Bob Cordell's Power amplifier book

[Bob Cordell]

Originally Posted by michaelkiwanuka

I am afraid i disagree: the loop gain sims tell the whole story from my perspective.

For those who have the time and would like to explore further the component values i gave earlier may help; here:

Bob Cordell's Power amplifier book

Mike,

For the life of me I just can't understand why YOU don't just go ahead and do it. Uuuggghhh!!!

Bob

[ostripper]

Originally Posted by JPV

Well as Roberge is explaining in his book, although TPC is unconditionnally stable, it can become conditionnally stable in a describing function way when facing clipping or saturation. This non linear event can trigger jump resonnance which is very bad and exhibits hysteresis meaning that the entry condition is not the same as the exit one. TPC is still a powerfull technique but you must avoid saturation by all means.

JPV

In an "unscientific way" , I put a traditional miller comp. amp up against the TMC one at the new years party. Since I am a prolific builder , I could care less about blown OP devices. We , at the party .. clipped , overloaded , actually tried to blow out 2 pairs of amplifiers. The traditional one sounded terrible well into clipping as it should .. we blew a fuse on them but everthing survived with a new fuse. The TMC enhanced blameless actually had a less fatiguing sound as it clipped. Neither amp had ANY hot devices in the IPS/VAS (the OPS was a nice hand warmer) no tweeters blew , no warm zobles. We tried to FRY these amps for the "hell of it". This is a good testament that Bob's amplifier examples are robust and reliable and that TMC does not detract from those strengths AT ALL. One amp we even shorted out at full power ... it actually survived with a blown fuse.

Below- the "sacrificial amp" , 80W...

OS

[jcx]

I don't suppose a direct sim of robustness to added output delay will cause anyone to rethink?

As I stated earlier in this thread a sticking point seems to be people wanting to believe that the outer loop gain/phase margin Bode plot of TMC looking like CMC means you managed greater loop gain/distortion reduction - "for free"

I find it hard to believe the resistance to the TANSTAAFL principle among a engineering oriented community

Specifically I have pointed to Bode's Integral relations and BJ Lurie's work that shows how this "conservation law" for feedback is applied

My earlier output conductance sim clearly established that the added feedback is externally visible - simply by "tugging" on the output node you can tell TMC from CMC

the Bode Integral relations lead me to expect more feedback has a cost in loop gain slope and phase margin - the extra distortion reduction performance of TMC - is not "free"


Now I have another sim - this time I packaged up Bob's amp, bringing out the VAS connections so compensation can be seen as external components

I've added a pure delay to the output of the amp - between the loaded output and the normally output connected feedback components to sim variable "excess poles" - which Bob mentions in his book can be approximated as delay

To give a baseline I have a "CMC" amp on top and the TMC compensated amp on the bottom

I step the delay from 0 to 100 ns in 25 ns steps

You should see that the "simple Middlebrook" gain probe is in the "outer loop" in both sims - and "step 1: tau = 0" does make TMC look very similar to CMC as can be read from the cursors

however as output stage delay is added the crossover frequency jumps upward in the TMC sim and gain margin collaspes - not like your 1st order loop


other things you can do with the sim - plot V(out_Xmc) in .AC to see the TMC peaking with added delay

replace the sine source with a small amplitude step (with ~ 1 us delay) and run a .tran (very high resolution 10 us, 1 ns t_step_max) - and see the TMC sim break into oscillation at higher delays


In sum: You Have been bitten by those inner loops - TMC is not as stable as CMC, the gain/phase measured in the outer loop is not informative in the TMC case

[Edmond Stuart]

Originally Posted by wahab

Hi, Edmond

On second thought, if you increase the LTP current, you mechanicaly increase the slew rate, that s why i did also use the variable degeneration route to check the cause of IMD.

cheers

Hi wahab,

In principle, you are right: the maximum SR is increased. But under the given circumstances, the signals are by no means limited by the max. SR:
IPS output current: 0.58mApp
VAS output current: 0.11mApp
These values are way below the tail current, respectively VAS-CCS. So I still believe that the IPS non-linearity is the culprit.

Cheers,
E.

[Bob Cordell]

Originally Posted by jcx

I don't suppose a direct sim of robustness to added output delay will cause anyone to rethink?

