Last century the French Yannick Mahé showed an allmost the same circuit.Perhaps interresting ?
Mona
Mona
Attachments
Last edited:
Interesting that the Mahe version drives the 300B's with just 6SN7's everywhere. Of course it depends on what level input signal is spec'd. I estimate the 6SN7 input with 47K loading (Einstein version) is getting around 17 X gain. With the 100K load in the Mahe amp, that would increase to around 18.5 X maybe.
I can't read the R values on the differential stage, but they're probably a little higher (? maybe not, with the low R tail) than the 47K in the Einstein amp. too. The tail resistance is rather lower too. Half Mu gain there (for a CCS tail, but the low tail R here will give closer to Mu), so maybe 15 X gain there. Looks like the low tail R is what makes the real difference. So a full blown CCS tail might be a bit of a close call on gain.
Need to know what primary Z is used in the OT to determine exactly how much drive level is needed for the 300Bs. But probably something like 140 V pk-pk needed. A higher Zprimary requires less drive signal, so some wiggle room there.
I can't read the R values on the differential stage, but they're probably a little higher (? maybe not, with the low R tail) than the 47K in the Einstein amp. too. The tail resistance is rather lower too. Half Mu gain there (for a CCS tail, but the low tail R here will give closer to Mu), so maybe 15 X gain there. Looks like the low tail R is what makes the real difference. So a full blown CCS tail might be a bit of a close call on gain.
Need to know what primary Z is used in the OT to determine exactly how much drive level is needed for the 300Bs. But probably something like 140 V pk-pk needed. A higher Zprimary requires less drive signal, so some wiggle room there.
Last edited:
One interesting point I notice about the Einstein amp using a 400 Ohm tail resistor under the drivers (or the 170 ? Ohm R for the Mahe amp), this is close to 1/gm of the 6SN7s. That would put it in the range of the M. V. Kiebert 3rd harmonic fix for triodes. Just an observation. Would require an FFT analyzer to see if the 3rd harmonic dist. is nulled. (would depend on signal amplitude too)
The driver load resistors for the 6SN7s driving the 300B grids probably want to be a little lower than for driving 6B4Gs. More grid input capacitance expected. ....... Hmmm, not really, after looking at the datasheet, about the same.
The driver load resistors for the 6SN7s driving the 300B grids probably want to be a little lower than for driving 6B4Gs. More grid input capacitance expected. ....... Hmmm, not really, after looking at the datasheet, about the same.
Last edited:
The cathode resistance in the "differential" stage doesn't play any role.The voltage at the cathodes don't change.The two tubes are driven by the same signal with opposite fase from the fase splitter.If the current in one tube rises it drops by the same amount in the other.As if there was a big capacitor there.
A -Vg of 80V is noted for the 300A, max driving voltage 160Vtt before the grid gets positive.If you want 0.7Veff input sensitivity a 80x gain is needed, no problem.
Mona
A -Vg of 80V is noted for the 300A, max driving voltage 160Vtt before the grid gets positive.If you want 0.7Veff input sensitivity a 80x gain is needed, no problem.
Mona
Indeed.
I was assuming the Einstein amp was gain budgeted reasonably, so it would then need another 40% gain for the 300B. Now I have to wonder if Einstein went deaf. Maybe there were some plans for global NFdbk originally.
Kiebert has a paper on the low value R tail for removing the 3rd harmonic distortion. It only requires a very small amount of tail V variation to work.
I was assuming the Einstein amp was gain budgeted reasonably, so it would then need another 40% gain for the 300B. Now I have to wonder if Einstein went deaf. Maybe there were some plans for global NFdbk originally.
Kiebert has a paper on the low value R tail for removing the 3rd harmonic distortion. It only requires a very small amount of tail V variation to work.
Last edited:
In the original drawing on the Make Magazine site (link is at Tubecad site about the Einstein amp) the amp had a single tube phono stage in front of the input 6SN7 tube with a potentiometer between them. I bet the amp is pretty sensitive.
