Heathkit W5M restorations and mods ideas

[gkargreen]

Has Dave G. written anything on this topic or are there any good ideas for these amps before I dig in? I seem to recall that Dave thought quite highly of these amps. Thanks

 

[primosounds]

I have 4 of these amps and i rebuilt 2 of them but each was done differently. In one i tried a mod using 6cg7 instead of 12au7. Unfortunately the results were underwhelming. I followed some guys schematic i found searching HK W5 mods or something like that. The other amp i rebuilt to stock and that amp sounds really good. So its up to you what you want to do but the stock rebuild sounded best to me. The rebuilds were total replacements of all parts only reusing the sockets and transformers. So the comparison is pretty much based on the circuits.
The schematic i followed with 6cg7 was designed by some guitar guy and i am not sure how adept he was with tube circuits. The rebuilds were done over 5 years ago and i have not had the impetus to return to the other amps to rebuild. I did have a problem with one of the amps having a bad power trans and another one of the amps had a bad output trans. This really bummed me out about these amps, that 2 of the 4 i had purchased with blown trannys, no pun intended. Be sure to check both the power and output trans of your units before you proeceed. Anyways, good luck with yours.

 

[Brice]

I rebuild 2 of those with great success. They sounded very good, surprisingly good to me.

As primo said, the stock schematic is quite good, when you replace all parts with quality component.

If I remember correctly I did the following modifications:

- Increased the caps after the choke to improve filtration. There is not enough IMO.
- Added 100 ohm screen resistors to prevent arcing.
- Increased the cathode resistor to lower the bias current. It ran the tubes pretty hot. I think I went 390 Ohm and relocate it away from the AC shielded wire, which is a crazy idea.
- Very IMPORTANT: increase the voltage of coupling caps 1uF 400V to 1uF 600V. As startup, those caps get the full B+ on it. 400V is not enough and over time they get damage, start to leak DC and then take down the output transformer. That's why so many OT failure happens.
- Added 2 SS rectifier prior the tube rectifier (pin 2 and 8) to prevent damage in case of a rectifier failure.

Me too I had a busted power transformer. Those get hot. Lowering the bias current did help.
I used a SS weber rectifier with dropping resistor to get the B+ on specs, so I saved the wasted rectifier heater power.

I played with a 12bh7 instead of the second 12au7 with mixed result. I did not like the sound so I went back to stock.

Other than that, this amp once redone is a very sweep amp.

 

[trobbins]

I'd also recommend a HT CT secondary fuse. PSUD2 indicates that the W5M could use a 250mA F IEC 60127 fuse. A 315mA F UL 248-14 would be needed as it can only operate up to 80% of rated value, and CT rms current is likely to be just under 250mA. The fuse needs to be compliant to a standard to ensure the rating is going to work ok. A 250VAC rated fuse (M205 or3AG) is ok for this current limited application.

 

[soundmotor]

One major issue with the W-5M is the power transformer. At some point Heathkit changed the rating of the 1st version to a higher B+ current one. They even issued a service bulletin as well suggesting the use of a Stancor PC8412 as the replacement if there was a failure. The HV taps on it were rated for 200mA. So be mindful that the PTs, unless the uprated versions, might pop.

 

[gkargreen]

Thanks, all, that is useful help! I do have a couple of other questions. 1) I am thinking of using a pair of 20 ufd 900vdc film caps for the power supply instead of two 'lytics in series, that should be better than fine! 2) was measuring the taps of the output tranny, one side across the tube sockets is fine, but from the center tap to the tube sockets is .515 meg, so probably measuring something else. From that I conclude that the center tap to one side of the output tranny is gone whereas the winding from the plate to screen is OK, so it looks as though that W5M has a blown output tranny, correct?

 

[Brice]

Yes 1) will work nicely. Just add a couple in parallel so you have increase in capacity.
2) Yep, looks like one leg is almost open. Sorry.

 

[gkargreen]

Thanks, Brice, damn, I'm gonna have to beat on Doc for selling me this piece he said it was working! I know Steve will fight me on this even those he is an MD and not poor at all, surprised he tried to **** me with this, maybe it is his mistake and he will refund my $300 or at least enough for a new output tranny. BTW, you guys wouldn't know Steve, he is a local radio club member. And yes, I think the film caps in the supply will work great and eliminate the need for balancing resistors!

 

[soundmotor]

 

[primosounds]

Thanks, all, that is useful help! I do have a couple of other questions. 1) I am thinking of using a pair of 20 ufd 900vdc film caps for the power supply instead of two 'lytics in series, that should be better than fine! 2) was measuring the taps of the output tranny, one side across the tube sockets is fine, but from the center tap to the tube sockets is .515 meg, so probably measuring something else. From that I conclude that the center tap to one side of the output tranny is gone whereas the winding from the plate to screen is OK, so it looks as though that W5M has a blown output tranny, correct?

It is best if you measure the output trans out of circuit to be sure. Measuring something in circuit especially if it is attached to a cap will give you weird results.
If i were to do it all over again , i would probably make the amp choke input filter, using twin chokes in a LCLC configuration with a 5ar4 rectifier. That way you could stay with 500v Panasonic dc link filter caps and it would clear up a lot of space underneath with the series capacitors and balancing resistors. The 5ar4 would take care of the surge voltage. The trick would be sourcing the first choke as i don't think the original choke would be adequate for input duty. You would have to play with PSUD to see what you would need. But, just off the top of my head, something around 3H x 250ma would work (or higher would be better).

