Noob cap question
Kristian, Thu Mar 29 2007, 12:46PMFirst off, I would like to introduce myself and say hello to the group. My name is Kristian or just Kris. I just started work on my first coil a few months ago. I think I am coming along quite well, but I do have a few questions.
I have 15,000V 60mA to work with. I have figured about .01uF should do. I would like to use Cornell-Dubillier 942C20P15K-F Capacitors. They are rated at 2000VDC, .15uF. Is it necessary to quarter the DC voltage rating to get an accurate AC rating with these caps for this application? If thats the case, I would need 60 of them to get 60,000VDC .01uF. I don't want to blow them out, but could I get away with lower voltage, without putting too much stress on them. I will have the gaps set appropriately and will not be over powering the transformer. I'm asking because its going to be hard to explain to my wife why I need to spend another $250 on a project that is already looking like it will be a few hundred over the discussed budget (at usual).
So, what are you guys using with success? Should I use 60,000V or can I get away with say 30,000V? That would quarter my cost.
Re: Noob cap question
Marko, Thu Mar 29 2007, 01:22PM
Hi!
Most people just seem to ignore the AC rating on pulse caps and with 942C's it seems to work well. Your cap will never be charged above the transformer's peak voltage (21,2kV) so something like ~25kV MMC would be right, considering you don't use resonant charging wich would be damaging for your transformer anyway! I think 30kV is completely fine voltage to work with. I would actually remove a cap or two to get a bit higher-than resonant value.
Stacking 60 of these into a 60kV monster capacitor would be crazy if you ask me!
Marko, Thu Mar 29 2007, 01:22PM
Hi!
Most people just seem to ignore the AC rating on pulse caps and with 942C's it seems to work well. Your cap will never be charged above the transformer's peak voltage (21,2kV) so something like ~25kV MMC would be right, considering you don't use resonant charging wich would be damaging for your transformer anyway! I think 30kV is completely fine voltage to work with. I would actually remove a cap or two to get a bit higher-than resonant value.
Stacking 60 of these into a 60kV monster capacitor would be crazy if you ask me!
Re: Noob cap question
Hazmatt_(The Underdog), Thu Mar 29 2007, 06:19PM
1. Your tank circuit will ring down at about 2X the input voltage, so you want to be aware of that. You can even see it on simulation which is pretty handy. I'd go with at least a 30KV rating, and probably closer to a 40kV rating.
2. Your cap value .01uF, this is resonant, your transformer terminal voltage will rise until dielectric breakdown and kill your transformer.
10.6nF is resonant, I've done some efficiency analysis on tranformer-capacitor matching and my LTR (larger then resonant) cap value maintains terminal voltage of the transformer and doesnt load down the current too much. The multiplier is about 1.414-1.5, which gives you 16nF for your capacitor.
3. Even if the terminal voltage is maintained to nameplate rated value, the transformer is not ment to operate in this range. They are designed for short circuit conditions, eg. to a discharge tube. So the voltage that they are meant to operate at is .707 * 15KV = 10KV. So we must realize that even operating a transformer at 15kv stresses the transformer somewhat.
4. Try to put some bleeder resistors on your cap strings to discharge them safely. I've had some nasty little surprises from some of my caps before.
5. I would probably add potential equalizing resistors in parallel with each cap, but since I've never made a 942C MMC, you're going to have to see if any of the other members recommend this.
Hazmatt_(The Underdog), Thu Mar 29 2007, 06:19PM
1. Your tank circuit will ring down at about 2X the input voltage, so you want to be aware of that. You can even see it on simulation which is pretty handy. I'd go with at least a 30KV rating, and probably closer to a 40kV rating.
2. Your cap value .01uF, this is resonant, your transformer terminal voltage will rise until dielectric breakdown and kill your transformer.
10.6nF is resonant, I've done some efficiency analysis on tranformer-capacitor matching and my LTR (larger then resonant) cap value maintains terminal voltage of the transformer and doesnt load down the current too much. The multiplier is about 1.414-1.5, which gives you 16nF for your capacitor.
3. Even if the terminal voltage is maintained to nameplate rated value, the transformer is not ment to operate in this range. They are designed for short circuit conditions, eg. to a discharge tube. So the voltage that they are meant to operate at is .707 * 15KV = 10KV. So we must realize that even operating a transformer at 15kv stresses the transformer somewhat.
4. Try to put some bleeder resistors on your cap strings to discharge them safely. I've had some nasty little surprises from some of my caps before.
5. I would probably add potential equalizing resistors in parallel with each cap, but since I've never made a 942C MMC, you're going to have to see if any of the other members recommend this.
Re: Noob cap question
Kristian, Thu Mar 29 2007, 10:52PM
I was thinking 60,000V was a little excessive. Recalculating using the before mentioned caps and the new info I have received from you guys, I have come to two strings of 19. for 38,000V .0158uF . Pretty close to the 1.5x resonant value.
If the peak voltage of the transformer will reach 21.2KV, and the tank circuit rings down at 2x that, should I increase the voltage to 42,000V? It looks like that would give me .0143uF which is a little light, going by the 1.414 - 1.5x resonant capacitance.
