Primary strike protection for DRSSTCs / Tesla Coils / Forums

Forums

4hv.org :: Forums :: Tesla Coils

« Previous topic | Next topic »

Primary strike protection for DRSSTCs

1 2 next

Move Thread

LAN_403

mbd

Sun Jun 15 2014, 10:06AM

Registered Member #43278 Joined: Sat Feb 22 2014, 09:18AM
Location: Sydney
Posts: 9

Looking around a bit, I have found three different protection schemes to protect against primary strikes:

What I will call the "Steve Ward" approach of adding a 0.1uF cap from DC bus negative to ground (C3 in the schematic at http://stevehv.4hv.org/DR_plexi/general%20wiring%20scheme.pdf)
What I will call the "Eastern Voltage" approach of adding a 0.1uF cap from an H-bridge output to ground (the bottom right cap in the schematic at http://www.easternvoltageresearch.com/datasheets/flexiBrute10shuntcap_ds.pdf)
What I will call the "Terry Fritz" approach of adding bandpass filters on each H-bridge leg (http://drsstc.com/~terrell/notes/DRSSTCprotec.pdf)

The first and third come from this earlier thread: http://4hv.org/e107_plugins/forum/forum_viewtopic.php?136375.post

I am not sure that I really understand any of the three protection schemes properly.

The Steve Ward approach

Steve Ward explains his approach on this on http://stevehv.4hv.org/DRSSTC4.htm:

Notice that if the coil arcs to the primary system, either the IGBTs or diodes (which are always there) will let the current go safely back into the DC bus caps, where it can couple through C3 (a 100nF PP film cap) to the ground connection of the secondary coil. If the capacitor was not there, then it would likely blast through to the heatsink and destroy the IGBTs insulation. Or, it will find some other way. The idea here is simply to provide an excellent path for the streamer current to take, simple.

So, let me see I have understood this protection scheme:

the reference to the IGBTs can be ignored because one cannot be certain that the "correct" IGBT will be conducting at the necessary time; and
that being the case, the free-wheeling diodes are key to the scheme - if the primary strike drives the H-bridge leg negative, the free-wheeling diode across the negative driving IGBT will conduct and discharge the strike across "C3". if the primary strike drives the H-bridge leg positive, the free-wheeling diode across the positive driving IGBT will conduct and discharge the strike across the bus by-pass capacitor ("C2") and "C3".

Question 1: Have I understood the "Steve Ward" approach properly?

The Eastern Voltage approach

On the one hand, the "Eastern Voltage" approach seems pretty straightforward - a straight discharge path from the H-bridge output leg to ground, but:

Question 2: Doesn't this scheme load up the H-bridge? (If Fres is 200kHz, Xc = 8 ohms. If Vbus is 400VDC and ground is at approximately 200V, then the peak current across the cap would seem to be 200/8 = 25A)

Question 3: Why is a cap only needed on one leg if a full bridge is used? See http://www.easternvoltageresearch.com/images/gallery_fbfbassy04_lg.jpg

The Terry Fritz approach

Question 4: If one adopted the "Terry Fritz" approach, would one also need to adopt the "Steve Ward" approach or the "Eastern Voltage" approach?

Question 5: Is the "Terry Fritz" approach widely used?

Question 6: If so, are the values of Cpro and Lpro in the paper used? Or do people adjust them? And, if so, what values does one pick to allow for tuning?

Thanks for your thoughts
MBD

PS This is my first post here so I here tender, in advance, my apologies for anything I have done wrongly.

Steve Conner

Sun Jun 15 2014, 11:36AM

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

I use the "Steve Ward" approach and my understanding of how it works is the same as yours. I don't much like the other two methods.

Goodchild

Sun Jun 15 2014, 04:38PM

Registered Member #2292 Joined: Fri Aug 14 2009, 05:33PM
Location: The Wild West AKA Arizona
Posts: 795

I have also been using the "Steve Ward" approach on my DRs and have been doing so for the last couple years now. I have found that it works very well. Generally you only need one capacitor from the bus to the heatsink (RF ground) because the main bus cap and/or your snubbers are so low impedance at TC frequencies that C2 is not really needed.

Just an FYI the EVR method is "Steve Wards" method, it's just a copy.

I feel that a slight improvement on the scheme is to have one cap going from RF ground to each side of the bridge rail, one from +V and one from -V to RF ground. This give two parallel paths, rather than one series path.

Steve Conner

Sun Jun 15 2014, 06:26PM

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

Goodchild wrote ...