Re: Mona K
"The cathode resistance in the "differential" stage doesn't play any role.The voltage at the cathodes don't change.The two tubes are driven by the same signal with opposite phase from the fase splitter. If the current in one tube rises it drops by the same amount in the other.As if there was a big capacitor there."
Ohhh, I get you now. I'm suffering from splitters on the head lately. Signal is already split, equal complementary drives to the "diff" stage, so no gain loss except for plate loading.
----------------------------
This Einstein amplifier could use some seriously lower Mu tubes.
"The cathode resistance in the "differential" stage doesn't play any role.The voltage at the cathodes don't change.The two tubes are driven by the same signal with opposite phase from the fase splitter. If the current in one tube rises it drops by the same amount in the other.As if there was a big capacitor there."
Ohhh, I get you now. I'm suffering from splitters on the head lately. Signal is already split, equal complementary drives to the "diff" stage, so no gain loss except for plate loading.
----------------------------
This Einstein amplifier could use some seriously lower Mu tubes.
Thanks for taking the time to do that, some of those values were unreadable. I really appreciate it!
Mona K,
Thanks for the clarified schematic.
So a 470 Ohm tail on the "diff" and 27K loads.
---------------------------------------------------
One could play around with this setup. A 6BL7 "diff" driver running at say twice the idle current of the 6SN7 would give an authoritative driver and a more reasonable Mu around 13. (unfortunately also 3X the filament current of the 6SN7)
Lets say we put in some 300B "simulants" (a TV sweep tube with a V dropping zener from plate to g2), so the rest of us can build one:
36LW6 ($11) 140V Zener
6CB5A ($5) 100V Zener
26HU5 ($16) 150V Zener
36MC6 ($17) 150V Zener
These can handle a lot more plate current than the 300B, so the OT primary Z can be dropped in half or more (down to 1/4 X for 1/2 the B+). And the B+ could be dropped by 1/2 too. ( these are all like Mu 4 tubes, so the g2 transconductance is around twice the 300B's 1/Rp. Dropping the B+ will allow reducing the dropping Zener V, or even eliminating it altogether)
These would have about twice the grid 1 transconductance of the 300B, so we can reduce the drive requirements and Mu's further. Something like a 6EW7 (9 pin), 6DE7 (9 pin), or 10EG7 (octal) dissimilar triodes, with Mu 17.5 and 6.0. Two bottles, with the high Mu sections for the front end and splitter, and the low Mu sections for the "diff" driver. ( Like $3 to $7 each.)
Join the party....
Thanks for the clarified schematic.
So a 470 Ohm tail on the "diff" and 27K loads.
---------------------------------------------------
One could play around with this setup. A 6BL7 "diff" driver running at say twice the idle current of the 6SN7 would give an authoritative driver and a more reasonable Mu around 13. (unfortunately also 3X the filament current of the 6SN7)
Lets say we put in some 300B "simulants" (a TV sweep tube with a V dropping zener from plate to g2), so the rest of us can build one:
36LW6 ($11) 140V Zener
6CB5A ($5) 100V Zener
26HU5 ($16) 150V Zener
36MC6 ($17) 150V Zener
These can handle a lot more plate current than the 300B, so the OT primary Z can be dropped in half or more (down to 1/4 X for 1/2 the B+). And the B+ could be dropped by 1/2 too. ( these are all like Mu 4 tubes, so the g2 transconductance is around twice the 300B's 1/Rp. Dropping the B+ will allow reducing the dropping Zener V, or even eliminating it altogether)
These would have about twice the grid 1 transconductance of the 300B, so we can reduce the drive requirements and Mu's further. Something like a 6EW7 (9 pin), 6DE7 (9 pin), or 10EG7 (octal) dissimilar triodes, with Mu 17.5 and 6.0. Two bottles, with the high Mu sections for the front end and splitter, and the low Mu sections for the "diff" driver. ( Like $3 to $7 each.)
Join the party....
Last edited:
- Status
- Not open for further replies.