[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: help me put my coil on a diet; its capacitor appetite is expensive!



Original poster: "Gerry  Reynolds" <gerryreynolds@xxxxxxxxxxxxx>

Hi Nick,

Sorry about not answering before this. Your email got lost in a sea of other email and I just now rediscovered it.

I will presume your 100amp ballast is with 240V so you are running at 24KVA. If your Cp is 100nf and you are running at 120 BPS sync and the timing of the SRSG is set normally, the voltage will be 43KV. However, if the timing is incorrectly set, the voltage could go as high as 87KV. This is not a case where you adjust for maximum voltage. I think it is very likely that the caps were overvolted. Cres for a 25KVA PIG is 320nf so if you want to run at 120 BPS sync, you will need a Cp of about 900 nf. Alternatively, you could go async and shoot for a high BPS. With 100nf and 400BPS, some of the firings will get to 39KV. I do think you need to have a larger Cp. With 200nf and 400BPS, some of the firings will get to 24KV.

Hope this helps,
Gerry R.

Original poster: "seanick" <edgarsbat@xxxxxxxxxxx>

Hi Gerry,
SG: SRSG
4 electrodes in the disc,
1 stationary.
All 3/8" tungsten carbide drill blank electrodes.
 9" diameter x 1/2" thick fiberglass rotor
BPS: 120  - 4 electrodes in the disc,  spinning at 1800 RPM.
Ballast point: Craftsman stick welder, output leads shorted, ballast point set to about 100 Amps (max is 200 so thats halfway, I think. if memory serves it was actually set slightly lower.)

Cp : .06 uF for the 2 x .03 - Maxwell 37667's in parallel, (strings of 1)
       .04 uF for the 4 x .01 - CDE 940C30P1K strings of 10

For the maxwell's death, I didn't get to see the amps being pulled on the LV side. but for the MMC failure I was prepared. Clip on amp meter display (which is questionable due to the pulsed operation) was 30 at initial startup, then rose to about 50 after the caps had obviously started failing but right before the breaker tripped, then on a later run did the same thing, rose up to 50 before something popped. the second time it wasn't the breaker though.rather, it stunk like capacitor and the pop had been from inside the cap box right before the smoke started rolling out of the cap box's vent.

About the BPS -
I was using a 3600 RPM motor before, but that motor was too underpowered and it no longer gets up to synchronous. I tried cleaning and re-greasing the bearings to no avail so I suspect that the windings are bruised and thus no longer delivering enough repulsion to spin the rotor at 3600. It won't even spin 3600 with no load! As that motor was no longer effective, I was looking at getting a new one.

While it was working still, but between runs, I glanced at the rotor and realized only 2 of the 4 electrodes were oxidized on the tips, and only 2 locations of the rotor looked like they had been heated up at all. I observed this effect a few times and was beginning to think that 4 electrodes was no different than 2 for 3600 RPM since 2 of them were at the zero-cross point and thus had zero-energy transfer. since I already have 4 electrodes I figure 1800 RPM is the same thing but using all 4 electrodes. Is that not the case? Is there additional benefit to run at 3600 and have more than 4 electrodes? Aside from the fact I did not have any powerful 3600 RPM motors floating around, another problem is that even if I had such a motor, at 3600 RPM any difference in electrode weights is quite a bit more noticable in vibration noise (and assembly travel across the floor) than with only 1800 RPM.


I also saw responses from others. I am not actually on the list anymore, just on the digest, but lately the digest isn't getting to my email very frequently. messages 370-374 for instance, I never received . So I am combining my responses here since I actually copied gerry's reply from the pupman.com list archive.

Dr. R,
thanks for the advice with the gloves, I probably would have forgotten them but it is a good idea and if I do try this I will remember to wear gloves. I also plan on doing this in a fairly clean environment, but I don't really have such an environment available so I am not really sure how that will pan out. I was planning on large overlap, and multiple layers of each. Some suggest using oil-soaked paper layers between the polyprop and the foil, but still others suggest leaving it dry altogether. I would probably err on the side of oiling them and will let you know if I do go that route and need the oil.

Would still like to know the part number of the maxwell's that you use, as well as the configuration, do you use 2 or more in series or just parallel them individually?

-

Also I saw the post about the 940C20P47K which noted that the rating of it was 960V/uS DV/Dt, 451A Ipk, and 13A Irms. the only difference really is the DV/Dt versus the 942C20P15K which has 2879V/uS DV/Dt, 432A Ipk and 13.5A Ipk. I was wondering if 20 of the 941C20W1K (1 uF, 2000 v) which are rated as 754 V/uS DV/Dt, 754 A Ipk, and 19.4 A Irms in series would work... they are $10.10 at newark.com, which would be less in total than 100 of the 942C's and slightly higher in every rating other than the DV/Dt. is the DV/Dt number significant at tesla coil frequencies?

I am guessing the difference is that between a full voltage being reached in discharge or charge, within 5 microseconds for the 942C's or 15 microseconds for the 940C's. the "total energy transfer time" is 80 microseconds, it doesn't matter that the DV/Dt is slightly slower for the 940C's since it is still 6 times faster than it needs to be in the voltage rise time. Am I way off my rocker and misunderstanding that rating, or are the higher capacitance 2000V 940C's (which have the same Ipk and IRms as the 942C's) plenty good for the job?

