MMC confusion
Gordie Orange, Fri Mar 18 2016, 06:59PMI have a MMC with 6 strings of 11 with each cap being 1.5Kv and .047uf. (25.6uf 16.5Kv)
My NST is a 10Kva 50ma I used TeslaMap to work it out but heard someone say TeslaMap works out the MMC incorrectly, is this correct?
Re: MMC confusion
Sigurthr, Fri Mar 18 2016, 07:52PM
I've always used DeepFriedNeon's site (google it) and multiplied the capacitance by Phi (1.618). Has worked beautifully every time. No idea how TeslaMap performs.
Sigurthr, Fri Mar 18 2016, 07:52PM
I've always used DeepFriedNeon's site (google it) and multiplied the capacitance by Phi (1.618). Has worked beautifully every time. No idea how TeslaMap performs.
Re: MMC confusion
Goodchild, Sat Mar 19 2016, 07:17PM
Capacitor in series and parallel follow the same but inverse equations for that of resistors in series/parallel circuits.
So for each string:
For total capacitance 0.047uF / 11 = 4.27nF
For total voltage 1.5Kv * 11 = 16.5Kv
For the whole bank:
Capacitance 4.27nF * 6 = 25.62nF (not uF)
For voltage we follow Kirchoff's good old law, voltage on parallel elements must be the same on all elements in the network. So your bank voltage is the same as the string voltage.
For more in depth info on caps in series and parallel:
@Sigurthr I don't know what you are getting at with multiplying by phi you don't need to do that. DeepFriedNeon's MMC page uses the same math as I have shown above.
As an additional note, this is a very small capacitance for a coil <200KHz You may end up with a very large primary and very high tank impedance if you keep your current MMC configuration.
Goodchild, Sat Mar 19 2016, 07:17PM
Capacitor in series and parallel follow the same but inverse equations for that of resistors in series/parallel circuits.
So for each string:
For total capacitance 0.047uF / 11 = 4.27nF
For total voltage 1.5Kv * 11 = 16.5Kv
For the whole bank:
Capacitance 4.27nF * 6 = 25.62nF (not uF)
For voltage we follow Kirchoff's good old law, voltage on parallel elements must be the same on all elements in the network. So your bank voltage is the same as the string voltage.
For more in depth info on caps in series and parallel:
@Sigurthr I don't know what you are getting at with multiplying by phi you don't need to do that. DeepFriedNeon's MMC page uses the same math as I have shown above.
As an additional note, this is a very small capacitance for a coil <200KHz You may end up with a very large primary and very high tank impedance if you keep your current MMC configuration.
Re: MMC confusion
Sigurthr, Sun Mar 20 2016, 01:14AM
The reasoning behind multiplying by Phi is it guarantees a Larger Than Resonant value that should prevent potentially-destructive resonant rise during a runaway condition (spark gap failure for example).
Sigurthr, Sun Mar 20 2016, 01:14AM
The reasoning behind multiplying by Phi is it guarantees a Larger Than Resonant value that should prevent potentially-destructive resonant rise during a runaway condition (spark gap failure for example).
Re: MMC confusion
Goodchild, Sun Mar 20 2016, 01:42PM
That doesn't make sense. Making the capacitor larger won't prevent resonant rise, it simply lowers the frequency at which it occurs in the primary tank.
I also can't fathom a spark gap failure. Are you expecting the gap to fail short? If this happened you would not have a runaway current in the primary, because there is simply no way for the NST to add additional energy (it's short with a gap failure). In my opinion this is one of the disadvantages of a SGTC vs a DR. Once the gap fires the only energy you have to resonate with is what is already stored in the tank, from that point on it's a decaying oscillation.
So in consultation by making the capacitor larger like that you are not doing yourself any favors nor are you preventing resonant rise.
Goodchild, Sun Mar 20 2016, 01:42PM
Sigurthr wrote ...
The reasoning behind multiplying by Phi is it guarantees a Larger Than Resonant value that should prevent potentially-destructive resonant rise during a runaway condition (spark gap failure for example).
The reasoning behind multiplying by Phi is it guarantees a Larger Than Resonant value that should prevent potentially-destructive resonant rise during a runaway condition (spark gap failure for example).
