Unless a draw is present, a choke I/P filter charges up to the peak voltage, just like a cap. I/P filter does. Good regulation, at approx. 0.9X VRMS, sets in, when the critical current is drawn. A good approximation for the critical current, in mA., is given by V/L. Placing a bleeder resistor across the 1st filter cap. that guaranties the critical current is drawn is a common practice. For a 10 H. choke, a 10,000 Ω resistor is an appropriate bleeder part. If the bleeder is present, a 600 WVDC 1st filter cap. will not fail. The bleeder needs to be high wattage.
If you refer to an older version of the ARRL handbook, you will find that LCLC is recommend for critical current (choke I/P) filters.
If you refer to an older version of the ARRL handbook, you will find that LCLC is recommend for critical current (choke I/P) filters.
It may not be just the first filter cap that needs high voltage rating (if you want to avoid large power ballast resistor or some form of power zener). And don't forget that mains may go at least +5% on what you may presently be measuring (especially in the middle of the day if all your neighbours are exporting solar power).
450V large electro's (eg. 470uF) are soooo common due to smps, so if you have the room and need that level of filtering then it can be quite cheap. Alternatively, a much smaller value cap can be used (ie. cheaper) because you have choke input for starters and the ripple current spec is unlikely to be exceeded.
If your screen and preamp demand is small, then you could try to run them off the mid-point.
450V large electro's (eg. 470uF) are soooo common due to smps, so if you have the room and need that level of filtering then it can be quite cheap. Alternatively, a much smaller value cap can be used (ie. cheaper) because you have choke input for starters and the ripple current spec is unlikely to be exceeded.
If your screen and preamp demand is small, then you could try to run them off the mid-point.
Thanks guys!
It is a 500mA+ western electric power transformer and a Hammond 193Q 10H @ 500mA choke. I suspect the ripple to be fairly low off the choke before the caps. I have even considered a group of five or so 10uf 800V poly caps.
I have another 7H choke for the driver tube supply.
It is for a quad of KT88s, so there should be enough current.
It is a 500mA+ western electric power transformer and a Hammond 193Q 10H @ 500mA choke. I suspect the ripple to be fairly low off the choke before the caps. I have even considered a group of five or so 10uf 800V poly caps.
I have another 7H choke for the driver tube supply.
It is for a quad of KT88s, so there should be enough current.
The Hammond 193Q may not be up to serving as the 1st inductor in a HIGH current choke I/P filter PSU. Unless an inductor is explicitly rated for choke I/P service, it is necessary to derate its current handling capacity. Noise and overheating are what occurs, when an inductor is "in over its head". 
You may be able to get a handle on things by installing a 0.47 μF. "fudge factor" cap., in what would be the 1st position of a CLC filter. Use a film snubber rated for at least 2500 WVDC. Truly NASTY kick back spikes can be present.
If vacuum rectification is to be employed, the 5R4 is only common tube whose PIV rating is "tall" enough. You will need a pair of 5R4s to handle 500 mA. of B+. Another possibility is pair of 6CJ3 damper diodes. If you use 6CJ3s, dedicate a filament trafo to them and tie the cathodes to 1 end of the heater. Make sure that the filament trafo you use has an outstanding hipot rating.
You may be able to get a handle on things by installing a 0.47 μF. "fudge factor" cap., in what would be the 1st position of a CLC filter. Use a film snubber rated for at least 2500 WVDC. Truly NASTY kick back spikes can be present.
If vacuum rectification is to be employed, the 5R4 is only common tube whose PIV rating is "tall" enough. You will need a pair of 5R4s to handle 500 mA. of B+. Another possibility is pair of 6CJ3 damper diodes. If you use 6CJ3s, dedicate a filament trafo to them and tie the cathodes to 1 end of the heater. Make sure that the filament trafo you use has an outstanding hipot rating.
Large amounts of AC are present in the 1st inductor of a choke I/P filter. To cope with that state of affairs, robust construction and altered gapping technique are required.
