KSTR said:
Yes, this would be good. I tend to like the idea of back-driving tests. One test I have thought of, which I think is pretty cruel to an amplifier, is to drive it forward with 20 Hz and drive it backward with 19 Hz. Sort of like a very LF version of CCIF. The 1 Hz beat note will really stress the amplifier and will cause a considerable excursion over SOA conditions.
Of course, even a 19Hz + 20Hz forward LF CCIF test would likely be revealing. The power supply current and output power would go from very low to full power in concert with the 1 Hz beat note.
Cheers,
Bob
Bob Cordell said:
Bob,
Just a quick one. One of my favoured designers JLH would use a Or22 resistor on the output line for a variety of reasons. In one amp he designed (mosfet follower output stage with ample stability margins - important) he omited the output inductor and it's damping resistor all together. Apparently it can be done with BJT's in some designs but is not straight forward.
Obviously the damping factor is adversely affected but JLH says two things. It can help damp the back EMF from the speakers such as dynamically delayed echo effects. And recently someone mentioned the following remark.
"The final point of note on the simplified diagram is the 0.22 ohm resistor in the output to the loudspeaker. This serves to prevent DC negative feedback within the amplifier, which stabilises the DC working point voltage, from disappearing in the event of an inadvertant short circuit across the output.
But having said that, I seem to recall that he later also made the point that Salas makes. I can't find the quote exactly but I think it was to do with reducing the amount of EMF that got back into the amp. I'll try and chase the reference up." (taken from another post). PS. I don't have the full JLH articles with his explanations etc, so I am going by what other people say.
My point is even back driving an amplifier with another amplifier will probably not show this. Have you tried this and observed using the R and speaker with a scope (my inclination is JLH would of and seen some difference doing this). I can't see him just putting it there.
Any chance you could do some testing of this? or give some insight into this. It's not as simple as Zout is 0.001% etc. JLH could design amps to do this easily.
Best Regards
Kevin
and other relevant posts.Fanuc said:
Gentlemen?
I could not find the remark by Salas (reference please?) and would not claim to be able to remember any great portion of this informative thread, but something jars here:
We are talking about loudspeaker drive. I understand the argument, but how much 'back'drive' are we considering from a typical loudspeaker? I seem to miss that the lowest loudspeaker impedance must be the voice coil dc resistance, which for an 8 ohm system usually comes to about > 5 ohm minimum. (OK, I have heard that in practice some funny designs go down to 2 ohms - must we entertain those??)
Thus taking say 5.6 ohm resistance, the maximum real damping factor according to basic electricity is 8/5.6 = 1.43. Adding another 0.22 ohm makes the loudspeaker resistance 5.82 ohms, then giving real DF as 1.38 .... etc; not to go off topic and argue DF here. Only the said resistances/impedances are in the circuit for all frequencies (to simplify), which as I see it makes loudspeaker influence something of the nature of (amplifier output impedance)/(total circuit resistance) - a large ratio under any condition, even including phase differences. (I hope we are talking about decent amplifiers with negligible phase shift in the audio band plus some.)
I apologise if I have missed something - but what would that be, exactly? Any calculation I make shows a minute effect from (loudspeaker) back drive under typical conditions.
Fanuc said:
Hi Kevin
That is right, and it’s the same with all amps using GNFB.
Please note that I’m not against GNFB, but I think it’s important to look at the GNFB loop as a loop from the input to the output and back to the input, thereby as I commented in another thread you can not look at the FB loop as a “light speed” thing.
Sorry if this was OT.
Cheers
Fanuc said:
Sometimes a little bit of resistance can influence the sound in a way that is subjectively pleasant. A 0.22 ohm resistor will reduce the damping factor to below about 36. This is beginning to get into the range of tube amps. I actually built a SS amp with a resistor in series with the output to make it sound more tube-like. That amplifier also had a soft clipping circuit. It was about a 45-watt amplifier. It sounded quite good, and rather similar to a KT-88 - based 35-watt tube amp I built.
Note that such a resistor will have an effect on frequency response into typical loudspeakers; just look at the frequency response of some amplifiers as measured in Stereophile into their test loudspeaker load.
Cheers,
Bob
Bob Cordell said:
Thanks for your reply Bob but dissapointed in your answer!. Let me show you some real scope outputs where back EMF whether voltage or current is being used as an input signal for a motional feedback subwoofer signal. It has nothing to do with sounding like a tube amp or the amplifier/speaker response variatons (I accept the technical arguements/reasoning for this). A 180 degress phase shift (reactive) in a amplifier output stage will dissipate nothing into the load etc.
Need to find the relevant mag with technical explaination (could be all over the place! lol). Moving house etc.!
