If you are usig a modern source the gain from the EL86 is excess and it can just be removed.
A CCS at the bnottom of the long tail pair. Abd ECC81 (12AT7) can be substituted for the ECC83, its transfer curve is nicely in opposition the the EL84 and you get a more linear vtransfer curve.
Have a look at Elin Dutton’s El Cheapo for an evolution of this circuit.
We built an EL84 triode Class A variation thatrivels my nSIT-3 — but onlyoutputs 3.2 w at clipping.
dave
A CCS at the bnottom of the long tail pair. Abd ECC81 (12AT7) can be substituted for the ECC83, its transfer curve is nicely in opposition the the EL84 and you get a more linear vtransfer curve.
Have a look at Elin Dutton’s El Cheapo for an evolution of this circuit.
We built an EL84 triode Class A variation thatrivels my nSIT-3 — but onlyoutputs 3.2 w at clipping.
dave
As planet10 says (in part): "If you are using a modern source the gain from the EL86 is excessive......" I agree. See the link to the Tubelab Simple Push-Pull. Go to post 895 of this link: https://www.diyaudio.com/community/threads/tubelab-simple-p-p.148694/page-45 The Simple Push-Pull is easy to build and has more than enough gain for modern sources. Output transformer selection isn't super-critical. Low amounts of NFB, AND, it sounds good too! If you don't want to use a tube rectifier, use SS diodes. Tube rectifier: Power trans 300-0-300 HT secondary, SS rectifier use 275-0-275 HT secondary.
Cheers, S.
Cheers, S.
No, you can't use a 12AT7 without modification. This is a high-feedback amplifier, so ANY changes you make, including the output transformer, will require significant changes in the design. There really is no point building a design that you don't have the tubes or the proper transformers for. There are plenty of PP EL84 projects out there, like the one Steve links to.
grovergardner,
AI 6n2p writes that I can use the lamp with a small modification. Is this possible?
AI 6n2p writes that I can use the lamp with a small modification. Is this possible?
Steve Morley's advice is good. I'd follow it.
6N2P is a Russian twin-triode similar to a 12AX7 (or ECC83) but with a 6.3V heater (not 12V). Not a replacement for 12AT7 (or ECC81).
6N2P is different enough from a 12AT7 that it will not take the same circuit values. If you design for a 12AT7 you will not be able to simply swap in a 6N2P and have it work properly.
Using a 6N2P would require its own re-design of the circuit.
6N2P is a Russian twin-triode similar to a 12AX7 (or ECC83) but with a 6.3V heater (not 12V). Not a replacement for 12AT7 (or ECC81).
6N2P is different enough from a 12AT7 that it will not take the same circuit values. If you design for a 12AT7 you will not be able to simply swap in a 6N2P and have it work properly.
Using a 6N2P would require its own re-design of the circuit.
I suppose since the original schema in the first post used ECC83 (12AX7) then it would probably be a different story, and the 6N2P could be a sub in that case.
There is no gain stage in the first schema, so although you don't need all the gain of the EF86, I think you would need some. The EL84 looks to need about 20V peak to peak for max output. The ECC83 is a LTP phase splitter, so about unity gain. Your source is likely line level, about 2V peak to peak, so you would need 10x gain with no feedback, more with feedback.
This could be a candidate for an alternative solution: Useful PP Amplifier. The phase splitter is a concertina, so the other half of the triode would be giving you the gain you need. It is a design for an ECL86, but the triode part is like half an ECC83, so the passive components should be similar, and the pentode part you can keep from your original design (but anyway the ECL86 pentode is similar but lower power).
You do not need the bells and whistles, tone control and input selector. Hope it is not confusing.
There is no gain stage in the first schema, so although you don't need all the gain of the EF86, I think you would need some. The EL84 looks to need about 20V peak to peak for max output. The ECC83 is a LTP phase splitter, so about unity gain. Your source is likely line level, about 2V peak to peak, so you would need 10x gain with no feedback, more with feedback.
This could be a candidate for an alternative solution: Useful PP Amplifier. The phase splitter is a concertina, so the other half of the triode would be giving you the gain you need. It is a design for an ECL86, but the triode part is like half an ECC83, so the passive components should be similar, and the pentode part you can keep from your original design (but anyway the ECL86 pentode is similar but lower power).
You do not need the bells and whistles, tone control and input selector. Hope it is not confusing.
This is a myth that came about cos the original Mullard 5-10 had very high gain to allow for passive tone controls on the input.
But this version has MUCH less gain; 57x closed loop gain for the OP's Elektor 8R version. This is almost perfect for a Red Book CD player (2V at FS) via a 10k log pot passive preamp. This is true for the original Mullard 5-20 too.
Also the Elektor uses the EF86 in triode mode. This is nearly the same as half an ECC83 so that's my answer if you want to keep to the 'original' circuit. Check the anode voltages of the ECC83 phase splitter aren't changed too much.
Last edited:
You're right, in the original schematic in this thread, there is an EF86 pentode wired in triode, direct-coupled to an ECC83 (12AX7) long-tailed pair. While you can't simply remove the EF86 and use the rest of the design (the ECC83 would not be biased correctly if you do that) you can make a push-pull EL84 amp using nothing more than a 12AX7 and a pair of EL84s.
Actually, the gain of a 12AX7 LTP should be about half the gain of the same triode configured as a common cathode amplifier. You'd probably get about 25X gain from each triode's plate. This is how Eli Duttman's El Cheapo PP 6V6-oid amplifier was designed, except using a 12AT7 instead of a 12AX7. That could easily be adapted to EL84 output tubes.
You can also use a 12AX7 (or 12AT7) as a 'cathodyne' or a common cathode amplifier coupled to a 'split load' phase splitter. The Tubelab Simple PP design uses this type of design. That would give you more gain and is probably easier to get right than a push-pull amp design using an LTP input stage.
Another twin-triode that would work is the 6N1P (lower gain than 6N2P). Also 6N3P, 6DJ8, etc.
There are lots and lots of very effective and simple push-pull EL84 designs on diyAudio and elsewhere.
Tubelab SPP
Eli Duttman El Cheapo
Broskie simple PP EL84 design (I can't find the link, but here's the schematic):
There are many more.
The reason there are so many push-pull EL84 amp designs is that it's pretty easy to make a good sounding one.
--
EF86 is in triode mode, take a look, so the gain of that stage is not huge.
the problem is high impedance source of the driver stage with ecc83 and in the other sch ecc81
EF86 in triode mode has probably about 6 x gain
.
Use ecc82 for the driver stage or some double triode with lower Ri as ecc82.
.
EL84 does not deserve much input signal for max power.
it is biased about 10V Ug
so arround 20Vp-p is for maw power
if the input is 2Vrms that is arround 5.65Vp-p
You need only 4 x gain from input for max power.
.
I suggest to use 1:5 step up transformer followed with ECC82 as the buffer and driver for EL84.
in the same time this transformer is a phase splitter...
Without feedback.
Instead of the triode connected EF86 you can use one half of the ecc82; Cathode R becomes 2k2. If possible vary ths resistor to get 60v at its anode. But even when it isn't exactly 60v the amp will work just fine.
As mu g2g1 of the EF 86 is 38 and the mu of the ECC82 is 17 the feedback becomes a little less. I once had to repair a French power amp that also uses this ECC82\ECC83 combination as driver ( followed by KT88"s).
As mu g2g1 of the EF 86 is 38 and the mu of the ECC82 is 17 the feedback becomes a little less. I once had to repair a French power amp that also uses this ECC82\ECC83 combination as driver ( followed by KT88"s).