Hi,
i would like to ask the community about suggestions to build an tube phono preamp using exclusively the highly praised russian 6H30 tube. Some commercial designers seems to having changed their designs from ECC88/6DJ8 to 6H30. But the gain of 6H30 is only 15 compared to 33 of 6DJ8.
We need at least 40dB of gain (60dB minus 20dB RIAA attenuation). I think this precludes a two stage design. A possible circuit may be a CCDGC (constand current draw grounded cathode) three stage design with split RIAA. What do you think?
i would like to ask the community about suggestions to build an tube phono preamp using exclusively the highly praised russian 6H30 tube. Some commercial designers seems to having changed their designs from ECC88/6DJ8 to 6H30. But the gain of 6H30 is only 15 compared to 33 of 6DJ8.
We need at least 40dB of gain (60dB minus 20dB RIAA attenuation). I think this precludes a two stage design. A possible circuit may be a CCDGC (constand current draw grounded cathode) three stage design with split RIAA. What do you think?
What's the advantage of the 6H30? From the data I've seen, it's a current hog with unexceptional transconductance; ECC88-oids do much better. And mu is pretty low for a phono stage first hole.
Ok, this is a citation from BAT:
http://www.balanced.com/products/line/Vk-50SE/6H30P.html
"The trial results surpassed all our expectations. The 6H30P proved to be better than the 6922 in virtually every respect - whether in its measured performance, or more importantly, in the quality of the resulting sound. From that moment on, our goal became clear - to obtain this new SuperTube and bring it to the West for use in our flagship products. "
http://www.balanced.com/products/line/Vk-50SE/6H30P.html
"The trial results surpassed all our expectations. The 6H30P proved to be better than the 6922 in virtually every respect - whether in its measured performance, or more importantly, in the quality of the resulting sound. From that moment on, our goal became clear - to obtain this new SuperTube and bring it to the West for use in our flagship products. "
I agree with SY here (as usual). The first tube in an RIAA stage really should have both high transconductance and relatively high mu (some call this desirable combination a high “figure of merit” tube). The high gm gives good noise performance in the first stage and the high mu also improves overall S/N by raising the signal higher above the noise sources in subsequent stages. Also the high mu gives you a good head start on the needed gain for RIAA, particularly when passive equalization is used. The good old 6DJ8 is fine, but real thoroughbreds for this application include the 417/5842, 6C45, and the triode-connected 7788 (my favorite right now). If you find a 437A or an EC8020, those would be great, although I’d sell them on eBay, use a 7788, and buy more LPs with the profits 🙂
Brian: From what I see, the tube's major virtue is a low plate resistance. That's useless in an RIAA stage since the plate resistance will presumably be padded upward for the RIAA equalization...
I certainly have no specific knowledge, but I'm willing to bet that BAT got a pile of them on the cheap.
I certainly have no specific knowledge, but I'm willing to bet that BAT got a pile of them on the cheap.
Marketing hyperbole from BAT aside, the 6H30 is probably an excellent tube for many other applications, such as for line stages, power amp inputs/phase splitters, etc. I have not used it yet, but I don't doubt its merits.
But one tube does not fit all applications, and it's just not the best choice for the critical first stage of an RIAA amp with a mu of only 15 (which is too low here ) and a gm of 18mS, IIRC (which is just OK here). The equivalent noise resistance of a triode is Req=2.5/gm. Hard to get around that little rule.
But one tube does not fit all applications, and it's just not the best choice for the critical first stage of an RIAA amp with a mu of only 15 (which is too low here ) and a gm of 18mS, IIRC (which is just OK here). The equivalent noise resistance of a triode is Req=2.5/gm. Hard to get around that little rule.
Not even that good. Steve Bench reports the gm at 20mA is something like 12mS. At 10mA, it's under 8mS.
Ughh. As you say, Victor Khomenko, whose products I do respect, probably got a pallet-load of these tubes cheap. Audio Research seems to like them too (but not in RIAA stages) .
