What are the limitations of this voltage multiplier?

[nsled]

Digging through a shed this come up, what would you do with it?

It's a 3 stage full wave multiplier, constructed with 9 10KV 0.1uF caps and 12 10KV 1A diodes. I just checked the prices and it's about $400US worth of components.

What would this even be good for and what limitations does it have? I've looked at multipliers on youtube but they are always constructed with much smaller components (although this forum category is opening my eyes).

Am I right in thinking this will be limited on the frequency it can be used at due to larger caps? What advantage do larger caps hold in this configuration? Does 10KV mean I can hook a 10KV neon transformer (for a bit shy of 60KV?) up to it or does the rating mean you should leave a margin below it?

[MRMILSTAR]

The 10 KV capacitor rating means that the input AC power supply probably shouldn't exceed about 5 KV peak with a 50% margin. Larger capacitors allow for a lower charging frequency power supply. Since you know the components, plug the values into this CW multiplier design tool and see what it can do. The main specs are input voltage, output voltage, voltage droop, output current, and charging frequency.

https://www.extremeelectronics.co.uk/calculators/cw-voltage-calculator/

[nsled]

Ok thankyou, but while the input is AC isn't effectively being used in DC "mode" in a multiplier? Or is that a passing comment and a 50% margin should be used for DC?


I'm not sure what the calculator is telling me, for example if I put in

3 stage
50Hz
0.1uF
0.1mA?
2000V

It's telling me the output is 16.8KV?

When I thought it's meant to follow:
2000>4000>6000>8000

Why is it twice the expected?

I'll assume the ripple etc values are incorrect since the calculator is for a half wave mp.

[Luca c.]

Hi, calculating the voltage of a CW multiplier is harder than it seems, on the internet you can find different formulas, the most common is that Vout=n*Vin (n number of stages) that you can find on Wikipedia, as is stated there in the formula Vin is voltage Peak to Peak.
So with Vin=2000v we get 6000v out, but it is important to note that the voltage here is measured as Vp-p, if you use the Vp instead the formula is Vout=2*n*Vin.
The site suggested by MRMILSTAR (that works really well, and I have confirmed experimentally) uses Vrms, which is very different, Vrms=Vp/1,41, so if you use this instead the formula becomes Vout=Vrsm*2*1,41*n and if you use this with Vrms=2000v you get 16920v, that is larger than the stated 16,8kV this is because in reality the site even counts for losses, a good formula for keeping in mind losses is the following:

from https://josepheoff.github.io/posts/diode-capacitors-volts-pt7
where Vf is the forward voltage of your diodes.
Another option is to disassemble and reconfigure 8 of the 9 capacitors and part of the diode to get 4 stages and a larger voltage output.
The margin is a question of how much you don't want to destroy your equipment, if you don't want to risk damaging it, it is better to keep the Vp of the source to half of the rated voltage of the caps, or a percentage as high as your availability to break the capacitor or diode. In this case, the capacitor rating is DC but it doesn't mean the multiplier is using DC, you certainly can push 10kv ac and it can work for some time, but you would never know if that one would be his last spark.
Lastly, if you want to do only big sparks it is ok to use a 50Hz source, but as you can see the ripple is immense (even for a full wave multiplier), so if you need to use it for something more precise, I suggest using a flyback transformer with a frequency in the range of KHz, with that big capacitor you could get a really low ripple voltage output.

[nsled]

Ok thankyou for the explanation.

I did find a comment that supports MrMilstars reply about the voltage drop for 50 Hz and giving example of the same 3 stage 0.1uF caps but tested at 9v:

"I did many of my experiments using 100nF capacitors at 50Hz. I found that the output begins to "sag" quite badly with just 3 stages. The voltage when using just the "load" of my oscilloscope (1MOhm) was noticeably lower than when measuring the output voltage with a voltmeter (10MOhm.)

To get usable voltage multiplier at 50Hz, you'll probably want to go with 1µF capacitors. The alternative is to use a higher frequency."

From https://electronics.stackexchange.com/questions/510390/voltage-multiplier-does-not-work

[nsled]

Reading more, here is a good article that gives measurements comparing 50Hz with 25KHz into a half wave multiplier using 0.1uF capacitors.

https://ijarcce.com/upload/2016/june-16/IJARCCE%2050.pdf

In summary, the 25KHz signal reached a stable high voltage very quickly, (under 100ms) the 50Hz signal was much slower (4 stage still rising after 4sec) and didn't reach full voltage.

3 Stage
25Hz: 60ms to reach 3KV
50Hz: 1.6s to reach 2.8KV

I guess the question I have is what practical effect this would have in the real world when charging a capacitor? Would it just take 1.6 seconds longer to charge to 200v lower? Or does adding a load to the multiplier make things a lot worse for the 50Hz?

[klugesmith]

...does adding a load to the multiplier make things a lot worse for the 50Hz?

Depends on the load. 
For charging a relatively large capacitor, would you consider 500 times slower to be "a lot worse"?

Without simulation or fancy analysis, you could get a ballpark estimate by figuring the charge stored in one VM capacitor at its working voltage.
What's the current if that much charge is transferred 50 times per second, and 25000 times per second?

[nsled]

So instead of the 10kVDC 0.1uF capacitors couldn't one just use the 2.1kVAC 1uF microwave capacitors to make a 4 or 5 stage multiplier fed from the MOT? I've been looking for examples of this but have come up with nothing even though it seems logical. I realise its purpose is to act as a doubler in microwaves but why can't it be pushed further?.