Homemade high voltage secondaries
Commercial high voltage coils, especially those from TV, have high turns ratio, which makes dozen of volts enough to induce tens of kilovolts, but remind that coils aren't supposed to work in such high voltages. The most common consequence are inter-layer ruptures, arcing to core or primary coil. Those problems can be minimalised via i.e. immersing coil in transformer oil, connecting core and layers that are most close to core, to ground (thanks to that, those layers don't arc to core), but other problems remain. It made me to decide to try and build own coil, which will work better in those hard conditions, better than those commercial ones
. My target - 20kV RMS and dozen of miliamps. Will I achieve it? Read.. ;-))
On 4hv.org forum about high voltages and other similiar subjects I've found an interesting topic, where person wanted to made an homemade high power and high voltage transformer. He was using this scheme:
The scheme is simple, two coils are connected in series and placed on core in a specific way, so they have -10kV and +10kV potentials. Potential difference is 20kV. A connection of two secondaries was grounded. There I've done a little modification, I connected core to ground. Layers that were close to core, were grounded also. Those connections solves problem with arcing to core - between core and first layers of secondarie's potential different is too small to rupture. But three problems remain. Primo, between primary and -10kV or +10kV a arcing can occur. Secundo, outer layers can arc themselves, 20kV can easily ionise in air at quite length. Thirdo, -10kV and +10kV can arc to core (those first layers won't do, but outer, yes). Then, what do we do in order to evade those problems? Have big core and good isolation ;) I've bought high voltage transformer on stock-exchange for 5 euro and it came from colour Sony TV and had internal diode. The core's size is bigger than majority of TV HV transfrormers. A insulation matter remained. After few days of reading discussing forums, I've came to clue that I might try presspaper and epoxy resing Epidian 53 with hardener Z1. Presspaper, as I've learned out later, doesn't have a good dielectric parameters. But! We can impregnate it with transformer oil. Presspaper has quite good soaking attribute, so I think this solution is a great idea - presspaper holds a big amount of oil, which has a very high dielectric strength - pure 1mm of oil is enough to hold 30-40kV (although, I don't know at which frequency - the high frequency is, lower dielectric strength is). After preliminary calculations, I've done a first coil. Construction in really simple - cut presspaper to proper width, wrap few times around core, glue with superglue, wind emmanelled wire (there I've used 0.15mm fi one) and keep distance between wire and edges of presspaper layer - you have to keep distance from core, so it won't arc. I had 5mm distance and it was enoguh for middle supply voltage (ZVS flyback driver and 30-40V). My fist secondary has 6 layers of wire, each has average 35mm width. 35mm divided by 0.15mm gives 230 coils, so the whole coil has 6x230 = 1380 coils. Of course, it's only a estimating. With primary consisting 5 coils, ratio is equal to 1380:5 = 275:1. If we'll apply 30V with high frequency, in output I'd have 275*30V=8.3kV. With second secondary I'd twice more, 2x8.3kV=16.6kV. If we increase supply voltage a bit, we will achieve 20kV RMS :) A photo of ready secondary and second secondary in process:
As bonus, I've uploaded a movie, where with autotransformer and ZVS flyback driver I'm testing slowly, if internal arcing will occur. Oh well, sink or swim! I must say that test ended with big success :) (be aware that there isn't sound)
Perhaps we should toast a sausage? Download it [around 10MB]
Actually second coil is prepared and I'm confronting
with problem: too small space between secondaries. It would be best to use biggest core with a lot of space for secondaries. Unfortunately, from what I've noticed, the thing about ferrite cores is that there's proportional link between a total size and core's diameter. They're just fat
. The biggest high quality class Ferroxcube cores are expensive and there isn't much more space for secondaries. At this moment, I have two options: try to mount two coils on ferrite and optionally immerse it in oil (it has big chance to work and i will probably do that if second option fail). Second option is to find second identical core, or the one with same diameter, so I would put coils, and then make arrangement like a picture shows below:
Projecd made in AutoCAD 2007
For me, it has alot of advantages: the maximum power doubles - electromagnetic field divides to two halfes, each for each core. There is a lot of space between secondaries and primery. I think there's only one danger I see: arcing to core by high voltage leads. But it's easy to solve - just keeping a proper distance between wire and core or immersing in oil is enough. Unfortunately, there's a small chance to find identical core, but I will try. To be continued.. ;-)