That is a discussion about a valve head amp for LMC, you could use 4 6C45P also as I have wrote but I also warned its not an easy affair. A second "opinion":
"The question begged is, “Can I run an MC in directly?” Mmmm, perhaps, but at a noise penalty. Let’s say that you choose a particularly quiet specimen of tube; the very best have input-referred equivalent noise resistances in the 60-100R range. Cherry pick to get a 60R. Compared to the cartridge thermal resistance (15R), that represents a 7dB degradation of the already-marginal signal-to-noise ratio.
Massive paralleling of input tubes might prove efficacious, at a stunning penalty. Four tubes in parallel will drop the equivalent resistance to 15R, reducing the degradation to 3dB, which is still not quite there. Transformer starting to look better?
Things get worse: to get that low noise out of a tube, the transconductance will be quite high, and high transconductance means oscillation is but a moment away. The solution is generally grid-stopper resistors, but they have to be 100R-1k before doing any good, providing yet another noise source, one big enough to swamp the tube’s noise. Maybe you can wave the magic ferrite beads on the grid leads and keep things stable (I can’t). But that just takes us back to the tube noise problem. Nope, if we want tubes, we want MC, and we want quiet, we need a transformer."
Stuart Yaniger
His Master's Noise: A Thoroughly Modern Tube Phono Preamp
V1.2 measured 48dB but I will give that a margin down to 45dB worst case as there is a mu range in those valves spec per sample. That is why we move them around to match the channels gain in the end. 6SN7s have some tolerance too. We move both because the second stage is the gain multiplier and there is no global NFB in this circuit. So they help in gain matching even if they got less Mu play as types per se. DV-10X5 has been measured as 2.8mV on American test records where 5cm/sec is RMS when in a Japanese JVC test record is peak. 1/1.41 standards difference. 5/1.41=3.5cm/sec.
Back to my sweet little cheapy www.lundahl.se/pdf/1550.pdf then. I should be able to use it both in 6 and 12 dB step up to get modern standard output voltages dont you think? Seems like they even appreciate some grid capacistance.
Hi,
Those figures for Req are not related to LF so the point is rather moot.
Whatever you do all the //ed sections need to be matched and selected for low noise. The better they're matched the better the result. And that's true whenever valves are //ed.
Re: the 6C45: great animal if you can tame it.
The MC headamp I proposed has been working flawlessly for years, it does not oscillate even without gridstoppers.
It may not be the best choice for the lowest of LOMCs but you certainly won't hear any noise.
Anyway, I just suggested it as a cheaper alternative to a SUT. One way to make a SUT even more quiet is to use silver wire BTW.
I'm still reading through this thread so forgive me if I repeat some things that have been said before.
Ciao,
Those figures for Req are not related to LF so the point is rather moot.
Whatever you do all the //ed sections need to be matched and selected for low noise. The better they're matched the better the result. And that's true whenever valves are //ed.
Re: the 6C45: great animal if you can tame it.
The MC headamp I proposed has been working flawlessly for years, it does not oscillate even without gridstoppers.
It may not be the best choice for the lowest of LOMCs but you certainly won't hear any noise.
Anyway, I just suggested it as a cheaper alternative to a SUT. One way to make a SUT even more quiet is to use silver wire BTW.
I'm still reading through this thread so forgive me if I repeat some things that have been said before.
Ciao,
It would take a differential input to implement some CMRR, when a balanced impedance SUT primary offers in spades, to completely emulate its goodness means 3dB noise penalty thus a further tubes doubling to restore. It can be done but its not something I would lightly recommend for the casual builder. The whole gm and noise matching, the layout, the oscillation traps, the shielding, the CMRR preservation, the rail filtering, takes a fully equipped bench and experienced effort in low noise vacuum state circuits. A rare specialization few had mastered.
That, and a sale at a small boutique in a nice nothern town called Falun settled it. Sorry for OT
Last edited:
Hi,
That's one thing I don't understand: the 6N2P wasn't designed to run at Vg=0 (grounded cathode) so while the grid leak resistor may just about stop it from drawing current the bottom tube still isn't biased properly. (IMHO)
Hence, if you bypass the SUT you absolutely need to put something between cathode and ground be that a resistor (is RIAA still correct then?), a LED or a diode.
I also noted that the RIAA filter uses a 35.2 nF cap to ground. Is there a particular reason for this odd value? The calculated value for a 22K breakout resistor is close to 34nF.
Sorry for all the questions,
That's one thing I don't understand: the 6N2P wasn't designed to run at Vg=0 (grounded cathode) so while the grid leak resistor may just about stop it from drawing current the bottom tube still isn't biased properly. (IMHO)
Hence, if you bypass the SUT you absolutely need to put something between cathode and ground be that a resistor (is RIAA still correct then?), a LED or a diode.
