On Saturday, 2 March 2013 15:20:22 UTC+11, Alex Pogossov wrote:
"patrick-turner" wrote in message ... Recently we heard about Mr Pogossov's "boosted cathode follower" using a bjt between the cathode and 0V rail where the bjt has base current from an anode R between anode and B+.. Current production of bjt is regulated by Vg-k of the follower tube, and all sort of positive claims are made about lower Rout, lower THD etc, but not one measurement of Vac signals possible or distortion have been offered to prove the claims are valid. In the fullness of time, the voltages and THD measurements may follow, we might hope. Meanwhile, instead of using a bjt as an active current source in the cathode circuit, we can always use the bjt as a passive current source, as seen in my preamp circuits at Fig4 Schematic at page http://www.turneraudio.com.au/Line-preamp-2003.htm
Alex: Of course we can use a current source, but it is a more or less standard solution. You can also use just a 10K...20K resistor from the cathode of 12AU7 to GND and bias the grid to +40...80V.
***Indeed, but the CCS not only gives reduction of THD/IMD, but avoids the cut off distortion at highist Vo levels if the load is as low as say 10k, which is OK for 1/2 a 12AU7 to drive.
It is easy to drive a hi-Z load, but what if someone wants to drive a 600ohm load? Or a line coupling transformer which is nonlinear at low F?
***Well I would not expect a 12AU7 to drive such a horribly low load. Standards for professional equipments stipulate all signal handling gear have no less than 10k Rin, and no more than 600 Rout. Anyone wanting to drive 600 ohms with 1/2 12AU7 is a "Pharkahn Idiot".
Then your follower on the lousy 12AU7 will have gain of 0.5 instead of 0.95 plus several % of THD.
*** You are quite correct. But nobody in their right mind will use a 1/2 12AU7 to power 600 ohms. So if one really must power 600ohms, use an OPT, or a mosfet source follower, or maybe a power tube like EL84 as a follower and with 30mA at least for idle Idc so you can get about 11Vrms max, and to keep it linear you may need to boost the gain around the open loop, but it gets complex, so better is the OPT with 10:1 step down ratio, so the BQ5 in triode can work with 6k0 while output load becomes 600 ohms, and Rout at sec is Ra / ZR = 20 ohms so DF with 600 ohms RL = 30. For 12Vrms output, Va = 120V, Vg = 7.5V, so gain overall is 12V / 7.5 = 1.35 approx, and the 6BQ5 acts like a follower but it ain't a follower, because it has slight gain. THD into 6k0 RL is not corrected though.
Boosted cathode folower with Rout about 5 ohm or lower will just laugh maintaining almost the same 0.95 gain and low THD.
***But the bjt is doing the work if you have a 12AX7 follower up to the feeble current limits. At this point, we need to consider SPECIFIC schematics and all items in them to consider driving 600 ohms rather than make general statements to make your booster idea look better than it is. Please try to be scientific at all times.
If you just rewire your venerable MJE340 from a CCS into a booster common emitter configuration, you will vastly improve performance and NOT increase parts count.
***Well, the preamp where those MJE240 exist has been working fine since 2003, and located over 1,000km away, and I don't need to offer the owner an upgrade to something that already sounds so dreamily damn good.
***Please Pardon the human element within scientific discussions.
(You will be able to get rid of -125V rail as well.) Something for nothing? Sounds too good to be true, you might think.
***Next time I find myself farnarkling around with a cathode follower load, I'll try your idea, it shouldn't take more than 30 minutes to install a bjt and anode R etc and do a test for THD.
The last thing -- no one forces you to use this or that circuit. The website is only for "educational" purpose. Perhaps it might give some reader some hints.
*** My website has become +20dB more educational now that I have officially retired from the "Force Of Slaves", that army full of people desperate for munny, and forced to make hand crafted gear for customers who only pay peanuts instead of a socially just wage. Now, when they cry tears to get something nice that I might make, I say just "Make it yourself, its all the easy to follow info is at my website." They never ever offer me a financial deal that's better than the dole or the old age pension, and its quite unthinkable for them to ever offer $32.00 an hour which is now the gross mean wage in australia before tax. Self employed tradesmen usually charge $80 per hour, and the last medical specialist charged $320 for a discussion lasting 35 minutes. So this explains why there will probably not be too many "Pogssov Boosted Followers" built in Australia any time soon, but all that does not take away the interest I have in such matters.
