Yep, somewhere about in that range

??

isnt 10^-6atm = ~ 0.00076000000012 torr ??

thats pretty far away from 10^-6 torr, no ?


wiring strings of unmatched diodes in parallel will likely have result of one string.

 

I do know that diodes in parallel have problems when they are unmatched. One heats up, which causes it to draw more current, which heats it up more and so on.

The thing is, these diodes will only experience an extremely short pulse of electricity, so as Lon as they all start at the same temperature, I don't think they will have time to heat up, so the only difference in current would be given by their difference in voltage drop, and while that alone may cause some mayhem, it is only one pulse.

 

two strings of 20 unmatched diodes rated at Vf=1.5v each diode

20 x 1.5 = 30v

but, not every diode is exactly Vf, so lets say string-1 is off by -1% and string-2 is off by +1%, so now you have a diff between strings of 2%.
ok, you can run the math/physics from here.

its the same issue why we dont parallel two full bridge rectifiers, better to use bigger diodes, etc.

 

Perhaps a strong magnetic field applied along the long axis of the nested tubes could trigger a spark and the diodes can be dispensed with. Neodymium are cheap and plenty powerful. Just juice it up and pass the magnets by. I believe this would depend on the voltage gradient and distance between electrodes putting them "close" to the insulation breakdown point already? Any defect in the principal I'm suggesting but don't see? I think it could be worth trying to rid the circuit of the weak link diodes.

 

I'm not sure what you mean, are you saying that 2% is too much difference or too little?

Nevertheless I do know that for continuous current, the diodes will not be enough, but what about a pulse? Imagine say, in a power supply, you have two parallel diodes, but you only turn that power supply on for a few milliseconds. I don't think the diodes would have enough time to heat up that destruction starts to happen.

Anyways the idea behind having 20 parallel diodes was that at least a couple of them share equal amounts of current.

 

I'm not sure what you mean, are you saying that 2% is too much difference or too little?

well if string-1 does Vf at say 29.7v and string-2 does Vf at 30.3v, do you see an issue with this?

i might suggest something like a model rocket engine igniters, put one in the center between the two electrodes, ignite it, watch cap discharge? even a very thin aluminum wire can work, thin enough to vaporize in the vicinity of the electrical field that exists between the two cap electrodes.

 

Oops, sorry Kermit2 I didn't see your answer.

How would that magnetic field thing work? To be honest, my knowledge doesn't go that far...

As you say, it would be amazing to be able to get rid of the diode, and that's exactly what I am trying to figure out.

 

DC Kid:

Nope? I guess... I'm not being rude I promise, I just don't have the experience / knowledge to know if that would pose a problem. I think that the difference wouldn't be large enough to cause destruction in such a short time.

 

when string-1 conducts before string-2, string-2 does nothing because the voltage across the strings will be lower than the Vf of string-2, hence no current in string-2.

 

Hmm you've sparked a thought train mentioning the rocket igniters and thin wire.

If I use a thin piece of wire to short out both electrodes inside the vacuum, so that the 320v cap discharges through the wire, if it is thin enough it might vaporize and form plasma, sparking a chain reaction which results in cap discharge!!

Brilliant! Many thanks for the idea, as it doesn't require diodes nor high voltages.

A question though. We are talking about a cap 5400uF at 320v.
How thin would the wire need to be? Could I use one single strand from regular stranded wire?

 

i would look for something thinner than that given the relatively low amount of joules in your cap.
maybe use another low voltage cap way bigger in uF and place a thin wire near your HV cap electrodes, use low V cap to burn up the thin wire.

get some ~28 or 30ga stranded wire and pull out one strand, this should burn up easily. you could also put some fine metal powder on the wire to assist in the process (literally just sprinkle metal powder on the wire then put wire into chamber).

its in a vacuum, so you could perhaps grab a tungsten filament from a ~12v incandescent light bulb and just use a benchtop power supply to get the filament very hot and the e-field from you HV cap should be able to find a path.

its just not clear to me if you are attempting to create a plasma arc for HV cap to follow, or just an electrical path. if you want a plasma then look to burning up a thin wire or some form of chemical igniter like model rocket stuff, etc.

 

The idea is that the electrodes in the chamber are at a distance, and I want to be able to discharge 320v across them, generating a plasma. The important bit is that the discharge is done through a plasma.

If I grab a very thin piece of wire, and use it to short (that is, no air/vacuum gap) both electrodes, and then discharge the 320v 5400uF cap across that wire, would it burn and generate a plasma? Thus discharging the rest through the plasma.

If I were to use a lower voltage, much higher capacitance cap to "vaporize" a thin piece of wire into plasma, would that allow me to discharge the 320v? Through said plasma.