To summarize your opnions, the ECC88/6DJ8/6922 is a better choice for the first gain stage.
Audio Note uses 6072 tube in their RIAA stage, it has a much higher plate resistance and a lower transconductance than 6DJ8. What do you think of 6072?
Audio Note uses 6072 tube in their RIAA stage, it has a much higher plate resistance and a lower transconductance than 6DJ8. What do you think of 6072?
The 6072 has a gm of 1.75mS, about the same as the 12AX7, which is pretty low. Compare to the 6DJ8 at 12.5mS, or 6C45 and 7788 at around 40-50mS. Look, a lot of well known RIAA designs have used low gm tubes in the first stage, especially the 12AX7. This is OK if you have a high output MM cartridge and don’t mind a little hiss in the background. I’m just saying there are easy ways to make it better at least from a noise perspective.
Konnichiwa,
You mean overhyped. As any valve, it has applications where it is not at all bad, but to Phonostages are not one.
There where several other points raised....
First, the first stage needs above all high gain, high transconductance is desirable as it reduces the circuit impedances, but if you use a high impedance RIAA Network it is not an essential point.
As for "High Gm reliably = low noise", hogwash, correctly applied a good quality ECC83 in the first stage of a Phonostage has about the same output noise level as any good quality ECC88, BUT usually has more gain, so in reality better S/N ratio, for a given input. The topic is rather more complex, much depends upon specific items of the design.
There are many circuits out there, the ANJ ones sounds good, many others do quite well too. Using Cascodes and/or Penthodes in the first stage is a really good option if you want something less usual.
Sayonara
docali said:
You mean overhyped. As any valve, it has applications where it is not at all bad, but to Phonostages are not one.
There where several other points raised....
First, the first stage needs above all high gain, high transconductance is desirable as it reduces the circuit impedances, but if you use a high impedance RIAA Network it is not an essential point.
As for "High Gm reliably = low noise", hogwash, correctly applied a good quality ECC83 in the first stage of a Phonostage has about the same output noise level as any good quality ECC88, BUT usually has more gain, so in reality better S/N ratio, for a given input. The topic is rather more complex, much depends upon specific items of the design.
There are many circuits out there, the ANJ ones sounds good, many others do quite well too. Using Cascodes and/or Penthodes in the first stage is a really good option if you want something less usual.
Sayonara
Don't waste money 😀
Look at that low priced E/PC900 6/4HA5.
With CCS load it performs really well as the input stage.
Low noise, Hi gain, low microphonic (at least the 6 units I've checked, all from Philips UK).
Ok, Ok, I admit it's not the best choice for the second one 🙄
http://www.dissident-audio.com/RIAA_PC900/Page.html
Yves.
Look at that low priced E/PC900 6/4HA5.
With CCS load it performs really well as the input stage.
Low noise, Hi gain, low microphonic (at least the 6 units I've checked, all from Philips UK).
Ok, Ok, I admit it's not the best choice for the second one 🙄
http://www.dissident-audio.com/RIAA_PC900/Page.html
Yves.
Yves: Yes, the 6HA5 and variants ought to be useful here. Another tube is the 6GK5. These variable-gain RF tubes shouldn’t exhibit much of a linearity problem since the signal levels are so low in the first stage of a RIAA preamp.
KYW: Hogwash? – not at all. Assuming decent circuit design with low values of resistors in the grid and cathode, and quiet supplies, the major noise contributor will be shot noise. Since triodes do follow the Req=2.5/gm rule fairly reliably (especially when comparing tubes with large differences in gm), you can expect the 6DJ8 to have an Req of about 200 ohms. In fact, it is specified at 300 ohm, still very low. The 12AX7 will have an Req somewhere around 1600 to 2000 ohms. The 6DJ8 will have a S/N that is better by about 8 dB. If the next stage is suitably quiet, the gain advantage of the 12AX7 will be trumped by the superior S/N of the 6DJ8 in determining total S/N. I too would prefer a higher mu than what the 6DJ8 offers in this application, hence the other tubes mentioned.
