CW multiplier

[Jake Wells]

In my last post I have discussed the use of microwave capacitors. I was told that they would not work. If they do not work then why is everybody using them?
Jake Hecla power supply


Robert Tubbs Power supply
http://www.youtube.com/watch?v=PtqRrVTjHCg

Will Jack power supply

[Dennis P Brown]

Simple electronics 101 - capacitors in series add in their voltage tolerance; so yes, two 8 kV caps in series will withstand 16 kV, and so on to fusor voltage levels of 35 kV (i.e. four in series as long as you ignore RMS voltage issues - which you can't.) That said, they are low voltage compared to fusor voltage levels and should not be used except using precautions. Since these caps were designed for 'only' 8 kV, and one has to be very careful for shorts between adjacent caps and corona issues - I understand that most people that use these low voltage rated caps put them under oil for higher voltage use. Very critically, if one shorts out internally, they all will go and it can get violent with these very high voltages. Since these caps were not designed to handle those voltages there is no inhert safety margins like 35 kV caps would have. So I'd think that these are not the correct caps for amateurs. I would also speculate that for most people these issues would not be well understood/appreciated so it would not be good to uses these in fusors unless these and other issues are carefully addressed.

Also, in series, the capacitance goes down so you will get less output current. As in the circuit drawing you are showing, the voltage will build up using that simple voltage multiplier and if you have enough caps, some acceptable current level will be obtained for part of the cycle. However, there are no free lunches - the total power of the multiplier stack is somewhat less than the NST alone but you do get the voltage increase you need at the cost of some current performance.

[Jake Wells]

these caps were only designed for 2kv. these capacitors were designed for 8kv http://www.ebay.com/itm/NWL-S00686-30uf ... 23114b18b7. could I align them in 2 series powered by a MOT

[Dennis P Brown]

I am no electronics expert - I am building a 400 kV multiplier stack using 20 kV caps/30 kV diodes BUT I do not need any significant current (2-3 micro-amps) for my application. Also, they will be under oil in a special case and if a failure occurs, I can accept the total burn up of the system. If you want to use 2 kV, that can be done as can 8 kV caps - the issue is building the voltage multiplier correctly - use an internet site to calculate the capacitance you will need as well as the diodes (max voltage, response time performance, and current tolerance.) Also, RMS performance is critical if you want to avoid failure.
If you need more current - either get bigger capacitance caps or use them in parallel. Again, be careful since these caps are not designed for high voltages and precautions must be observed.

[Rich Feldman]

In my last post I have discussed the use of microwave capacitors. ... If they do not work then why is everybody using them?

1. They're fine for demo fusors, though why bother? I bet the designs you pointed to are early steps by people who went on to better HV supplies before getting neutron counts. Name your target voltage and current, then show us a design calculated to meet them using MOC's. You can't keep putting off doing a bit of electrical engineering.

2. Could we say microwave _oven_ capacitors? My experience with microwave capacitors relates to parts like these: http://www.mouser.com/new/Kemet-Electro ... tcbrmlccs/ . That is to say, capacitors rated and characterized for performance at frequencies higher than 300 MHz. MOT's are not the same as microwave transformers.

3. Thanks for capitalizing and punctuating your most recent post it really does make it easier to read more readers will take you seriously now

[Dennis P Brown]

You should check out Chris' post on using two or more transformers and voltage multipliers to get both high fusor voltage and the required current (see the previous post and Chris' answer to my question.) The design is ingenious and is a great way to use NST's to power a real fusor.

[prestonbarrows]

You need to be careful when putting multiple capacitors in series in order to increase the overall voltage rating. Caps typically have very loose tolerances, +/- 10 or 20% or worse is pretty common for the bigger/cheaper ones.

I = C*(dV/dt). In series, the current through each cap is equal. This means that if current is flowing, the range of actual capacitance for each cap due to manufacturing tolerances leads to uneven changes in voltages over time for each cap. The total voltage across the stack is not equally shared by each cap, those with smaller capacitance will take a higher voltage drop while dis/charging.To be safe, you have to de-rate the full stack so that even caps at the worst end of the tolerance spectrum will still never drop more than their rated voltage across them.

In general, if you are stacking many non-linear components like diodes or loose-tolerance components like big caps, it is best to have a resistor divider in parallel to clamp each component to a solid voltage reference. High precision resistors are comparatively cheap. This will add losses through drain currents though.

[Rich Feldman]

More on Preston's point.
Microwave oven capacitors have built-in bleeder resistors,
which could have widely varying values in a mixed lot (even if nameplate C values match).
In a series string, the DC voltages will ultimately divide in proportion to the R's, not to the 1/C's.

In voltage multiplier applications, or even the original MO voltage doubler,
DC voltage is what the capacitors are there for.
AC voltage (across a capacitor) just represents ripple and voltage drop.