Super, thanks, Dave!
Untapped Potential: Heath's W-5M
- Thread starter dcgillespie
- Start date
Super, thanks, Dave!
Oh, and which package? TO-220, TO-220FP or TO-247? Sorry, I have very little experience with SS components!
to-220 is metal back, to-220FP is plastic back, TO-247 is a chunker but it has a metal back. The TO-220 and the TO-247 would need to be insulated from chassis with a mica insulator and hat bushings for the screw, or a plastic screw. The -FP one you can just bolt down since its already insulated.
The part number I gave is the one assigned to the plastic (FP) piece, so specifying that part number will automatically get you the correct piece.
Dave
Dave
Ah, good, thank you.
maxhifi
Addicted Member
Dave: Thanks you for this excellent write up. I am imagining a book some day, with a chapter for each of your rebuilds... W2M, W3M, Dynacon etc.
Like you did I have a single W5M waiting for its partner who may never appear. I'm wondering, do you prefer the Acrosound equipped W3M, or the W5M?
Like you did I have a single W5M waiting for its partner who may never appear. I'm wondering, do you prefer the Acrosound equipped W3M, or the W5M?
That's a tough question for me to answer because for me, there's a number of different issues involved. For example, which is better from a sonic standpoint? Or which is better from a measured performance standpoint? Circuit design standpoint? And are you talking about their original stock design, or modified? Because I've been involved in all of these aspects of these amplifiers, a simple question like which do I prefer doesn't easily generate an easy answer.
From a sonic standpoint, I need more time to assess the modified W-5, but it is certainly special sounding -- notably so right off the bat. But honestly, the W-2, W-3AM, and W5 -- all in modified form -- sound really excellent, so I could easily live with any of them once modified from a sonic standpoint. I don't have the ability to perform any true double blind testing to really test for any true sonic differences, but from extended listening I've already done with the modified W-2 and W-3, and now what I am starting to gather the W-5, they are all really excellent.
But the measured performance and circuit design elements play into my preference thoughts as well. So when you mix all of this together, my thought process would take this route:
In stock form the W-5 is out for me, because I know of the potential damage it's doing to itself with every application of power. That leaves the W-2 and W-3AM. In stock form, the W-2 is at best barely stable, so it's out too, leaving the W-3AM as my preference in stock form. It also has the turnoff pulse, but with a more robust output transformer, it's more of a Timex watch in this case, so the W-3AM gets the nod.
In modified form however, things get turned upside down. The W-3AM has the least desirable HF transient and load stability of the bunch when all three output taps are considered. The W-2 has wonderful HF transient and good load stability regardless of output configuration, but with the HF power loss, while the modified W-5 now has wonderful HF transient stability, zero HF power loss, and absolute stability. Load stability is hugely important to me, producing (I think) one of the most audible differences between amplifiers of marginal, and exceptional load stability capability.
The only way in which the modified W-2 and W-3AM have better measured performance than the modified W-5 is they have lower 20 Hz distortion, but the W-5 produces more 20 Hz power than the other two, so this advantage isn't as much of an advantage as it may seem, and goes completely away at 25 Hz or greater.
Then there's the issue of the 50% tap versus the genuine UL specification, versus the partially cathode coupled output stages. In the 6L6 class of tubes, I don't know that there was ever a transformer developed of true UL specifications with individual 4, 8, and 16 Ohm output leads, that had a flat power response with low distortion, identical supersonic performance on each tap, would allow for absolute stability, and could produce an extended frequency response down no more than 1 db at 40 kHz or greater, and could produce a flat top 10 kHz square wave. In all of my experience, I only know of only a very few transformers that are capable of that level of performance. One is the Dynaco A-470 transformer in the Dynaco ST-70 (for EL34 tubes), the unbelievably stellar Peerless UL transformer in the Heath W-6 series (for 6550 tubes), and of course McIntosh and 1980s and later ARC transformers. There may be others of this capability, but they don't readily come to mind.
Ultimately, because of my direct work with 3 partially cathode couple output stage applications (and knowledge of a 4th), I am rather quickly coming to the opinion that this type of output stage is capable of better performance than a tapped screen or UL specification output stage is because it inherently produces much less distortion, is easier to drive, and is easier on the tubes as well. But of course, that mode of operation is only possible with an output transformer that has individual 4, 8, and 16 Ohm output leads, and that has identical HF performance at each output tap. The Acro transformer gets automatically ruled out of this type operation, as any attempt to do so will only result in a power oscillator. The early Peerless transformers won't work either because only one output impedance is available at any one time. In this regard then, the W-5 transformer really shines.
