• The move to the new server is done. There are some software and database maintenance updates in process. This has us passing the hat around to help out. We appreciate any donations. Seriously, even a dollar helps. The payment page may be found here - https://www.audiokarma.org/support.html

Untapped Potential: Heath's W-5M

I'll jump into the discussion with these thoughts:

1. So much depends on the quality of coupling within the transformer. Dynaco and Peerless transformers are coupled well enough to be able to use them in a partially cathode coupled output stage, whereas others (classic Acrosound examples) simply turn into a power oscillator due to insufficient coupling. With good coupling, what happens in the lower half of the secondary happens "well enough" in the upper half by transformer action, even though the load may not be connected to the upper half.

2. The coupling isn't perfect however. In developing the modification for the W-5, I noted that the amount of local feedback produced in the output stage reduced distortion exactly as theory indicated it would, but the global NFB loop did not. It was still effective enough to produce 88% of the projected reduction, so I ran with it anyway. In the original amplifier, NFB was taken from the full secondary, while with the modified design, it is only taken from the top half. I took this to be the reason that the global loop was only 88% effective.

3. I'm struggling with Gadget's comment that any signal (i.e. back EMF) input into the lower half of the 16Ω winding will appear out of phase in the upper half. If (for example) a 1 volt signal is injected into the Com-4Ω terminals of the secondary winding, then 2 volts will appear between the Com-16Ω terminals because the windings are internally connected/constructed series aiding, so they must be in phase. Maybe I'm not understanding his comment correctly.

I know your question wasn't regarding NFB, but I do believe what I observed regarding it is related to your question: In my case, the FB signal was only receiving what was produced in the bottom secondary by transformer action, not a direct connection. Your question Rust is effectively dealing with the same concern -- and my answer is again, that the coupling isn't perfect, but certainly good enough to still be quite effective and not cause any problems (ala Acrosound). In other words, any back EMF injected into the bottom portion of the secondary will also largely appear in the top part as well, making for a symmetrical reflection back to the primary. It does beg the question however as to just how effective is NFB taken from the 4Ω tap simply because that's where the load is connected to in a conventional output stage. Modern thinking says that taking the global FB from the tap the load is connected to is the way to go -- but is it? With the 16Ω tap there is of course no issue, but with the intermediate taps? That might be an interesting study to make.

Having a separate balanced winding to facilitate the partially cathode coupled output stage is no doubt the way to go even if the winding resistance is not accounted for. Ultimately though, I've found that given an adequate transformer, using a balanced 16Ω secondary load winding to provide for partial cathode coupling produces benefits that are well worth the effort to implement it.

I hardly consider myself the last word on these thoughts, so all comments are welcomed!

Dave
 
dcgillespie said:
3. I'm struggling with Gadget's comment that any signal (i.e. back EMF) input into the lower half of the 16Ω winding will appear out of phase in the upper half. If (for example) a 1 volt signal is injected into the Com-4Ω terminals of the secondary winding, then 2 volts will appear between the Com-16Ω terminals because the windings are internally connected/constructed series aiding, so they must be in phase. Maybe I'm not understanding his comment correct
Click to expand...

essentially the secondary would be operating like a center tapped choke. Any signal between the center tap and one end appears in opposite phase at nearly the same voltage between the center tap and the other end. Since the 4 ohm terminal is grounded that becomes the reference point, not the C terminal.

If you measured from C to 4 and C to 16 it would be double at the 16 ohm since its essentially working like an autotransformer.

I've used that center tapped choke thing in old radios with dead inverter transformers. When the primary goes open, it gets subbed for a plate resistor and a coupling cap is tied to one of the tube grids. The secondary provides inversion and the other tube gets out of phase grid signal. Not quite perfect but its good enough for a 1930s AM radio.
 
Ah — We’re both effectively saying the same thing then. Thanks for the clarification. Of course both ends of the windings are out of phase with each other — I was taking your comment to mean that the two halves of the secondary winding would somehow be out of their normal phase relationship with each other…..

Dave
 
This is gold even for a non-guru. Is there a parts list so as not to miss something on the order. Thanks
 
How might one measure an OPT to determine if was suitable for use with partial cathode coupling? Measure from the 4 ohm connection separately to the 0 and 16 ohm connections and compare the responses? I remember an article in Audio Amateur or Glass Audio years ago where the author did this to a Dynaco ST-70 I believe.
John
 
John — The easiest way is to look at a 1 watt 10 kHz square wave on the output of the amplifier. Other than for amplitude, the display should appear absolutely identical on each output tap — not similar — identical. If it does, then the transformer will remain stable when used in a partially cathode coupled output stage.

I hope this helps!

Dave
 
I have Heatkit AA100 outputs, Eico ST70 and Dynaco ST70 along with some Fisher and Scotts. I think the Dynacos would fit this scenario but am curious about the others. One of these days I will set up a test and see what I find.
 
