Re: NST filter
Terry Fritz, Tue May 09 2006, 11:11PM

Oh!! "mine" is better

The component values are not critical at all and are commonly +-100%

The only significant difference is that Greg uses spark gaps for NST over voltage protection and I use MOVs "and" a safety spark gap up front.

Spark gaps are cheap but you "must" set them right or they do no good. MOVs are expensive, but you can't set them wrong ) I use the spark gaps to take normal over voltages and the MOVs are the last ditch protetion incase the safety gaps are set wrong or fail. If the MOVs burn up, they fail shorted. So in mine, "nothing can go wrong" enough that the NST is damaged. But MOVs are pricey so Greg's is the next best choice on a budget.

Be sure to put a fuse on the NST too since there is a core saturation issue that "might" blow an NST now and then.

With my filter and a fuse, the reports I have gotten of NST fialures are "zero" A few folks complain that the resistors get hot and that is why they are so over rated. Power resistors can reach 300C in "normal" operation which was freaking people out, so I way oversized them. There have been a few people that have fried the MOVs. But the MOVs died saving the NST from the user there

The reason I went with MOVs is because there is a certain percentage (a high one) of folks that "just can't" get the safety gaps set right. The MOVs solved that, but at a price...

If Digikey is sold out of the MOVs, use a few more of the next voltage size down... I hate making things like this and every body orders the stuff and DK goes sold out for three months...

Guess why I have not mentioned the PN of my new Super IGBTs for the Sidac/IGBT gap

Pick your poison, but mine is best

Cheers,

Terry

Re: NST filter
..., Tue May 09 2006, 11:54PM

I should point out that littlefuse will sample mov's from the little smd ones to huge bot down monsters. Use them wisely, but they will usually send you as many as you need... (not going to say anything more)

I haven't done any damage to my nst (7.5kv/30ma) with no filter (not even a saftey gap), but I will admit that I don't really abuse it with my micro coil...

Have fun!

Re: NST filter
Brian, Wed May 10 2006, 01:06AM


Where in the circuit and what sort of fuse would be needed?

BTW - Thank you Terry for you ever unbiased opinion. I like the idea of idiot proof though.

Re: NST filter
Terry Fritz, Wed May 10 2006, 05:36AM

Hi,

The fuse should be a fast blow type. One of those 1-1/4 x 1/4 inch cartridge fuses is fine (Radio Shack). The value is hard to guess without knowing anything about you NST, gap type, etc. but it should probably be about twice the VA rating of your NST divided by the input voltage. So if your NST is a 12/60:

2 x 12000V x 0.060A / 120V = 12 amps

If the fuse blows too much, go higher.

The fuse is directly on the NST 120V low side. On the Hot wire is best but not a big deal.

The reason for the fuse is that if an NST is driving a capacitor, the output current can go too high and saturate the core and shunts. Then it looses its current limiting and the currents can go far too high. They make a very loud bad sounding buzz when that happens. The fuse simply blows in that case preventing damage.

Cheers,

Terry

Re: NST filter
JimmyH, Wed May 10 2006, 07:01PM

I don't get this whole RC filter thing. It's unnecessary loss.

People used to use LC filters for a while, but realised that the ringing they created was just as bad. Terry tested that out once, if I remember correctly.

So once people realised that, they seemed to have gone RC, since it's safer, even though it wastes a lot of power.

Why not both ?

Instead of a huge power resistor, why not use a smaller resistor with an inductor in parallel with it? If it's set to be critically damped, then there is no voltage spiking to worry about.

At 60hz, the inductance provides much less reactance, so current passes much easier, plus whatever VA flowing in the inductor is purely reactive, so it really cuts down on loss.

Say for example we want to replace our RC filter with a critically damped RLC filter at 360khz (which is 1/RC for Terry's filter). Since the mains frequency is 60hz, and the resonant frequency of the LC part of the filter is 360,000hz, the frequency, and therefore reactance (Xl = 2*pi*F*L) varys by 6000x.

say we wan't to have the same rejection that our RC filter had. That means Rnew has to be 2*Rold, and Xlres = Rnew. Xlmains (hope you understand my notation) is .0003 Rold. The voltage drop when charging the capacitor is now .2% of what it was, and reactive (if the inductor is lossless).

Power = V^2/R = (0.0003^2)Vold/(2*Rold), or ~0*old power dissipation. This doesn't count any high frequency noise which has to be eaten up, or any inductor loss, but the point remains that it is a big power saver.