As I stated earlier in this thread a sticking point seems to be people wanting to believe that the outer loop gain/phase margin Bode plot of TMC looking like CMC means you managed greater loop gain/distortion reduction - "for free"

I find it hard to believe the resistance to the TANSTAAFL principle among a engineering oriented community

Specifically I have pointed to Bode's Integral relations and BJ Lurie's work that shows how this "conservation law" for feedback is applied

My earlier output conductance sim clearly established that the added feedback is externally visible - simply by "tugging" on the output node you can tell TMC from CMC

the Bode Integral relations lead me to expect more feedback has a cost in loop gain slope and phase margin - the extra distortion reduction performance of TMC - is not "free"


Now I have another sim - this time I packaged up Bob's amp, bringing out the VAS connections so compensation can be seen as external components

I've added a pure delay to the output of the amp - between the loaded output and the normally output connected feedback components to sim variable "excess poles" - which Bob mentions in his book can be approximated as delay

To give a baseline I have a "CMC" amp on top and the TMC compensated amp on the bottom

I step the delay from 0 to 100 ns in 25 ns steps

You should see that the "simple Middlebrook" gain probe is in the "outer loop" in both sims - and "step 1: tau = 0" does make TMC look very similar to CMC as can be read from the cursors

however as output stage delay is added the crossover frequency jumps upward in the TMC sim and gain margin collaspes - not like your 1st order loop


other things you can do with the sim - plot V(out_Xmc) in .AC to see the TMC peaking with added delay

replace the sine source with a small amplitude step (with ~ 1 us delay) and run a .tran (very high resolution 10 us, 1 ns t_step_max) - and see the TMC sim break into oscillation at higher delays


In sum: You Have been bitten by those inner loops - TMC is not as stable as CMC, the gain/phase measured in the outer loop is not informative in the TMC case

Nice work, JCX. This is the kind of analysis and direction we all learn from. Sometimes there is no free lunch, but sometimes there is a tastier one :-).

Let me chew on this for awhile.

Cheers,
Bob

[SpeakerScott]

Originally Posted by Bob Cordell

Mike,

For the life of me I just can't understand why YOU don't just go ahead and do it. Uuuggghhh!!!

Bob

Mr. Cordell,

As someone who's bought your book (and the first, fourth and 5'th editions of Mr. Self's ), read both this and the negative feedback threads start to finish, and found the vast majority of the discussion to be extremely educational...I completely understand your frustration. I am extremely confused by the points that Mike is trying to make, and his method of communication.

I just hope it isn't so frustrating to you and Mr. Stuart that the valuable discourse on compensation and amplifier design in general is shut down.

So, I thank you for your patience and your contributions to the DIY community.

Scott

[jcx]

This sim moves the simple Middlebrook gain probe to the “inner” feedback loop, measuring all of the gain around the output stage

I think you can see it is a much more useful way to view the stability margin of TMC

To look at the phase margin numbers with the added delay we have to convert the delay time into phase – since delay has linear phase with frequency the phase depends on the frequency

For the CMC loop 100 nS delay at the 977KHz gain intercept is:

100ns * 977kHz * 360 degrees = 35 degrees added phase shift

the cursor reading gives 77.8 degrees phase margin for the CMC at 0 delay,

42.6 degrees at 100 ns (append @5 to both V in the waveform formula, ie V(d_cmc)@5/V(b_cmc)@5)

77.8 - 35.2 = 42.6 which agrees with the idea that the added delay is a fair measure of phase margin


the numbers clearly don’t work so neatly in the outer loop gain/phase plot of TMC because the gain intercept is a moving target


the Inner loop Bode gain/phase measurement of TMC below Is well behaved and we can do the same calculation:

100ns * 1.69MHz * 360 = 60.8 degrees added phase shift

the cursor reading gives 58.7 degrees phase margin for the CMC at 0 delay,

-2 degrees at tau = 100 ns, step 5

58.7 – 60.8 = -2.1 which again agrees and predicts that the TMC loop will be unstable with 100 ns added delay – as the tran sim pic above shows

as a further check note that 100nS is just over the 0 degree line and the previous step at 75ns delay has poor but positive phase margin – and the tran sim shows it does settle out instead of oscillate

(Thanks to Mike Engelhardt of Linear Technology – you may have noticed the stepped waveform colors didn’t agree in the previous post Bode plot but they do match in this sims pic – the 1st time I’ve had a bug fixed over lunch! )

I assume anyone wanting to verify can mod the .asc in my previous post

[Bob Cordell]

Originally Posted by ostripper

In an "unscientific way" , I put a traditional miller comp. amp up against the TMC one at the new years party. Since I am a prolific builder , I could care less about blown OP devices. We , at the party .. clipped , overloaded , actually tried to blow out 2 pairs of amplifiers. The traditional one sounded terrible well into clipping as it should .. we blew a fuse on them but everthing survived with a new fuse. The TMC enhanced blameless actually had a less fatiguing sound as it clipped. Neither amp had ANY hot devices in the IPS/VAS (the OPS was a nice hand warmer) no tweeters blew , no warm zobles. We tried to FRY these amps for the "hell of it". This is a good testament that Bob's amplifier examples are robust and reliable and that TMC does not detract from those strengths AT ALL. One amp we even shorted out at full power ... it actually survived with a blown fuse.