Magnetics are heavy. International shipping charges are quite oppressive. Try a domestic vendor, perhaps Heyboer, for inductors capable of withstanding the rigors of choke I/P service.
Somewhere around 1 μF., in the "fudge factor" position, behavior changes from choke I/P to cap. I/P and regulation becomes POOR. A 0.47 μF. part will take some pressure off the choke, but will not ruin critical current behavior.
Magnetics are heavy. International shipping charges are quite oppressive. Try a domestic vendor, perhaps Heyboer, for inductors capable of withstanding the rigors of choke I/P service.
Somewhere around 1 μF., in the "fudge factor" position, behavior changes from choke I/P to cap. I/P and regulation becomes POOR. A 0.47 μF. part will take some pressure off the choke, but will not ruin critical current behavior.
All very good advice. Personally, I would always use a dedicated transformer for the tube rectifier section. There's a lot of action going on in that rectifier, just another source for hash to enter the audio path.
This is a great application for Hammond's affordable split bobbin transformers.
Thank you Eli! You are certainly correct on that. I've looked at it on my scope when I start adding small caps before the choke. The 193Q is a monster choke weighing around 20 lbs. I would certainly not mind at all strapping a .47uF across the first position if it will help the cause and save the expensive chokes!
Blair
Blair
This would usually be how I would do things in general. I almost exclusively use different transformers for plate and filament. These are big potted cubes from Western Electric. I was in hopes of using them aline for this project.
Blair
Blair,
You are planning on passing approx. 500 mA. through your 193Q. Put the 0.47 μF. "fudge factor" part in, from the very beginning. Even then, be prepared to run like a striped tail ape, should the choke start "singing" loudly. If you get a small amount of noise, install an additional 0.1 μF. in the "fudge factor" position. Perhaps that will stop the "singing". It's "cut and try" to find the max. amount of capacitance that does not cause critical current behavior to cease.
You are planning on passing approx. 500 mA. through your 193Q. Put the 0.47 μF. "fudge factor" part in, from the very beginning. Even then, be prepared to run like a striped tail ape, should the choke start "singing" loudly. If you get a small amount of noise, install an additional 0.1 μF. in the "fudge factor" position. Perhaps that will stop the "singing". It's "cut and try" to find the max. amount of capacitance that does not cause critical current behavior to cease.
This is what Hammond had to say about operating the 193Q at 500mA as a choke input:
We did a software run to ensure that this part will be happy used as an input power supply and it is. As a DC choke, there is mostly DC current with resistive losses but no (very little) core loss because the frequency is zero. When used as an input choke, there is AC current with resistive losses, as above, but it now has core losses because the current is at 60 Hz. The calculated temp rise at 500 mA, 60 Hz (so full load) is 45 C degrees which is right in line with our power transformers at class A.
I will still use a "fudge factor" cap because any extra safety I can get, I will use.
Also, I will probably just run a single PP KT88 off this PS and use a beefy cap filtered supply for my sweep project since it may require more current than the 193Q can handle adequately.
Thanks Eli!
Oh, and FWIW, Heyboer quoted me a VERY good price for a 1.2KV 10H choke. I will not quote it because it is probably subject to change.
Blair
We did a software run to ensure that this part will be happy used as an input power supply and it is. As a DC choke, there is mostly DC current with resistive losses but no (very little) core loss because the frequency is zero. When used as an input choke, there is AC current with resistive losses, as above, but it now has core losses because the current is at 60 Hz. The calculated temp rise at 500 mA, 60 Hz (so full load) is 45 C degrees which is right in line with our power transformers at class A.
I will still use a "fudge factor" cap because any extra safety I can get, I will use.
Also, I will probably just run a single PP KT88 off this PS and use a beefy cap filtered supply for my sweep project since it may require more current than the 193Q can handle adequately.
Thanks Eli!
Oh, and FWIW, Heyboer quoted me a VERY good price for a 1.2KV 10H choke. I will not quote it because it is probably subject to change.
Blair
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.