Dave and I were conversing about the Eico ST70 output transformers- as I had the same idea, in using them with cathode feedback, to reduce the amount of global feedback needed in the ST70 "Hot Rod" amp design. He indicated that they gave essentially identical high frequency response from the 0-4 ohm taps, as from the 4-16 ohm taps- the transformer had the same high frequency characteristics at all the taps. This means they should work well in a cathode-feedback configuration.

I think he also mentioned that the A-470 from the Dyna ST70 also exhibited identical stability and high frequency response from every output tap, as well- but I can't find an attribution for that, readily. If anyone knows where that was discussed, it would be appreciated to know where that was.

Regards,
Gordon.
 
I know that Audio Research made an amp with the A470 opts that used the secondaries for cathode feedback and I think I remember Dave mentioning this. I'm also curious about the Heath AA-100 outputs suitability for cathode feedback.
 
Curious about how to determine the resulting PP load as seen by the tubes when using the balanced cathode feedback from the secondary winding. I did some digging in RDH4 but didn't find or recognize what I was looking for.
Thanks,
John
 
This might answer your question. I had basically the same one a while ago, though at this point I don't quite remember what it was about. I never did anything with it though.

**edit**

might answer it better if I pasted the link. Derp.

do cathode feedback windings affect primary Z ?

Not really sure how to phrase this question in a subject line, so apologies if it makes no sense. I've gotten my hands on a couple of Hammond output transformers that contain dedicated cathode windings. I haven't gotten around to ringing these out, so I have absolutely no idea what the ratios...
audiokarma.org

mind thats for a dedicated cathode feedback winding. I don't know if it applies in quite the same way if the winding in question is both cathode feedback and speaker output.
 
Last edited:
Hi John -- When I was studying the use of partially cathode coupled feedback with my ARC D115/V70 project a couple of years ago, I went looking for anything on the subject as well, and found exactly..... nothing, zilch, nada. I'm not saying I overturned every stone out there, but enough of them that I concluded that there was nothing readily available on the topic. So, I ended up developing the answer for myself.

When the secondary is used for PCC feedback, then effectively the secondary must be considered as part of the primary, since plate current is now flowing through the secondary as well. Mathematically then, it breaks down like this: Since the traditional turns ratio for a given transformer is usually expressed in a value of XX:1 -- with the "1" representing the turns in the secondary -- then "1" must also be added to the traditional value of XX to accurately represent the new configuration.

As an example, consider a transformer with a designated 10KΩ primary and 16Ω secondary. With conventional analysis, this represents a turns ratio of 25:1. But when used in a PCC feedback arrangement, the reflected impedance then becomes the result of a 26:1 turns ratio. Now, the same 16Ω load that reflected 10KΩ in a conventional setting, reflects 10.816KΩ with the new output stage configuration.

ARC, who used this concept throughout all of their vacuum tube designs, maximized the concept by accounting for the increase in reflected impedance that the arrangement creates. As a result, the reflected impedance of their OPTs is optimum only when used in a PCC design. When working with transformers whose specifications are based on conventional output stage arrangements, then you just have to accept the increase in reflected impedance that employing a PCC arrangement creates.

Again, this is based on my own tests and analysis -- not to draw any attention to myself, but to emphasize that it's my own analysis because I too could not find anything to draw on regarding the change in relected impedance that the concept creates. Even just the fact that the configuration alters reflected impedance is not even recognized anywhere I could find. I welcome any input from others on this topic, but unless someone can show otherwise, I am rather certain that the analysis is accurate, as born out in the testing conducted.

I hope this helps!

Dave
 
Last edited:
I've just started to tear down the first of a pair of W5's for a friend. I'm going to rebuild according to this thread. The tube sockets are all pretty crispy looking so I will install new sockets. Question is, do the 12AU7's really need shields? I have a large bag of new Belton 9 pin sockets but without the shield bases. The original sockets have integral bases for the shields but none of the previous W series amps had tube shielding although they all used 6SN7's and maybe they are less prone to noise pickup or there just weren't any shields available for that size tube.
Thanks,
John
 
Working on the 1st of two W5’s and have a question about the zener string and power dissipation. Since they are 5W rated it makes me suspect some heat and how Dave mounted the in the air. I’d like to use some perf board to make a small module but not sure if close spacing and having them against the board is a problem. This is my mockup
 

Attachments

  • IMG_4181.jpeg
    IMG_4181.jpeg
    103 KB · Views: 33
nerdorama, I'd suggest your layout may not provide sufficient cooling of each zener. Cooling is best done by conduction through the wire leads to 'somewhere else', and/or by free air convection into the coolest air locally available. Dave's photo in post #15 shows he doesn't have much for the leads to conduct to, but he does use free air convection. The main hassle with zener strings is that the '5W' capability of each part can dramatically fall to say 1-2W if there is little cooling available - people look at the 5W as if its a resistor part, but they have very different survival outcomes.
 
You must log in or register to reply here.
Back
Top Bottom