I have to agree that transistors with a higher gain sound better in circuit.
Also that Low gain HV transistors aren't that great for amps.
I tried to make push-pull and other amps with HV 400V transistors from computer powersupplies. If I remember correctly, HFE = 15
VCE= 400V Ic = 8A
It was harder to correctly bias these and get rid of distortion.
Using output transistors with a higher beta IMO works better, and sounds better. Having high beta drivers also, I would highly recommend, especially if you do use low beta outputs.
Also that Low gain HV transistors aren't that great for amps.
I tried to make push-pull and other amps with HV 400V transistors from computer powersupplies. If I remember correctly, HFE = 15
VCE= 400V Ic = 8AIt was harder to correctly bias these and get rid of distortion.
Using output transistors with a higher beta IMO works better, and sounds better. Having high beta drivers also, I would highly recommend, especially if you do use low beta outputs.
You can use them in the GEM amplifier, being the extra NPN, the one that provides
class A current all time long.
In that position, the gain will be not problem.... dissipation and current limit will be important...it is producing some crossover correction, conducting always.
I tested and worked fine....of course, not needing gain and linearity will avoid problems.
Having them at hand, why not to try for evaluation purposes?
regards,
Carlos
class A current all time long.
In that position, the gain will be not problem.... dissipation and current limit will be important...it is producing some crossover correction, conducting always.
I tested and worked fine....of course, not needing gain and linearity will avoid problems.
Having them at hand, why not to try for evaluation purposes?
regards,
Carlos
Are you saying that the driver stage must be able to handle the outputs after they become non-linear and the Hfe drops? The VAS would still have to produce nonlinear current in order to correct, IOW, create harmonics that cancel those produce by the output's nonlinearity. Possibly higer frequency harmonics? (mathmatically speaking)
Yes, of course the VAS still has to amplify an inversely distorted signal in order to cancel the non-linearities. The feedback loop of course causes this.
I certainly don't purport to this being the ultimate amplifier design; however, I'm of solid belief that a good amplifier -can- be made with the horizontal output transistors as long as the design of the circuit takes care of this.
I certainly don't purport to this being the ultimate amplifier design; however, I'm of solid belief that a good amplifier -can- be made with the horizontal output transistors as long as the design of the circuit takes care of this.
Now I have heard that to choose a VAS transistor that is very fast is not as good as higher gain because it is more likely to occilate. Only a rumer to me, but I suppose it would require less miller fixing to make it stable. Since the correction signal is generated by the diff by way of neg. feedback, would miller feedback or miller compensation on the VAS transistor cut the amount of correction fed to the outputs(and drivers)?
I personally think that both (Ft & Hfe) are good for VAS, if you can keep it stable without dampening the HF correction harmonics.
I personally think that both (Ft & Hfe) are good for VAS, if you can keep it stable without dampening the HF correction harmonics.
Well, what the miller stabilization tends to do is limit the high frequency response of the amplifier as a whole. Then again, if there are high harmonics in the output, the miller cap will change how they are corrected for somewhat.
Also, I have had the oscillation problem occur with faster transistors in the VAS, but if you can stabilize them without too much capacitance, then they seem a bit nicer.
There are lots of ways to get around things though.
Also, I have had the oscillation problem occur with faster transistors in the VAS, but if you can stabilize them without too much capacitance, then they seem a bit nicer.
There are lots of ways to get around things though.
I am making an amp using outputs with this gain curve to drive low impeadence, mostly lower audio frequencies. Certainly not a HOT, but I suppose it would still require some correction to linearize. Experimenting is the best part of DIY... DC SOA limits at 20A with 10V Vce and 9A with 20V Vce.
DC SOA limits at 20A with 10V Vce and 9A with 20V Vce.Attachments
Duo said:
Umm....It depends on your point of view. To me a Darlington is simply a high gain transistor with a Vce (sat) of 1.2 volts instead of 0.6 volts.
They are used exactly the same as a single transistor.
My point is; by adding another cheap transistor to the low gain output transistor you create a single high gain transistor.
Cheers
Yes...
I'm using darlington pairs here (BD139+MJE13005 Now). But... Will'nt the hfe curve problem of the power transister make problems inside a darlington...? NON-LINEAR (Hmm.. thats the problem, which wasn't actually a problem for me before posting and hearing to this thread ) current flowing from the emitter of the first transsiter to the base of second one...?
I'm using darlington pairs here (BD139+MJE13005 Now). But... Will'nt the hfe curve problem of the power transister make problems inside a darlington...? NON-LINEAR (Hmm.. thats the problem, which wasn't actually a problem for me before posting and hearing to this thread
) current flowing from the emitter of the first transsiter to the base of second one...?
Actually, in darlington, the output transistor will act more like the driver than it itself. I would have to perform some type of analysis on this, but I imagine that it might help with linearity problems of the output itself.
I do know that this will help with current gain anyway.
quasi: What I'm saying is that in an audio amplifier, the darlington configuration is one method to drive the outputs, as compared to CFB pairs or other methods.
I do know that this will help with current gain anyway.
quasi: What I'm saying is that in an audio amplifier, the darlington configuration is one method to drive the outputs, as compared to CFB pairs or other methods.
It seems to me that since in a darlington configuration, the overall beta is the product of the betas of each transistor so the nonlinearity will still appear even if the driver is linear in the operating region being operated in.
The input impeadence of the pair will be that much higher allowing the gain stage to correct for that nonlinearity with less current. Not 100% sure on this though, so experts feel free to correct this assumtion if it is wrong.
The input impeadence of the pair will be that much higher allowing the gain stage to correct for that nonlinearity with less current. Not 100% sure on this though, so experts feel free to correct this assumtion if it is wrong.