Hi,
from a scrapped Hammond S6 organ I have the power transformer. This is a rather heavy beast, it's plate secondary voltage is 290-0-290 Vac. Due to the big number of tubes in the S6, the heater winding is capable of 6.3 V/10 A. I'm about to use the transfo in a stereo power amplifier with two 807's per channel in PP with the output trannies of two Hammond K-100 organs (plate to plate impedance is 9.000 ohms). The PSU will consist of a silicon diode bridge, so the plate and screen voltages (the latter taken from the CT) would be 800 and 400 Vdc. Too much for 807's. So I'm thinking of choke input filters. But instead of the usual arrangement of the choke's placement between the rectifier output and the 1st filter cap, I'd prefer to put chokes between the transformer and the bridge. This would have the advantages of lacking DC current bias and regulation of both DC voltages. In addition, the laminations wouldn't need no air gap, hence inductance would be increased or less turns count were possible. Even a bifilarly wound dual choke for both AC lines is well within imagination.
My question is: Does this make sense? Or does a choke input rely on the choke's DC bias?
Best regards!
from a scrapped Hammond S6 organ I have the power transformer. This is a rather heavy beast, it's plate secondary voltage is 290-0-290 Vac. Due to the big number of tubes in the S6, the heater winding is capable of 6.3 V/10 A. I'm about to use the transfo in a stereo power amplifier with two 807's per channel in PP with the output trannies of two Hammond K-100 organs (plate to plate impedance is 9.000 ohms). The PSU will consist of a silicon diode bridge, so the plate and screen voltages (the latter taken from the CT) would be 800 and 400 Vdc. Too much for 807's. So I'm thinking of choke input filters. But instead of the usual arrangement of the choke's placement between the rectifier output and the 1st filter cap, I'd prefer to put chokes between the transformer and the bridge. This would have the advantages of lacking DC current bias and regulation of both DC voltages. In addition, the laminations wouldn't need no air gap, hence inductance would be increased or less turns count were possible. Even a bifilarly wound dual choke for both AC lines is well within imagination.
My question is: Does this make sense? Or does a choke input rely on the choke's DC bias?
Best regards!
The choke would effectively become transformer leakage inductance, and wouldn't support 100% choke conduction so the benefit of a choke-input style design to maintain a x0.9 output voltage level would not be available.
Practically you are likely to get 750Vdc from 290-0-290 unless your mains is high. You could wash some voltage off by using a CLC filter on each rail and use a lowish first C value (ie. circa 2uF) - the operating dc rail voltage is a bit touchy on first C value, but still quite practical to achieve circa 600V/300V. Another option is to use cathode bias of the output stages to wash circa 40V off.
The 807's were uprated to 400V screen for triode mode in the late 1940's, so are pretty rugged for slightly high fixed screen voltage. So you could just CLC filter the screen feed for lower voltage, and allow 750V for the anodes (which would have the benefit of lower hum via the screen supply).
With such separated anode and screen supplies, only the screen supply can be fused (at the CT), so over-current protection of the stereo output stage B+ levels may best be achieved by cathode end protection of individual valves as it would be tears if the B+ feed was opened only by a fuse.
Practically you are likely to get 750Vdc from 290-0-290 unless your mains is high. You could wash some voltage off by using a CLC filter on each rail and use a lowish first C value (ie. circa 2uF) - the operating dc rail voltage is a bit touchy on first C value, but still quite practical to achieve circa 600V/300V. Another option is to use cathode bias of the output stages to wash circa 40V off.
The 807's were uprated to 400V screen for triode mode in the late 1940's, so are pretty rugged for slightly high fixed screen voltage. So you could just CLC filter the screen feed for lower voltage, and allow 750V for the anodes (which would have the benefit of lower hum via the screen supply).
With such separated anode and screen supplies, only the screen supply can be fused (at the CT), so over-current protection of the stereo output stage B+ levels may best be achieved by cathode end protection of individual valves as it would be tears if the B+ feed was opened only by a fuse.
Ok, thank you. You're saying that the chokes need to be in the DC line, right? Anyway, I don't get where to place a screen supply choke in my arrangement. Remember that it is nothing else than a split PSU (CT'ed secondary, bridge, two capacitors) with it's negative output connected to ground.
Best regards!
Best regards!
