Awhile back there was a discussion regarding OPT replacement for a Heath W5 amp (Linked Here). During the discussion the topic veered away from the W5 to the W4 and its deficiencies esp WRT to stability and HF performance. During that discussion I commented on how I found the W4s I have are unstable with the stock circuit esp at HF and would actually break into sustained oscillation if a resistive load was paralleled with a 1uF cap. I also posted a schematic with my fix for the HF instability.
I decided to take one of the amps off the shelf and run some sq waves thru it to illustrate the problems I found with the stock circuit. I took pics of scope traces on both stock and modified circuits with 8 ohm load and 8 ohm || 1uF load. Signals were 1KHz and 10KHz sq waves. Horizonatal scale is in uncal mode to get a few cycles on the screen.
The first set of pics shows performance with straight resistive load, stock and with mods. Although the 1KHz patterns don’t look too bad, the stock circuit definitely has some delayed overshoot and ringing compared to modified. At 10KHz, differences are more apparent. The stock circuit appears to be ringing @ ~200KHz. Although I thought the stock circuit sounded edgy I probably would have accepted this level of performance if this was all there was to it. However……..
The second set of pics show what happens when 1uF is placed across the load resistor. As is apparent, the stock circuit has broken into sustained oscillation @ ~70KHz. Surprisingly, the sq wave is is still reproduced… somewhat. Without an input signal the amp oscillates @ 75KHz / 5V P-P. IMO, this is absolutely unacceptable performance and disqualifies the amp for use in a music system. Incredibly, Heath apparently was aware of the issue since they recommend that only low capacitance speaker cable be used and never use coax.
Although the modified circuit isn’t pretty, esp @ 10KHz, at least it isn’t oscillating. Ringing occurs @ ~ 30KHz . Stock response is specified as extending beyond 100KHz @ a couple of dB down. Modified response is nowhere near as extended being flat to ~20KHz and –0.3dB ~25KHZ but falling rapidly after that.
As I noted in the previous thread, my original objective with the W4 was unconditional stability and improved sq wave response. Sound was also improved as the edginess was greatly reduced. After revisiting these amps and the mods, if I were to redo things, I would probably accept a little more ringing to extend the HF a bit. I would also address another flaw in the W4 I discovered while putting together these images: a rather pronounced LF instability. Miniscule perturbations in line voltage or signal transients would produce a LF bounce in the output signal of ~0.5 V P-P that would continue for a couple of seconds after the disturbance had ceased. Although there are some changes that can be made to the circuit that can reduce the LF instability, as I opined before, I believe most of the problems with the W4 arise from use of an OPT that is simply not up to the task. If I replace the OPT with something markedly better (and probably mucho $$ more expensive) I probably no longer have a W4 so why bother?
Bottom line from my perspective: The stock W4 is poorly designed and spec’d and probably should not be a part of a real-world music system. This is esp so if the speaker load is reactive or if fancy, high capacitance speaker cables are to be used. The gold version is pretty in its own 50’s industrial sorta way and would make a nice conversation-starting display piece.
All of the above is of course JMHO.
I decided to take one of the amps off the shelf and run some sq waves thru it to illustrate the problems I found with the stock circuit. I took pics of scope traces on both stock and modified circuits with 8 ohm load and 8 ohm || 1uF load. Signals were 1KHz and 10KHz sq waves. Horizonatal scale is in uncal mode to get a few cycles on the screen.
The first set of pics shows performance with straight resistive load, stock and with mods. Although the 1KHz patterns don’t look too bad, the stock circuit definitely has some delayed overshoot and ringing compared to modified. At 10KHz, differences are more apparent. The stock circuit appears to be ringing @ ~200KHz. Although I thought the stock circuit sounded edgy I probably would have accepted this level of performance if this was all there was to it. However……..
The second set of pics show what happens when 1uF is placed across the load resistor. As is apparent, the stock circuit has broken into sustained oscillation @ ~70KHz. Surprisingly, the sq wave is is still reproduced… somewhat. Without an input signal the amp oscillates @ 75KHz / 5V P-P. IMO, this is absolutely unacceptable performance and disqualifies the amp for use in a music system. Incredibly, Heath apparently was aware of the issue since they recommend that only low capacitance speaker cable be used and never use coax.
Although the modified circuit isn’t pretty, esp @ 10KHz, at least it isn’t oscillating. Ringing occurs @ ~ 30KHz . Stock response is specified as extending beyond 100KHz @ a couple of dB down. Modified response is nowhere near as extended being flat to ~20KHz and –0.3dB ~25KHZ but falling rapidly after that.
As I noted in the previous thread, my original objective with the W4 was unconditional stability and improved sq wave response. Sound was also improved as the edginess was greatly reduced. After revisiting these amps and the mods, if I were to redo things, I would probably accept a little more ringing to extend the HF a bit. I would also address another flaw in the W4 I discovered while putting together these images: a rather pronounced LF instability. Miniscule perturbations in line voltage or signal transients would produce a LF bounce in the output signal of ~0.5 V P-P that would continue for a couple of seconds after the disturbance had ceased. Although there are some changes that can be made to the circuit that can reduce the LF instability, as I opined before, I believe most of the problems with the W4 arise from use of an OPT that is simply not up to the task. If I replace the OPT with something markedly better (and probably mucho $$ more expensive) I probably no longer have a W4 so why bother?
Bottom line from my perspective: The stock W4 is poorly designed and spec’d and probably should not be a part of a real-world music system. This is esp so if the speaker load is reactive or if fancy, high capacitance speaker cables are to be used. The gold version is pretty in its own 50’s industrial sorta way and would make a nice conversation-starting display piece.
All of the above is of course JMHO.
Attachments
Last edited:
