My tube guru friend evaluated several tubes for this purpose, and the 12AX7 won, I think because of microphonics. But I wouldn't rule out the EF86 equivalents made by JJ or EH (which are in current production). I put a JJ EF86 (EF806s I think they call it) in the front end of one of my guitar amps, and measured a voltage gain of 230. More gain than I expected or wanted. Having most of the gain in the first stage means minimal noise, and being a pentode also minimal miller effect. Wiring a pentode as a triode reduces the gain quite a bit, and I'm not sure how that affects miller cap or output Z variations. Then you could have a 12AX7 or 6922 do second stage gain and output follower buffer.
If you are using a MM cartridge, the input capacitance loading of the cartridge is likely to be the weak link. It needs to be right. I rec using a test record and variable cap (pulled from an old radio tuner) to get that right, then replace the variable cap with a fixed polyprop, polystyrene or mica cap. Very few people bother with this, and yet it makes a rather big difference in the high freq response.
If you are using a MM cartridge, the input capacitance loading of the cartridge is likely to be the weak link. It needs to be right. I rec using a test record and variable cap (pulled from an old radio tuner) to get that right, then replace the variable cap with a fixed polyprop, polystyrene or mica cap. Very few people bother with this, and yet it makes a rather big difference in the high freq response.
The A- and RIAA-weighted average noise of an optimally biased, triode-connected EF86 is about 10 nV/sqrt(Hz) (assuming a "healthy" valve, some of them have much higher noise). That is, an amplifying device with 10 nV/sqrt(Hz) white input noise will give you about the same A- and RIAA-weighted SNR as a good triode-connected EF86.
For moving magnet, this is definitely not state of the art, but it also is not so bad that you would actually hear any hiss at moderate volumes in a normal living room.
For moving magnet, this is definitely not state of the art, but it also is not so bad that you would actually hear any hiss at moderate volumes in a normal living room.
I actually got about the same values with a more conventional triode connection (screen grid and anode tied together, suppressor grounded). Maybe that's to be expected because the screen grid is closer to the control grid than the anode anyway.
I made a preamp using a circuit I found in an old issue of Audio Magazine which seems to be about the same as the Mullard version (two EF86 tubes per channel). I thought it sounded best in one room I liked, but when I tried it out in some other rooms it didn't sound the same. Recently, I've been using a TC-750 transistorized preamp with a Fisher 500-B stereo tube amp.
Vintage Radio and Electronics. Mullard Pre Amplifier
EF86 valve phono preamp. - UK Vintage Radio Repair and Restoration Discussion Forum
Vintage Radio and Electronics. Mullard Pre Amplifier
EF86 valve phono preamp. - UK Vintage Radio Repair and Restoration Discussion Forum
The "infamous" Mullard pre-amp....
Never mind the the noise from the EF86, the Johnson noise from all those large resistors will drown the poor pentode in the ocean of noise....
All those 1Meg resistors...yuck!!!
The first stage is using local feedback to it's input grid, thus lowering the input Z to a virtual ground... now you need a SERIES input resistor for cartridge loading, thus more Johnson Noise... Better to inject the feedback into the SCREEN of the EF86 from a latter stage direct coupled follower...this way the EF86 is "forced" to behave like a TRIODE ... DC on the heaters does help since it is non-humbucking heater..
The EF86 can be used in phono pre to sound respectable... but it's a lot of work
Never mind the the noise from the EF86, the Johnson noise from all those large resistors will drown the poor pentode in the ocean of noise....
All those 1Meg resistors...yuck!!!
The first stage is using local feedback to it's input grid, thus lowering the input Z to a virtual ground... now you need a SERIES input resistor for cartridge loading, thus more Johnson Noise... Better to inject the feedback into the SCREEN of the EF86 from a latter stage direct coupled follower...this way the EF86 is "forced" to behave like a TRIODE ... DC on the heaters does help since it is non-humbucking heater..
