Actually, using the same components allows a fair apples and apples comparison to be made, otherwise one would never know which compensation scheme is truly superior to the other.
It demonstrates, components being equal, that "TMC" applies the same amount of loop gain about the second stage and the output stage as TPC applies about the whole amplifier for the same unity loop gain frequency and the same stability margins.
Given this (entirely correct) information one should be able decide for oneself which compensation arrangement one prefers.
Incidentally, you should be aware that the algebraic expressions for the location of the zero restoring a single pole roll off in both TPC and the minor loop of "TMC" are identical. This constitutes one more piece of evidence of the relationship between TPC and "TMC".
It follows, therefore, that if component values selected for TPC result in a poor location of the zero restoring a single pole roll-off with TPC, causing instability or compromised stability margins, then one should expect the same instability or compromised stability margins within the minor loop with "TMC", as both techniques give a double pole single zero loop gain frequency response characteristic for the major loop in the case of TPC and the minor loop in the case of "TMC".
Last edited:
So-called power DACs can never approach the subtlety and resolution of a small signal DAC.
Digital to analogue conversion for true high resolution audio is, therefore, the province of small signal DACs.
This is one reason why linear amplification will be around for the foreseeable future.
Next the digital ear wireless connected to the internet Nerve Stimulation Highly Effective in Battling Tinnitus, Nature Study Reports | Professional Hearing Services Lakewood, Colorado it can be done.
I believe the real future will be developing the DSP technology to the point of being able to completely deconstruct any archival musical material -- Lexicon were going great guns on this, but stuffed up and the project was aborted -- effectively separating all the strands into as many tracks as seems reasonable. Then, just reassemble as suits the individual, feed into as many speakers as you like, remix all the classic albums into a version that sounds good to you. Everyone can become a studio recording engineer, able to play with the virtual microphone feeds to their heart's content ... 🙂
Hi Damir.
I didn't comment because I don't understand the post but I like to learn more.
You say it's a Middlebrook probe but I don't see that, and it is not clear where it is placed.
If you just clipped the circuit and the plot rather than the whole screen then there would be less clutter and you could include more of the relevant information.
Best wishes
David
Hi David,
Middlebrook probe is on the bottom(V6).
To get Loop Gain you do .ac (V6 should set AC amplitude to 1 and input signal V set to 0)
and you do the plot -V(vout)/V(A). A is lebel net in my shematic. Remember that all names here are fom my shematic, you could use others.
BR Damir
Attachments
I see the location of the probe now, it was a bit cluttered on the previous post and I missed it.
But I still don't think what you have is a Middlebrook probe.
There are actually 2 different Middlebrook techniques but usually people mean the 1975 version.
That requires 2 separate measurement and is more complex than what you have done. It is described in a copy of the paper on his website.
But it is probably not required here because the probe location is practically ideal.
Best wishes
David
What you have done is probably very accurate because the impedance on the output is very much lower than the impedance on the other side of the probe.
But it is not "Middlebrook".
We should be careful with his name because his work was exceptional and he kindly made it freely available, even after his death.
Best wishes
David
Last edited:
Most probably not. Because of economical issues. You will see power devices
for audio purposes disappear , as they are discontinued, one by one.
It doesn't pay off to invest in research anymore. The wonderful Thermal Track
BJTs didn't gain market share and were quite soon discontinued.
One big issue is the sources. Vinyl players CDs ...are all niche markets. No growth expected at best. Younger folks have "music off the net" or internet radio and these are the issue for active speakers equipped with cheap reliable
switching amps. Off the shelf.
I would call it just voltage loop gain measurement.
But I don't have these techniques entirely clear in my mind yet, happy if someone has a better idea.
There are neater ways rather than duplicate the circuit. It is possible to insert just one probe and then have it do two measurements so the results can be combined. This is what I would call a Middlebrook probe.
Two separate circuits and then combine the results I would call Middlebrook (1975) calculation.
But Tian probe is simpler in some ways and theoretically more accurate, so the name problem can be avoided.😉
Best wishes
David
Here's what I would like to see covered with your usual rigour in the next edition:
Current Dumping.
Feedforward
Error correction.
Non-Linn topologies.
A fully vetted practical design, 100W - 200W (any topology), with full output protection that could be built by a competent DIYer. PCB layout please.
Current Dumping.
Feedforward
Error correction.
Non-Linn topologies.
A fully vetted practical design, 100W - 200W (any topology), with full output protection that could be built by a competent DIYer. PCB layout please.
Dr. Middlebrook's Website
Easy to find if you make the effort to search, but too valuable not to post anyway.
And the paper is here
http://www.ele.tut.fi/teaching/ele-3100/lk0809/tehol1/MiddleBrook75.pdf
Best wishes
David
Last edited:
The design covered by D. Self in his book has nothing to do with Linn's design. The later was a single gain stage design, while Douglas's is in fact due to Thompson.
Good; I didn't like them. The ideal "Thermal Trak" device should have another BJT, and not diodes, on the same substrate as the power device. This would allow the user to design a near ideal bias multiplier.
I think it's Lin, not Linn. (From Harry C Lin who invented the topology while at Bell Labs in the 1950's)