That doesn't make sense. Making the capacitor larger won't prevent resonant rise, it simply lowers the frequency at which it occurs in the primary tank.
I also can't fathom a spark gap failure. Are you expecting the gap to fail short? If this happened you would not have a runaway current in the primary, because there is simply no way for the NST to add additional energy (it's short with a gap failure). In my opinion this is one of the disadvantages of a SGTC vs a DR. Once the gap fires the only energy you have to resonate with is what is already stored in the tank, from that point on it's a decaying oscillation.
So in consultation by making the capacitor larger like that you are not doing yourself any favors nor are you preventing resonant rise.
Re: MMC confusion
Gordie Orange, Sun Mar 20 2016, 07:22PM
Sorry I meant nF
as for the small capacitance what would you suggest I'm just going by what TeslaMap says is the optimum
Gordie Orange, Sun Mar 20 2016, 07:22PM
Goodchild wrote ...
Capacitance 4.27nF * 6 = 25.62nF (not uF)
As an additional note, this is a very small capacitance for a coil <200KHz You may end up with a very large primary and very high tank impedance if you keep your current MMC configuration.
Capacitance 4.27nF * 6 = 25.62nF (not uF)
As an additional note, this is a very small capacitance for a coil <200KHz You may end up with a very large primary and very high tank impedance if you keep your current MMC configuration.
Sorry I meant nF
as for the small capacitance what would you suggest I'm just going by what TeslaMap says is the optimum
Re: MMC confusion
woodchuck, Sun Mar 20 2016, 08:10PM
woodchuck, Sun Mar 20 2016, 08:10PM
Goodchild wrote ...
That doesn't make sense. Making the capacitor larger won't prevent resonant rise, it simply lowers the frequency at which it occurs in the primary tank.
I wonder if Sigurthr meant to say that he uses an MMC with a voltage rating phi times the required voltage rating.That doesn't make sense. Making the capacitor larger won't prevent resonant rise, it simply lowers the frequency at which it occurs in the primary tank.
Re: MMC confusion
Sigurthr, Mon Mar 21 2016, 11:24PM
Nope, I'd been using Phi * capacitance. Voltage rating was always at least 400% expected working peak voltage.
It's been too many years since I've done spark gap coils, but I did have static gaps fail open circuit. They just stop firing (usually a result of electrode ablation). Rotary gaps can of course fail open quite easily.
I can't remember the specifics of it any more, but a lot of the literature I had read when I was learning about and still doing SG coils talked about how when you perfectly impedance match the tank capacitance to the NST secondary impedance there's potential for resonant rise if the gap fails to fire, where the voltage across the secondary of the NST can rise beyond the winding insulation ratings, causing NST failure. The idea was to use a larger than resonant value so that the cap doesn't have time to fully charge (because it is a lower impedance than the nst, and thus loads the output of the nst to a lower voltage initially as well as taking longer to reach a certain voltage) within the expected break rate. This way skipped beats don't stress the NST secondary insulation.
Sigurthr, Mon Mar 21 2016, 11:24PM
Nope, I'd been using Phi * capacitance. Voltage rating was always at least 400% expected working peak voltage.
It's been too many years since I've done spark gap coils, but I did have static gaps fail open circuit. They just stop firing (usually a result of electrode ablation). Rotary gaps can of course fail open quite easily.
I can't remember the specifics of it any more, but a lot of the literature I had read when I was learning about and still doing SG coils talked about how when you perfectly impedance match the tank capacitance to the NST secondary impedance there's potential for resonant rise if the gap fails to fire, where the voltage across the secondary of the NST can rise beyond the winding insulation ratings, causing NST failure. The idea was to use a larger than resonant value so that the cap doesn't have time to fully charge (because it is a lower impedance than the nst, and thus loads the output of the nst to a lower voltage initially as well as taking longer to reach a certain voltage) within the expected break rate. This way skipped beats don't stress the NST secondary insulation.
Re: MMC confusion
Hazmatt_(The Underdog), Tue Mar 22 2016, 12:39AM
In spark gap coils, even if you maintain the rated terminal voltage, say 15kv, you are STILL exceeding the transformer's insulation rating.
