Nearly double quality on output device. Hefty drivers, a trade-off that might be moot. That makes an application question. If one uses 2sa1930 driver, then what is the amplifier's watts output? I suspect that this makes a small scale amplifier with big devices.
Circlophone
Call it a lucky accident for me that I happened across this wonderful design.
It's an "off the beaten track" design that achieves the ends beautifully and is highly 'tolerant'. My congrats to Elvee for the pioneering design and the others for the contributions to the thread, esp the builder's thread.
What am I waiting for, as I have most of the parts in the bin ... !!
Congrats again!!
Call it a lucky accident for me that I happened across this wonderful design.
It's an "off the beaten track" design that achieves the ends beautifully and is highly 'tolerant'. My congrats to Elvee for the pioneering design and the others for the contributions to the thread, esp the builder's thread.
What am I waiting for, as I have most of the parts in the bin ... !!
Congrats again!!
Thanks
Conservatively, this would translate into ~500W on a 4Ω load, and 250W on 2Ω, thus quite feasible.Nearly double quality on output device. Hefty drivers, a trade-off that might be moot. That makes an application question. If one uses 2sa1930 driver, then what is the amplifier's watts output? I suspect that this makes a small scale amplifier with big devices.
But the 2SA1930's would require a good heatsinking to achieve that.
Best solution would be to mount them on the main heatsink.
It seems like one could make a concert-sounding DJ amp sized circlophone.
Ah, I'm sorry to ask again, but maybe I didn't ask correctly.
I would wish to derate corresponding to value in watts when running 4 ohm speakers and the amp is run up well in to the 10% distortion range?
Let us be clear: you asked about the power the drivers could manage with MJW18020-type transistors, I gave you an answer from the driver's point of view.
But asking 500W from a single pair of transistors would be unwise, not to say completely insane.
The MJW's could conservatively manage 100W, they could deliver up to 150W without too many risks, but going beyond would be dangerous.
There is no problem with the distortion: they could deliver 500W at 0.005% THD, but the SOA is insufficient to do it safely.
Thank you very much. I had calculated 37+37vdc rails for it, if 4 ohm speakers are used. That's something like 150w@1%RMS, isn't it?
Question: Would parallel ouput devices put more stress on the drivers?
It is.
No, not particularly: with ideal output transistors, there would be no change at all, because n*(Ic/(n*Hfe))= Ic/Hfe.
But in reality, transistors have a gain droop at high current, which means the composite multiple transistor will have a higher Hfe than a single one, easing the task for the driver.
I don't mind in the least: it is an opportunity to hint at what this thread title refers to:
MLCD [My Little Cheap Dictaphone] : The tragic tale of a genius
I hope you don't mind having such a bucket of subculture thrown at you!
@Elvee
Then perhaps a minor addition like 'My little, cheap Circlophone' ....
Sorry about the 'comic relief' in the middle of serious discussion.
Genius will always be rewarded. Period.
I'm sure Alan Turing would disagree.
Good fun is one of the most serious and important thing in life, you have no reason to feel sorry.
I'm curious about the relationship between the driver and output HFE.
For example, a Tip41 (some of them are HFE 15) or MJW18020G output device. . .
And, then there's driver selection:
KSA1220AYS (BD140-16) or 2SA1930 (BD140-8)
Is it true that the high HFE driver will run hotter because of doing more work, and in this case we want to put the higher hfe driver with higher hfe output and likewise put the lower hfe driver with lower hfe output?
For ease in these selections, is there an ideal proportion between driver versus output?
There's been something going on for quite a while and the number of repeatable results samples has added up well enough to say that this is probably not subjective.
For example, a Tip41 (some of them are HFE 15) or MJW18020G output device. . .
And, then there's driver selection:
KSA1220AYS (BD140-16) or 2SA1930 (BD140-8)
Is it true that the high HFE driver will run hotter because of doing more work, and in this case we want to put the higher hfe driver with higher hfe output and likewise put the lower hfe driver with lower hfe output?
For ease in these selections, is there an ideal proportion between driver versus output?
There's been something going on for quite a while and the number of repeatable results samples has added up well enough to say that this is probably not subjective.
