calling Yves....

[Nanook]

Yves, I had done a search of the "giant killer". and have the schematic. Can this be easily modified to use the 6BM8 tubes (ECL82)? I have an old PP amp I can hack up.

As I understand it, it's basically the pin configuration that differs between the ECL86 as used and the ECL82 that I want to. I'll re-use the premap section (including phono) or modify it . Any ideas? The transformers are ok, the amp has been running.

thanks

stew

[Yvesm]

Hi Stew !

Are you calling me and talking about that ?

http://www.dissident-audio.com/PP_ECL86/Page.html

Yves.

[Shoog]

If you go from the ECL86 to the ECL82 expect it to sound harder and less rich. The ECL82 is an inferior valve to the ECL86. The pentode section is not very linear - even when triode strapped.

Shoog

[Nanook]

I believe so. I have a hand drawn version of the circuit which looks similar (gingertube added/changed it slightly)

Shoog, the question is in regards to the ECL82/6BM8 specifically. I understand that many are not fond of it, but as many are, and it has been successfully used in many amps.

I already have an old receiver that uses a 5GU4 (?) rectifier, and 4 6BM8s in PP configuration (2 per channel). I am just wondering if the iron and tubes can be reconfigured into something like the schematic Yves has provided. My intent is to remove the tuning section (it's a mess and not really required---the receiver is far from collectable in any way), and retain the preamp, including the phono stage.

I have been looking for something better out of the box, but it seems that many feel any tube amp/receiver is worth $$$ (ebay, etc). After that, my only alternative is to start fresh with all new parts and (hopefully) end up with something better. But at a cost of at least $300 (CAD)/$265(USD)...and that still probably wouldn't be enough to do a decent job of it.

[Yvesm]

Hi Nanook,

Originally posted by Shoog
If you go from the ECL86 to the ECL82 expect it to sound harder and less rich. The ECL82 is an inferior valve to the ECL86. The pentode section is not very linear - even when triode strapped.

Shoog

Yes, and both triode and penthode sections of the ECL82 have less gain.

I've already tried to convert this schemo for ECL82 and I've failed.
Other tried too w/o success.
For the ECL82s, prefer a concertina phase splitter, the triode is exaclty an half 6SL7.

ECL86 was probably the latest audio tube enginereed by Philips labs and is very good tube.

Yves.

[gingertube]

Yves lovely circuit requires 2 things for optimum results.
1) Good Gm in the output (pentode) section, because the shunt feedback is tading off gm for reduced rp.
2) High rp low current operation of the triode section because the triodes rp forms the bottom leg of a voltage divider for the shunt feedback.

The ECL86 (6GW8) fits these requirements well. The ECL82 (6BM8) is less suited to both the above requirements but has the advantage of being available from modern production.

A solution which will (probably) work with the ECL82 is to use the 2 triode sections in hybrid arrangement as a cascode diff amp with bipolar transistors (as the lower elements of the cascode). This would give more gain, better linearity, lower noise AND increased rp in the front end - all things we want.

Cheers,
Ian

[planet10]

Originally posted by gingertube
A solution which will (probably) work with the ECL82 is to use the 2 triode sections in hybrid arrangement as a cascode diff amp with bipolar transistors (as the lower elements of the cascode). This would give more gain, better linearity, lower noise AND increased rp in the front end - all things we want.

Excellent suggestion... i've been looking for an excuse to do something like Allen Wright's PP2C



You'd just need to borrow the hybrid cascode from the phono amp...



any suggestions on how you would bias this up?

dave

[gingertube]

Dave (and others),
Use the RTP3C Phono Front end as a template - The ECL82 triode go at the top, bog standard NPN transistors like 2N5551 or even BC547B could be used at the bottom. Match them for beta (hfe) on your multimeter if it has a hfe test.

The gain of a cascode is approximately equal to :
RL x gm where RL is the load on the top triode and gm is the gm of the lower device.
There is lots of gm to play with in the transistors. Re = 26mV/Ic where Ic is in mA and then gm=1/Re.
We would want to run about 600uA per side so that gives Re = 43 Ohms and gm = 23mA/V
I would add say 10 Ohms in each emitter linearise the transistors and improve balance. Effective Re then becomes 43+10 = 53 Ohms and gm is degenerated to 1/53 = 18.9mA/volt - you could even use quite a bit more degeneration than that if you wanted.

Aside: I've built a number of SS amps using differential transistors on the front and degenerating the gm always improved the sound BUT in a SS amp front end you can overdo it. I always tried 20%, 10% and 5% degeneration and generally settled on 5 or 10%. The 10 Ohms suggested above give approx 20% degeneration. Whether this will be the case in a hybrid cascode I don't know.

The transistor bases are at 0V DC wise so you need a negative supply to return the diffamp current source to. I've had best success with cascode biopolar transistors as the current source (2 more BC547B) with either cheap RED LED reference or if you have plenty of voltage to work with, use one more BC547B reverse connected and with the base left unconnected as a low noise approx. 6.8V zener. The output tubes need about -16V bias for fixed bias so use a common supply of say -24V.

The triode grids need to be tied to a +ve supply. You don't want that too high (reduces Vak of the triode section) but it needs to be high enough to allow the say 1 or 2 V pk-pk signals driving the triode cathodes without it letting Collector volts of thetransistors getting too close to the base voltage (0V). Around 6 to 8 Volts would probably do.

The beauty of the ECL82 is that VERY high Rg1 on the output pentodes is allowed. In fixed bias the max value is 1MOhm.

I would also run the output pentodes in Ultralinear Mode BUT thats personal taste. With Ultralinear Mode + the shunt feedback you can probably get away with zero global feedback. Ig you prefer pentode mode then there is enough gain to push up the shunt feedback and add some global feedback.

Cheers,
Ian

[planet10]

Originally posted by gingertube
I would also run the output pentodes in Ultralinear Mode... If you prefer pentode mode

In this case it is most likely determined by my donor amp, which, IIRC, doesn't have UP taps (an old Electra)

dave

[Nanook]

any chance helping me out directly? I am no amp designer, the Sparton is 2X's 6BM8 per channel, PP. uses the 5AR4 as the rectifier. A local audio buddy would be helping me out, and is experienced with high voltages and the pitfalls of being overly anxious --he's built several tube amps from kits, and from scratch (his designs), but has no real experience with the little 6BM8s.

Again, I want to utilize as much as I have as possible: tubes, sockets and Iron. The rest is cheap enough and readily available here as a walk in at local electronics suppliers, or I can always order from digikey. I have no idea what my trannies are, other than they work in the Sparton (I've listened to it and found it pretty enjoyable, but the preamp is noisy and I think I could improve it by reducing all the tuner hardware)

Thanks to both gingertube and Yves for responding to my inquiry.

stew