Help with rf/emi protection for QUAD MOTs (terry filter) / Tesla Coils / Forums

Forums

4hv.org :: Forums :: Tesla Coils

« Previous topic | Next topic »

Help with rf/emi protection for QUAD MOTs (terry filter)

Move Thread

LAN_403

Vigh holtage

Fri Dec 15 2017, 03:37PM

Registered Member #61739 Joined: Wed Aug 23 2017, 04:43PM
Location:
Posts: 44

Ok ok i know how to makr one for a small low current nst but what about a verry high current quad mot stack ????

Could i use 2 or more 1k ohm 150watt power resistors in parrallel to get the wattage up ? And what about capasitor to ground capaditance rating, s?? 500pF still good juzt go with large 30kv ceramic caps instead of a seriers string ? Should i leave of the mov's ???

Should i add a choke to the filter ? Lcr ? Or just rc filter?

I would like to protect my mots as seeing they are under 3gallons of oil and would be a pain to fix

profdc9

Fri Dec 15 2017, 10:47PM

Registered Member #58522 Joined: Tue Mar 15 2016, 08:33PM
Location:
Posts: 50

Here's my experience. I made a Terry filter for my 2 MOT power supply that uses a voltage doubler and it seems to be holding up pretty well:

I used 2200 pF 6.2 kV ceramic, 3 in series, 3 in parallel to make an effective 2200 pF by 18.6 kV, two sets each side to RF ground

560 ohm, 100 watt (you may need higher power rating for higher current)

I also used 10 X 1400 V MOVs in series:

I think the MOVs are worth it. I use rectifiers that are probably more sensitive that the transformers, but I think it couldn't hurt. So far I haven't had problems. My safety gap fires frequently though.

Dan

Vigh holtage

Fri Dec 15 2017, 11:49PM

Registered Member #61739 Joined: Wed Aug 23 2017, 04:43PM
Location:
Posts: 44

Awsome thank you ! The values are a bit higher than i thought they would be, i guess the mots need different protection than nst do.

Any one else have any other mot protection circuits ?

Vigh holtage

Mon Dec 18 2017, 07:18PM

Registered Member #61739 Joined: Wed Aug 23 2017, 04:43PM
Location:
Posts: 44

What about those metal case resistors on e bay ? Will those short out at HV or what , i can get 1k ohm 200watt ohmites but they are 30$ each the metal case ebay ones are 1k ohm 150watt and 10$ for 2

Vigh holtage

Fri Dec 22 2017, 12:54AM

Registered Member #61739 Joined: Wed Aug 23 2017, 04:43PM
Location:
Posts: 44

Any one ?? I have seen some older threads here discusding terry filters and i have made one for a single nst 15kv @ 30ma

But with the values i used to make that one, i am positive it will NOT holt up to my quad mots.

My problem is i am just enrolling into collage for EE batchlors and as of now i do not quite understand the math calculations involved in choosing the capasitance value and resistor R and wattage value to get the correct cut off freq to protect the mots and also still be used with a 15kv 120ma nst bank 2 x 60 ma..
However...
i do know though that large ceramic caps designed for rf with screw threads rated at a few kva are best if you can get them vs a string of regular ones so i am going that route, but i get conflicting info , some specs say 500pf some say more some say less. Does it matter that mutch ? I was just going to get 2 30kv 500pf 15kva rf rated ceramics and use one on each side

And how do i choose tje right resistor value and wattage without wasting too much power ? I see values ranging anywere from 1k ohm to 2.5k ohm ! Amd wattages from 50 to 200 watts each resistor. I dont want them to over heat but i also dont want to waste any more power than i have to so how do i calculate the R value and needed wattage ?

profdc9

Sat Dec 23 2017, 02:06AM

Registered Member #58522 Joined: Tue Mar 15 2016, 08:33PM
Location:
Posts: 50

Here is my understanding of the reasoning behind the Terry filter.

There are two kinds of RF spikes you are trying to protect against. The first is at the resonance frequency of the primary/secondary. This is a fairly low frequency, only 100-300 kHz or so, and therefore it may be difficult to make a filter for. When the spark gap is firing, the spark gap is providing a low impedance path for the resonating signal at this frequency, and therefore ordinarily under normal conditions this should be a relatively small contributor. However, because you still want to protect the transformers, both a static spark gap and a chain of MOVs are used to try to divert these spikes from the transformers. The MOVs are sort of a "last line of defense" that is going to be eventually degraded by repeated spikes, but is cheaper to replace than neon signal transformers.

