Inspired by John Broskie's article about supertriodes, a collection of TV-tubes, 2 BUZ901 mosfets and my love for LTP and CCS, I came up with this circuit.
'All' I did was add the CCS to the circuit, but imo, it make a great difference!
Because the output stage has a gain of 1 now I needed to add a driver stage. Also, the input (miller) capacitance is pretty high
The driver stage is an LTP (duh) build with EF184's in triode, 8k2 on anodes and about 35mA through that stage. Works off the same 285V supply.
With this stage, -3db points are at 5Hz and 500kHz (-1db: 10Hz-250Khz) Also, distortion is pretty low.
If you're no balanced-lover, or just want to use a tube less, you can drive the output stage from a single source and ground the other connection.
My questions:
1. I need to keep the output free of DC, on my prototype, I used a potmeter in parallel with R3/R4. But I want to preserve AC balance..
Other possibilities are replacing R3&R7 with one potmeter (and R4andR8).
Somebody has an idea for a servo perhaps?
2. There is a little dip in the frequency plot around 9kHz. Where is it caused?
If there is interest, I will post some measurements and perhaps full schematic.
Jeffrey
'All' I did was add the CCS to the circuit, but imo, it make a great difference!
Because the output stage has a gain of 1 now I needed to add a driver stage. Also, the input (miller) capacitance is pretty high
The driver stage is an LTP (duh) build with EF184's in triode, 8k2 on anodes and about 35mA through that stage. Works off the same 285V supply.
With this stage, -3db points are at 5Hz and 500kHz (-1db: 10Hz-250Khz) Also, distortion is pretty low.
If you're no balanced-lover, or just want to use a tube less, you can drive the output stage from a single source and ground the other connection.
My questions:
1. I need to keep the output free of DC, on my prototype, I used a potmeter in parallel with R3/R4. But I want to preserve AC balance..
Other possibilities are replacing R3&R7 with one potmeter (and R4andR8).
Somebody has an idea for a servo perhaps?
2. There is a little dip in the frequency plot around 9kHz. Where is it caused?
If there is interest, I will post some measurements and perhaps full schematic.
Jeffrey
Attachments
With the CCS tail it will have to run in class A. And will have to draw a lot of current thru the CCS for 8 Ohm output.
The usual technique for DC stability is some feedback path with more gain at DC than at AC. The source resistors could be made larger and then bypassed, or some DC/LP only network rigged back to the gates or grids.
The circuit scheme is interesting in that it removes the need for two floating HV B+ supplies in Circlotron mode. The remaining floating LV supplies could be removed too by going to a Twin Coupled or McIntosh like setup using two LV filament xfmrs for OTs (dual 120V primary). Without the CCS for the drains then, it can run class AB too.
Or go half way and just put a LV dual pri OT in the drain circuits with the secondary paralleled to the 8 Ohm output. Will still have to run in class A with a high CCS current though.
The usual technique for DC stability is some feedback path with more gain at DC than at AC. The source resistors could be made larger and then bypassed, or some DC/LP only network rigged back to the gates or grids.
The circuit scheme is interesting in that it removes the need for two floating HV B+ supplies in Circlotron mode. The remaining floating LV supplies could be removed too by going to a Twin Coupled or McIntosh like setup using two LV filament xfmrs for OTs (dual 120V primary). Without the CCS for the drains then, it can run class AB too.
Or go half way and just put a LV dual pri OT in the drain circuits with the secondary paralleled to the 8 Ohm output. Will still have to run in class A with a high CCS current though.
Last edited:
late edit:
OOPs, the CCS only has to handle the class A tube currents in the shematic. The OT in the drains variant case would require the CCS to handle the 8 Ohm current, so not practical. The two OT Twin -Coupled variant gets back to only the tube currents thru the CCS and a single LV supply.
OOPs, the CCS only has to handle the class A tube currents in the shematic. The OT in the drains variant case would require the CCS to handle the 8 Ohm current, so not practical. The two OT Twin -Coupled variant gets back to only the tube currents thru the CCS and a single LV supply.
The CCS affects only the input tubes because the LV supplys are both floating.
It's kind of a hassle to BIAS this kind of output stage tough.
1. Choose bias current for mosfets, and look up what Vgs we'll need.
2. Look up plate curves of the triode, we want Vgs of the mosfet to be same as -Vg of the tube.
3. Chose your operating point on the given curve. (OR choose another valve)
4. Now we can calculate the anode and kathode resistors, and the CCS resistor.
The circuit as drawn gives 40W in 8ohm.
It's kind of a hassle to BIAS this kind of output stage tough.
1. Choose bias current for mosfets, and look up what Vgs we'll need.
2. Look up plate curves of the triode, we want Vgs of the mosfet to be same as -Vg of the tube.
3. Chose your operating point on the given curve. (OR choose another valve)
4. Now we can calculate the anode and kathode resistors, and the CCS resistor.
The circuit as drawn gives 40W in 8ohm.
1.DC servo ! , huh very hard ,you have to made some analysis of the Bongiorno`s SS Circlotron Amps topology to maybe find the right solution.
Use two hi quality 220R trimpotmeter instead of R3 & 4 , with conected middle point of each potmeter direct to two 220R gate stop resistor to adjust Circlotron optimal DC balance (offset) & to adjust separate BUZ bias ,but leave R7 & 8 same as they are ( 1K fixed ) .
2. C1 & C2 bootstrap condenser can make that dip around 9Khz , play with different values .
Best Regards
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.