Or is the fact that there are more strings for the 942C's more important than the individual string ratings?

Thanks again for your help. I tried to address any questions I saw in any of the responses to my original mail, but if I missed any, please reply to me directly at edgarsbat at hotmail d o t commercial (I am sure you can figure out my email address from that.)
-NICK


-----------------------------------------------
Original poster: "Gerry  Reynolds" <gerryreynolds@xxxxxxxxxxxxx>

Hi Nick,


25KVA Wow!! As for the 940 caps, their peak current rating is not as good as 942 caps and you will need to make sure this spec is not violated else you'll blow the connections out. Maxwells are probably a better choice for this power level but others can give better advise on which ones are good and what stresses to avoid. One thing I can say is if you run STR you better have a very high BPS otherwise the voltage will become excessive. If you give me your BPS and Cp value and at what KVA level you ballasted your PIG, I will be happy to spice it for you. Say more on what the conditions were where your caps blew: type of SG, BPS, ballast point, Cp, number of maxwells in a string and number of 940's in a string, and capacitance per string.
Gerry R


Original poster: "seanick" <edgarsbat@xxxxxxxxxxx>


I have now toasted one of my two maxwell's, as well as one of my four strands of CDE 940C's. In the interest of not blowing up too many more capacitors, I am trying to do my research this time before ordering their replacements (and not using any of the remaining caps, for this coil anyway).
unfortunately finding a few roadblocks to that end:

1. I have a possible source of doorknob caps, in the form of another pacific northwest coiler who I recently bought a pole pig from, who is going to try to acquire some soon. However, I have heard doorknob caps might not be as effective, not to mention they are going to be HEAVY to get enough of them for a 25KVA pig powered coil, which is using a 120BPS synch rotary. 2. Have a rell.com shopping cart, containing 90x CDE 942C's (the -F version, which they have in stock, not the non -F which has a lead time of weeks). however, that is a $300 order that will still only have the same voltage rating as the 940C's which failed QUICKLY when I used them last week. A safer voltage rating would require more caps, and would still be less than the resonant cap voltage by about 1/2. .06 uF is what the 2 37667's I was using added up to, with great results, so that is what I was aiming for this time. LTR according to JavaTC is something closer to .48 uF (hmm... $2400 worth of 942C's, even if I only use strands of 15! I think I am going to pass on that option). 3. Maxwells- I would love to find out more about Dr. Resonance's better Maxwell caps, than the 37667's which are known to fail easily in a 25kva pig powered coil. (Between Aaron and I we have fried 2 of the 3 maxwells that I have ever seen in my life, both within the last month or two. ) Unfortunately all I can find is part numbers and ratings, nobody selling anything other than the 35 kv .03 uF 37667's, and even those are $225 where I have seen them. Does anyone know of a current source for better maxwell's, and/or have other recommendations in this area? 4. Any other options? I haven't found any that don't involve a custom order from CSI (which doesn't seem to take me seriously, judging from their lack of a response to my recent request for quote). I didn't even mention that I would be using it for a hobby!

So, in lieu of having many other options, I am starting to lean in the direction of building my own. I have tried this in the past but didn't know what I was doing, at all, last time. This time I am a bit more educated, prepared, and have a more impending need now that everything else in my coil seems stable including for the first time ever a synchronous rotary. For building this cap, I have a vacuum pump that will pull 29+ inches of mercury, and a t-shirt press which may help between the two of them to remove air bubbles. I was thinking that if I ordered a roll each of these two, I might be able to make enough caps that are both fairly efficient, and high capacitance/ voltage, that I could replace the maxwell/ MMC with something not too much larger, in the overall scheme of things. <http://www.papermart.com/Templates/47-0-40.htm>http://www.papermart.com/Templates/47-0-40.htm (a 100' roll of 30" wide metallic .8 mil polypropylene sheet) <http://www.papermart.com/templates/47-0-15.htm>http://www.papermart.com/templates/47-0-15.htm (a 500' roll of 30" wide clear polypropylene 1.2 mil sheet) (multiple layers of, between each layer of the above.)

I figure if I do a careful job in the overlapping, and use some mineral oil or something to pot it once the whole thing is rolled up in a tube and wires attached, I would be able to make it fairly robust. I even thought about making it in a sandblasting chamber that was completely cleaned out, and had an argon tank hooked up to it or something to purge the air (and corresponding particulate) out of the chamber after putting the parts in it, then sealing it before working on assembly. but that is way overkill considering it wouldn't even fully do the job- better just to use multiple layers of clear polyprop between each metallized layer, and leave large areas of clear overlapping all 3 edges of the metallized, and most of all removing every single air bubble by pulling a full vacuum for a few days after filling the tube with oil. then maybe a little more oil, an over-pressure safety valve for when the oil gets hot enough to raise the pressure inside it, and it should be decent enough for a run or two of the coil, right?

Definitely has to be better and smaller than a fishtank with 2 half-racks of saltwater filled/ motor oil covered corona beer bottles in it.

please give me feedback if you think I am nuts or could do this better another way or using something off the shelf. Regardless of what option I choose, I only budgeted myself about $400 to spend on replacement caps and I am hoping to use them more than once. I also don't feel like waiting more than a couple weeks for something which is on backorder so that rules out at least one option.
thanks in advance,
NICK