Refurbishing some Ampex el34 monoblocks

I

indosmokejon

Active Member
A local audio acquaintance let me borrow his ampex el34’s and after a very short listen I was hooked. Not sure if my vandy 2ce sig II’s have ever sounded as good. They liked the power, as I mostly run smaller/weaker tube amps. I have owned many other large tube amps that put out 40-60 watts, from dynaco mkiii’s, st70, quicksilver kt88, etc, but I never heard any of them stand out, so they were sold. The ampex just grabbed my attention. The acquaintance wanted a hefty $1100 for his pair and so I passed, and found a rough pair for less. I bought them sight unseen, had a friend grab them, then waited 8 months due to Covid to finally get them this week. They are rough, and someone tried to update them, did it wrong, and they clearly haven’t worked since. One of the filter caps positive lead goes to ground! I’m not an amp engineer, but I am good at polishing a turd. These will be polished. I need help on the engineering side. These sounded amazing to me but I’m sure they could sound better, and I have questions. These mono’s are the true el34 variant, with the opt that are larger than the pt, solid state rectification, 40W per, and 16ohm secondaries.
 
Register to hide this ad
Well, you likely didn't save much money by buying the broken set. Most of us who refurbish these things for a hobby, arent making any money doing so. When you see a pair like that restored for eleven Bill's, you must realize that the labor rate had to have been under $2 per hour.

Anywho

Restoring these yourself is likely to teach you a lot about analog electronics. Furthermore you'll have a really nice set of monoblock amps which you can keep for decades.
I dont think theres an all encompassing thread on these yet, so you get to break ground there...

Best of luck, I'll be following along.
These have always piqued my curiosity but the only ones ive ever seen were priced beyond my budget at the time.
 
I knew it wouldn’t save me much to do the rebuild myself. I have done lots of other builds. When I can save a little money though, I usually go that route. It will cost me time, which I don’t have a lot of either, but I will end up with amps that I know in and out, and are tailored looks-wise to my tastes.
Some initial questions
1. As linked above, I think I need to find the proper capacitor to be included with the NFB resistor for better HF stability?
2. I would like to omit the 500k ohm potentiometer at the signal input included in the original build? Seems like my signal running through an extra pot isn’t going to help things and if I could just use a single resistor that this would work better?
3. There is an additional 100ohm trim pot between the output tubes for biasing. What ma in volts would I even be looking for to be correctly biased? What leg of the el34’s would I check? Can I incorporate test points, like in a Heathkit w4-am, that would make it easier to quickly test bias?
4. These amps have 16ohm secondaries. Initially this worried me as my Vandy’s are 6.8ohm. They sounded so good when I paired them together though that I figured what the heck. Am I missing anything regarding this mismatch?
 
Also, the 3 most critical capacitors in the signal path are shown in the schematic as 100k pf, aka 0.1uf. In all the amps like these that I have seen, they were supplied with 0.1uf 600v axial non-polarized types. Do these caps really need to be rated to 600v? It doesn’t look like there is near that much voltage flowing thru that portion of the amp? Am I missing something there?
 
Looks like a Mullard 5-20 amp but with slightly lower plate voltage. The 5-20 does closer to 30 watts, but whatever. I expect this will also be in the 25-30w range at reasonable distortion levels.
1) if you know the feedback components are original, chances are the values are correct. If they have been changed and you suspect its wrong, go with the schematic as a starting point. If the engineers who designed this in the first place were even halfway competent, the stock values should produce reasonable stability. No guarantee, but its definitely where I'd start.
2) dump the 500k pot for a 470k to ground and you're good. Maybe add a 1k between the grid and the 470k just for a bit of stability insurance.
3) Its cathode bias, so it more or less takes care of itself. If I read the schematic properly it calls for "approx 36v" at the cathode. If my math is right, the total cathode to ground circuit is 175 ohms, so thats a total of 205 ma. You'd see 20.5 volts across the 100 ohm resistor if so. The pot is a current balance. Adjust for minimum hum, or you can put a 10 ohm resistor between pin 8 and the 100 ohm resistor on each tube and adjust the pot so the voltage across each of those 10 ohm resistors is equal.
4) its going to make power with 6.8 ohms on a 16 ohm tap. It won't damage anything, but its not going to perform quite as well as it would into a 16 ohm load. Basically your options are replace the speakers or just enjoy it as-is. Personally I'd just roll with it if it sounds good.


630v caps are very available. The thing with the rating is you have to consider DC voltage difference from plate to grid, along with the AC signal on top of it. There may also be a surge voltage at power-up before the tubes conduct fully that has a larger DC voltage difference. Honestly I don't even buy caps that are under 630v, and I use them for everything. Makes it easier to maintain stock vs having multiple voltages in the same value.

*edit*
forgot this link to a Mullard 5-20 in case you want to know a little more about the design topology here
http://www.r-type.org/articles/art-003d.htm
 
I agree with all that gadget73 said and would only add that you should get them working first and modify later.
 
gadget73 said:
Looks like a Mullard 5-20 amp but with slightly lower plate voltage. The 5-20 does closer to 30 watts, but whatever.
Click to expand...

@gadget73 -- Always enjoy your insights. Couple of questions:

The Mullard 5-20 shows a plate voltage of 430V on the EL34 plates -- whereas the Ampex 40W indicates 465V.

Did you mean to say "Looks like a Mullard 5-20 amp but with slightly higher plate voltage. The 5-20 does closer to 30 watts..." [...so all else being equal to the Ampex should output about 35W.] ?

I figured that by taking the ratio of the square of the plate voltages times the 30W... Does that sound right?

My other question is re: <<its going to make power with 6.8 ohms on a 16 ohm tap. It won't damage anything, but its not going to perform quite as well as it would into a 16 ohm load. Basically your options are replace the speakers or just enjoy it as-is. Personally I'd just roll with it if it sounds good.>>

With a nominal 6.8 ohm voice coil on the 16 ohm winding and with the output load reflected back to the transformer input, doesn't that more than double the current on the primary of the output transformer? And doesn't that cause some risk to output tubes and transformer primaries due to that higher current?
 
Last edited:
Thanks for your insights gadget, and your questions Coastinhome. I will rebuild the amps to start with and then revisit the question of modification. I will post to this thread as I refurbish them.
 
CoastinHome said:
The Mullard 5-20 shows a plate voltage of 430V on the EL34 plates -- whereas the Ampex 40W indicates 465V.
Click to expand...

looked like it says 405 on the schematic when enlarged a bit, bit of fuzz makes all the difference. So yeah in that case you'd get a bit of a bump in power output.

The half impedance does increase current flow at peak output, but unless you're running it full out all the time it won't bother anything. The cathode resistors will limit how much it can flow anyway. It would be more of a concern if it was fixed bias. Odd as it sounds, having a significantly higher impedance is more likely to damage the tubes. The voltage swings are worse and its more likely to arc internally.
 
gadget73 said:
The half impedance does increase current flow at peak output, but unless you're running it full out all the time it won't bother anything. The cathode resistors will limit how much it can flow anyway. It would be more of a concern if it was fixed bias. Odd as it sounds, having a significantly higher impedance is more likely to damage the tubes. The voltage swings are worse and its more likely to arc internally.
Click to expand...

Thank you @gadget73. That does make sense.

@indosmokejon -- I don't get around to many of these projects but I did complete a full rebuild of an identical pair of amplifiers about three years ago. I posted my experience and some photos in some threads here about Ampex consoles. You should be able to find the threads by searching for "Ampex 40W".

I restored them in basically their stock configuration other than the choice of capacitor technologies used and some attention to improving grounding schemes. Beyond that the only thing I've changed so far was splitting the output tube common cathode resistor (and bypass capacitor) so that quiescent currents of the output tubes could be matched by measurement. Once adjusted the cathode resistors and capacitors are again paralleled.

Here's a RadioCraftsman 500 output stage which provided the inspiration:

RadioCraftsman 500 Output Stage.png
 
Personally I would leave the input pot.
True you could eliminate it without consequence, but I find on my MC30s that you can turn down at the amps and find the sweet spot where they run silent during the rests. Wide open, you tend to get more static and extraneous noise than when you attenuate by 10-20% at the amplifier.
Plus you will be left with a hole to fill IF you take it out.
The Pots I used on my MC30s were sealed Bounes pots from Mouser. Not cheap but also not noisy or flimsy.
How about some photos.
We all love photos.

IIRC these have a nice industrial look to them, as if in another life they were hidden behind a walnut pannel...
 
probably does depend somewhat on the input sensitivity. The extra series resistance with a pentode won't negatively affect HF response like it does with a triode.

On my Rauland-Borg amps, one of the pots was bad and the other was a bit sketchy so I ditched both. The triode input stage sits outside the feedback loop on that amp and it had too much gain to be really useful without an input attenuator which caused a drop in HF response. I ended up changing the input stage to drop the gain so it was at a reasonable level. I guess I could have done an attenuator and compensated it with a cap, but changing a plate and cathode resistor vs adding a pair of resistors and a cap to the input is about the same effort, and I was replacing the plate and cathode resistors anyway.
 
Some pics. The cleaner of the two amps is my buddies, the ones I originally auditioned which were the inspiration for finding a pair. I borrowed it to use as a roadmap to help figure out what was maybe wrong with mine. Anyone recognize the yellow caps someone dropped into my unit at some point? Only identifier on them is CAT224-6C. My plan is a complete tear down and rebuild at this point. It’s how I’ve approached other projects like this before and it’s worked well. Like I mentioned before, good at polishing a turd, not so great with the engineering.6F71AA6F-1902-49D4-9910-039BAB44938F.jpegC35A2773-62D9-4881-9D47-A593DD4F6DC5.jpeg62FB1174-3981-41DD-81AA-9705FBB87C89.jpegE621C2CB-4696-4488-B846-3E90BDACB313.jpeg
 
Also, those are 6L6 output tubes in my amp, which is how they were sold to me, so not the correct tube. Someone tried to rebuild these, missed with one wire, a positive cap to ground, and the pair must have never worked again, and I bet they gave up on them. My pair must be an early version, as they don’t have the Triad stickers, they are black instead of grey, and they are all cloth wiring, whereas my buddies cleaner pair have all plastic coating on the output transformers wiring. The OPT in the old unit is marked 658-0027, the new is 658-0027B, for anyone who is curious.
 
Yellow caps look fairly generic. I've used plenty of similar ones myself. Probably metalized polypropylene like all the rest of the boring yellows. They actually work pretty well but you're never going to win cool points with them.
 
I ordered some k42y-2 Russian pio coupling caps to replace those yellow jobbies, but then I rummaged through my stash and found some vintage USA caps. The yellow are Pyramid brand, the green just say Fast? Are these also PIO caps? Would they be comparable to the Russian caps? Would you guys consider using these in an otherwise fresh build?
Also, I am having a hard time finding 300ohm 7watt replacement resistors. Can I just use 10watt instead? They seem a lot easier to find.
 
could be PIO, glass seals on the end would be a better clue. Leak check them at full voltage if you're going to consider using them.
 
You must log in or register to reply here.
Back
Top Bottom