Let us try Otala's waveform (that one that draws 6 times the current from the O/P stage) with various speaker simulated loads including the old quad els57, magneplanar ribbons and tamer loads - something that will actually back drive a power amplifier. It's important that this is not about peak power etc into a load. and things blowing up.
Reactance can affect phase margin stability. This shouldn't be confused with what my point is or what JLH has made.
Gotta build four amps this week, getting all the parts ordered etc. Nout like it!
All the best
Kevin
EMF, GNFB and JLH 0.22 Ohm output resistor
post #26
http://www.diyaudio.com/forums/showthread.php?postid=1879538#post1879538
post #28
http://www.diyaudio.com/forums/showthread.php?postid=1879580#post1879580
post #26
http://www.diyaudio.com/forums/showthread.php?postid=1879538#post1879538
post #28
http://www.diyaudio.com/forums/showthread.php?postid=1879580#post1879580
Fanuc said:
Thanks for your reply Bob but dissapointed in your answer!. Let me show you some real scope outputs where back EMF whether voltage or current is being used as an input signal for a motional feedback subwoofer signal. It has nothing to do with sounding like a tube amp or the amplifier/speaker response variatons (I accept the technical arguements/reasoning for this). A 180 degress phase shift (reactive) in a amplifier output stage will dissipate nothing into the load etc.
Need to find the relevant mag with technical explaination (could be all over the place! lol). Moving house etc.!
Let us try Otala's waveform (that one that draws 6 times the current from the O/P stage) with various speaker simulated loads including the old quad els57, magneplanar ribbons and tamer loads - something that will actually back drive a power amplifier. It's important that this is not about peak power etc into a load. and things blowing up.
Reactance can affect phase margin stability. This shouldn't be confused with what my point is or what JLH has made.
Gotta build four amps this week, getting all the parts ordered etc. Nout like it!
All the best
Kevin
Hi Kevin,
I'm sorry I missed your point(s).
I agree that in some amplifiers without an output coil that the presence of a 0.22 ohm resistor might help HF stability a bit into some loads.
The back EMF from a loudspeaker results in a current, not a voltage (as long as the output impedance of the amplifier is small compared to the voice coil resistance). As such, it will flow into the output stage with or without a 02.. ohm resistor in series. Any damping of back emf provided by the 0.22 ohm resistor is simply the result of lowering the damping factor of the amplifier so that the output impedance of the amplifier is a bit more resistive than it otherwise would have been.
All of these concerns underline the need for the amplifier output stage to supply high output currents under certain conditions. The numbers can get ugly fast with loudspeakers with small minimum impedances.
I think the particular waveform you are referring to was the one that I concocted in my AES IIM paper available on my website. Self shows a near-replica of that waveform in a later revision of his book.
Cheers,
Bob
stinius said:
Well if your point was that gnfb makes the '5 inputs' even worse, that's wrong too.
Gnfb generally makes those '5 inputs' less sensitive to disturbances. For instance, if your power supply ripple/noise rejection ration (PSRR, the two supply 'inputs') is 60dB at some freq, and you apply 30dB of feedback at that freq, you have now increased your PSRR to 90dB.
And if your output (another 'disturbance input') rejects externally impressed disturbances at a certain freq by 40dB, and you apply 30dB of gnfb at that frequency, your rejection at that output is now increased to 70dB.
Winning on all fronts, thanks to gnfb.
Jan Didden
Member
Joined 2009
Paid Member
Bigun said:
I have always been a big fan of bi-amping and tri-amping for a multitude of reasons, including the good one you point out.
I started my love of audio in the tube days in the Sixties. Tube amps had their limitations. I was tri-amping with three home-built stereo amplifiers when I was in high school.
Another reason I tri-amped was that I did not like winding crossover inductors! For me, active crossovers were simpler, although I must admit unsophisticated at the time.
Each of my Athena active loudspeakers has four 125-watt MOSFET power amplifiers built into it, with a 3-1/2 way active crossover. People may think I'm crazy putting a 125-watt amplifier behind a tweeter, but I've seen 30-volt peaks across the tweeter on some material (yes, my favorite crest-factor buster, Rickie Lee Jones, Getto of my Mind).
Cheers,
Bob
maybe some others will believe, now that you have put your weight/opinion forward.Bob Cordell said:
I keep repeating that the target peak SPL of each driver should all be about the same level.
If one wants 110dB from the mid driver then the system design should at least try to achieve that same peak SPL from all the drivers.
Bob Cordell said:
I’m also a big fan of active x-over.
I like to use power amps that are a lot more powerful than the loudspeaker; the loudspeaker can handle a pure signal that is too large better than a distorted or clipping signal at lower level.
OTOH I don’t like to put the power amp inside the loudspeaker cabinet; there will always be some micro phonic issues to consider.