Maybe middle of the road would be the way to go, 40,000V .015uF. That looks more reasonable than the 60,000V I had calculated before.
Thanks for the help fellas, and thanks for not eating me for lunch, Hasmatt.
Kristian, Thu Mar 29 2007, 10:52PM
I was thinking 60,000V was a little excessive. Recalculating using the before mentioned caps and the new info I have received from you guys, I have come to two strings of 19. for 38,000V .0158uF . Pretty close to the 1.5x resonant value.
If the peak voltage of the transformer will reach 21.2KV, and the tank circuit rings down at 2x that, should I increase the voltage to 42,000V? It looks like that would give me .0143uF which is a little light, going by the 1.414 - 1.5x resonant capacitance.
Maybe middle of the road would be the way to go, 40,000V .015uF. That looks more reasonable than the 60,000V I had calculated before.
Thanks for the help fellas, and thanks for not eating me for lunch, Hasmatt.
Re: Noob cap question
Marko, Fri Mar 30 2007, 09:16AM
I may be misunderstanding you, but in a damped tuned circuit voltage on the capacitor will *always* be highest in first half-cycle, no matter what you do. Cap will never see more than highest voltage the transformer charged it to, simply due to conservation of energy.
In this case it will (should?) be max of 21,2kV and 30kV is almost 50% safety clearance.
The only problem I see may be arranging the single string of these caps for optimum capacitance. Maybe you could buy some smaller CDE capacitors and still be cheaper?
Marko, Fri Mar 30 2007, 09:16AM
1. Your tank circuit will ring down at about 2X the input voltage, so you want to be aware of that. You can even see it on simulation which is pretty handy. I'd go with at least a 30KV rating, and probably closer to a 40kV rating.
I may be misunderstanding you, but in a damped tuned circuit voltage on the capacitor will *always* be highest in first half-cycle, no matter what you do. Cap will never see more than highest voltage the transformer charged it to, simply due to conservation of energy.
In this case it will (should?
) be max of 21,2kV and 30kV is almost 50% safety clearance.The only problem I see may be arranging the single string of these caps for optimum capacitance. Maybe you could buy some smaller CDE capacitors and still be cheaper?
Re: Noob cap question
Hazmatt_(The Underdog), Fri Mar 30 2007, 09:38PM
firkragg, as the cap is charged to 21kv and the system rings, the energy is transfered back and forth between L and C, from + 21kv to -21kv, so 2x is peak to peak, 42kv peak to peak, so 40kv rating may not be good enough. Remember this is 100% reversal in the ringing circuit.
Before you go and buy more caps, ask some of the guys who build MMC's and see if 40kv is going to be good enough. I don't want to be responsible for advising someone to buy a bunch of stuff they don't need, and costing them extra money.
My MMC's are rated for 40KVDC, but they're doorknob caps, so they may be more robust then the poly caps, I don't know for sure yet as I haven't blown either up yet.
*CHUCKLES* there's always next time....kris
Hazmatt_(The Underdog), Fri Mar 30 2007, 09:38PM
firkragg, as the cap is charged to 21kv and the system rings, the energy is transfered back and forth between L and C, from + 21kv to -21kv, so 2x is peak to peak, 42kv peak to peak, so 40kv rating may not be good enough. Remember this is 100% reversal in the ringing circuit.
Before you go and buy more caps, ask some of the guys who build MMC's and see if 40kv is going to be good enough. I don't want to be responsible for advising someone to buy a bunch of stuff they don't need, and costing them extra money.
My MMC's are rated for 40KVDC, but they're doorknob caps, so they may be more robust then the poly caps, I don't know for sure yet as I haven't blown either up yet.
*CHUCKLES* there's always next time....kris
Re: Noob cap question
Marko, Sat Mar 31 2007, 01:09PM
Afaik, voltage reversal 'attacks' the capacitor by mechanical stresses, and I always tought of it as a problem only with really big pulse capacitors. I didn't think it's really important with MMC's in pulsed operation as this.
I would like to be proven wrong though, and am also interested to see what others think.
For example, steve charged a 22,5kV cap to 20kV or so without any failures for quite a while
while some people still advise 2xpeak input voltage + safety clearance...
I guess it's all about price vs. reliability.
Nevertheless, overcurrent and overheating seem to be much more important things to be considered... I actually don't remember where I seen good polypropylene caps fail just from overvoltage even if they were slightly underrated.
Marko, Sat Mar 31 2007, 01:09PM
Afaik, voltage reversal 'attacks' the capacitor by mechanical stresses, and I always tought of it as a problem only with really big pulse capacitors. I didn't think it's really important with MMC's in pulsed operation as this.
I would like to be proven wrong though, and am also interested to see what others think.
For example, steve charged a 22,5kV cap to 20kV or so without any failures for quite a while
while some people still advise 2xpeak input voltage + safety clearance...
I guess it's all about price vs. reliability.
Nevertheless, overcurrent and overheating seem to be much more important things to be considered... I actually don't remember where I seen good polypropylene caps fail just from overvoltage even if they were slightly underrated.
Print this page