????
Can-we say that the performance of a formula one should not be dependent of the motor used ?
I use a Tian probe. I also tried the Middlebrook method (duplicate the schematic, etc...), to exactly the same results (as expected).
I can't be bothered to redo everything only for your convenience of rescaling. The results, as they are, are more than conclusive.
Hi Waly,
I have run the simulation again with your inordinately large value of 100ohms instead of my 22ohms and I've found that the phase shift at unity minor loop gain frequency without the 100R resistor is 153 degrees (green trace), and with the 100R resistor in situ the phase shift is 145 degrees. A slight improvement, but an improvement nevertheless.
It would appear an excessively large value of TIS degeneration resistor is required to reduce the minor loop transmission significantly so as to effect a marginal improvement in phase margin.
Certainly, with the more usual 22R resistor there is no such improvement in stability margins.
Note that the Middlebrook method you refer to gives errors at HF. Also see attachment below for my derivation of that particular method.
Also see the following:
http://www.diyaudio.com/forums/solid-state/25438-phase-gain-plot-will-oscillate-5.html#post463340
Attachments
Last edited:
This analogy doesn't work, but what do I know.
The point I was trying to make, clearly not very succesfully if your post is anything to go by, is the perfomance of a feedback amplifier should not be dependent on the transistors used.
For instance using MPSA18 and Zetex1056 for the TIS stage should not make a significant difference to things like major and minor loop transmission.
It's not a point you are going to get across because it's not true:
1) The distortion performance of an output stage depends on the beta characteristics of the drivers and output devices.
2) Try using a power transistor as a VAS. Shocking!
Both these effects are demonstrable in simulation without difficulty.
3) Using the wrong transistor type in the input current-mirror can cause perplexing problems with distortion- not sure if that will simulate.
All these issues are dealt with in the 6th edition.
If larger can one have the best of both worlds ? For example if 1 uf would it coincide with the area which the book illustrates that distortion rises . I recommended to a friend 27 R + 1 uF for him to simulate . This was for a secondary reason as I fell it suits the LTP current mirror better as it raises the VAS Vbe a little ( about 0.2V at 8 mA ) . The Vbe can be as low as 0.55 V which is very close to the saturation voltage of the mirror if including swamping resistors .
I can see if the stability is better with lets say 100 R a capacitor might make it worse . Worst of all worlds perhaps ?
Actually I was refering to "small signal" transistors, and then only in respect of minor and major loop transmission..
Last edited:
Don't forget operating conditions & currents ... another point that (pseudo?) guru Kiwanuka seems to regard as trivial.
Students of the art may wish to look at the currents of VAS, drivers, pre-drivers & outputs in his myriad examples to learn his greatest skill .. how to muck up someone else's design.
_____________
Michael, you have poured scorn on various people's BJT SPICE models. Can you tell us how you have validated YOUR models? As some have pointed out, they appear to have some 'fantastic' properties.
How have you checked that these fantasms are real? Did you check? Or are these just the models that support your theories ...
Facts can always be adjusted to fit theories
Last edited:
Students of the art may wish to look at the currents of VAS, drivers, pre-drivers & outputs in his myriad examples to learn his greatest skill .. how to muck up someone else's design.
Inutterable nonsense.
Where and when precisely have I "poured scorn on various people's BJT SPICE models"?
Get a grip.
You gonna tell us how you validate your models?
Some (all?) of us 'don't trust them' (your words IIRC used with reference to Bob's models. My facility with English as she is spoken may have missed some subtle nuance of yours.)
But rather than focus on who said what, telling us how you validated your models will actually increase the store of human knowledge ... and increase your credibility too. Win - Win
Do I look like I care what you think of my models?
I suggest you cogitate and ruminate exhaustively on the following post before posting anything else addressed to me:
http://www.diyaudio.com/forums/soli...-self-wants-your-opinions-89.html#post3530109
In particular note the following:
Have you actually tried to replicate or otherwise perform my simulations with the "correct" circuits and models of your choice? Hell No!
Indeed, all you've done in response is captiously cling to extraneous and wholly impertinent minutiae such as whether the quiescent TIS current differs by a few milliamps and the like, seemingly entirely oblivious to the fact that an amplifier's major loop and minor loop stability should be completely insensitive to TIS quiescent current or the degeneration resistors used in the current mirrors and other irrelevant trivia.
Ciao!
Douglas, I can't wait to get my hands on your sixth edition. I hope Amazon will deliver my copy before the end of the first week of July.
Ah! If it only were so! Unfortunately, it isn't, and I do not mean some special batch of once only super duper transistors made by whoever.
Here's a small test you can try out for yourself. Use a classic Darlington, or a cascode if you prefer, arrangement for a VAS stage. I purposely mention the VAS because it's neither at the beginning, or the end of an amp, but right in the middle.
Make it easy on yourself and use transistors from the same family and manufacturer, suspecting that they are probably at least similar. So, make the stage with ANY gain you like, from 20 to 80 dB, your choice, using a say BF471 or 472 (1.8W, 300V) for the "lower" transistor. Do what you need to do and make the proper measurements.
Then, changing nothing else, substitute them for BF 720/721. Lord Electron will smile upon thee, pilgrim. The immediate results are that classic static measurements like THD drop, and not insignificantly, while your bandwidth incrases. Then, if you make sure your 720/721 run with at 5-6 mA of current, where they are the at their best, again better results.
