Your opinions are sought on Audio Power Amplifier Design: 6th Edition. Douglas Self

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Possibly "common mode lockup" is known by different names in different parts of the world. This is a case where the signal is faster than the output transistors and the second transistor turns on before the first has a chance to turn off. The result is the output transistors become a short across the power supply rails.

My understanding is that quasi complimentary outputs are more susceptible to this lockup condition than full complimentary.

The transistors were house numbered RCA 40411 which have ft = 800 kHz, very slow by any standards today.

I was increasing the volume to maximum with an open input on phono to listen for noise using headphones.

We call it "shoot through" in power electronics.
 
Possibly "common mode lockup" is known by different names in different parts of the world. This is a case where the signal is faster than the output transistors and the second transistor turns on before the first has a chance to turn off. The result is the output transistors become a short across the power supply rails.

I don't see output device conduction overlap as a plausible failure mode. The limitation on amplifier speed is almost always the slew-rate limit, which is set by Miller compensation. I think that would be true even with ancient and very slow output devices. The problem might have been latch-up (though you could certainly call it lock-up) triggered by some hard clipping. I have seen this before with unconventional amplifier configurations, but I would have thought a Heathkit would use conventional circuitry. Can you show us a schematic?

It could of course just have been failure of an overstressed small-signal transistor, the VAS being the prime candidate unless it is fitted with current-limiting to protect it during negative clipping.

My understanding is that quasi complimentary outputs are more susceptible to this lockup condition than full complimentary.

I don't think I have never heard that before, if we're talking conduction overlap. I suppose that there might be a difference in how fast the charge carriers are sucked out of the base at turn-off. I can think of no good reason to use quasi-complementary output stages in this day and age, so not a high priority for research.
The transistors were house numbered RCA 40411 which have ft = 800 kHz, very slow by any standards today.

I was increasing the volume to maximum with an open input on phono to listen for noise using headphones.

Not sure what you hoped to learn by that. Sounds hazardous, and so it proved.
 
Administrator
Joined 2007
Paid Member
That's the second time in a couple of days I've encountered the term "common mode lockup" and "common mode conduction", the first being on a Carver amp that has a strange transistor added between the bases of the final output devices. Apparently it is to prevent "common mode conduction", and certainly a technique I have not seen used before.
 

Attachments

  • Capture.PNG
    Capture.PNG
    98.3 KB · Views: 449
.....
, but I would have thought a Heathkit would use conventional circuitry. Can you show us a schematic?

It could of course just have been failure of an overstressed small-signal transistor, the VAS being the prime candidate unless it is fitted with current-limiting to protect it during negative clipping.
.....
I can think of no good reason to use quasi-complementary output stages in this day and age, so not a high priority for research.
.....

Schematic attached. Circuit from the VAS forward is almost the same as an RCA document from maybe 1967.

I don't think the VAS went out, just drivers, outputs, and emitter resistors. It was rebuilt and has been working for over 40 years.

You are correct, not a high priority for research. I looked at redesign several times but this unit also has a single supply and capacitor coupled output and with no center tap on the transformer, there is no way to redesign into anything worthwhile. Thanks for looking but don't spend too much time on this.
 

Attachments

  • Screen shot 2014-09-08 at 5.21.46 PM.png
    Screen shot 2014-09-08 at 5.21.46 PM.png
    83.4 KB · Views: 434
Schematic attached. Circuit from the VAS forward is almost the same as an RCA document from maybe 1967.

I don't think the VAS went out, just drivers, outputs, and emitter resistors.

Well, there's certainly a fairly big resistor in the VAS emitter that may have kept it safe. Hard to say.

Many thanks for showing us the schematic. It looks like an early 4-stage amplifier, which is fairly unusual. There is no sign of Miller compensation, which is also unusual, and several components whose purpose is rather hard to fathom. It would be interesting to analyse it, but I can't say it would be my first priority in advancing the state of the art.
 
Last edited:
Well, there's certainly a fairly big resistor in the VAS emitter that may have kept it safe. Hard to say.

Many thanks for showing us the schematic. It looks like an early 4-stage amplifier, which is fairly unusual.

I've never seen an input stage like that. The RCA circuit this was taken from uses a more typical differential IPS.

There is no sign of Miller compensation, which is also unusual,

C214, lag compensation, may have been the order of the day.

and several components whose purpose is rather hard to fathom.

You are probably looking at the Zener diodes. That was an early, ill fated attempt, at short circuit protection. I believe they are 4.7 V.

It would be interesting to analyse it, but I can't say it would be my first priority in advancing the state of the art.

Quite correct. I often wonder about those that restore some of this stuff without doing some redesign. Again, being single supply, this is not worth the effort. I have a simple Java simulator (Circuit Simulator Applet) and entered the schematic. Due to charging the output capacitor, there is a 5 volt pulse to the speaker at turn on and a very low frequency decaying oscillation. Additionally, this oscillation appears on application or removal of a large signal. It's inaudible but obviously is not good. This analysis has turned me against any capacitor coupled output which includes a number of highly regarded Fisher receivers.

The 3rd trace is after the 5 V spike so the vertical scale is expanded. The 4th trace shows the DC shift with a signal.

The later Heathkit AR-1500 used a more conventional design and a split supply. Bob Cordell starts with a similar circuit and with simple additions, most of which you also describe, proceeds to reduce the distortion by 3 orders of magnitude.
 

Attachments

  • HeathAR-15-1401 copy.png
    HeathAR-15-1401 copy.png
    98.1 KB · Views: 361
Hi Mr Douglas Self

I have just finished a integred amplifier, which is based on your "blameless amplifier". it work very well!
I'm not an expert in electronic, and your book (the fith edition) was a precious help for me.
When a french version???

Thanks

EGS100, l'intégré DIY sur le forum Amplification du site Homecinema-fr.com - 30053539 - 1056

I'm glad your project was successful. The Blameless amplifier has a well-deserved reputation for being a docile and tractable design.

There have been Spanish and Chinese version published, but I am afraid there are no plans for a French edition as far as I know. I will see what my publisher thinks about it.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.