My brother has discovered vinyl, so naturally I've offered to build him a phono preamp. I figure tubes would be more exotic than any opamp based design, so I'm going with tubes. He'll like the nice glow-in-the-dark thingies sticking up out of the enclosure.
I'm going to use a power transformer salvaged from a Dyna PAS3, so I'm limited to about 15mA plate current. I've been looking at Eli D's RCA Phono Redux. Both use grid leak (contact potential) bias on their second stages. (I can't use Thorsten Loesch's DC Phono, as its second stage alone draws 12mA or so.)
I've simulated them both in LTSpice, and of course they work fine. In Eli's phono pre, when I look to see what the second stage grid bias is, it says something like -190mV. Huh?? Shouldn't that show about -1V, since the cathode is shorted to ground, so can only be at 0V?
Maybe SPICE can't properly calculate grid leak bias. Does anyone know if SPICE models show the correct current draw from a 12AX7 (ECC83) using grid leak bias?
Since the plate of the grid leak biased 12AX7 is DC-coupled to the gate of the output MOSFET, I'm wondering if plate current (and therefore the conditions for that DC-coupling) will vary wildly between different 12AX7's. I'd like the circuit to be as easy to use and reliable as a tube circuit can be.
Do most all 12AX7's bias up the same when used with grid leak bias? Or do they vary quite a lot between samples?
I'm going to use a power transformer salvaged from a Dyna PAS3, so I'm limited to about 15mA plate current. I've been looking at Eli D's RCA Phono Redux. Both use grid leak (contact potential) bias on their second stages. (I can't use Thorsten Loesch's DC Phono, as its second stage alone draws 12mA or so.)
I've simulated them both in LTSpice, and of course they work fine. In Eli's phono pre, when I look to see what the second stage grid bias is, it says something like -190mV. Huh?? Shouldn't that show about -1V, since the cathode is shorted to ground, so can only be at 0V?
Maybe SPICE can't properly calculate grid leak bias. Does anyone know if SPICE models show the correct current draw from a 12AX7 (ECC83) using grid leak bias?
Since the plate of the grid leak biased 12AX7 is DC-coupled to the gate of the output MOSFET, I'm wondering if plate current (and therefore the conditions for that DC-coupling) will vary wildly between different 12AX7's. I'd like the circuit to be as easy to use and reliable as a tube circuit can be.
Do most all 12AX7's bias up the same when used with grid leak bias? Or do they vary quite a lot between samples?
Well, I am not the person to give the definitive and correct opinion, but I think grid leak bias IS very dependent on differences between different tubes, This as the bias is developed by stray electrons hitting the grid, so many factors are involved, such as the quality of vacuum, spacing between grid and cathode, and the coating and temperature of the cathode.
But the circuit by Eli is used by many, and they agree that the circuit works and sounds good 🙂
But the circuit by Eli is used by many, and they agree that the circuit works and sounds good 🙂
Have a look at this circuit also, maybe this can satisfy your needs. The text is in finnish, but the circcuit and notes are in english. Urarltone.fi have a nice explanation with point-to-point wiring diagrams also 🙂
Tee itse RIAA-korjain, Mauri Pännärin Revenge MM -rakennusohje | AudioVideo.fi
12 mA total current for the B+
Tee itse RIAA-korjain, Mauri Pännärin Revenge MM -rakennusohje | AudioVideo.fi
12 mA total current for the B+
An externally hosted image should be here but it was not working when we last tested it.
Last edited:
The grid leak biased 2nd gain block in the tweaked RCA setup runs the grid at approx. -2 V. The 20 Megohm resistor allows VERY few contact electrons to bleed off into ground.
Frankly, I'd like to see the 12 VAC filament winding used exclusively by a regulated DC heater supply that energizes the 12AX7 heaters.
If you want to vacuum rectify the B+, use an inexpensive Loctal 14Y4 and a $4.99 "Rat Shack" 273-1385 to energize its heater.
Frankly, I'd like to see the 12 VAC filament winding used exclusively by a regulated DC heater supply that energizes the 12AX7 heaters.
If you want to vacuum rectify the B+, use an inexpensive Loctal 14Y4 and a $4.99 "Rat Shack" 273-1385 to energize its heater.
Thank you Eli for your input 🙂
But I was of the impression that the stray electrons hitting the grid made up the negative bias needed to keep the tube in a linear part of working condition. Or is it so, the electrons keep the grid at a negative potential thanks to the high ohms resistor, that dont bleed of the charge then?
But I was of the impression that the stray electrons hitting the grid made up the negative bias needed to keep the tube in a linear part of working condition. Or is it so, the electrons keep the grid at a negative potential thanks to the high ohms resistor, that dont bleed of the charge then?
Hi Eli, thanks for jumping in.
Well, if a real-life 12AX7 settles with a grid-to-cathode voltage of -2V, then the SPICE models are not showing it. That's OK. Real world measurements trump modelling!
I intend to take the PAS xfmr's 11V AC winding and use it to get a regulated 12V DC, if I can. Probably with a diode bridge made from Schottkys, a 6800uF cap and a good ol' LM317.
I'll use UF4007s for the plate supply rectification. No need to use a tube rectifier. I could also use one of the many old 6.3VCT transformers and 6X4 tubes I have lying around, if I needed to. But I don't think I'll bother.
Hey, I have a question for you... How awful would it be to use a 12AT7 cathode follower instead of the source follower on the output?
I was also thinking I can get 300V for B+, easy.
Something along these lines? (attached)
Attachments
If you want to use a triode as the DC coupled voltage follower, be my guest. IMO, the enhancement mode MOSFET is superior. The FET requires no heater power, occupies little space, does not have a heater to cathode potential limit, and is very easy to set up.
Well executed voltage followers, regardless of active device technology, are highly transparent. Sloppy execution yields BAD sound.
Well executed voltage followers, regardless of active device technology, are highly transparent. Sloppy execution yields BAD sound.
100% agreed on the superiority of the MOSFET in this application. I'm just looking to put glowing tubes on view.
Thanks Eli! That prompted me to go back and make sure the 12AT7 follower was in a good spot. It wasn't. A 12AU7 is more appropriate there, no other values need to change. Distortion should be reduced by 3X.
Changing the cathode load resistor to 60k would put the 12AT7 in a better place, but then the plate current is only 1.6mA or so. The 12AU7 drawing about 3mA is a better choice for this particular application.
I'll try the Zetex MOSFET follower too. I wonder if I'll hear a difference... I expect the MOSFET follower would work noticeably better into a solid state amp or receiver.
--
Then vacuum rectify the B+ with a 14Y4, as I suggested previously. The Loctal bottle will be "impressive". L_RD knows the tube is very inexpensive. AAMOF, the Loctal socket costs more than the tube. 😀 Electrically, the 14Y4 and 12X4 are very close.
- Status
- Not open for further replies.