 

[tcdriver]

Thanks to all that have posted to this thread with good ideas for a W-5M rebuild.

 

[Retrovert]

The subject of inrush current causing transformer burnout in the Heathkit Williamson amplifiers was fully analyzed here:

https://audiokarma.org/forums/index.php?threads/mystery-device-found-in-heathkit-w4-am.862840/​

Included are reference to service bulletins issued by Heathkit and commentary from a repair tech at Heathkit.

The basic issue is the excessive current from heater inrush rapidly heats the primary which then burns out when the B+ kicks on. Heathkit's fix was to add a surgistor, a time-delay thermistor, to limit inrush current and thus the initial heating. No need to re-discuss.

 

[Tubologic]

Heathkit was well aware of and aknowledged this issue since the (later) W-7A amplifier included the same Surgistor in the kit right from the start (not a service update). The problem is finding a replacement if the Surgistor ever fails, since these are no longer manufactured. The CL-90 ICL has about the same "cold" resistance (120 ohms @ 25°C vs 100 ohms for the Surgistor used by Heathkit) but very likely a much shorter "ON" switching delay time and (unlike the Surgistor) is "non-resettable" when hot switching, thus it offers less protection, but still better than nothing. I remember some old McIntosh (tube) amps used similar NTC inrush current limiters in the primary circuit side of the PT ,very likely for similar purposes (protection).

 

[8xlaxx]

@Tubologic

Pardon my ignorance but could you better define or provide an example of hot-switching?

Would 'hot switching' be illustrated as an occurrence where a commercial power outage has taken place and then a quick flicker of commercial power is resumed? Like a quick on then off, then on again?

Apologies for the poor examples on my part.

 

[Tubologic]

@Tubologic

Pardon my ignorance but could you better define or provide an example of hot-switching?

Would 'hot switching' be illustrated as an occurrence where a commercial power outage has taken place and then a quick flicker of commercial power is resumed? Like a quick on then off, then on again?

Apologies for the poor examples on my part.

Yes, that would be one possible definition. ON-OFF-ON in a very short sequence, which leaves too little time for the ICL to cool down and regain his high value "cold" resistance. (well-designed) electronic time delay circuits (and the surgistor) are called "resettable" since in the event of a short power interruption they'll restart a new delay cycle.

 

[8xlaxx]

Yes, that would be one possible definition. ON-OFF-ON in a very short sequence, which leaves too little time for the ICL to cool down and regain his high value "cold" resistance.

Thank you.

Have a pair of these and am looking to rebuild/restore both. Been putting off the rebuild due to fear of OPT/PT failure.

 

[Retrovert]

Hot-switching isn't the cause. The real problem here, and the best focus of one's attention, is the general inrush current. Cold heaters are a dead short, and that heats the transformer core to the point that when the B+ winding kicks on (rectifier conducting) the additional heat sends everyone to Valhalla.

Reduce the inrush current and the spot heating issue goes away. Heathkit's modification resolved the issue, and joy returned to Mudville.

The ideal solution would be a MOSFET or relay with an RC cutout, to delay the heaters for tens of seconds to permit a slow warmup to eliminate the intense spot-heating. Better for the tubes, better for the primary, better for your wallet, better for everyone except those selling replacement transformers.

 

[8xlaxx]

The CL-90 ICL has about the same "cold" resistance (120 ohms @ 25°C vs 100 ohms for the Surgistor used by Heathkit) but very likely a much shorter "ON" switching delay time and (unlike the Surgistor) is "non-resettable" when hot switching, thus it offers less protection, but still better than nothing.

Reduce the inrush current and the spot heating issue goes away. Heathkit's modification resolved the issue, and joy returned to Mudville.

Are the CL-90 and CL-120 (either/or) adequate for this application?

 

[Retrovert]

Are the CL-90 and CL-120 (either/or) adequate for this application?

I am dubious that the few seconds delay from these thermistors is adequate to properly ameliorate this particular problem. Given the well-known issues with inrush current for this particular amplifier line, it would be best to err on the side of safety.

The surgistor delayed application of full current for approximately ten seconds. That greatly reduces the heating which, we must remember, varies as I^2 x R. So halving I will greatly reduce the heating by a factor of four.

A simple delay can be constructed using a 555, a transistor, and a relay. The current flows through a resistor, limiting the value, until the 555 fires closing the relay.

 

[8xlaxx]

I am dubious that the few seconds delay from these thermistors is adequate to properly ameliorate this particular problem. Given the well-known issues with inrush current for this particular amplifier line, it would be best to err on the side of safety.

The surgistor delayed application of full current for approximately ten seconds. That greatly reduces the heating which, we must remember, varies as I^2 x R. So halving I will greatly reduce the heating by a factor of four.

A simple delay can be constructed using a 555, a transistor, and a relay. The current flows through a resistor, limiting the value, until the 555 fires closing the relay.

Be really nice to see a schematic of this and where to place the circuit inside the chassis of the W5M. There seem to be plenty of room for it. Thank you for suggesting this @Retrovert