 

well, you should read this
http://www.dtic.mil/dtic/tr/fulltext/u2/723107.pdf


but in general, yes, a thin wire that vaporizes quickly should create a local plasma field big enough for 320v to use as a path. if you short your cap with thin wire some of your cap energy is being used to vaporize the wire (longer paths means more energy to do such, etc). you can try it with one strand from say a stranded 30ga wire, which uses 44ga strands to make the 30ga wire. use a strand in the 44-47ga range, see what happens.

hence to see all your HV cap energy cross in a plasma, burn up the wire without shorting it to your cap. a 44ga strand plugged directly into USA 120vac 60Hz will vaporize very fast, etc.

 

Testing time!

Thanks for the document, pretty helpful data in it. Especially since I do quite a few things in vacuum.

 

I have no idea how to quote so I'll instead paste the message:

"hence to see all your HV cap energy cross in a plasma, burn up the wire without shorting it to your cap. a 44ga strand plugged directly into USA 120vac 60Hz will vaporize very fast, etc."

There are various options for burning the wire without discharging the main cap, for example, the one you mention. However, for protection, both the cap and the "ignition" circuit must be isolated from each other, so that the cap doesn't mess up the ignition, and vice versa. Hence the diode problem. Although, it might be easier to solve now that the wire doesn't need high voltage to burn, which reduces the risk of overvoltaging the main 450v cap.

(As an unrelated idea, what if instead of a thin wire, I tried a small, low current fuse, like the ones that come inside a glass tube, but smashing the glass so that it acts as a bare conductor?)

 

the thickness of a fuse is fairly thick. wire and fuse are both "wires" and "fuses", etc. i think a thin wire will vaporize before a fuse of any size will.

to vaporize the wire, just use a separate power supply and do not connect the wire to the cap. however, the same plasma could also allow some amps to flow from this seperate supply, hence why i did also suggest another cap with low voltage but higher uF, so it is more like a pulse to use up the energy, and thats it.

what exactly are you studying here with this cap in vacuum setup?

 

The problem is, there are only two electrodes inside the vacuum chamber, and the cap must be connected across them, so that it discharges when a plasma is formed by the thin wire. To burn said wire, it would need to be connected to those same electrodes.

What I am tying to study is the thermal energy characteristics of plasmas. One of the experiments for example, involves discharging x amount of joules into a plasma and measuring how that changes the temperature of a fixed mass of metal in the vicinity.

 

can you have more stuff in the chamber?

a fine hair filament shorting the cap is a good 1st try to see what happens.
if it is attached to one electrode and long enough to reach the other, the efield might be strong enough to just pull the end into the electrode and short.

 

I might be able to squeeze in a third electrode, which would simplify the problem quite a bit, with the ignition and the cap only having common ground. However, due to some bad design choices on my part, it will require quite a lot of work. Not to say its not doable, but other solutions would be prefferable.

I ran some tests today with very thin wire. (Not sure which AWG but I got it from a piece of stranded wire which was allready very thin itself.)
In a separate test chamber, which has 4 electrodes, I connected the wire directly between two electrodes, and plugged it to mains. (230v here in Spain). The breaker instantly popped, as expected, and the thin wire was nowhere to be seen, I suppose vaporised. I tried filming it in slo-mo, but the flash prevents me from seeing anything.

Taking that as a succes, I tried to put the wire inside the actuall vacuum chamber, shorting both pipes. I charged the main capacitor to about 230v, DC this time, and when I was closing the chamber I accidentally struck the glas tube against the table and dented it. (SH*******T!) Anyways, untill I get another glass tube I won't be able to test anything in that vacuum chamber. I still have the other one tho, and the tube shouldn't take too long to arrive.

One of the things I am trying to work out now, is burning the wire without popping the breaker. I plan on using a separate capacitor, which recieves rectified mains, and charges to 320v. Then discharges on the wire, without being plugged in. I haven't tried yet, but I suppose that, say, 100uF at 320v will be able to vaporise the wire?

I also have to test it with the full 5400uF cap. But I need to figure out a way to connect the capacitor to the wire when it is charged. Maybe a big manua switch? Or a badass IGBT? (I think I have some laying arround capable of continuous currents of 60A)

 

well, 230v made the plasma from the wire and lots of amps flowed from the mains. maybe put a fuse of appropriate size inline with the 230v (one just smaller than your mains breaker) so that fuse blows and not breaker. or, use a 12v car batt instead (or two batts in series for 24vdc). or try the chemical igniter method.

and, we really need to know what diameter of wire you used, any caliper will suffice to measure it.

the arc from 230v should have died away at zero crossing (AC) and extinguished itself. but it sounds like the arc stayed long enough to trip your mains breaker (many AC cycles, etc). what low pressure was it during this test?