*** Presumably, one could set up a power mosfet with small current sensing R in drain circuit, then cap couple the drain to a bjt base of a darlington pair power BJT in the source to -ve rail circuit, and Id at idle was say 1Amp, then
performance would drive a speaker, or anything not needing more than +/- 1A of peak current. Other ideas can what is dervived from a Ziclai circuit where you have a small signal npn bjt in common emitter gain mode driving a PNP power bjt with its collector output all applied to npn input bjt emitter, and with a CCS or just resistance to a -ve rail. Many SS power amps have used this principle with such ziclai pairs in complementary form. So thus the drive to the power bjts is determined by the current in Q1 collector RL, and the output bjt current forces the Ve of Q1 to chase after the Vb.
But there's no reason why the Q1 of the ziclai pair can't be a vacuum tube and the RLa generates a voltage to drive the base of Q2, a PNP and thus the high NFB involved gives lower THD and lower Rout.
IMHO, Q1 could be say 12AX7 paralleled, and with suitable small value RLa to give a suitable small drive voltage of maybe +/- Vpk to a gate of Q2 NP small power mosfet ( or darlington pair bjt) set up in class A with Idc = 100mA, and with its source going to a +10V rail, and with drain load of 100 ohms (or CCS) to -10V. Drain is also directly connected to AX7 cathode, and the 100r source load ( or CCS )offers the I sink for Ia of tube. The drain voltage output will be at same Vdc as 12AX7 Ek, so cap output coupling is safest, although Eg1 of AX7 may have adjustable Vdc to get Eout to be close to 0Vdc. Then the output will have no trouble driving all loads needed up to about +/- 8Vpk and +/- 0.1A pk, (32 ohm headphones would be fine,) and all action of the mosfet is rigidly controlled by 12AX7.
Another alternative for driving 600 ohms might be to use a pair of EL84 arranged as white follower, with top follower '84 in triode and bottom '84 in pentode, so that bottom unit tends to do most of the work into RL, but controlled with NFB of the top '84 follower. But you still end up with a tube driving a load under say 1k0, not good, and likelihood of oscillations, unless you have an OPT. As load ohms come down, and expected Vout goes higher, and as human expectations increase, then complexity, hardware and expense goes up. So possible solutions need to be assessed with a specific function in mind.
I did once make a headphone amp in a guitar amp for non critical monitoring of levels and using a single 6BM8. The input triode gave some gain and drove the pentode as a triode CF with choke load and about 30mA dc. Even with 32 ohm phones, there was enough Vo. But for better fidelity, there needs to be more voltage gain ( or current gain ) in devices used, which may be a complementary pair of darlington pair bjts. MJE340 and MJE350 with small TO92package bjt drivers will work fine in class A where Ic = 50mA, and with +/- 12V rails. A tube can be used as an input follower if needed, and its cathode load is the high Z load of the solid state arranged as darlingtons. Expect THD 0.01% at a few volts output into say 100 ohms.
Yet another way is to use a pair of input triodes in an LTP with CC tail.
The RLa of V1 is taken to an N channel mosfet source follower,( or to P and N complementary pair ). V1 anode goes to B+ equal to Ea of V1. Input is to V1 grid. Source output from mosfets idling at 0Vdc is connected to grid of V2 for NFB and thus Vout and V1 Vg1 follows after Vin V1 g1. Again, the mosfets don't need to be mosfets and can be pair of NPN and PND darlington pairs.
I'm rather not inclined to ever allow bjts to do the lion's share of the work when an audio signal is involved. Passive CCS are OK because they don't allow the bjt to do anything to the signal level, the tube does it all, and better because the supply of Idc is constant. Mosfets are more forgivable, IMHO, and I like every device used anywhere working in class A and its most linear region whatever the signal level is. I allow the human in me to interfere with the purest scientific rational because the sound of tube circuits justifies their irrational use. Apologies to fragile egos of any readers...
Patrick Turner.