The EF86 can be used in phono pre to sound respectable... but it's a lot of work
The use of a EF184 or 6Z51 russian triode connected give a gm 15 mA /v and RP of 3.7 kohm.
The gain is 55.
EF184 is better than 6z51 as general specs, tested on Sofia; practically the thd is lower in 184.
The noise figure of a gain stage with this configuration is very good. The Ranode can be about 15-18kOhm with 7-12 mA current; the Vdc can be in the range of 220-260 Vdc.
RK from120 to 220 ohm
I used many nos type that I found at reasonable price; the TFK and Mullard are the best also for selection.
The microphonics are not low but with some counter measure is possible to reduce the effect.
The overload is very high also.
Bye
Walter
The gain is 55.
EF184 is better than 6z51 as general specs, tested on Sofia; practically the thd is lower in 184.
The noise figure of a gain stage with this configuration is very good. The Ranode can be about 15-18kOhm with 7-12 mA current; the Vdc can be in the range of 220-260 Vdc.
RK from120 to 220 ohm
I used many nos type that I found at reasonable price; the TFK and Mullard are the best also for selection.
The microphonics are not low but with some counter measure is possible to reduce the effect.
The overload is very high also.
Bye
Walter
I had a problem with hum until I modified the power supply to provide DC for the filaments. I tried a hum balance control, but only one channel would balance at a time. Also, its not recommended to put the power transformer on the same chassis as the preamp because it might pick up hum. National 6267/EF86 (made in Russia) tubes sounded better to me compared to Svetlana and some Sylvania 6267/EF86 tubes.
http://www.diyaudio.com/forums/tubes-valves/75972-my-favorite-preamp.html
http://www.diyaudio.com/forums/tubes-valves/71300-photo-gallery-169.html#post1240498
See
"Audio Designer's Handbook" Audio February, 1960, pg 36-50
http://www.diyaudio.com/forums/lounge/269140-old-electronics-magazines-line.html
The pages are missing from the issue here
AmericanRadioHistory.Com - Documenting the History of Radio TV and FM broadcasting
http://www.diyaudio.com/forums/tubes-valves/75972-my-favorite-preamp.html
http://www.diyaudio.com/forums/tubes-valves/71300-photo-gallery-169.html#post1240498
See
"Audio Designer's Handbook" Audio February, 1960, pg 36-50
http://www.diyaudio.com/forums/lounge/269140-old-electronics-magazines-line.html
The pages are missing from the issue here
AmericanRadioHistory.Com - Documenting the History of Radio TV and FM broadcasting
Last edited:
It's RIAA A.Bokarev
Wow... I have not seen that schematic before...but that is very similair to what I was describing as driving the screen with a follower ....This way the pentode is running in a "forced" TRIODE mode.... In some ways that could be called mu follower ...since the pentode gain is now the mu of it's own triode connection by follower, Haha....
I believe that circuit suffers from non linear gain, ie depending on the input amplitude level, the Gain will vary slightly.. ..needs another feedback loop to keep the gain stabilized...
The circuit I built uses a 6CB6 into a 6C4 follower....
Anyone else have experience with 6CB6 in phono stages ???
DC feedback only to the screen, and it's definitely running as a pentode direct coupled to the second stage. Operating points controlled by a fine balance of characteristics between the two tubes. The RIAA implementation is quite interesting and relies on the high rp of pentode mode operation.
Your right.... my bad....
I initialy did not notice the cathode bypass cap....
The circuit I made using 6CB6 and 6C4 does not use bypass cap and drives the screen with AC signal to force pentode into triode mode....The 6CB6 in Triode mode is very linear....
Last edited:
I recall Wireless World did such an article... I know because I built it.... using EF86 for phono. Must have been in the 1960-70 time. Low distortion and noise, too.
THx-RNMarsh
If you can trust the data for such a smal difference, with the 6N14 you can only have 6mA (7mA ecc) or going higher with the supply voltage.