When all of this is considered then, and with what I now know, I prefer the modified W-5M, followed by the modified W-2M, with the modified W-3 bringing up the rear. Don't get me wrong. The modified W-3 is a very good amplifier, and I'm hardly ready to part with mine. But if you ask me which I prefer, then because of all the issues cited, the modified W-5 rather handily takes the top spot. At this point it is at least as good as the other two are audibly, likely better, and does all of the measured elements better as well with the exception of the 20 Hz caveat mentioned, which I can easily live with.
I hope this helps!
Dave
From a sonic standpoint, I need more time to assess the modified W-5, but it is certainly special sounding -- notably so right off the bat. But honestly, the W-2, W-3AM, and W5 -- all in modified form -- sound really excellent, so I could easily live with any of them once modified from a sonic standpoint. I don't have the ability to perform any true double blind testing to really test for any true sonic differences, but from extended listening I've already done with the modified W-2 and W-3, and now what I am starting to gather the W-5, they are all really excellent.
But the measured performance and circuit design elements play into my preference thoughts as well. So when you mix all of this together, my thought process would take this route:
In stock form the W-5 is out for me, because I know of the potential damage it's doing to itself with every application of power. That leaves the W-2 and W-3AM. In stock form, the W-2 is at best barely stable, so it's out too, leaving the W-3AM as my preference in stock form. It also has the turnoff pulse, but with a more robust output transformer, it's more of a Timex watch in this case, so the W-3AM gets the nod.
In modified form however, things get turned upside down. The W-3AM has the least desirable HF transient and load stability of the bunch when all three output taps are considered. The W-2 has wonderful HF transient and good load stability regardless of output configuration, but with the HF power loss, while the modified W-5 now has wonderful HF transient stability, zero HF power loss, and absolute stability. Load stability is hugely important to me, producing (I think) one of the most audible differences between amplifiers of marginal, and exceptional load stability capability.
The only way in which the modified W-2 and W-3AM have better measured performance than the modified W-5 is they have lower 20 Hz distortion, but the W-5 produces more 20 Hz power than the other two, so this advantage isn't as much of an advantage as it may seem, and goes completely away at 25 Hz or greater.
Then there's the issue of the 50% tap versus the genuine UL specification, versus the partially cathode coupled output stages. In the 6L6 class of tubes, I don't know that there was ever a transformer developed of true UL specifications with individual 4, 8, and 16 Ohm output leads, that had a flat power response with low distortion, identical supersonic performance on each tap, would allow for absolute stability, and could produce an extended frequency response down no more than 1 db at 40 kHz or greater, and could produce a flat top 10 kHz square wave. In all of my experience, I only know of only a very few transformers that are capable of that level of performance. One is the Dynaco A-470 transformer in the Dynaco ST-70 (for EL34 tubes), the unbelievably stellar Peerless UL transformer in the Heath W-6 series (for 6550 tubes), and of course McIntosh and 1980s and later ARC transformers. There may be others of this capability, but they don't readily come to mind.
Ultimately, because of my direct work with 3 partially cathode couple output stage applications (and knowledge of a 4th), I am rather quickly coming to the opinion that this type of output stage is capable of better performance than a tapped screen or UL specification output stage is because it inherently produces much less distortion, is easier to drive, and is easier on the tubes as well. But of course, that mode of operation is only possible with an output transformer that has individual 4, 8, and 16 Ohm output leads, and that has identical HF performance at each output tap. The Acro transformer gets automatically ruled out of this type operation, as any attempt to do so will only result in a power oscillator. The early Peerless transformers won't work either because only one output impedance is available at any one time. In this regard then, the W-5 transformer really shines.
When all of this is considered then, and with what I now know, I prefer the modified W-5M, followed by the modified W-2M, with the modified W-3 bringing up the rear. Don't get me wrong. The modified W-3 is a very good amplifier, and I'm hardly ready to part with mine. But if you ask me which I prefer, then because of all the issues cited, the modified W-5 rather handily takes the top spot. At this point it is at least as good as the other two are audibly, likely better, and does all of the measured elements better as well with the exception of the 20 Hz caveat mentioned, which I can easily live with.
I hope this helps!
Dave
maxhifi
Addicted Member
This is a very thoughtful and interesting answer, thank you!
Any thoughts on the W4 and how it compares to the W2, W3 and W5? I know you've done some work on it, and I have a copy of that revised schematic that I need to quit being lazy and build a second amp around. Modified the W4 clone I've had for years and it responded exactly as indicated with the Chicago iron. Mine are Chicago BO-13 rather than the Heath branded ones, but so far as I understand its the same transformer in a different can.
Even more interesting, since the W-4 has multiple output taps, could you use the same cathode-coupling technique there? You'd have to calculate the screen voltage, I guess, and fiddle with the feedback, but I would imagine the potential is there.