Yes the choke-input filter is positioned either in the rectified pos line or the CT line, where the choke continuously conducts DC current throughout the AC mains frequency period. The diodes commutate the DC current alternately to the two half PT secondary windings.
Additional leakage inductance in each PT secondary winding causes a commutation glitch at the instant where one diode stops conducting and the other starts. In small power supplies like valve amps those glitches are not too noticeable, especially for valve diodes, but are a concern for ss diodes, and are quite a concern for larger industrial application transformers that have substantial leakage inductance and where ss diodes need to be protected from dV/dt induced failure.
The typical power supply circuit I think you are referring to is below, as was used by Philips back in the 1960's. The screen supply feed uses the CT and the two bridge diodes connected to 0V. That screen supply could use a CLC filter, with the L being a low current rating and hence add a few volts DCR drop, as well as significant ripple reduction.
I reckon a choke could even be inserted in the CT feed to make a choke input filter for the screen and hence provide a significantly lower screen/preamp voltage assuming the choke was large enough to maintain continuous conduction. I can see that continuous choke conduction would force one of those two bridge diodes to always be on, but that should cover the time when the other two bridge diodes provide commutation charging for the plate supply cap-input filter. But there could be some quirks in operation, such as if choke current were to reach zero, but maybe an extra steering diode in the choke feed could prevent any quirky reverse current situation.
Additional leakage inductance in each PT secondary winding causes a commutation glitch at the instant where one diode stops conducting and the other starts. In small power supplies like valve amps those glitches are not too noticeable, especially for valve diodes, but are a concern for ss diodes, and are quite a concern for larger industrial application transformers that have substantial leakage inductance and where ss diodes need to be protected from dV/dt induced failure.
The typical power supply circuit I think you are referring to is below, as was used by Philips back in the 1960's. The screen supply feed uses the CT and the two bridge diodes connected to 0V. That screen supply could use a CLC filter, with the L being a low current rating and hence add a few volts DCR drop, as well as significant ripple reduction.
I reckon a choke could even be inserted in the CT feed to make a choke input filter for the screen and hence provide a significantly lower screen/preamp voltage assuming the choke was large enough to maintain continuous conduction. I can see that continuous choke conduction would force one of those two bridge diodes to always be on, but that should cover the time when the other two bridge diodes provide commutation charging for the plate supply cap-input filter. But there could be some quirks in operation, such as if choke current were to reach zero, but maybe an extra steering diode in the choke feed could prevent any quirky reverse current situation.
Attachments
Yes, this excerpt is the PSU of a Philips EL6425 and the like and exactly what I'm intending (besides the capacitance values). So I could place a single choke either between CT and the capacitor to regulate the screen voltage (does this work at all?) or between the bridge's negative terminal and gnd to regulate both voltages?
Best regards!
Best regards!
I'm less sure of operation with a choke in the negative arm of the bridge rectifier to the filter cap negs. With all diodes alternately conducting, the screen supply cap is charging through the same duration as the anode supply cap. However the voltage waveform across the choke would not be the same compared to the typical situation where the screen supply and the anode supply were being charged through their own separate chokes. An LTSpice sim may be enlightening, and then a test circuit with variac.
Note that typical choke input filtering only implies a lowered B+ voltage when there is sufficient load current to maintain continuous conduction through the choke. Choke input filtering the screen/preamp feed has the likelihood of achieving a better regulated screen voltage, and down at or just below the datasheet 300V level, as the choke inductance can be conveniently made large given the low load current. There is also nothing wrong with deploying a psuedo choke input filter using a CLC network where the first C is low in value and helps manage forced diode commutation (as per a pure choke input filter), and raises the load voltage sufficient to hit a 300V level. For 807's, a high anode voltage is not a concern, given the OPT impedance/loadline is acceptable and the OPT insulation can support the higher voltage.
Note that typical choke input filtering only implies a lowered B+ voltage when there is sufficient load current to maintain continuous conduction through the choke. Choke input filtering the screen/preamp feed has the likelihood of achieving a better regulated screen voltage, and down at or just below the datasheet 300V level, as the choke inductance can be conveniently made large given the low load current. There is also nothing wrong with deploying a psuedo choke input filter using a CLC network where the first C is low in value and helps manage forced diode commutation (as per a pure choke input filter), and raises the load voltage sufficient to hit a 300V level. For 807's, a high anode voltage is not a concern, given the OPT impedance/loadline is acceptable and the OPT insulation can support the higher voltage.