Not so much wanting to start a new thread discussing the pros and cons of DCMB in general or is it really (signal)capacitorless at all but rather to ask opinions of different driver tube choices to use in the topology. Ari Polisois ( Ari I hope you'll jump in as well) seems to use/prefer the "evergreen" 6SN7 exclusively, while I think, that some newer, miniature tubes, with lower Ri could have the added benefits of lower output impedance and greater current feed capability. Wondering about 12BH7, 5687, ECC99 6N6P etc.... What’s your take on this and what tubes would be suitable?
Regards,
argo
Regards,
argo
argo,
Well I give you a hand, I am building the Simplex ; I wanted a SE amplifier was so excited about this design, and the 6C33C-B are gorgeous...
From Ari Polisois: in the Simplex article
{the first triode's plate connected directly, that is, without any blocking
condenser, to the grid of the second triode. The potential of the cathode must be higher, to ensure the necessary bias (between - 7 and - 9 V, depending on the actual characteristics of the tube used). }
Regarding that last point I think that a full hand of dual triodes can be used in this configuration.
But as we need the circuit to eliminate the bias, in the case of the 6C33C-B it has support 60 to 100 more volts than usual, and that could be the max. Voltage Plate of the tubes used that could eliminate them...
Alain.
Well I give you a hand, I am building the Simplex ; I wanted a SE amplifier was so excited about this design, and the 6C33C-B are gorgeous...
From Ari Polisois: in the Simplex article
{the first triode's plate connected directly, that is, without any blocking
condenser, to the grid of the second triode. The potential of the cathode must be higher, to ensure the necessary bias (between - 7 and - 9 V, depending on the actual characteristics of the tube used). }
Regarding that last point I think that a full hand of dual triodes can be used in this configuration.
But as we need the circuit to eliminate the bias, in the case of the 6C33C-B it has support 60 to 100 more volts than usual, and that could be the max. Voltage Plate of the tubes used that could eliminate them...
Alain.
Thanks Argo for calling me in.
Alain is right.
I used so far the 6SN7GT (mainly) but also the ECC82/12AU7 and recently the 6H30.
When it comes to supply a DCMB bias of high value ( 45 up to 125 V) you need a driver triode that works with a PS voltage gradient of at least 375 V to ground.
The best results with the 6SN7GT ( and the equivament ECC82) have been obtained with a B+ of 400-410 V.
From this high voltage you must deduct the bias voltage, across the load resistor, meant for the output tubes, i.e. around - 90 V for a 300B or a 6C33C-B (working respectively at 425 V and 270 V DC supply) and about 60 Volts drop on the cathode bias resistor .
Therefore, between the plate and anode of the driver you typically find : 410 - 90 - 60 = 260 V, which is acceptable for the above driver tubes. With a plate current of, say, 9-10 mA, you get the required bias using a 9k - 10k load resistor.
The plate dissipation if below the recommended limits.
This relatively low value of the load resistor can preserve the high frequencies delivered to the output tube(s) even if they have a relatively high miller-increased anode-grid capacity.
In this right moment I am working on a driver unit for a King III amplifier ( 3 parallel 6C33C-B with 50 W output ). Said driver has a frequency range of 5 to 120kHz at -3dB attenuation points (measured with a HP 334A Distortion Analyzer - I do not know how many pF it has at the input), but surely you cannot expect to keep it all the way out, unless you use low mu output tubes with low cg-a capacity and , of course a good quality OPT.
The circuit for the 6SN7/ECC82 uses :-
Anode of V1 (triode) 180 k-ohms - 2W
Cathode bias resistor of V1 : adjustable from 330 to 830 ohms
( this adjusts the bias from - 35 to - 120 V - depending on the
driver tubes used - Note : some of them fail to do so ).
Anode of V2 resistor : 10 k - 4 W
Cathode bias resistor : 2//2k7 - 4W = 1,35 k
Auxiliary resistor feeding this bias resistor from the B+ line:
10 K - 50 W - with heat sink.
Overall gain of the driver more of less 60x .
Both triodes' load resistors are connected to the 410 V B+ line, without RC filter in between.
At 50 V rms on V2's load resistor the distortion is below 2%
( with the Sovtek 6SN7s I used ).
It could be better, but some authors say that the power stage often "compensates" somewhat this figure that scares the purists.
Cheers
Alain is right.
I used so far the 6SN7GT (mainly) but also the ECC82/12AU7 and recently the 6H30.
When it comes to supply a DCMB bias of high value ( 45 up to 125 V) you need a driver triode that works with a PS voltage gradient of at least 375 V to ground.
The best results with the 6SN7GT ( and the equivament ECC82) have been obtained with a B+ of 400-410 V.
From this high voltage you must deduct the bias voltage, across the load resistor, meant for the output tubes, i.e. around - 90 V for a 300B or a 6C33C-B (working respectively at 425 V and 270 V DC supply) and about 60 Volts drop on the cathode bias resistor .
Therefore, between the plate and anode of the driver you typically find : 410 - 90 - 60 = 260 V, which is acceptable for the above driver tubes. With a plate current of, say, 9-10 mA, you get the required bias using a 9k - 10k load resistor.
The plate dissipation if below the recommended limits.
This relatively low value of the load resistor can preserve the high frequencies delivered to the output tube(s) even if they have a relatively high miller-increased anode-grid capacity.
In this right moment I am working on a driver unit for a King III amplifier ( 3 parallel 6C33C-B with 50 W output ). Said driver has a frequency range of 5 to 120kHz at -3dB attenuation points (measured with a HP 334A Distortion Analyzer - I do not know how many pF it has at the input), but surely you cannot expect to keep it all the way out, unless you use low mu output tubes with low cg-a capacity and , of course a good quality OPT.
