supplies? Could you post a clear schematic, or the .asc file?
You haven't answer the question: brand and model, please. To the extend that I am aware of, this is the highest ULGF commercial bipolar amplifier I have ever heard of. BTW, I personally don't give a rat's *** on review journals, unless they publish relevant, verifiable measurements.
What makes you believe that a "VFA" with the same ULGF will do worse?
I disagree.
A CFA topology amp is a CFA topology amp. Period.
Changing the feedback resistor value changes the gm. That does not happen on a VFA.
In the final analysis, the CFA performance may be similar in a power amp, a lot of this due to the Ft of the output stage (but this subject we have not explored fully here in my view) but there are fundamental differences in operation.
Manso, is this the 'simple' amplifier whose .ASC you promised to post ?
I too would like to know the brand and model of this excellent amp. Also to play with the .ASC
Your jpgs are not quite readable even enlarged many times. 🙁
_________________________
Waly, the aim of this thread is to get something useful. If this is obtained by better understanding different input stages, this is still more useful than semantic pedantic arguments with Mikek.
[deleted: my slim 59 pg monograph on what is a CFA ]
_________________________
My present stance .. which is changing even now as the pearls emerge from the Golden Pinnae noise .. is that simple CFA i/ps like with VSSA have some very useful properties for power amps which may be unavailable to VFAs. I've yet to see any properties in a Diamond i/p that I want. But I'm open to what their fans might bring up.
One of these advantages is the 'compressive' vs 'expansive' behaviour on overload of a LTP vs simple CFA i/p that Marshy has posted. If you look at the numbers on the graph, you'll see that at similar currents & Gm, the 'expansive CFA' i/p has lower THD at similar voltage modulation levels and also drops faster with reducing level (this being subject to loadsa caveats about the operating conditions around the i/ps bla bla ..)
Also the 'expansive' distortion is in the opposite sense to most other distortions which tend to be 'compressive' so may cancel some other evils.
Of course I'm not putting this forward as the ONLY explanation for the nice properties of CFA i/ps but ... 🙂
I too would like to know the brand and model of this excellent amp. Also to play with the .ASC
Your jpgs are not quite readable even enlarged many times. 🙁
_________________________
Waly, the aim of this thread is to get something useful. If this is obtained by better understanding different input stages, this is still more useful than semantic pedantic arguments with Mikek.
[deleted: my slim 59 pg monograph on what is a CFA ]
_________________________
My present stance .. which is changing even now as the pearls emerge from the Golden Pinnae noise .. is that simple CFA i/ps like with VSSA have some very useful properties for power amps which may be unavailable to VFAs. I've yet to see any properties in a Diamond i/p that I want. But I'm open to what their fans might bring up.
One of these advantages is the 'compressive' vs 'expansive' behaviour on overload of a LTP vs simple CFA i/p that Marshy has posted. If you look at the numbers on the graph, you'll see that at similar currents & Gm, the 'expansive CFA' i/p has lower THD at similar voltage modulation levels and also drops faster with reducing level (this being subject to loadsa caveats about the operating conditions around the i/ps bla bla ..)
Also the 'expansive' distortion is in the opposite sense to most other distortions which tend to be 'compressive' so may cancel some other evils.
Of course I'm not putting this forward as the ONLY explanation for the nice properties of CFA i/ps but ... 🙂
Last edited:
Lets all also agree that for modern fast bipolar output devices (Ft circa 30 MHz) the limiting ULGF is 3 MHz with an output inductor and Zobel. A better design target is 1.5 to 2 MHz
Think also about th circuit parasitics at that frequency.
Think also about th circuit parasitics at that frequency.
It looks like the Bonsaï NX amp topology if you remove the cascode on input stage, the buffer before the VAS and one output driver.... Even the same transistors are used....
Fab
It is quite similar to the Revox B750 set apart the VAS driver that
are inverted in polarity as well as a non cascoded input amplifier stage.
Compensation is also shunt in the Revox but with much lower impedance
and it use only a classic double EF output stage.
The schematics arent partial or there wouldnt be any results would there. What is not shown is a feedforward circuit that is now used with updated design.
Are you not familiar with shunt comp ??
Divulgence of manufacturer, claim of ownership or origen and dilvulgence of any documentation is prohibited by contract. The design gets sold lock stock and barrel. I had to ask permission just to show this simplified version although its just a typical CFA design optimised for max performance.