Why? Because when loaded to a Neon tube (a diode) it drops the terminal potential of the transformer to its rated insulation.
Don't believe me? Call up Franceformer and find out for yourself. I was surprised after I called with a battery of questions.
Hazmatt_(The Underdog), Tue Mar 22 2016, 12:39AM
In spark gap coils, even if you maintain the rated terminal voltage, say 15kv, you are STILL exceeding the transformer's insulation rating.
Why? Because when loaded to a Neon tube (a diode) it drops the terminal potential of the transformer to its rated insulation.
Don't believe me? Call up Franceformer and find out for yourself. I was surprised after I called with a battery of questions.
Re: MMC confusion
Goodchild, Tue Mar 22 2016, 02:01PM
I'm sorry but what you are saying really doesn't make sense to me. If the spark gap just stops firing the NST is more or less going to run open circuit, and because the NST is putting out 60Hz the tank capacitor is only going to pass a small amount of current (limited by Xc @ 60Hz in series configuration) Xc will be so high at 60Hz that for all intensive purposes the NST will be open circuit as Hazmatt suggest.
Regardless, even if you did mange to somehow match the NST impedance to the tank impedance at Fo (unlikely) you would generate the resonate high voltage at the L and C node of the tank, not across the NST as you suggest. A DR is a perfect example of this, when we create resonate rise in a DR tank the voltage on the H-Bridge output never rises above that of the drive voltage, however the L/C node will rise to many thousands of volts during normal operation. Hence why you have to rate the MMC to Ipk * Z but the bridge is only rated to 600V/1200V.
Goodchild, Tue Mar 22 2016, 02:01PM
Sigurthr wrote ...
Nope, I'd been using Phi * capacitance. Voltage rating was always at least 400% expected working peak voltage.
It's been too many years since I've done spark gap coils, but I did have static gaps fail open circuit. They just stop firing (usually a result of electrode ablation). Rotary gaps can of course fail open quite easily.
I can't remember the specifics of it any more, but a lot of the literature I had read when I was learning about and still doing SG coils talked about how when you perfectly impedance match the tank capacitance to the NST secondary impedance there's potential for resonant rise if the gap fails to fire, where the voltage across the secondary of the NST can rise beyond the winding insulation ratings, causing NST failure. The idea was to use a larger than resonant value so that the cap doesn't have time to fully charge (because it is a lower impedance than the nst, and thus loads the output of the nst to a lower voltage initially as well as taking longer to reach a certain voltage) within the expected break rate. This way skipped beats don't stress the NST secondary insulation.
Nope, I'd been using Phi * capacitance. Voltage rating was always at least 400% expected working peak voltage.
It's been too many years since I've done spark gap coils, but I did have static gaps fail open circuit. They just stop firing (usually a result of electrode ablation). Rotary gaps can of course fail open quite easily.
I can't remember the specifics of it any more, but a lot of the literature I had read when I was learning about and still doing SG coils talked about how when you perfectly impedance match the tank capacitance to the NST secondary impedance there's potential for resonant rise if the gap fails to fire, where the voltage across the secondary of the NST can rise beyond the winding insulation ratings, causing NST failure. The idea was to use a larger than resonant value so that the cap doesn't have time to fully charge (because it is a lower impedance than the nst, and thus loads the output of the nst to a lower voltage initially as well as taking longer to reach a certain voltage) within the expected break rate. This way skipped beats don't stress the NST secondary insulation.
I'm sorry but what you are saying really doesn't make sense to me. If the spark gap just stops firing the NST is more or less going to run open circuit, and because the NST is putting out 60Hz the tank capacitor is only going to pass a small amount of current (limited by Xc @ 60Hz in series configuration) Xc will be so high at 60Hz that for all intensive purposes the NST will be open circuit as Hazmatt suggest.
Regardless, even if you did mange to somehow match the NST impedance to the tank impedance at Fo (unlikely) you would generate the resonate high voltage at the L and C node of the tank, not across the NST as you suggest. A DR is a perfect example of this, when we create resonate rise in a DR tank the voltage on the H-Bridge output never rises above that of the drive voltage, however the L/C node will rise to many thousands of volts during normal operation. Hence why you have to rate the MMC to Ipk * Z but the bridge is only rated to 600V/1200V.