I also noted that the RIAA filter uses a 35.2 nF cap to ground. Is there a particular reason for this odd value? The calculated value for a 22K breakout resistor is close to 34nF.
Sorry for all the questions,
Hi,
Those figures for Req are not related to LF so the point is rather moot.
Whatever you do all the //ed sections need to be matched and selected for low noise. The better they're matched the better the result. And that's true whenever valves are //ed.
Re: the 6C45: great animal if you can tame it.
The MC headamp I proposed has been working flawlessly for years, it does not oscillate even without gridstoppers.
It may not be the best choice for the lowest of LOMCs but you certainly won't hear any noise.
Anyway, I just suggested it as a cheaper alternative to a SUT. One way to make a SUT even more quiet is to use silver wire BTW.
I'm still reading through this thread so forgive me if I repeat some things that have been said before.
Ciao,
1/F region is a sore point where sample to sample variation is 10s of dB apart many times. I think I vaguely remember that pre-pre of yours from posts in the old years. Was single ended, right? Had NFB too?
Hi,
That's one thing I don't understand: the 6N2P wasn't designed to run at Vg=0 (grounded cathode) so while the grid leak resistor may just about stop it from drawing current the bottom tube still isn't biased properly. (IMHO)
Hence, if you bypass the SUT you absolutely need to put something between cathode and ground be that a resistor (is RIAA still correct then?), a LED or a diode.
I also noted that the RIAA filter uses a 35.2 nF cap to ground. Is there a particular reason for this odd value? The calculated value for a 22K breakout resistor is close to 34nF.
Sorry for all the questions,
Yes that is what we did for direct MM or HMC, we put something. You will have to look in the further builds links. The little differences in treble EQ were practical for offsetting a drooping certain SUT. We recommend some customization play there.
P.S. The bottom tube managed 1.3mA with the SUT in that original build.
Hi,
Absolutely.
RE the pre-pre, I tried to revive it somewhat lately with little success.
I thought about going differential with it, it will add twice the amount of tubes and I am rather uncertain about the noise cancelation of that in the end.
If you feel inclined to take a look, the circuit is re-posted here:
http://www.diyaudio.com/forums/analogue-source/252588-mc-head-amp-stage-s-using-tubes.html
As you can see it uses a J-FET (obsolete one for which I'd like to find a good alternative) configured as a CCS, a cap multiplier using a MPSA14 preceded by a standard 24V Vreg.
Two //ed ECC88 in a grounded cathode config. No global feedback.
Alternative tubes are 6S3 (Russian), D3a (and equivalents). Basically any triode that is linear at Vg=0.
In my case I stuck to the ECC88 since it shouldn't overload an existing MM stage. If you design from scratch then that would make life a little easier.
Ciao,
Absolutely.
RE the pre-pre, I tried to revive it somewhat lately with little success.
I thought about going differential with it, it will add twice the amount of tubes and I am rather uncertain about the noise cancelation of that in the end.
If you feel inclined to take a look, the circuit is re-posted here:
http://www.diyaudio.com/forums/analogue-source/252588-mc-head-amp-stage-s-using-tubes.html
As you can see it uses a J-FET (obsolete one for which I'd like to find a good alternative) configured as a CCS, a cap multiplier using a MPSA14 preceded by a standard 24V Vreg.
Two //ed ECC88 in a grounded cathode config. No global feedback.
Alternative tubes are 6S3 (Russian), D3a (and equivalents). Basically any triode that is linear at Vg=0.
In my case I stuck to the ECC88 since it shouldn't overload an existing MM stage. If you design from scratch then that would make life a little easier.
Ciao,
Oh, yes that was the pre-pre. Its the CCS (JFET in your case) that forces the bias, that happens in the μ follower too, the upper half is a CCS also and that is why it biased in the original Itch. Try a DN2540 for a modern alternative if you need run them hotter at 5mA each. Else, BF245A is in production.
Hi,
Thanks for the reply.
As I understand it the BF245A is also obsolete but I can still buy them in EU.
My main concern in view of what other people will have to use is it's limited availability worldwide.
The LS270K should do as well but it isn't exactly cheap.
You mean the CCS force the tube to self-bias, correct?
A decent MKP cap won't hurt there for sure. At the time those values simply did not exist in that tech.
Thanks for the help and should you ever feel the "Itch" for a differential phono pre, I am all for it.
Cheers,
Thanks for the reply.
As I understand it the BF245A is also obsolete but I can still buy them in EU.
My main concern in view of what other people will have to use is it's limited availability worldwide.
The LS270K should do as well but it isn't exactly cheap.
You mean the CCS force the tube to self-bias, correct?
A decent MKP cap won't hurt there for sure. At the time those values simply did not exist in that tech.
Thanks for the help and should you ever feel the "Itch" for a differential phono pre, I am all for it.
Cheers,