The inductor would have to be around 870uH and the resistor 2k to be a drop in replacement for Terry's filter. The inductor could just be a 2x3" coil with a couple hundred turns of 28awg for example. The DC resistance would be a mere 6.8 ohms compared to the 1000 ohms that it had to push through before when charging the capacitor.

So what do you think?

Re: NST filter
ArcLight, Wed May 10 2006, 07:41PM

It seems that I remember reading someplace that MOVs clamp faster than a sparkgap can strike an arc. But the sparkgap can sink much more power so a combination of the two makes good sense.

Re: NST filter
Hazmatt_(The Underdog), Wed May 10 2006, 10:07PM

Consider this:

My Tesla system running with a 12KVAC 60mA transformer and 3000 ohm load resistors have a 300VAC drop across them at maximum power pull. This has been measured with my potential divider.

So for each 3000 ohm load, we have 300*.06 = 18 W for peaks. And since they get warm during a 15 minute run, they're probably dissipating 25W each.

Say I'm loosing 100W in the resistor. 760W - 100W / 760W * 100 = 87% power transfer, meh, I'm happy with that. And if I'm really loosing 50W, then I have 93% which I can live with too.

The losses I care about are not having good enough coupling, long leads, bad capacitors, bad leads, bad sec. gnd. connection. etc. etc.

Re: NST filter
Steve Conner, Wed May 10 2006, 11:53PM

JimmyH: The DC resistance of a NST secondary winding seems to be about 5-10k ohms. There's an awful lot of real thin wire in there. So personally I wouldn't bother adding inductors to decrease the DCR of the filter.

For a pole pig or high frequency solid state power supply, I guess your idea would be good. Mainly because it saves having to buy such big resistors and manage the heat from them. (I guess that argument depends on how much of the heat is RF and how much is DC.)

Re: NST filter
HV Enthusiast, Thu May 11 2006, 01:23AM


The reason is because the resistance of the secondary winding of an NST is considerably greater than the RC filter most would use for an NST filter. Any savings would be quite negligible. Similar to trying to increase the Q of the secondary when its the primary Q that is the limiting factor of Q in the system.


EDIT: Oops. I see Steve beat me to it . . .

Re: NST filter
JimmyH, Thu May 11 2006, 06:28PM

Hmm.. I guess I was thrown off by there being 2 100W resistors on the schematic, instead of much smaller ones, and reading stories about the resistors getting really hot (IIRC, that was Gary Lau, who had saturated his shunts, and was pulling too much current anyway, so he could have gone less R more C).

I guess it's just a question of whether you'd rather wind a couple inductors, or buy a couple big power resistors.

If you don't mind winding more turns, you can get away with a lot higher noise rejection before power loss becomes a problem, so you can get some extra safety, and possibly skip out on the MOVs instead. =\

Re: NST filter
Terry Fritz, Thu May 11 2006, 07:21PM

Hi,

I studied RLC filters back in 98:



Also RC filters thet the present "terry filter" is based on:



Today, with computer modeling, it might certainly be possible to "tune" the mess to get better performance out of the RLC filter. But the present RC design seems to work an just about any system as is. Back then, there were no nice models and the knowledge was limited. Today, there may be many more options as Jimmy mentions!!

One does need to take into account the enormous inductance of the NST secondary and its resistance too.

I think 90% of the good it does is in the spark gap and MOVs absolutely prevent overvoltage though.

It's claim to fame is basically that "it works really well" even though it is costly, burns power, and might not be the brightest design... There is plenty of room for inovation though espicially if it brings the cost down.

Cheers,

Terry

Re: NST filter
HV Enthusiast, Thu May 11 2006, 09:01PM

If you are truly concerned about efficiency, just eliminate the filter altogether. Remember, engineering is all about trade-offs, and it may be worthwhile to increase efficiency at the expense of reliability, should efficiency truly be an issue. Personally, i have never used filtering on my NST coils and never had any NST failiure or other problems.

Also, one needs to look at the entire system. The amount of losses introduced into the system by a filter may be swamped out by the losses in the primary circuit (i.e. spark gap) so any improvement here really won't matter much in the end.

And as Terry nicely said, the claim to fame is basically that it "works really well" at the expense of cost and power losses. The goal of the filter is to gain protection and reliability. If efficiency is a concern, just eliminate it completely, of course at the expense of reliability.

Re: NST filter
Hazmatt_(The Underdog), Wed Aug 23 2006, 01:54AM

I just wanted to post a picture of my NST line filter.

The capacitors were chosen as 10pF because throughout the simulation, the larger value capacitors such as 200pF and larger did not arrest large enough transients. The 10pF capacitors return the line surges back to normal operating nameplate current.