Re: MMC confusion
Sigurthr, Tue Mar 22 2016, 07:27PM
I found the old reference I used: See:
Sigurthr, Tue Mar 22 2016, 07:27PM
I found the old reference I used: See:
Re: MMC confusion
Hazmatt_(The Underdog), Wed Mar 23 2016, 03:25AM
I did quite a bit of work on this, but nobody ever really seems to give a damn, but I'll go ahead and post it one more time,
Hazmatt_(The Underdog), Wed Mar 23 2016, 03:25AM
I did quite a bit of work on this, but nobody ever really seems to give a damn, but I'll go ahead and post it one more time,
Re: MMC confusion
Sigurthr, Wed Mar 23 2016, 05:29AM
Interesting, Hazmatt! I was unaware of that thread and your work therein. It looks like my Phi multiplier is close to your exacted value.
Sigurthr, Wed Mar 23 2016, 05:29AM
Interesting, Hazmatt! I was unaware of that thread and your work therein. It looks like my Phi multiplier is close to your exacted value.
Re: MMC confusion
Perezx, Wed Mar 23 2016, 07:29AM
What type of capacitors do you have? They are soooo different...
And tell us details on your primary/secondary configuration, please
Perezx, Wed Mar 23 2016, 07:29AM
What type of capacitors do you have? They are soooo different...
And tell us details on your primary/secondary configuration, please
Re: MMC confusion
Uspring, Wed Mar 23 2016, 11:54AM
Hazmatt, I can't quite follow the math in the older thread you referenced. What would be your suggestion for the 10KV 50ma NST of the OP?
Uspring, Wed Mar 23 2016, 11:54AM
Hazmatt, I can't quite follow the math in the older thread you referenced. What would be your suggestion for the 10KV 50ma NST of the OP?
Re: MMC confusion
Gordie Orange, Wed Mar 23 2016, 02:11PM
These are the caps I used
http://www.ebay.co.uk/itm/251504542592?_trksid=p2057872.m2749.l2649&ssPageName=STRK%3AMEBIDX%3AIT
I'm in the UK so the NST is 10kv, 50ma at 50hz (am i right in thinking a 6kv 100ma will give me bigger sparks?)
I have a 6" secondary thats about 30" high and a 6x24" toroid and a static gap of 6 copper tubes (i'm looking into a rotary gap but my brain is melting)
Gordie Orange, Wed Mar 23 2016, 02:11PM
These are the caps I used
http://www.ebay.co.uk/itm/251504542592?_trksid=p2057872.m2749.l2649&ssPageName=STRK%3AMEBIDX%3AIT
I'm in the UK so the NST is 10kv, 50ma at 50hz (am i right in thinking a 6kv 100ma will give me bigger sparks?)
I have a 6" secondary thats about 30" high and a 6x24" toroid and a static gap of 6 copper tubes (i'm looking into a rotary gap but my brain is melting)
Re: MMC confusion
Hazmatt_(The Underdog), Wed Mar 30 2016, 03:25AM
Hi Uspring,
1. Find your Z first
2. Calculate for C using Z, this is going to give you a resonant value for C
3. Then just multiply C by 1.4, this gives you your LTR C value.
It turns out to be 1.4 because this relates to the peak current in the charging cycle. (should really be 1.414, but 1.4 or 1.5 is really close)
The actual importance of the calculations was an accurate and faithful derivation of C LTR, which was verified in simulation of course because it converges everything to impedance's, so naturally they should match up.
Hazmatt_(The Underdog), Wed Mar 30 2016, 03:25AM
Hi Uspring,
1. Find your Z first
2. Calculate for C using Z, this is going to give you a resonant value for C
3. Then just multiply C by 1.4, this gives you your LTR C value.
It turns out to be 1.4 because this relates to the peak current in the charging cycle. (should really be 1.414, but 1.4 or 1.5 is really close)
The actual importance of the calculations was an accurate and faithful derivation of C LTR, which was verified in simulation of course because it converges everything to impedance's, so naturally they should match up.