The circuit for the 6SN7/ECC82 uses :-
Anode of V1 (triode) 180 k-ohms - 2W
Cathode bias resistor of V1 : adjustable from 330 to 830 ohms
( this adjusts the bias from - 35 to - 120 V - depending on the
driver tubes used - Note : some of them fail to do so ).
Anode of V2 resistor : 10 k - 4 W
Cathode bias resistor : 2//2k7 - 4W = 1,35 k
Auxiliary resistor feeding this bias resistor from the B+ line:
10 K - 50 W - with heat sink.
Overall gain of the driver more of less 60x .
Both triodes' load resistors are connected to the 410 V B+ line, without RC filter in between.
At 50 V rms on V2's load resistor the distortion is below 2%
( with the Sovtek 6SN7s I used ).
It could be better, but some authors say that the power stage often "compensates" somewhat this figure that scares the purists.
Cheers
Thank you Alan and Gigi(or Ari?) for your insight.
I was obviously overlooking, what influence the lowish value of load resistor of V2 has on frequency response.
Did a quick simulation with the 6SN7GTB and with the circuit values you gave above, and indeed the high frequency extends up to 100 kHz. THD 0.8% !?! (of course it's only simulation)
Still wouldn’t a bit more current be beneficial, especially if you want to parallel the output tubes?
The 6n30p is one hellave Russian "wunder kind" - can take the high plate voltage and put out gobs of current and only with slightly less mu of 15. They are'nt cheap though - the DR version can run up to 50 US money.
Now I wonder, if the low plate load of V2 helps the frequency response and additional current could came handy, what about using Russian 6N1P, with say 250V on the plate and 20mA cathode current?
Regards,
Argo
I was obviously overlooking, what influence the lowish value of load resistor of V2 has on frequency response.
Did a quick simulation with the 6SN7GTB and with the circuit values you gave above, and indeed the high frequency extends up to 100 kHz. THD 0.8% !?! (of course it's only simulation)
Still wouldn’t a bit more current be beneficial, especially if you want to parallel the output tubes?
The 6n30p is one hellave Russian "wunder kind" - can take the high plate voltage and put out gobs of current and only with slightly less mu of 15. They are'nt cheap though - the DR version can run up to 50 US money.
Now I wonder, if the low plate load of V2 helps the frequency response and additional current could came handy, what about using Russian 6N1P, with say 250V on the plate and 20mA cathode current?
Regards,
Argo
Hi Argo,
Dunno !
You should try it.
Are the pin connections the same as some of the valves I recommended ( because I tested them ) ?
Not having enough time to spare, I had no choice but to concentrate on few types of valves trying the get the most of them.
Regarding the increased plate current, as far as the 6SN7GT are concerned ( also branded 6H8 ), you are right. The THD drops significantly when you go from 9 to 10 mA and should probably even better beyond.
Keep also in mind that the signal at the second triode, if too strong, can cause distortion. It's a matter of compromising.
Good luck.
Ari.
Dunno !
You should try it.
Are the pin connections the same as some of the valves I recommended ( because I tested them ) ?
Not having enough time to spare, I had no choice but to concentrate on few types of valves trying the get the most of them.
Regarding the increased plate current, as far as the 6SN7GT are concerned ( also branded 6H8 ), you are right. The THD drops significantly when you go from 9 to 10 mA and should probably even better beyond.
Keep also in mind that the signal at the second triode, if too strong, can cause distortion. It's a matter of compromising.
Good luck.
Ari.
Sure Ari,
I will run some sims and definitely try it, if all looks good on paper. Not that I have anything against 6SN7(6H8C), quite opposite. It must be just this experimenting bug has bitten me.
Argo
P.S. I found a funny reference for 6N1P : (actually it,s a translation from Russian factory datasheet) sayng that - "6N1P is an analogue of 6N8C. The two tubes are interchangeable and are used in the same diagrams"
Anyhaow, it has a similar Ri of 11 but can be runned for a bit higher current and has mu of 35
instead of 20 of 6SN7. Might be useful in DCMB driver?
I will run some sims and definitely try it, if all looks good on paper. Not that I have anything against 6SN7(6H8C), quite opposite. It must be just this experimenting bug has bitten me.
Argo
P.S. I found a funny reference for 6N1P : (actually it,s a translation from Russian factory datasheet) sayng that - "6N1P is an analogue of 6N8C. The two tubes are interchangeable and are used in the same diagrams"
Anyhaow, it has a similar Ri of 11 but can be runned for a bit higher current and has mu of 35
instead of 20 of 6SN7. Might be useful in DCMB driver? Dear Argo,
I checked che characteristics of the 6N1P and found that this tube is similar to the ECC81/12AT7.
I have tried it in the past and I do not remember why I discarded it. Maybe it was because the 6SN7GT had a lower plate resistance.
The ECC82/12AU7 is closer to the 6SN7GT and I have tried it (my impression is that the sound is slightly better, but it was not with the same output valves).
If you try the different types, make sure you supply their heaters the correct voltage.
I would be grateful if you could keep me informed of the results.
Please call if you need schematics, etc.
I checked che characteristics of the 6N1P and found that this tube is similar to the ECC81/12AT7.
I have tried it in the past and I do not remember why I discarded it. Maybe it was because the 6SN7GT had a lower plate resistance.
The ECC82/12AU7 is closer to the 6SN7GT and I have tried it (my impression is that the sound is slightly better, but it was not with the same output valves).
If you try the different types, make sure you supply their heaters the correct voltage.
I would be grateful if you could keep me informed of the results.
Please call if you need schematics, etc.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.