KYW: Hogwash? – not at all. Assuming decent circuit design with low values of resistors in the grid and cathode, and quiet supplies, the major noise contributor will be shot noise. Since triodes do follow the Req=2.5/gm rule fairly reliably (especially when comparing tubes with large differences in gm), you can expect the 6DJ8 to have an Req of about 200 ohms. In fact, it is specified at 300 ohm, still very low. The 12AX7 will have an Req somewhere around 1600 to 2000 ohms. The 6DJ8 will have a S/N that is better by about 8 dB. If the next stage is suitably quiet, the gain advantage of the 12AX7 will be trumped by the superior S/N of the 6DJ8 in determining total S/N. I too would prefer a higher mu than what the 6DJ8 offers in this application, hence the other tubes mentioned.
Konnichiwa,
Howash, I'm happy to say it again.
That may or may not (in fact not) be so, Simply make a pair of comparable stages, followed by RIAA I find that the ECC83 stage and the ECC88 stage have similar levels of noise in the AF Band, both with the same strong emphasis on lower frequency noise. Often which one is quieter has more to do with the actual manufatcurer and exemplar than with the type.
Now what theoretically should be the case is of little consequence to me, what is obvserved is however is of consequence.
Given a real choice I'd use neither ECC88 nor ECC83 in Phonostages (I use myself E810F Penthode as first stage and D3a as second), but with ECC83 vs. ECC88 (and 12AY7/6072A/12AT7/ECC81) it is in my experience more a question of circuit design, the 6072A/12AY7 is materially the quietest though.
A 12AY7/ECC88 cascode can be great in the first stage, do this right and you can use a really neat RIAA EQ based strictly on operating the input stage as transconductance amp into a variable load to attain the EQ....
Sayonara
Brian Beck said:
Howash, I'm happy to say it again.
Brian Beck said:
That may or may not (in fact not) be so, Simply make a pair of comparable stages, followed by RIAA I find that the ECC83 stage and the ECC88 stage have similar levels of noise in the AF Band, both with the same strong emphasis on lower frequency noise. Often which one is quieter has more to do with the actual manufatcurer and exemplar than with the type.
Now what theoretically should be the case is of little consequence to me, what is obvserved is however is of consequence.
Given a real choice I'd use neither ECC88 nor ECC83 in Phonostages (I use myself E810F Penthode as first stage and D3a as second), but with ECC83 vs. ECC88 (and 12AY7/6072A/12AT7/ECC81) it is in my experience more a question of circuit design, the 6072A/12AY7 is materially the quietest though.
A 12AY7/ECC88 cascode can be great in the first stage, do this right and you can use a really neat RIAA EQ based strictly on operating the input stage as transconductance amp into a variable load to attain the EQ....
Sayonara
Kuei Yang Wang said:
But taking the arguments from above the 12AX7 should be superior to 6072 because is has a much higher mu than 6072 with comparable transconductance, ok plate resistance is about one third. Audio Note uses 6072 in a cascode for the first stage 😉
Konnichiwa,
In practice the noise is notably lower than the 12AX7 and as noted, in the audio range the classic noise formulas seem to have limited applicability/reliability.
Yes, good idea too. Sounds good as well.
Sayonara
docali said:
In practice the noise is notably lower than the 12AX7 and as noted, in the audio range the classic noise formulas seem to have limited applicability/reliability.
Yes, good idea too. Sounds good as well.
Sayonara
Konnichiwa,
Actually, it is a good example of using a cascode, but not for the type of Equalisation which I recommended....
Sayonara
pedroskova said:
Actually, it is a good example of using a cascode, but not for the type of Equalisation which I recommended....
Sayonara
- Status
- Not open for further replies.