When you put the circuit to actual everyday work, you hear a faster, more precise and better focused sound. Not a revelation, mind you, no heavenly light on you, but you do hear it with even a decent system, let alone a very revealing one.
My feeling is that the more gain a stage has, the better you CAN hear and measure differences between transistors, and vice cersa, the more you degenerate a stage, the less of a difference you are likely to hear.
But either way, if you seriously want your amp to sound good, you need to be literally intimate with the transistors you use, you need to know the especially good sonic matches. Just reading the Data Sheets simply isn't enough.
Sorry for barging in, but Nige and I have been at this same discussion for about a year now, off and on.
Sorry gents, I have to say this.
Mr Cordell, thank you for your most excellent book. It's not at all often, about the rate of wisdom teeth, that one has an opportunity to encounter such a methodical and lucid piece of work.
I've been waiting for something like that ever since 1980 and the National Semiconductor Audio/Radio Handbook, and I daresay, it was well worth it. And trust me, I know what I'm on about, I published a book locally which sold out its three editions, and would have done even better if some local idiots hadn't started a series of local wars 1991-1996.
Mr Cordell, thank you for your most excellent book. It's not at all often, about the rate of wisdom teeth, that one has an opportunity to encounter such a methodical and lucid piece of work.
I've been waiting for something like that ever since 1980 and the National Semiconductor Audio/Radio Handbook, and I daresay, it was well worth it. And trust me, I know what I'm on about, I published a book locally which sold out its three editions, and would have done even better if some local idiots hadn't started a series of local wars 1991-1996.
For the avoidance of doubt, I will say it just ounce more: I was, in fact, referring to small signal transistors and only in the context of minor and major loop gain.
So, you can actually hear small signal transistors in an amplifier with large quantities of feedback, can you?
Last edited:
One thing I would like to ask is what is an ideal VAS ( or trans-impedance stage ) current ? A stupid question I know . Still I will make a stab at it . Often I say 8 mA . 5.6 mA seems to often choose itself in my designs . On the face of it the more the better . My ears say not true . Perhaps it is simply like a jigsaw puzzle ? It all falls into place when right . Fixing one parameter helps that is certain . One thing that is worth asking is if using output triples does one need less ? If so do typical amps still run about 5 to 8 mA because it is easy for them to do so ?
Sorry if my question is too mundane for the advanced levels . Like the optimum cylinder size of an engine it is the sort of thing worth knowing . 330 cc set by Citroen in the 1920's I think ?
I am laying bets if answered at all it will be some need this and others need that . I think that's where I am at already . I am hoping for a pearl of wisdom here . My guess is 0.5 to 1 watt dissipation sets the VAS current?
I think the VAS transistor will easily sound different . It's inherent capacitance is all there is to know . I think in the Handbook it says internal capacitance not as nice as adding it externally . Naturally all people here know that . However it is overlooked elsewhere . The classic example is people saying no VAS cap required if single input transistor . This is simply not true . It was the high capacitance transistors of old that made that possible . Gain is an obvious problem so I won't state it . My guess is it is like the high jump , clear the pole and the sound should be the same ?
Sorry if my question is too mundane for the advanced levels . Like the optimum cylinder size of an engine it is the sort of thing worth knowing . 330 cc set by Citroen in the 1920's I think ?
I am laying bets if answered at all it will be some need this and others need that . I think that's where I am at already . I am hoping for a pearl of wisdom here . My guess is 0.5 to 1 watt dissipation sets the VAS current?
I think the VAS transistor will easily sound different . It's inherent capacitance is all there is to know . I think in the Handbook it says internal capacitance not as nice as adding it externally . Naturally all people here know that . However it is overlooked elsewhere . The classic example is people saying no VAS cap required if single input transistor . This is simply not true . It was the high capacitance transistors of old that made that possible . Gain is an obvious problem so I won't state it . My guess is it is like the high jump , clear the pole and the sound should be the same ?
Hi Guys
Maybe it depends which time zone you are in? I don't know.
I don't see how ADAP6 will mail to you in the first week of July when both amazon.ca and amazon.com say the book won't arrive until the 13th. In my time zone, the 13th falls into the the second week of July - actually the last day of that week (Sat). I'm expecting my copy to arrive in the fourth week of July.
Really, if your time warp machine works very well you could get ADAP6 before Doug wrote it and tell him about all the mistakes he's about to make. That would change the time-line in this universe, but it would just shift us into a different quantum universe as all possibilities and outcomes exist and happen... according to those with strings...
The brass section keeps its own time.
Have fun
Kevin O'Connor
Maybe it depends which time zone you are in? I don't know.
I don't see how ADAP6 will mail to you in the first week of July when both amazon.ca and amazon.com say the book won't arrive until the 13th. In my time zone, the 13th falls into the the second week of July - actually the last day of that week (Sat). I'm expecting my copy to arrive in the fourth week of July.
Really, if your time warp machine works very well you could get ADAP6 before Doug wrote it and tell him about all the mistakes he's about to make. That would change the time-line in this universe, but it would just shift us into a different quantum universe as all possibilities and outcomes exist and happen... according to those with strings...
The brass section keeps its own time.
Have fun
Kevin O'Connor
No.Balanced drive to the output stage is essential for symmetrical slew rates. And no, I don't think 40V/us on a 150W amp is adequate.