Single supply, three-pole compensated class-B retro amp

Anchan

Member
2009-07-18 11:07 pm
A very brief, perhaps overly simplified explanation:

Most amplifiers have their open-loop gain dominated by a single pole; their open loop gain rises at 20dB/decade with decreasing frequency.

Less common are amplifiers with two poles; their open loop gain rises at 40dB/decade with decreasing frequency.

Very uncommon are amps with three pole compensation; 60dB/decade open loop gain. Hence, mention of that gain slope in LKA's opening post.

The opposing religion advocates amplifiers with no negative feedback. ;)
Great short explanation to help orient. Thanks!
 

LKA

Member
2014-12-28 5:31 pm
I measured the noise. (bandlimited to 20-20k) It's little higher than simulated. The input was shorted via 100R resistor.

UNW -89dBV
AW -95dBV
CW -92dBV
 

Attachments

  • H-ENGv3 noise.png
    H-ENGv3 noise.png
    29.1 KB · Views: 46
  • Like
Reactions: 1 users

LKA

Member
2014-12-28 5:31 pm
measurements - THD vs Freq at 1W 10W 50W 4R
output quiescent current 3mA and 10mA (10mA is the maximum without emitter resistors)
 

Attachments

  • H-ENGv3 THDvF 1W 10W 50W 4R 3mA.png
    H-ENGv3 THDvF 1W 10W 50W 4R 3mA.png
    24 KB · Views: 67
  • H-ENGv3 THDvF 1W 10W 50W 4R 10mA.png
    H-ENGv3 THDvF 1W 10W 50W 4R 10mA.png
    23.2 KB · Views: 63
measurements - THD vs Freq at 1W 10W 50W 4R
output quiescent current 3mA and 10mA (10mA is the maximum without emitter resistors)

I'm feeling a bit mean here, but I'm noticing a rising distortion characteristic at higher frequencies for the 1W/3mA plot - more so than for the higher-power traces.

If this is indicative of crossover distortion, then a plot for 100mW could be very revealing.

As I say, I feel a bit mean because I think the work you've done here is really useful.
 

LKA

Member
2014-12-28 5:31 pm
I can show you a simpler retro amp with very good performance, the DanWatt (class-b, current driven, designed by Danhard)
 

Attachments

  • h-eng-danwatt.jpg
    h-eng-danwatt.jpg
    615.9 KB · Views: 152
  • Dan-WATT-Mk-II.jpg
    Dan-WATT-Mk-II.jpg
    147.4 KB · Views: 168
  • Dan-WATTmk-IIR-THDv-F-1-W-10-W-50-W-4-R-MJW3281-A.png
    Dan-WATTmk-IIR-THDv-F-1-W-10-W-50-W-4-R-MJW3281-A.png
    20.3 KB · Views: 153
  • Dan-Watt-Mk2-p2.jpg
    Dan-Watt-Mk2-p2.jpg
    512.2 KB · Views: 120
Last edited:
  • Thank You
  • Like
Reactions: 1 users
Lots of expectations ... it's a simple retro low-end amp. THD10K 0.001% at 0.1W/4R is bad ? imho not at all
Well, the fact that the increase in thd at 100mW is modest, relative to 1W, seems to me to show that crossover distortion is being controlled remarkably well. So I would have thought it completely inaudible for all programme material - by quite a margin.

I wasn't being critical, by the way - just interested in what seems to me to be one of the main reasons for wide open-loop bandwidth. Obviously, I wasn't expecting zero crossover distortion, as the hf feedback is there to reduce, not remove. Nevertheless, who needs those many exotic topologies if we have such a relatively straightforward solution?
 
On the other hand, we have verified directly coupled topologies with input differential pair that do not depend on bootstraps, output electrolytic capacitors and NFB hard fight with crossover distortion. I do not understand why to return to all those circuit issues that we were trying to avoid in designs after 1970. Nostalgy?
 
On the other hand, we have verified directly coupled topologies with input differential pair that do not depend on bootstraps, output electrolytic capacitors and NFB hard fight with crossover distortion. I do not understand why to return to all those circuit issues that we were trying to avoid in designs after 1970. Nostalgy?

Well, if it measures well enough (in all areas of concern) and is relatively simple, cheap, and easy to source, then why not? By the way, I think output capacitors are a good thing, as they offer intrinsic protection of your expensive speakers. I have them at the output of my (dual-supply direct-coupled) amplifier regardless.
 

Anchan

Member
2009-07-18 11:07 pm
On the other hand, we have verified directly coupled topologies with input differential pair that do not depend on bootstraps, output electrolytic capacitors and NFB hard fight with crossover distortion. I do not understand why to return to all those circuit issues that we were trying to avoid in designs after 1970. Nostalgy?
For me I am attracted because of performance, simplicity of build, novel topology (already built several class-A and assembled class- D modules). I also highly value compactness. No 3U cases with ginormous heat sinks is a big plus.