Re: MMC confusion
Uspring, Sun Apr 03 2016, 02:01PM
Thanks, Hazmatt. I'm not sure about how to obtain Z. Assuming e.g. a 50ma rated NST, does that imply a 50ma short circuit current? If that is the case, one can obtain the approximate leakage inductance by V/I = 2*pi*f*L. This is approximate since, there is also a resistive component limiting the short circuit current.
I've found by some equation jiggling, that in the case of 2 burst per primary current cycle, the burst should best be around the zero crossing of primary voltage. That gives max voltage for the tank cap. For the choice of an optimal tank cap, voltage is maximized in the resonant case, but energy is somewhat larger for a higher capacitance value. This agrees with a preferred larger than resonant value.
The best value is the result of an transcendental equation, so it is not sqrt(2)*resonant value, but not too far away from it.
Uspring, Sun Apr 03 2016, 02:01PM
Thanks, Hazmatt. I'm not sure about how to obtain Z. Assuming e.g. a 50ma rated NST, does that imply a 50ma short circuit current? If that is the case, one can obtain the approximate leakage inductance by V/I = 2*pi*f*L. This is approximate since, there is also a resistive component limiting the short circuit current.
I've found by some equation jiggling, that in the case of 2 burst per primary current cycle, the burst should best be around the zero crossing of primary voltage. That gives max voltage for the tank cap. For the choice of an optimal tank cap, voltage is maximized in the resonant case, but energy is somewhat larger for a higher capacitance value. This agrees with a preferred larger than resonant value.
The best value is the result of an transcendental equation, so it is not sqrt(2)*resonant value, but not too far away from it.
Re: MMC confusion
Hazmatt_(The Underdog), Sun Apr 03 2016, 03:30PM
Uspring, just follow the link in my previous post, it calculates everything.
Hazmatt_(The Underdog), Sun Apr 03 2016, 03:30PM
Uspring, just follow the link in my previous post, it calculates everything.
Re: MMC confusion
Uspring, Sun Apr 03 2016, 05:40PM
It looks like your prescription is:
Z = Zbranch + Ztankcap
First, why should be this so? Z, i.e. Vtransformer/Itransformer is not a source impedance. A good transformer should have a near zero impedance.
Secondly, Zbranch and Ztankcap don't simply add up this way. Note that Zbranch contains inductive components and Ztankcap is a capacitor. You calculated Zbranch correctly (sqrt of sum of squares), but didn't add Ztankcap in the right way.
Thirdly, you used only 61H and 17k from one side of the transformers leg. Actually you have 2 legs in series with your tank cap, so Zbranch is twice the value you calculated.
A more sensible prescription would be
Zbranch = Ztankcap
since it matches the source impedance to the tank cap impedance. But that leads basically to the resonance condition and not to a larger than resonance C. Larger than resonant C can only be explained by taking into account the discharging of the tank gap through the spark gap. That is much more involved than calculations based on impedances.
Sorry, I've been sloppy using Z. Z is a complex number. Here only its absolute value is meant.
Uspring, Sun Apr 03 2016, 05:40PM
It looks like your prescription is:
Z = Zbranch + Ztankcap
First, why should be this so? Z, i.e. Vtransformer/Itransformer is not a source impedance. A good transformer should have a near zero impedance.
Secondly, Zbranch and Ztankcap don't simply add up this way. Note that Zbranch contains inductive components and Ztankcap is a capacitor. You calculated Zbranch correctly (sqrt of sum of squares), but didn't add Ztankcap in the right way.
Thirdly, you used only 61H and 17k from one side of the transformers leg. Actually you have 2 legs in series with your tank cap, so Zbranch is twice the value you calculated.
A more sensible prescription would be
Zbranch = Ztankcap
since it matches the source impedance to the tank cap impedance. But that leads basically to the resonance condition and not to a larger than resonance C. Larger than resonant C can only be explained by taking into account the discharging of the tank gap through the spark gap. That is much more involved than calculations based on impedances.
Sorry, I've been sloppy using Z. Z is a complex number. Here only its absolute value is meant.
Re: MMC confusion
Hazmatt_(The Underdog), Mon Apr 04 2016, 03:58AM
Start here, I = E / Z
Hazmatt_(The Underdog), Mon Apr 04 2016, 03:58AM
Start here, I = E / Z
Print this page