I feel that a slight improvement on the scheme is to have one cap going from RF ground to each side of the bridge rail, one from +V and one from -V to RF ground. This give two parallel paths, rather than one series path.

This is how I built my Odin bridge.

mbd

Sun Jun 22 2014, 08:21AM

Registered Member #43278 Joined: Sat Feb 22 2014, 09:18AM
Location: Sydney
Posts: 9

Thanks for the responses. I have now modified my TC to conform to the "Steve Ward" approach but with the two path/cap improvement.

I have a couple of follow-up questions:

Is there any magic in the choice of cap size (0.1uF)?
Following up on Eric's comment that 'the EVR method is "Steve Wards" method, it's just a copy', does he mean that the idea is derivative? (This would make sense to me.) Or, does he mean that it is electrically equivalent? (If so, I would be grateful if someone could explain how that can be.)

Hydron

Sun Jun 22 2014, 11:28AM

Registered Member #30656 Joined: Tue Jul 30 2013, 02:40AM
Location: UK
Posts: 208

mbd wrote ...

Thanks for the responses. I have now modified my TC to conform to the "Steve Ward" approach but with the two path/cap improvement.

I have a couple of follow-up questions:

Is there any magic in the choice of cap size (0.1uF)?
Following up on Eric's comment that 'the EVR method is "Steve Wards" method, it's just a copy', does he mean that the idea is derivative? (This would make sense to me.) Or, does he mean that it is electrically equivalent? (If so, I would be grateful if someone could explain how that can be.)

To me the "EVR" method seems different and inferior to the "Steve Ward" method regardless of whether it was intended as a copy or not. As you noticed, the displacement current through the capacitor is significant at TC frequencies and would both load the bridge somewhat and require a high RMS current capacitor.

To size the capacitor you need to know how much voltage you're happy to have across it and the current through it. I put a scope on my topload to measure groundstrike current for my 6.5" dia coil and the highest I saw was ~6A, so I suspect your 8 ohm reactance 0.1uF capacitor will be easily big enough to keep the voltage reasonable.

On my coil I have two bus caps in series for >450V operation, with a 0.1uF 2kV polypropylene capacitor from the center of them to ground, value chosen by what I had lying around. Fairly certain the coil has survived a primary strike, though I don't have it on video to be sure.

Uspring

Sun Jun 22 2014, 03:58PM

Registered Member #3988 Joined: Thu Jul 07 2011, 03:25PM
Location:
Posts: 711

I put a scope on my topload to measure groundstrike current for my 6.5" dia coil and the highest I saw was ~6A...

Interestingly low value. For e.g. 120kV voltage and 20pF topload, I get a time constant of 0.4us. Steve Conner quoted frequency components in the 10's of MHz for ground arcs. Something strange is going on.

Steve Conner

Mon Jun 23 2014, 10:23AM

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

Well, if the time constant were 0.4us, the output would be 3dB down at 400kHz, 60dB down at 4MHz, etc. But 60dB down on 100s of kV is still quite a large voltage, and the radiation efficiency goes up with frequency to compensate the fall-off in current.

wrote ...
Is there any magic in the choice of cap size (0.1uF)?

These capacitors are effectively from live to earth of the mains, so they have to be Class Y and should really be a maximum of 2.2nF. But we are building Tesla coils, not consumer electronics.

I used 68nF because I had a lot of 68nF capacitors lying around.

Uspring

Mon Jun 23 2014, 10:57AM

Registered Member #3988 Joined: Thu Jul 07 2011, 03:25PM
Location:
Posts: 711

You're probably right. One has to be careful, though, comparing noise spectra of e.g. van de Graafs or Marx generators to those of TCs. The voltages (per m of arc) are much higher, leading to a much faster breakdown in e.g. ns or tens of ns.

Goodchild

Thu Jun 26 2014, 10:56PM

Registered Member #2292 Joined: Fri Aug 14 2009, 05:33PM
Location: The Wild West AKA Arizona
Posts: 795

mbd wrote ...

Thanks for the responses. I have now modified my TC to conform to the "Steve Ward" approach but with the two path/cap improvement.

I have a couple of follow-up questions:

Is there any magic in the choice of cap size (0.1uF)?
Following up on Eric's comment that 'the EVR method is "Steve Wards" method, it's just a copy', does he mean that the idea is derivative? (This would make sense to me.) Or, does he mean that it is electrically equivalent? (If so, I would be grateful if someone could explain how that can be.)

To clarify, I mean it's exactly the same...

1 2 next

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