Rare HV to LV converter / High Voltage / Forums

Forums

4hv.org :: Forums :: High Voltage

« Previous topic | Next topic »

Rare HV to LV converter

1 2 3 4 next

Move Thread

LAN_403

GammaRay

Tue Apr 01 2014, 01:23PM

Registered Member #5323 Joined: Fri Jun 15 2012, 02:14PM
Location:
Posts: 104

I have exhaustively searched all my usual vendors for a HV to LV switching converter to use in conjunction with these large caps. Expended a lot of time in the search before approaching the collective wisdom of this forum.

The converter needs to accept input in the 10kVdc-20kVdc range and provide one 12vdc output and one 120vac. Or, it could be two independent converters with one converter providing a 12vdc output and the other converter providing a 120vac output. The extremely high input voltage makes this converter somewhat exotic. Assistance in directing me toward a vendor (or schematic) for this rare type converter will be appreciated. Thank you in advance.

Artlav

Tue Apr 01 2014, 02:32PM

Registered Member #8120 Joined: Thu Nov 15 2012, 06:06PM
Location: Moscow, Russia
Posts: 94

At what power level?
You are talking about something that is only used in HVDC transmission sort of applications, with power levels of megawatts, would be hard to come by in small scale in open sale, i think.

If you want to make one yourself, then all it would take is basically a 20KV rated switch or two, and a transformer.

If you only need a few dozen watt, then you can use a couple of cheap high voltage triodes, like

as switches. Make a push-pull converter out of them, connect a TV flyback backwards to it, and you should get 12VDC on the output, give or take a few complications and x-ray shielding.

For more you'll need SCRs or IGBTs of sufficient rating with carefully designed switch timing to make them work in series. No idea how to do that.

Mads Barnkob

Tue Apr 01 2014, 03:03PM

Registered Member #1403 Joined: Tue Mar 18 2008, 06:05PM
Location: Denmark, Odense C
Posts: 1968

What kind of application do you need this HV to LV converter for? Maybe someone know a alternative to your current idea.

Steve Conner

Tue Apr 01 2014, 03:11PM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

You bet it's exotic. It would also be very large and expensive, and stray capacitances would make it inefficient too.

Dr. Slack

Tue Apr 01 2014, 03:38PM

Registered Member #72 Joined: Thu Feb 09 2006, 08:29AM
Location: UK St. Albans
Posts: 1659

The power level is the key. If it's only to tickle some instrumentation at a watt or so, then option 1 is to buy a small off-line switcher that will accept 380v DC on the input, and drop to that with a watty resistor, protecting against overvolt with some sort of zener arrangement.

If you need MOAR POWA, then you're making another exciting side project that you probably don't need to.

You could drive an electrostatic motor with the HVDC, and couple it to a low voltage generator. But that's only going to be low power as well.

If you need isolated power at your caps, then rechargable batteries are hard to beat for convenience and isolation. What about a long nylon shaft, or bit of PVC plumbing (with end fittings!) connecting an ordinary mains motor, and a generator on your kit?

Converter topologies that allow you large amounts of leakage inductance, for instance for near-field charging, could allow you to put a thick sheet, or several thin sheets, of insulator between primary and secondary, but you have an obvious tradeoff between safety, thickness, actual safety, power levels, perceived safety and efficiency. Did I mention safety?

If you really want to make it a project, then the first thing that springs to mind is a sort of inverse TC, perhaps with an RSG to connect the HT to a high voltage resonant circuit, using an inductor that looks like a TC secondary, with a few 100pFs or even nFs discrete capacitor instead of a topload to get the frequency down, loosly coupled to an output circuit with much more C and less L, perhaps switched with HV FETs or IGBTs, to output at 100s of volts. Then a second stage using a commercial off-line line switcher as for option 1 above. Good luck.

2Spoons

Tue Apr 01 2014, 09:22PM

Registered Member #2939 Joined: Fri Jun 25 2010, 04:25AM
Location:
Posts: 615

Left field idea: inverse marx bank. Replace all spark gaps with controlled switches so you can charge all caps in series (switches closed) then discharge in parallel (switches open). That could get you down to more manageable voltages where standard topologies become viable. No i don't have a schematic - this is just a wild thought that popped into my head when i saw your query.

GammaRay

Wed Apr 02 2014, 05:32AM

Registered Member #5323 Joined: Fri Jun 15 2012, 02:14PM
Location:
Posts: 104

I realize efficiency isn't going to be great with any system converting extremely hv to lv, however, to answer some questions, I seek as much lv wattage as possible at the output. I'd like to power some small devices for a few minutes one at a time, laptop, lamp, fan, fuel cell to make hydrogen, etc mainly for the fun of it. Someone might reply, "just plug them into the wall outlet"... but that wouldn't be much fun and certainly no challenge. Theoretically, this cap bank can hold over 63,000 Joules when fully charged, that should be enough energy to power a few small devices if stepped down properly.

I've used a couple Dale 2M 40kV 35 watt resistors (below) when lighting 32 feet of florescent tubes but realize much of the energy is being wasted as heat.

The idea mentioned of using a 20kV switch and transformer sounds interesting and fairly inexpensive... could it be that simple? Pros/cons of using that technique?

1396416466 5323 FT162126 Hv Resistor A 35

Ash Small

Wed Apr 02 2014, 10:16AM

Registered Member #3414 Joined: Sun Nov 14 2010, 05:05PM
Location: UK
Posts: 4245

GammaRay wrote ...

The idea mentioned of using a 20kV switch and transformer sounds interesting and fairly inexpensive... could it be that simple? Pros/cons of using that technique?

I'm just 'bouncing' ideas, but if you connect a 'high value' resistor (maybe a water one) to the capacitors, then connect a 'low value' capacitor to the high value resistor, then discharge the low value capacitor into the primary of a step down flyback transformer using a high voltage triode, then charge a high value, low voltage capacitor with the output of the secondary, you have usable power.