As soon as I include the servo and input filter Im giving it kgrlee for review. He has asked long ago and seems impartial to the VFA CFA debate.
If this is to your knowledge the highest ULGF amp, your knowledge is very limited when it comes to what is available commercially but Im not surprised. For many the 50 year old blameless type amp is nirvana and dont look further.
Indeed a VFA can have the same or higher ULGF, such as topology using diamond buffers in differential mode. We have such a design and its also in use by marantz. Its not worse in performance but does have a noise penalty.
Lets see an example of the common-mode loops so we get a better appreciation of what you are describing.
Just about all basic "limitations' have been overcome in the newer IC CMA's. So, I am glad to read that a DIY'er has also improved the basic topology. What do you have to show us? Or, can you point to references to read.
Thx-RNMarsh
Just about all basic "limitations' have been overcome in the newer IC CMA's. So, I am glad to read that a DIY'er has also improved the basic topology. What do you have to show us? Or, can you point to references to read.
Thx-RNMarsh
This design uses the beta enhanced vas, same as the blameless for high loop gain.
Im not familiar with Revox amps but this is like I said a optimised but typical CFA design when it comes to audio. There is a dozen other manufacturers that have used very similar designs.
At its time no typical blameless amp could come close in performance. The use of TMC and TPC changed that but Im getting some interesting results with Edmonds DTMC which potentially evens matters again.
Bonsai has a design in the pipeline he claims has around the same performance. I have no doubt he can match it, there is nothing to it.
Many Assume the CM amps are unstable or less stable... perhaps because of the wide bandwidths they easily achieve. I have not found this to be the case. In fact, I find them very stable...
The headache of stablizing and lowering distortion at higher freqs, at the same time, is always a bit of finessing and doesnt always work out as planned from the SIM...
Thx-RNMarsh
The headache of stablizing and lowering distortion at higher freqs, at the same time, is always a bit of finessing and doesnt always work out as planned from the SIM...
Thx-RNMarsh
With 1K in series with 56pf there is no oscillation but PM (around 60 degrees) isn't great according to sims.
Optimise in the sims for PM/GM and transfer to real life and I get oscillations. The is no closed loop peaking in the sims. Not massive oscillationbut an extra 50mA is drawn and heat sinks get very hot.
Maybe in real life there is peaking in closed loop response but this isn't shown in the sims. I will be experimenting more tomorrow - I have a plan. 😉
PS. DTMC is superb. It's well worth investigating further. As I said before I got an order of magnitude improvement in THD with no reduction in PM/GM over straight shunt compensation in simulation. it is stable in real life too. The THD improvement was so great that the error correction could be removed.
Edit: edits due to cider consumption.
Last edited:
I ll give it a try in another simulator , here your schematic ,
not sure i drawned it correctly , as pointed by Waly it is quite
difficult to sort out the components values given the lacking
definition of the picture , i also add said Revox B750 wich date
back from the late 70s.
Attachments
Hello Richard, IMV the reasons for this are simple:-
The loop gain is generally lower, and as a result
The output stage pole falls in most cases easily higher in freq than the ULGF
The input buffer and level shifter stage +TIS can be configured for wider bandwidth than an LTP+TIS because of the latters higher gain
Ergo, in general wider loop gain bandwidths for CFA
Any others I did not catch?
Yyou cannot ignore the output stage Ft in this discussion - both VFA and CFA have to contend with this.
BTW, I was surprised to read in Ray Mancini's IC CFA TI article that the feedback resistor range for CFA OPA's is determined experimentally. I knew that this was limitied in OPA to a certain range, with gain being determined by the value of the lower leg of the feedback network, but nevertheless, no general ULGF formula as you have in VFA!
Last edited:
C1 (10m) and C7 (220ufd) are a deal killer.... all the circuits in this time/era required a large value cap(s) somewhere to stop Dc from the output. The CMA had already been published as a design concept but it could not be made in IC form until new processes allowed. So, you only had descrete designs then.
The direct-coupled design and the dc servo were conceived to eliminate this practice. Lets move forward to the present time (2013).
Thx-RNMarsh
Last edited:
Ref TI? What year? Its about 7 years old and based on infor older than that..... and Mancini did a fair job of making it useful to app engineers but didnt understand it well himself or just oversimplifed too much.
Things changed fast and processes in IC manufacturing changed so that better control and matching can be done more easily.
-RNM
Last edited:
I would be interested to know which transistors are better than bc55xC in input stage...
Fab