Lower current = lower transconductance = less gain
And no mariage no more .
Mona
Last edited:
Hmm, that's not a well thought out circuit. I can see why they used a cascode input stage (low Cmiller), but the ECC84 has quite poor linearity, and cascode connection exacerbates this even more. I'm guessing it was prone to oscillation too, since it has an enormous 3.3k grid stopper and the cathode resistor is unbypassed, which will generate far too much noise for a phono amp. For some unknown reason the RIAA network is connected between the HT and anode, thereby making its minimal PSRR even worse! By contrast, the 6N2P is a highly linear but low-curent valve, making it a very poor choice as a cathode follower output driver, but it does have low Cgp! The ECC84 would work better in place of the 6N2P where it would make a better cathode follower and would also benefit from some harmonic cancellation (probably). It also needs output over-voltage protection. The 6N2P would work better in the input stage, possibly as a parallel gain stage.You could try with half German (or Dutch, Philips !) ECC84 and half Russian 6N2P.Everybody
Mona
And are those diodes between the earth pin and the chassis? I hope not... OK maybe that's just an ambigous drawing. Must try harder.
Last edited:
Someone ought to stand up for small signal audio pentodes in pentode mode, so here goes. EF184s in particular have fine gain and aren't noisy overall in context of that IME. Besides, flicker noise probably dominates in audio and triode mode doesn't buy out of that. EF86s have relatively low microphony for the gain. So yes it's quite feasible to design good phono preamps with pentodes in pentode mode........plus with MM it's easy to escape Miller C input loading.
Well if you think so don't build it
As a somewhat calmer version of the ECC88,the ECC84 not bad at all.
The first triode has a voltage gain of practicaly zero, only current (anode clamped by the cathode above), no miller and 3k3 stopper not enormous.The transconductance of the triode (with Ia=7mA) more or less 4mA/V, giving a cathode impedance 250 Ω.The 68 Ω extra makes not much difference,but improves linearity.
The upper triode in common grid config.gives the same current from the lower one but at very high impedance.That give a rather welldefined input resistance for the RIAA filter, the Ra of 20k.
By puting the filter at the anode the signal is passed to the next stage at a high impedance.I you put the filter near the grid of the second stage there has to be a very low impedance connection to the anode+20k and the filter sees the supply in series with the 20k.Or an extra series resistor and make the filter at a higher impedance (more noise,more loss=extra gain needed and yes smaller C's !).
As for the cathode follower, the high operating voltage and low signal level makes it strong enough for most tube amp's.
Mona
As a somewhat calmer version of the ECC88,the ECC84 not bad at all.
The first triode has a voltage gain of practicaly zero, only current (anode clamped by the cathode above), no miller and 3k3 stopper not enormous.The transconductance of the triode (with Ia=7mA) more or less 4mA/V, giving a cathode impedance 250 Ω.The 68 Ω extra makes not much difference,but improves linearity.
The upper triode in common grid config.gives the same current from the lower one but at very high impedance.That give a rather welldefined input resistance for the RIAA filter, the Ra of 20k.
By puting the filter at the anode the signal is passed to the next stage at a high impedance.I you put the filter near the grid of the second stage there has to be a very low impedance connection to the anode+20k and the filter sees the supply in series with the 20k.Or an extra series resistor and make the filter at a higher impedance (more noise,more loss=extra gain needed and yes smaller C's !).
As for the cathode follower, the high operating voltage and low signal level makes it strong enough for most tube amp's.
Mona
Hmm, I would call 1uV of Johnson noise from a 3.3k resistor far too much, as it is probably greater than the equivalent input tube noise.Well if you think so don't build it
3k3 stopper not enormous.
You may be forgiven for thinking so, but the linearity is sadly far worse than the ECC88. Admittedly this is less important in a phono amp where the signals are very small.
I meant that the RIAA capacitors should be connected to ground, not to the HT, so that they improve rather than worsen the PSRR at high frequencies.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.