Hi all,
I've just received the transformer for my filament supply. It's rated at 8V in parallel configuration. With no load (DC reg boards and tubes) I get over 10V.
Is this because the 8V rating is for a 115V primary and mine sometimes being over 121V boosts the secondary relatively, or will this come down once the full load is connected? I doubt the latter and don't want to stress the regulated DC circuits over their input tolerances?
What is the best way to fix this? Resistors before the primary or after the secondary? How do I calculate this in order to get a useful selection of resistors? I assume I'll need about 10W?
Why do I get 3V between the matching phase dual secondary terminals? Shouldn't they be just above 0V? I shorted out the transformer secondary briefly while I was probing with the DMM. Can this cause serious damage? It was on for a couple of seconds.
That's more than three, huh? Whoops, there's another.
thanks,
gary
I've just received the transformer for my filament supply. It's rated at 8V in parallel configuration. With no load (DC reg boards and tubes) I get over 10V.
Is this because the 8V rating is for a 115V primary and mine sometimes being over 121V boosts the secondary relatively, or will this come down once the full load is connected? I doubt the latter and don't want to stress the regulated DC circuits over their input tolerances?
What is the best way to fix this? Resistors before the primary or after the secondary? How do I calculate this in order to get a useful selection of resistors? I assume I'll need about 10W?
Why do I get 3V between the matching phase dual secondary terminals? Shouldn't they be just above 0V? I shorted out the transformer secondary briefly while I was probing with the DMM. Can this cause serious damage? It was on for a couple of seconds.
That's more than three, huh? Whoops, there's another.
thanks,
gary
You're getting 10V for both those reasons. If your mains measures 121VAC (and that's still in spec! I always design for 130Vmax) then your 8Vnominal at 115V becomes 8.4V.
The primary and secondary windings will have resistance which will drop some volts as you start pulling amps. You can measure these easily and factor them into the equation.
You can use resistors to drop the volts, 1 ohm drops 1V at 1A which is probably all you need.
However, what about running a DC filament. The rectifier diodes will drop around 0.6V each regardless of current and then you can filter and get all that mains rubbish off your filaments.
In any case make sure whatever you use is rated for 1W minimum - I'd use 5W devices for longevity.
Measuring between windings will give strange results due to leakage currents and the high impedance of the meter. I wouldn't pay much attention to this. As for shorting the windings, not recommended but you'll be fine for short periods. The winding resistance will limit current and most transformers are short tested as part of UL safety testing.
Why didn't your fuse blow?
The primary and secondary windings will have resistance which will drop some volts as you start pulling amps. You can measure these easily and factor them into the equation.
You can use resistors to drop the volts, 1 ohm drops 1V at 1A which is probably all you need.
However, what about running a DC filament. The rectifier diodes will drop around 0.6V each regardless of current and then you can filter and get all that mains rubbish off your filaments.
In any case make sure whatever you use is rated for 1W minimum - I'd use 5W devices for longevity.
Measuring between windings will give strange results due to leakage currents and the high impedance of the meter. I wouldn't pay much attention to this. As for shorting the windings, not recommended but you'll be fine for short periods. The winding resistance will limit current and most transformers are short tested as part of UL safety testing.
Why didn't your fuse blow?
Iain,
thanks for the info. I like the diode idea. The transformer is first feeding a pair of dc regulating boards which work nicely up to 12-15 Volts but produce more noise and heat the higher the AC goes. Ideally I want to be feeding them 7.5 V to get my 6.3. (My line stage uses 4 6SN7s.) I have 4 10W 0.47uF resistors to dial this in but without loading the circuit with the tubes they don't do anything, my DMM still reads 10.1V. (I don't understand this, shouldn't the voltage drop be reflected even without any current?) As for the fuses, I haven't added them yet as I am still connecting the entire PS together on my desk to test it before I solder and fasten everything together.
Thanks again for the help. I take it I will need about 5 Ohms of resistance to get 2.5 V down at 2.4 amps?
gary
thanks for the info. I like the diode idea. The transformer is first feeding a pair of dc regulating boards which work nicely up to 12-15 Volts but produce more noise and heat the higher the AC goes. Ideally I want to be feeding them 7.5 V to get my 6.3. (My line stage uses 4 6SN7s.) I have 4 10W 0.47uF resistors to dial this in but without loading the circuit with the tubes they don't do anything, my DMM still reads 10.1V. (I don't understand this, shouldn't the voltage drop be reflected even without any current?) As for the fuses, I haven't added them yet as I am still connecting the entire PS together on my desk to test it before I solder and fasten everything together.
Thanks again for the help. I take it I will need about 5 Ohms of resistance to get 2.5 V down at 2.4 amps?
gary
Tom Bavis said:
Not sure what you mean here Tom - possible typo?
Anyway, you're right. Once the transformer is loaded the voltage may be just fine. Gary, You need to calculate or measure the loaded output voltage. Tweekers suggestion will work to measure it.
If you have regulators already then just make them beefy enough to handle the power dissipation and you're done. Most series regulators I've used need at least 2V to guarantee regulation. You say yours prefer ~1V? Maybe a simple LM317 or emitter follower stage would give you better noise performance? I'm surprised the regulators aren't better than you describe.
edit- Oh I see you say they work fine to 12V. You're good to go then! No wasteful resistors necessary
Thanks guys,
so much of this stuff is relatively esoteric (to me that is) that I forget the simplicity of Ohm's Law. I'll run the whole circuit tonight and report back with my results.
I'd like to better understand the series/parallel configuration better. I'll search the forum but if anyone has time to burn I wouldn't mind some pontification. The transformer powering the heaters gives 16 V up to 2.7 A in series and 8 V at 5.4 A in parallel. Does series/parallel require the center tap?
thanks,
gary
so much of this stuff is relatively esoteric (to me that is) that I forget the simplicity of Ohm's Law. I'll run the whole circuit tonight and report back with my results.