I assume you'd probably need some X-ray shielding around the triode.

Mads Barnkob

Wed Apr 02 2014, 10:43AM

Registered Member #1403 Joined: Tue Mar 18 2008, 06:05PM
Location: Denmark, Odense C
Posts: 1968

Series connection of switches is properly the way you want to go.

Series IGBT:

More ideas:

Gate Commutated Thyristor:

IGDT?:

Avalanching transistors used for triggering, just for inspiration:

and

Dr. Slack

Wed Apr 02 2014, 12:36PM

Registered Member #72 Joined: Thu Feb 09 2006, 08:29AM
Location: UK St. Albans
Posts: 1659

OK, so here's the deal. This is an inverse rotary Marx.I haven't nutted-out all the issues yet, but I don't see why it shouldn't be made to work in some fashion. I can't be bothered to model it, that would show whether it could work.

It is a physically large insulating disc, which carries a large number of uniformly spaced conducting studs on the periphery, which have suitable dimensions, tolerance and material to make contact with several fixed electrodes. Between each pair of terminals is a bipolar capacitor, of say 500v or thereabouts, perhaps a film motor-run cap, only a few of which are shown in the diagram. There have to be a few more than 3x as many caps as the HV/LV ratio.

The broad-brush principle is that the caps are charged in series from the HV supply, and discharged individually at low voltage. There is some devil in the detail. Some of which I will address, some of which is left as an exercise for the reader, and some of which may be just plain wrong and torpedo the idea below the waterline. That's a risk I take with premature publication, be kind in your flames.

Let's ignore how it reaches steady state, just assume that it's got there.

0) Steady state. The HV and ground contacts are in contact, and define the upper and lower voltages. The down-side caps are all equally charged to HV/n volts, let's call it 500v. The upside caps are charged traingularly with c0 uncharged, c2n-1 at nearly 500v, and all the caps in between to a linear interpolation of these voltages.

The following things happen in order

1) The HV electrode withdraws (assume positive for this discussion, though it needn't be), and the wheel turns one stud in the direction of rotation. Now C3n-1 has its positive terminal grounded, so the terminal with C0 is now at -500v with respect to ground.

2) C3n-1 is now discharged to 0v by a diode-stopped inductor into a +ve DC output bus, say +300v with respect to ground. If you are not familiar with this principle, then look it up elsewhere. Suffice it to say that it empties the cap completely of charge, with efficency of nearly 1, and allows the DC bus to be any practical voltage, find something on buck/boost converters to get a hint at the principle.

Note that all the other 3n caps will have their charge adjusted slightly by this process as their series string is in parallel with C3n-1. The delta charge will be small, and the effect should simply be a slight increase in the apparent capacitance of C3n-1. The HV terminal is not in contact, so the nodal voltages can do what they want.

3) The same thing has happened to cap C2n-1. It has moved from the up-side to the down-side, so is now charged in the 'wrong' direction. Its charge is reversed by thumping a diode stopped inductor across it. This process is lossless, in the limit of a low resistance inductor. This can be floating with respect to the HV terminal. The same arguments apply that small adjustments happen elsewhere in the string of caps. Now all down-side caps are charged in the same direction, and all their voltages add up in series.

Note that (2) and (3) can happen in either order, or simultaneously, as they are independent processes.

4) The situation is now that a 500v cap has been removed from the upside, and replaced with an uncharged one, neglecting what has happened on the down-side. This is the bit of detail I hope someone else can fill in, by modelling, or with more detailed reasoning. I feel that as capacitors are lossless, it shouldn't matter too much exactly what happens in detail. It looks like the downside is charged as it was, having lost and gained a cap each charged to 500v. However, it's in parallel with the upside string of half the capacity (twice the length) which has lost 500v. Therefore my first order approximation says that once these two things have happened, the stud next to contact the HV terminal is at HV-167v.

5) Finally, the HV terminal makes contact with the next stud. The potential difference of 167v is small enough with respect to the total voltage that a series resistor to limit the current will cause very little efficiency hit, it's really not worth using an inductor. Both sides of the string will charge up.

Conservation calculations

If the disc houses 3n caps of capacity C, then the input capacitance at the HV terminal is 3C/2n, so the charge that flows on each stud turn is LV/3 (the volts) * 3C/2n (the farads) = C.LV/2n. As all of this flows at a potential of HV, so the energy absorbed from the HV supply is C.HV.LV/2n.

The charge delivered at the bottom end is just C.LV, for each stud. However, the action of the inductor means that the mean voltage it's delivered at is LV/2, so the energy out is C.LV.LV/2.

If we now write n = HV/LV, we see that energy in = energy out, which is good, there's no obvious foul-up.

I wouldn't mind somebody checking this, and perhaps doing it another way as well.

Practicalities (like if you've got this far you should worry)

The sequences of connection and disconnection may be a little difficult to acheive at both HV and ground terminals with static electrodes. It might need synchronised spinning things, with the sort of gearing you find in a sewing machine. But you can have that in thought experiments.

Maybe that as once steady state has been acheived, potential differences are so low that ordinary small-gap switches can be used without the sort of arc extinugishing problems that you get with RSGs.

Background

Incidentally, this downconverter idea didn't spring up directly as a result of this thread, I've always entertained the idea of a rotary Marx upconverter as being a novel way to get HV. This is where the caps would get charged as the wheel turned, until something broke over from top to ground. I don't recall ever seeing a machine like this described, probably for good reason. The upconverter would be far simpler than its inverse to make, not needing the inductors or such complicated switching. Its operation would be charge the bottom, turn, repeat until bang.

1 2 3 4 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob