Old Italian amp with ELL80-6HU8

[Robert M. Martinelli]

Hi folks,
Ok, what I have here is an old Italian amp ( belonging to a friend ) I am currently restoring/repairing, everything is working great now, but I haven't changed one of the tubes, a double pentode ELL80 ( 6HU8 ) used in the PI section; I have used a 5814 instead of the original ECC82 at the final stage' s input and a couple of Russian 6P36 as power tubes instead of the original EL500/504 ( 6GB5 ).

I have been asked anything from 40 to 75 Euro for a NOS ELL80 tube, so I' m starting thinking about a suitable replacement - The first thought would be to put in a couple of EL95 ( way cheaper, and the ELL80 is indeed a "double" EL95 ), but I would have to extensively modify the amp, adding two 7-pin sockets and changing the layout and wiring. This ELL80 is used as a ( double ) triode, as the screens and the plates are tied together, so I' m wondering about substituting it with a double triode with a rather high plate dissipation ( another ECC82-5814? a 12BH7? They look pretty close to a triode-connected ELL80 )....any thoughts?

Here' s the final stage' s schematic, any suggestion welcome.






Thanks in advance

Best regards

Bob

[Robert M. Martinelli]

"I' m a poor lonesome cowboy....."

C' mon boys!

Ok, I must admit I simply don' t have the right legs to play the defenseless girls crying out for help, neither the "phisique du role" ( French sounds better ), but is there anybody out there?

Bob

[6267]

Why was an ELL80 being used in the PI? Does the amp drive the power tubes past grid current? Why triode-wire a tube that could have been replaced with a dual triode?!? All very strange...personally, I would try to critically examine *why* the designer chose to use a pentode wired as a triode there as opposed to a cheaper tube like 12au7 with another gain stage in front or whatever. I am assuming it's an LTP type PI and not a cathodyne or paraphrase, but I kind of wonder if this is a driver tube *after* the PI allowing the gb5s (a tube with which I unfortunately have little familiarity) to be driven past grid current. In that case the signal would be taken off the cathodes of the ELL80s to the grids of the 6gb5s. If this is what is going on, a 6cg7 might be a good replacement. If the ELL80 is the phase inverter, and not a driver, I have to assume that it's either paraphrase or LTP, and LTP would make more sense (the power resistor for paraphrase with what is essentially an output tube would be an onerous design choice in my opinion).

Anyway, tubesandmore.com has the ELL80 for $40 US. Do you have a schematic?

[Steve Conner]

Hi Bob, all,

I took a quick peep at the schematic. There are coupling capacitors to the power tube grids, so the driver will hardly be able to push any grid current. Hence, that explanation is out.

Next I noticed that the bias feed resistors R4, R5 have very low values. So maybe the EL504s they were using had a lot of grid leakage current, and they had to use this low value to stop thermal runaway. In that case, the meaty PI tube could just be to drive these bias feed resistors.

So, if your replacement power tubes have a better grid leakage spec, I'd change these for 100k, change R6 and R7 for higher values (33k? 47k? 100k?) and replace the ELL80 with a 12AT7 or 12AU7, readjusting VR3 for balance. When changing the bias feed resistors, you should change the coupling capacitors too, to keep the same RC time constant. If the amp has negative feedback, messing with the time constants inside the loop can cause motorboating.

Otherwise, I'd use that big oversized ECC-thingy that JJ make. The ECC99, I think:
http://www.jj-electronic.com/images/small/ECC99.jpg

I don't know how much it costs, but it's got to be cheaper than an ELL80.

Then again, maybe they used the low bias feed resistors to let it drive a little more grid current. I don't know if that is significant.

Finally, I guess the EL504 is a TV sweep tube? John Chambers has an interesting amp that uses the EL360. He uses 220k bias feed resistors and drives 12 power tubes off a 12AU7 PI! Maybe those would work in your amp too?

http://www.chambonino.com/construct/const8.html

[Robert M. Martinelli]

Hi Steve, Hi 6267,
for what I can see it' s not a long tail...( or the tail's been cut ) as the cathode resistor is low and so is the cathode voltage.....LTPIs have a quite high value resistor between the cathodes and GND.

Thanks Steve, for you have reinforced my belief that a 12AU7/5814/6189 or a 12BH7 could do the job with just a little tweaking....I' ll also take a look at that ECC99 you' re talking about.

As to try to understand the reasons why they did what they did, I think they used EL504s/6GB5s because those tubes were used in TV sets ( and were thus cheap ) and the ELL80/6HU8 as a PI because during the '60s it was widely used in radio/phonographs so it is possible that they had a huge stock and wanted to do some house cleaning ( I' m just guessing but I could be right ).

I found an ELL80 on ebay at 40 USD too, I will ask my friend' s intentions ( get some ELL80s or modify the amp ) and act accordingly....

6267, talking about your question about the schematics, are you referring to the pinouts? ( the amp schematic is linked to my first post so I thought you've seen it already )

Thanks

Take care

Bob

[6267]

Sorry...missed the schematic, it's been a long week . It's not a cathode follower output, looks like an adjustable paraphrase to me...the signal is tapped off the anode of one side of the ELL80 and fed through a voltage divider (adjustable with pot) to the grid of the other. I imagine you could use any number of twin tubes here, but the important consideration is to have the same amount of gain. I'm not sure if the ELL80 is particularly high gain though, so ECC99 may work really well here. The grid circuit resistance is only 33k, so that would explain the use of what is essentially a power tube.
This is a pretty strange circuit with the sweep tube outputs, since it appears they need a beefy driver. You'llhave to tweak that voltage divider to the other grid after moving pin connections around, the idea is that the voltage divider removes the gain after the first side, which is made back up by the second side, giving an equal but inverted signal. There are calculations available on the internet (check out valvewizard and 'fun with tubes' via google).

Sorry about the delay in responce, it's been fairly busy over here.

[Robert M. Martinelli]

Hi Steve, Hi 6267!
Yeah, EL504 need to be driven hard, as to the gain, one of the reasons that made me think about the possibility of replacing the ELL80 with a "beefy" double triode was its low gain ( and the fact that in the schematic the ELL80 is indeed connected as a triode ; sorry I didn't update the thread but this has been a busy week here as well....

I talked to my friend and he preferred to keep the amp as original as possible, so I purchased a NOS Telefunken ELL80 in Germany for about 40 Euro. The SS rectifier allowed me to "beef up" the filter caps, and I've been very careful about GND connections because we were very concerned about hum and noise.

I had to work on the preamp section too ( which oddly is solid-state, with only a couple of BC108, after all this was a no-effect one-channel amp ) because it was scratchy, I had to change several caps and switched to BC109s because of their lower noise figure.

I biased the two 6P36S at about 9W quiescent plate dissipation, and we just tested the amp connecting it to a 2x12 Framus cab, and it really surprised me with a very good presence, bass register and dynamic behavior. The amp is very quiet, at a 3/4 volume it' s so silent it seems off, but when you strum on the strings it blows you away - I guess with a good Telly my friend could do without a drummer, given the way the amp pumps....

I' ll keep your precious advice for future reference, who knows if in the future we' ll be able to find another ELL80, so possibly the mods are only postponed.....

Thanks!

Best regards

Bob