The basic requirement for the combination driver+OP is that the product of the Hfe's must be sufficient.
Using basic arithmetic, it is easy to make a quick and simplified estimation: the OP stage has to provide a peak current dependent on the supply voltage and the load: for example, +/-40V and 4 ohm mean 10A.
The VAS has a quiescent current of ~13mA. If we don't want to exceed a 50% imbalance, this means the available current excursion is limited to 6.5mA.
If we divide 10 by 0.0065, we get ~1540: that is the total gain required from the CFP to achieve the target power.
You can distribute that gain the way you want, an obvious method is to take the square root, in this case about 40 for each.
But the dissipation in the driver only depends on the OP's Hfe, not its own Hfe. But of course, if the OP has a low Hfe, the driver will need to compensate for it to meet the above calculation.
Note that the calculation is rather pessimistic: not only can you use the full 13mA of the VAS, but in case it is still insufficient, the bias servo will intervene and raise the average current level to meet the demand. This is why the Circlophone is able to drive obscenely low loads, but the linearity is somewhat degraded in those conditions.
The proportion itself is not important, provided the global Hfe is sufficient. What is important is the linearity of the Hfe vs. current.
The combination driver+OP has to be as flat as possible, this will in general mean that both the drivers and OP are flat themself, although some compensation of non-linearities would be in theory possible (although difficult to achieve in practice)
Let's see if I understood.
It is a current function.
Any excess driver gain serves like a spare tire--When the op gain drops when/because it is delivering more or even peak current, the driver makes up the difference only during that condition. And we don't want that to be abrupt.
If the op is poor, with a fluctuating shift, the op's sonic signature is heard during low volume playback, but when the op falters, the drivers kick in much harder with abrupt change to class a tones for high volume playback and dynamics.
or
If the op is good, with a level behavior, then the "hand-off" is gentle and transparent, with a well balanced sonic signature, and the intended function of a Circlophone. Oh, that OP short-list was important, wasn't it?
P.S.
If someone really wants parallel outputs, it seems they should consider matching some MJW18020G.
It is a current function.
Any excess driver gain serves like a spare tire--When the op gain drops when/because it is delivering more or even peak current, the driver makes up the difference only during that condition. And we don't want that to be abrupt.
If the op is poor, with a fluctuating shift, the op's sonic signature is heard during low volume playback, but when the op falters, the drivers kick in much harder with abrupt change to class a tones for high volume playback and dynamics.
or
If the op is good, with a level behavior, then the "hand-off" is gentle and transparent, with a well balanced sonic signature, and the intended function of a Circlophone. Oh, that OP short-list was important, wasn't it?
P.S.
If someone really wants parallel outputs, it seems they should consider matching some MJW18020G.
It is a picturesque description, but that's about it.
The OP transistor type is decisive, it is the single component having the most effect on the quality of the amplifier.
Note that the "original" Circlophone has been completely tested and characterized with 2N3055's, and it isn't bad. But it can be even better with better transistors.
Matching is a luxury, but it can help. It also allows for absolute minimal degeneration resistors, which is a good thing.
Maybe true for EF outputs, I dunno about that statement for Circlophone.
OP is the VAS and slowest stage, quite possibly an advantage for stability.
You really need all transistors aside from VAS to be in stable local loops
with much higher knees. And that's the challenge. Else you are having
to sandbag even the best VAS transistor down into the audio band to
avoid an overlapping phase shift.
For stability, open loop gain often wants to roll just above 1K. Does
not hurt closed loop performance at or below 1K. But everything to
be heard above the open knee is far less correct. And this is no fault
to blame an OP transistor, especially if you are forced to add CDOM
to stabilize. But we prefer this knee should be well above audio.
Any OP VAS with natural knee above 20K will do. What goes before is
what determines how much open loop bandwidth can be made stable.
Why Circlophone should do extremely well even with 2N3055.
Last edited:
I was talking mainly in the context of linearity, Hfe flatness versus collector current, not dynamic characteristics.
But the bandwidth will also play a role. If the OP are too slow, the drivers will need to provide larger current excursions to drive them properly.