No -- The W-4 transformer won't work with the modified W-5 design. It is a 10KΩ transformer with 50% screen taps similar to other Heath transformers, but it too has an unequal response at each output tap just like the Acro TO-300 does (which I assume it's basic winding pattern follows), which means using it in a partial cathode coupled output stage would also result in a power oscillator as well. It is precisely because of the dissimilar response that the Acro and W-4 OPTs produce at each output tap that makes them require a dual tap NFB network to achieve best results, and also makes them unsuitable for partially cathode coupled designs.
The W-4 was a good economy offering by Heath, basically being a poor man's W-3AM. The W-5 was designed to put Heath back front and center again in the High Fidelity industry, but it was the W-6 series that was the real force to be reckoned with.......
Dave
The W-4 was a good economy offering by Heath, basically being a poor man's W-3AM. The W-5 was designed to put Heath back front and center again in the High Fidelity industry, but it was the W-6 series that was the real force to be reckoned with.......
Dave
Ah, got it!
A quick followup after some repeated extended operating time: After 4 hours of operation with the cages in place in an open air 73˚F ambient environment (no forced air cooling), the power transformers of the modified amplifiers reached 105˚F on the back and outside sides, 106˚F on the top, 121˚F on the front side, and 125˚F on the side next to the rectifier tube and can caps. The numbers show that the transformers are accepting heat of course on the front and rectifier sides of the transformer, while only showing an ~ 35˚F rise above ambient from internally generated heat on the outside sides of the transformers. With an average operating temp then of ~ 115˚F, and 10 more degrees added in to arrive at a more realistic core temp, it indicates an average core temp of 125˚F, which is a very cool operating temp for these transformers indeed!
Dave
Dave
Ironic...
That should basically last forever, save for a lightning strike or something ridiculous like that.
Regards,
Gordon.
Dave, I have a question about how the speaker back EMF may effect the local feedback, see your schematic below.
Let's assume a 4 ohm speaker is hooked up as shown. The back EMF from it will be impressed onto the 0-4 part of the winding but not on the 4-16 section and may effect the feedback signal asymmetrically. Is this a concern or in this case not an issue since the local feedback is operating into such a low impedance?
Second part of the question, assume the feedback connections are similar to the McIntosh MC275 Mark IV shown. In the case of the Mac, there is a center-tapped winding on the OPT to allow symmetrical feedback signals to the input tube which is setup for a balanced connection. I have a project in mind where there is no center tapped feedback winding and an OPT with both a single output winding (no taps) for the speaker, and another winding for gnfb use. Such transformer is the Unity Coupled one from Vanderveen, datasheet attached. If I did my math correctly, the "3.5%" gnfb winding should have the same amplitude as the "5 ohm" secondary. The specifics here are presented as a reference, sorry if slightly off topic.
Simply put, will hanging a speaker load on one side of (assumed identical) windings, affect the feedback signal operating into a higher impedance load, i.e. the cathodes of the input tube. I can and will test such an arrangement but for now I am curious your thoughts here.
Let's assume a 4 ohm speaker is hooked up as shown. The back EMF from it will be impressed onto the 0-4 part of the winding but not on the 4-16 section and may effect the feedback signal asymmetrically. Is this a concern or in this case not an issue since the local feedback is operating into such a low impedance?
Second part of the question, assume the feedback connections are similar to the McIntosh MC275 Mark IV shown. In the case of the Mac, there is a center-tapped winding on the OPT to allow symmetrical feedback signals to the input tube which is setup for a balanced connection. I have a project in mind where there is no center tapped feedback winding and an OPT with both a single output winding (no taps) for the speaker, and another winding for gnfb use. Such transformer is the Unity Coupled one from Vanderveen, datasheet attached. If I did my math correctly, the "3.5%" gnfb winding should have the same amplitude as the "5 ohm" secondary. The specifics here are presented as a reference, sorry if slightly off topic.
Simply put, will hanging a speaker load on one side of (assumed identical) windings, affect the feedback signal operating into a higher impedance load, i.e. the cathodes of the input tube. I can and will test such an arrangement but for now I am curious your thoughts here.
Attachments
Last edited:
Asymmetric? I don't see it. The secondary is, but the feedback isn't - though it's slightly less effected by the load. Mac has to use symmetric feedback if they're bringing it to a balanced stage. A separate winding gives less load effect, but can't correct for the effect of winding resistance.
Anyway, the back EMF shows up in the impedance and phase of the load - it has to less than the driving voltage.
Anyway, the back EMF shows up in the impedance and phase of the load - it has to less than the driving voltage.
any back EMF would also appear from the 4 to 16 terminal out of phase with what appears from 0 to 4. The transformer itself would essentially be operating as a phase inverter at that point so there would be a very minor, if any, difference in the feedback signal. Arguably the only time you might get some asymmetric feedback is when using the 8 ohm taps, but I still suspect its not any significant problem.