The ground pad connects on both sides of the fiberglass board for a firm connection. I could possibly add a spark gap on top of the capacitors, but the simulation results are showing normal operation of the transformer.

The board is coated with a silicone spray to protect the copper from 'aging' and will hopefully last many years into the future.

Re: NST filter
HV Enthusiast, Wed Aug 23 2006, 02:34AM

Actually, my filter is better. This is the filter I use and i have never had an NST fail during Tesla coil use. I've used NSTs ranging from small 4kV/20mA units, up to a paralleled 15kV/60mA systems.


Dan's Filter

Re: NST filter
williamn, Wed Aug 23 2006, 02:43AM

Even though EVR's NST filter (using a radical mid point ground protection system) may be a bit complicated for the average coiler, I too have used it with great success. I have never had a NST fail while using it.

Re: NST filter
Dr. Drone, Wed Aug 23 2006, 03:21AM

Re: NST filter
Hazmatt_(The Underdog), Wed Aug 23 2006, 08:49AM

That's great that some of you guys have never lost a NST, and I say more power to you.

Some of us have lost NST's due to dielectric breakdown. I have lost 2 12/60's at different times. The latest being the victim of performance testing. Basically I wanted to find the LTR multiplier myself and verify the findings first hand, and doing this more scientifically with simulation and data.

SO, if you're going to push every watt out of an NST, I would suggest using a filter. There is no appreciable loss in power, and mine will eliminate the large reflected current pulses.

Re: NST filter
Terry Fritz, Thu Aug 24 2006, 07:41PM

That's great that some of you guys have never lost a NST, and I say more power to you.

I should note than most of these people are probably using the larger "LTR" cap values that naturally resist over voltage. The LTR caps have really made a big difference in saveing NSTs too!! Back 7 years ago, they were all running resonant and some people made piles of dead NSTs...

With LTR caps, filters, fuses... If the NST dies then, it was a "natural" death... The LTR caps give bigger sparks so no reason not to use them. The fuse is cheap and easy...

One could ask if the filter is needed for LTR coils? Sync LTR coils actually drop in voltage if the gap does not fire... "Assuming" everything is wired right, then the filter only blocks RF from the outer windings. I have tested that and most of the HF voltage drop is on the outer windings for sure... With people running coils so hard these days, that might be a real concern...

But it sort of depends one your budget and situation. If you just bought a brand new 15/60 NST for $250, the choice is easy. But if your whipping together a little coil and can get another NST easily enough, then the choice might be easy too...

Cheers,

Terry

Re: NST filter
Hazmatt_(The Underdog), Thu Aug 24 2006, 09:02PM

I'm having to take a lot of precautions because I am trying to get the exact right LTR cap value for my difficult to replace 9/120.

In my case it is no throw-together coil. I want it designed to be a show coil that lasts for 50 years, including possible abuse. So this means simulation, calculations, and real work.

I'm trying to pull maximum currnet and voltage possible from this NST without exceeding either nameplate value.

I am getting there, but it is a long process of validating exactly what value for LTR. And in this case I am not going to just look at other people's numbers, I need to verify my numbers for power, efficiency, and make sure all the hard questions are answered with the closest figures.

Matt

Re: NST filter
HV Enthusiast, Thu Aug 24 2006, 11:22PM


Then definitely get rid of the NST. Get a nice hearty potential transformer, or even a small pole transformer.

Re: NST filter
Hazmatt_(The Underdog), Fri Aug 25 2006, 08:15AM

EVR I don't think you've considered the portability/power limit aspect of this design.

The NST offers the power, weight, and Maximum Power Draw allowed. No PT is going to do that.

This has to be an entierly self contained unit capable of being run off an outlet at maximum draw without tripping breakers. That is why we use an NST in this design.

Now sure, I have a PT sitting out in the garage right now, but you know what, it tripps out the breakers. And I might add that I am building a 5kw stepdown for the PT now.

Long story short, we are going the efficiency route with what we have available, we are pulling maximum power without tripping breakers, a PT or pig exceeds our space allotment (footprint), weight is an issue, and the NST is the right answer for this situation.