Below- the "sacrificial amp" , 80W...

OS

Hi OS,

Nice report! Sounds like you had fun! That's encouraging to hear that the TMC design sounded less fatiguing in clipping.

Can we deem these results credible in spite of the amount of liquor that was probably flowing?

Cheers,
Bob

[Bob Cordell]

Some of you may recall that I introduced what I called Bridged-T Compensation (BTC) in Chapter 9 of my book, figures 9.4 and 9.5. It is TPC with a small capacitor bridging the "T", connecting directly from the VAS output to the VAS input. This capacitor that I will call C3 here is no more than a couple of pF. It eliminates the very large peak that we have seen here in the TPC loop gain characteristic that often occurs in the audio band, making the TPC loop gain characteristic look much better behaved (see Figure 9.5 in my book). Although it is not clear that such a peak causes serious problems in the closed loop response, it seems to me that it cannot be a good thing.

Accompanying this peak is the sudden phase drop to a value not far from 360 degrees, which also causes some consternation. The bridging capacitor also pretty much cures this.

At the same time, the introduction of the bridging capacitor does not increase distortion by more than 10%.

We have also seen here that if one examines the total loop gain seen by the output stage in a TMC design, we also see pretty much the same peak. Placing a small bridging capacitor across the TMC network in the same fashion as BTC yields the same benefit in what we might now call bridged TMC (BTMC).

The attached zip file contains a folder of six ready-to-run TMC loop gain analyses, three for TMC and three for BTMC, where a 0.5 pF bridging capacitor has been added.

The amplifier is the same one that I have presented earlier, but it has been further loaded by 0.1uF and 1 ohm in series to reduce phase margins.

In one pair of simulations the conventional global feedback loop is shown with the loop opened. For this loop, the gain crossover is at 1.3 MHz with a phase margin of 55 degrees and a gain margin of 20 dB (roughly the same with and without the bridging capacitor)

In another set of simulations the local compensation loop that encloses the output stage and VAS is analyzed for loop gain. For this local loop, the gain crossover frequency is 750 kHz with a phase margin of at least 73 degrees and a gain margin of at least 26 dB.

In a third pair of simulations the total loop gain seen by the output stage, due to both the global and local loops, is shown. For the total loop gain enclosing the output stage, the gain crossover frequency is 1.6 MHz with a phase margin of 29 degrees and a gain margin of 22 dB. Bear in mind the challenging load being placed at the output of this amplifier.

Without the bridging capacitor, the phase dip in the TMC total output stage loop gain goes down to a margin of 10 degrees and is accompanied by a loop gain peak of 15 dB. With a value of 0.5pF, the phase dip only goes down to a margin of 27 degrees and there is no longer any peak in the loop gain characteristic. At 1pF, the bridging capacitor keeps the phase from dipping to less than amargin of 37 degrees.

The bottom line is that the addition of the small bridging capacitor can improve the behavior of both TPC and TMC insofar as the loop gain peaking and (in my mind perilous) mid-band phase dip.

Please bear in mind that this technique, and these results, do not by themselves make a case for or against TMC vs. TPC.

Cheers,
Bob

[wahab]

Originally Posted by Bob Cordell

Some of you may recall that I introduced what I called Bridged-T Compensation (BTC) in Chapter 9 of my book, figures 9.4 and 9.5. It is TPC with a small capacitor bridging the "T", connecting directly from the VAS output to the VAS input. This capacitor that I will call C3 here is no more than a couple of pF. It eliminates the very large peak that we have seen here in the TPC loop gain characteristic that often occurs in the audio band, making the TPC loop gain characteristic look much better behaved (see Figure 9.5 in my book). Although it is not clear that such a peak causes serious problems in the closed loop response, it seems to me that it cannot be a good thing.

Bob

Hi, Bob

In fact, a cap as low as 0.2pF is enough to get rid of the TPC OL gain
peaking, so in real implementation, it s unprobable that such peakings
occur as the parasistic layout and components leads caps are enough to
create this cap, while not decreasing available gain.

Though, as soon as the loop is closed, the usual step response
peaking will remain, since the TPC loop becomes then dominant.

As i already pointed, a TMC network pondered with a TPC one
is generaly the optimmum in respect of the global behaviour,
although i m not still sure if there isnt undesirable by products.

cheers,
w