JC,
I will tell you how I stumbled across CMode operation of a FB circuit ----?
Early when SS circuits were taking over from tube designs, many did not like the sound of the SS circuits and preferred the tube sound. Now there were many reasons for the differences But, one was SS used high neg fb and tubes did not. Tube amps generally had modest nfb amount. I wanted to know if I did a SS circuit with same amount of global fb as the popular tube circuits, would it sound more like tube designs sound. Or to what degree or similarity.
In doing so, i found a sweet spot that had unusual characteristics. We know this as a current-mode amp/operation.
So, how did it compare in sound to straight ahead high global nfb? I'll just say it is better for todays sources. IMO.
THx-RNMarsh
I will tell you how I stumbled across CMode operation of a FB circuit ----?
Early when SS circuits were taking over from tube designs, many did not like the sound of the SS circuits and preferred the tube sound. Now there were many reasons for the differences But, one was SS used high neg fb and tubes did not. Tube amps generally had modest nfb amount. I wanted to know if I did a SS circuit with same amount of global fb as the popular tube circuits, would it sound more like tube designs sound. Or to what degree or similarity.
In doing so, i found a sweet spot that had unusual characteristics. We know this as a current-mode amp/operation.
So, how did it compare in sound to straight ahead high global nfb? I'll just say it is better for todays sources. IMO.
THx-RNMarsh
I agree with you Richard. I think that you found an optimum combination of linearity and feedback that sounds better than typical hi feedback designs.
I wanted to know if SS would sound like tube circuits sounded if both had same nfb etc. After that experiment, I never went back to tubes BTW.
I have stayed with current-mode operation for all amps --- modest amounts of global nfb. Trying instead to make inherently linear circuits with ultra low distortion without the use of high nfb.
I have mentioned a great new book on inherently linear amplifiers here before. IMO, that is a better approach for modern high resolution audio systems.
-RNM
Last edited:
Yes, I confess that gives us a head start; and don't forget vinyl.
But I fail to see how that becomes a bragging right for you?
Hi Richard,
What is going on with the CFA amp in Bangkok? That is an amp with quite high NFB level, more then 80 dB in whole audio band.
Best wishes, Damir
Hi Richard, I have been working on this for three years:
This is an ultra-simplified diagram, not even the required coupling caps are shown, the key is those four resistors. They basically control the tube and tube rolling is suppressed, so that all EL34s virtually sounds the same.
This uses local feedback near 20dB. But the key here is that the feedback is in real time and hence unlike normal loop nfb. But the down-side of this is that it is a beast to drive, the gain of the tubes is only 1.5x and hence the voltage swing required is hugely greater than before, like near 700V p/p is required to get 50 Watt into 8 Ohm. The above circuit has 24dB negative gain.
How does it sound? The least electronic of any amplifier I know. BTW, the Anode/Plate impedance is way below that of a Triode, probably around 25-30% and hence the nomenclature "Beyond Triode". Since the local feedback is also around the extremes of the Primary, that much lower impedance drives the capacitances and leakages of the Primary, so that a normal decent output transformer being almost flat a 20KHz can now have a flat 100KHz bandwidth.
Can the idea be used with SS devices? Maybe some can figure it out, but tubes have gain and normal SS outputs are unity gain, so I doubt it.
This is an ultra-simplified diagram, not even the required coupling caps are shown, the key is those four resistors. They basically control the tube and tube rolling is suppressed, so that all EL34s virtually sounds the same.
This uses local feedback near 20dB. But the key here is that the feedback is in real time and hence unlike normal loop nfb. But the down-side of this is that it is a beast to drive, the gain of the tubes is only 1.5x and hence the voltage swing required is hugely greater than before, like near 700V p/p is required to get 50 Watt into 8 Ohm. The above circuit has 24dB negative gain.
How does it sound? The least electronic of any amplifier I know. BTW, the Anode/Plate impedance is way below that of a Triode, probably around 25-30% and hence the nomenclature "Beyond Triode". Since the local feedback is also around the extremes of the Primary, that much lower impedance drives the capacitances and leakages of the Primary, so that a normal decent output transformer being almost flat a 20KHz can now have a flat 100KHz bandwidth.
Can the idea be used with SS devices? Maybe some can figure it out, but tubes have gain and normal SS outputs are unity gain, so I doubt it.
Last edited:
Is it important? 😉 Most here know better and those that don't are either trolls or hopeless cases 🙂
Hi Damir,
Now that I am here, I intend to replace all the output devices with more rugged ones but the pin out is different. It has not been reliable enough to go anywhere with it. I will find out why/how the outputs keep blowing up. BUT...
When it is operating, the imaging layers, stability of the imaging and specific size and location is amazing and its best feature.
I attribute this to the fact that distortion does not change with Z load, freq or power level.
But, I never brought the amps to Calif to hear on my own familiar system. Now I need to get the JBL M2 speakers here with those amps. I have a new electronic cross-over bought for it so i can get rid of the amps from JBL.
I will be contacting Mark about shipping them in a month or two... Got to pay my taxes etc first. Property and income Ouch! I already spent $9500 USD to ship a few things over here. Those M2 are going to cost a lot, also. Takes about 3 months by boat.
I have a big screen acoustically treated video demo room to play them in. Right below my office. 🙂
THx-RNMarsh
Last edited:
One more reason to add passive networks (RC+RLC) right at the speakers, to linearize their impedance in the all bandwidth.
Last edited:
Not the particular case. But in general, I think we need to react to fact degradation. It's important for the future.
//
Hi Richard, it is not just a question to use more rugged mosfets, pin out is not different for TO247AC mosfet package, but the capacitances are much higher and it could be a problem to drive them.
I understand this amp is with no protection?
Do you still use the same email address, I will mail you?
BR Damir
Possibly the TO264 package --- see Exicon.
C Input Capacitance 1850 pF
C Reverse Transfer 55 pF
Capacitance
t Turn-On Time 150 ns
V = -20V
I = 7A
V = 0
V = -10V
f = 1.0MHz
C Output Capacitance 850 pF
Exicon Lateral MOSFETS - Our Product Range
Exteamly reliable. Excellent compliments.
Single die may be better for driving but the IRF devices just dont seem to have the safe operating area that the specs imply when cooking at optimum bias.
So would like to try the Exicon.
yes, same e-mail
THx-RNMarsh
Last edited:
Those are laterals, not possible to use in this amp, it should be redesigned.
With those laterals it's possible to design simpler amp (no thermal run away)and I have 120W version in this forum but not built by me(I built with TO3 Hitachi laterals left from old JLH 80W amp and my daughter use it).
For me Exicon laterals where to expensive and I just showed schematic and layout for others to build.
BR Damir
I wrote a short article about the history of feedback here
Feedback: A Short History
We should not assign magical properties to something that is well understood Dan.
Feedback: A Short History
We should not assign magical properties to something that is well understood Dan.
- Status
- Not open for further replies.