The other, and generally more damaging spikes are the high frequency spikes. These occur because the spark gap, rather than simply making a connection and then cleanly breaking it, might several times reconnect and disconnect over the course of discharging the capacitor. When the spark disconnects the contacts, there is still current flowing through the primary spiral coil. The primary coil has a self resonance frequency because it has both an inductance and a capacitance that are in parallel with each other. After the spark gap disconnects, the primary coil oscillates at its self-resonance frequency, often in the 10-20 MHz range, and this is added to the voltage still on the capacitor. Therefore the voltage can be as high as double that of the initial voltage on the capacitor. Because the self-resonance frequency of the inductor is much higher than the primary resonance frequency, 50 to 100 times higher, one can use a simple RC circuit to filter it. The capacitors though must be built to sustain this doubling of the voltage, and so need to be rated probably 3 times the voltage that the primary capacitor is charged up to.

Doorknob caps are excellent for this purpose, but they are expensive. I have used combinations of cheap ceramic capacitors in series and parallel, and so far it seems to be working ok. I am using a voltage doubled dual MOT system so it has the same peak voltage, it just only has half the duty cycle. So what I put together might work for you.

I do not think the principle of the Terry filter is any different, but if you have 500 mA of current and 600 to 1000 ohm resistors, you are talking 150 W of dissipation per power resistor. You might want to have more capacitance and less resistance to keep the power dissipated by the series resistors down. I have about half that current, and I use 600 ohm resistors and 2200 pF @ 18 kV per side, so that the RC constant is 1.3 microseconds, or corresponding to about 1 MHz.

By the way, if you keep your MOTs in oil, keep the can tightly sealed, as water can diffuse into the oil somewhat and lower its dielectric strength.

Dan

Plasma

Sat Dec 23 2017, 02:24AM

Registered Member #61406 Joined: Thu Jan 05 2017, 11:31PM
Location:
Posts: 268

One thing I found, was after the arc, the voltage multiple would blow the diodes, a series restinace fixed that, but with the microwave caps having a 10Mohm resistor made a voltage divider network also make a vd network with multiple meter to messure the final voltage.
I found for a vd from a mot at 50hz 5 microwave caps in series, for impendence limited, but more for 3 or more stages.
As for capacitor storage don't use microwave capacitor the in built resistor makes it difficult to deal with.

Vigh holtage

Sun Dec 24 2017, 03:21PM

Registered Member #61739 Joined: Wed Aug 23 2017, 04:43PM
Location:
Posts: 44

Ok to keep water out i have Large dessican packs from the sand blaster sand at work. We take em out of new sand and toss em , lol so i keep them . I can add one or two on the shelf inside to help also this can seals air tight with a rubber seal so i will probably hot glue everything shut on the inside and add the dessicant pack it is 5in by 6 in large

The moc caps wont last long that i am shure so i do plan on getting 3 or 4 pulse caps to make the doubler
Also i am not useing mo diodes i am useing 20kv 2a diodes

So i should stay around 500 ohm to 1k ohm and less than 100watts per side

Also i have a source for ceramic caps that is reasonable so i can use those to make things simple say 15kv and 940pf per side ?

profdc9

Mon Dec 25 2017, 08:02PM

Registered Member #58522 Joined: Tue Mar 15 2016, 08:33PM
Location:
Posts: 50

I would aim to get RC above at least 300 ns at least. So if you have 1000 pF per side and 500 ohms per side, that is 500 ns, which should be ok I think, but there is a lot of guesswork here. I think being in oil will help make the transformers somewhat less susceptible to arcing.

For four transformers, you will have on the order of 8600 VAC @ 500 mA if you are driving two of the transformers from one circuit and two from another to get 4.3 kVA, or 8600 VAC @ 250 mA (or somewhat less voltage because of the ballasting) if you ballast the circuit so 2.2 kVA. 500 mA into 500 ohms will dissipate 125 W. You will need a hefty wirewound to handle that, probably with more like 200 W capacity at least. If you're only running at 2.2 KVA then 100 W will likely be adequate.

I think that more capacitance is probably better here, probably more like 2000 pF per side, you could lower the resistance to 250 ohms then and half the power dissipation.

Make sure you do your own calculations and verify my figures.

Dan

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