Re: NST filter
Sulaiman, Fri Aug 25 2006, 10:40AM

As a bona-fide NST destroyer I think an NST is not a good choice for a reliable SGTC because;

1) An NST is not meant to operate at continuous high voltage
they produce high voltage to start the neon tubes
but a much lower voltage during operation
operating mainly in reactance-limited current mode due to the flux shunt
the output characteristic of an NST into it's intended load is equivalent to
a high voltage with large series inductance
e.g. a 1500V 60 mA NST 'Ideally' gives
15 kV @ 0mA, 60 mA @ 0V, 7.5 kV @ 30 mA etc.
(i.e. max output is 7.5kV x 30 mA, NOT 15 kV x 60 mA)
2) Operating at or near resonant loading for best TC performance
puts a continuous high voltage stress on the insulation
made worse by operating at high temperature due to high I^2.R heating.

So, A question to all coilers -
Who has a SGTC powered by an NST that has been running for a long time?
ANYONE ?

Re: NST filter
Michael W., Fri Aug 25 2006, 01:51PM

1) An NST is not meant to operate at continuous high voltage
they produce high voltage to start the neon tubes
but a much lower voltage during operation
operating mainly in reactance-limited current mode due to the flux shunt
the output characteristic of an NST into it's intended load is equivalent to
a high voltage with large series inductance
e.g. a 1500V 60 mA NST 'Ideally' gives
15 kV @ 0mA, 60 mA @ 0V, 7.5 kV @ 30 mA etc.
(i.e. max output is 7.5kV x 30 mA, NOT 15 kV x 60 mA)

Does that mean the calculations for my MMC, for a 900 watt NST are wrong? (calculated for 15KV @ 60mA)

Re: NST filter
Dr. Drone, Fri Aug 25 2006, 03:07PM

Re: NST filter
Sulaiman, Fri Aug 25 2006, 07:06PM

Michael, no it doesn't mean your calculations are incorrect
If you look at Christopher's BEAUTIFUL coil
he is a using slightly larger than resonant (LTR) capacitor value
as you have been reccommended.
I was (evidently wrongly) explaining why I consider NST's in SGTC unreliable.

Christopher - you really have done some beautiful work - kudos to you man!

Re: NST filter
Dr. Drone, Fri Aug 25 2006, 07:20PM

Re: NST filter
Hazmatt_(The Underdog), Fri Aug 25 2006, 08:52PM

If you talk with any of the engineers that design NST's like I did with Franceformer, you will know that:

The transformer is designed to output full rated current in a short circuit condition with no terminal voltage.

The transformer does not strike the tube at a higher voltage then drop down. The older core-and-coil transformers are not capable of this task. Rather they are designed (insulation class and all) to operate at half-wave rectified voltage levels. This means that a 10kvac NST is designed, insulated and all, for 7070vac. The reason behind this is that the luminous tube is actually a diode. So the engineers insulate transformers based on the fact that the source will be running continuously at 7070 vac at a particular current (determined by how many and how long the gas tubes are).

You will also know that a transformer (any transformer) will not likely output its nameplate value under no-load. So if you go and measure a 15/60 with no load, you should not be surprised to find its output 10-20% higher then the nameplate value. This should be obvious of any transformer if you consider the losses (core and copper) do not truely affect performance until a load is present.

If you want you can even go to the transformer article in our wiki, where I have a worked example showing how you can extrapolate your transformer data and actually find the 'ideal source' before the losses. this is very helpful in design and performance.

Re: NST filter
Michael W., Mon Sept 04 2006, 03:13AM

Due to the lack of large power resistors in my neighbourhood, could I not wind some chokes in combination with my capacitors, like gregs filter or just chokes? If so how many uH or better, if someones already done this, your specs. (transformer : 15KV 60mA NST)

Re: NST filter
Terry Fritz, Tue Sept 05 2006, 09:29PM

Due to the lack of large power resistors in my neighbourhood, could I not wind some chokes in combination with my capacitors, like gregs filter or just chokes? If so how many uH or better, if someones already done this, your specs. (transformer : 15KV 60mA NST)

Hi Michael,

Due to a parts ordering goof, I have some 50W 500ohm resistors here. They will get pretty warm but nothing bad. If you want a couple, mail me your address and I'll send you two free.

Chokes are useless in this case:





Cheers,

Terry

Re: NST filter
IamSmooth, Mon Oct 16 2006, 06:21PM

Terry, what is the purpose of the 1000 ohm resistors for your filter? Is it to limit the current in a shorted condition when the NST can't current limit itself?

Re: NST filter
Terry Fritz, Tue Oct 17 2006, 12:05AM

Hi,

The resistors along with the caps form a low pass filter that keeps the high MHz plus noise out of the NST. See this:



Cheers,

Terry

Re: NST filter
IamSmooth, Tue Oct 17 2006, 02:27AM

Yes, I had read that. Nice work. I thought that the caps are all that were necessary to filter the high frequency. How do the resistors aid in this behavior?