I'd like to better understand the series/parallel configuration better. I'll search the forum but if anyone has time to burn I wouldn't mind some pontification. The transformer powering the heaters gives 16 V up to 2.7 A in series and 8 V at 5.4 A in parallel. Does series/parallel require the center tap?
thanks,
gary
Many will try to deny it in subtle ways but we are all governed by the laws of physics. Yes, it is that simple! V=IR is true everywhere.
Regarding the series/parallel question. That is a bit more complex. The power loss aspect is true and this is why our electric co's distribute power at 100,000's of volts - a megawatt is only ~10A! You can tolerate a lot of cable resistance with this arrangement.
In a multi-tube amp design I believe there are other considerations. I'll admit to being a solid-state kinda guy here and bow to experience but I think you need to consider what happens if one of the filaments goes open/short in a series/parallel design. Seems to me the fail condition is to put excessive power in the other tubes filament possibly causing a domino effect failure mode. Don't know what this does to your speakers.
Any experienced vacuum-state guys care to elucidate?
edit_ no center tap needed here - that's a balanced audio thang.
Regarding the series/parallel question. That is a bit more complex. The power loss aspect is true and this is why our electric co's distribute power at 100,000's of volts - a megawatt is only ~10A! You can tolerate a lot of cable resistance with this arrangement.
In a multi-tube amp design I believe there are other considerations. I'll admit to being a solid-state kinda guy here and bow to experience but I think you need to consider what happens if one of the filaments goes open/short in a series/parallel design. Seems to me the fail condition is to put excessive power in the other tubes filament possibly causing a domino effect failure mode. Don't know what this does to your speakers.
Any experienced vacuum-state guys care to elucidate?
edit_ no center tap needed here - that's a balanced audio thang.
Iain,
perhaps what Tweeker was getting at is running each side of the transformer to each pair of tubes. That is to maintain two discreet 8 V supplies each supplying 2.7 A to one pair or channel of tubes. I assume all this takes is to not merge the phased pairs of secondaries (legs might be the right word) before sending them to the dc boards. Ostensibly this gives 2 8V supplies. In my case they differ in voltage by a few dozen mVs. Not sure if that's too kewel.
Thanks again for the help, I hope the recent fires didn't cause you too much alarm. It's wet and cool here in SF.
gary
perhaps what Tweeker was getting at is running each side of the transformer to each pair of tubes. That is to maintain two discreet 8 V supplies each supplying 2.7 A to one pair or channel of tubes. I assume all this takes is to not merge the phased pairs of secondaries (legs might be the right word) before sending them to the dc boards. Ostensibly this gives 2 8V supplies. In my case they differ in voltage by a few dozen mVs. Not sure if that's too kewel.
Thanks again for the help, I hope the recent fires didn't cause you too much alarm. It's wet and cool here in SF.
gary
We prepared to evacuate twice this summer but the decisive efforts of the firefighters kept us safe. The way they dealt with the wildfires was incredible - thank you firefighters! (if any firefighters also DIYaudio)
back on thread
I'm not sure that 2V is anything to worry about on a filament supply let alone 50mV. They are, after all, self regulating to a degree. Voltage goes up, current goes up, filament gets hotter, resistance goes up, current comes down.
Based on Olofs experience where the filament fails open, you'd want to have all filaments in parallel to maintain the same voltage across each regardless of the state of the individual tube. Series/parallel does reduce the total current flow but if one tube fails, the current through the remaining filaments will go down. Will anything happen? probably not. (As Geek always says: transistors have specs, tubes have guidelines.) Back in the day, electronic components were very low tolerance and tubes worked just fine. I think we're being overly paranoid here.
Build it and see. If you're in San Francisco, you should come to Burning Amp on October 19th (www.burningamp.com) and bring the amp. We can all see (and maybe even hear) it in real life.
back on thread
I'm not sure that 2V is anything to worry about on a filament supply let alone 50mV. They are, after all, self regulating to a degree. Voltage goes up, current goes up, filament gets hotter, resistance goes up, current comes down.
Based on Olofs experience where the filament fails open, you'd want to have all filaments in parallel to maintain the same voltage across each regardless of the state of the individual tube. Series/parallel does reduce the total current flow but if one tube fails, the current through the remaining filaments will go down. Will anything happen? probably not. (As Geek always says: transistors have specs, tubes have guidelines.) Back in the day, electronic components were very low tolerance and tubes worked just fine. I think we're being overly paranoid here.
Build it and see. If you're in San Francisco, you should come to Burning Amp on October 19th (www.burningamp.com) and bring the amp. We can all see (and maybe even hear) it in real life.
I just mentioned series parallel as a way to attain the resistance to simulate 4 parallel (assumed this was the layout) 6SN7 heaters with whatever parts are to hand.
For DC heaters series arrangement is more efficient due to diode and regulator losses. Also, if you have the reserve in the heater windings (both in terms of voltage and current) full wave is also more efficient at really low voltage than bridge rectification, again due to diode losses. It has lower transformer utilization, but diode losses can dominate. I suppose the rectifiers are on the boards, so this may not be relevant.
The cleanest DC would be obtained by burning off the excess voltage after rectification by an RC arrangement, before the regulator.
For DC heaters series arrangement is more efficient due to diode and regulator losses. Also, if you have the reserve in the heater windings (both in terms of voltage and current) full wave is also more efficient at really low voltage than bridge rectification, again due to diode losses. It has lower transformer utilization, but diode losses can dominate. I suppose the rectifiers are on the boards, so this may not be relevant.
The cleanest DC would be obtained by burning off the excess voltage after rectification by an RC arrangement, before the regulator.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.