High Voltage Forum
Tesla coils => Dual Resonant Solid State Tesla coils (DRSSTC) => Topic started by: TDAF on March 29, 2018, 03:15:11 PM
-
So, I've started working on my QCW DRSSTC
Here's what i came up with,
do the numbers check out??
if not, what can i improve??
-
So, I've slightly tweaked some parameters to get a higher Fres
Attached is a shot from JAVATC
-
I can only comment on this from a DRSSTC point of view, where I think that your topload has a too small minor diameter and too large major diameter. You will loose some of the tight field control around the secondary / primary coil when the edge of the topload is so far away. Not sure if its needed in QCW-land :)
-
Thank you for the reply!!
So, the top load was made that large as I've read that qcw coils with large toploads work well
So, what might a solution be to the field shaping problem?
I can't change the minor diameter due to certain monetary issues
Would adding another intermediate toroidal with a smaller major diameter between the main top load and the secondary help?
-
I'd agree that this may not be a big issue for a QCW coil - topload voltages are much lower and the spark growth seems to be driven by the field at tip of the arc. If the topload is already finished then I'd test it on the finished coil before worrying about whether it needs to be changed.
-
The thing is I haven't even started :P
I wanna make sure everything is just right before proceeding with the construction
-
Now, I know that turns ratio isnt of much importance in a tesla coil but a 1:10 ratio as is the case here Would it produce any problems??
also, will the low secondary impedance have any negative effect??
-
So, for the MMC
Will CBB capacitors work?
I'm thinking of something like 100 caps 10 in series 10 in parallel to give about 10nf of capacitance
https://www.ebay.com/itm/10Pcs-CBB-Capacitor-100V-101-683-0-1nF-68nF-101J-683J-Pitch-5mm/192143864676?hash=item2cbcaab364:m:miP5zeN54Fp-RSicta7nMyQ (https://www.ebay.com/itm/10Pcs-CBB-Capacitor-100V-101-683-0-1nF-68nF-101J-683J-Pitch-5mm/192143864676?hash=item2cbcaab364:m:miP5zeN54Fp-RSicta7nMyQ)
Something like these capacitors...
-
I think this has already been covered on the forum - complete mystery caps are a bad idea, especially when you don't even know what dielectric they have. If you can find a datasheet that shows them to be polypropylene with reasonable RMS and peak current specs then maybe, but without that or someone having tested them then it's not recommended.
-
I think this has already been covered on the forum - complete mystery caps are a bad idea, especially when you don't even know what dielectric they have. If you can find a datasheet that shows them to be polypropylene with reasonable RMS and peak current specs then maybe, but without that or someone having tested them then it's not recommended.
OK
But I've seen people use similar caps in large drsstcs of the yore
-
I think you can assume that they are similar to the old Panasonic capacitors, I got some datasheet and derived data on them here: http://kaizerpowerelectronics.dk/theory/good-mmc-capacitors/
They are old by design and compared to modern MKP power film capacitors, so I would advise on getting something better.
-
I agree that they look similar to the old Panasonic caps, but unless you know at least if they are polypropylene I wouldn't assume anything about the actual equivalence.
It also looks like you've already made a thread about these exact capacitors here: https://highvoltageforum.net/index.php?topic=252 - please try and not ask the same things in different places!
-
I'm sorry
I thought I asked that on 4hv :P
Really sorry
So, what in your opinion might be a good mmc capacitor for this use?
-
If you can't find a good surplus source, I suggest these caps (which aren't very expensive anyway and work very well if you operate at Vpk <= VAC rating):
https://www.mouser.com/ProductDetail/EPCOS-TDK/B32672L1333J?qs=sGAEpiMZZMv1cc3ydrPrF17RCV4vUJ9zc%2fmHTzNoF2E%3d
I find that the smaller capacitors offer far better IRMS/Capacitance ratings which is desirable for QCWs as the required C is usually small, but the RMS current can be large.
Your toroid capacity looks to be "more than adequate". I might suggest smaller toroid and more turns on the secondary. The design trade off here is that a lower impedance secondary necessitates higher secondary current to charge that capacitance to the useful range of 50-80kV that typical QCW coil will operate at. This higher secondary current will mean higher losses in the copper wire. On the other hand, the large top capacitance gives frequency stability when the plasma grows, which has the advantage of more efficient energy transfer from primary to secondary as the tuning does not drift as far. It's difficult to identify an optimum selection of coil parameters, even if you understand the theory and math. I did notice particularly nice results when using a 16"x3" toroid instead of an 8"x2" toroid (different secondary wire awg, however, to keep Fres similar). But, it's not as if the 16x3 toroid was twice as good at twice the diameter... maybe just 20% better than the small toroid, and so it never seemed practical to me.
This line of thinking led me to experiment with extra capacitors inside of my secondaries to have both a small toroid and big capacitance. With modest additional capacitance in the 20pF range this did seem to have a nice effect when using a "small" 8" x 2" toroid. Using too much capacitance resulted in disappointment as the charging current can become unacceptable, and basically you won't see the best spark length to primary current ratio.
-
Your toroid capacity looks to be "more than adequate". I might suggest smaller toroid and more turns on the secondary. The design trade-off here is that a lower impedance secondary necessitates higher secondary current to charge that capacitance to the useful range of 50-80kV that typical QCW coil will operate at.
how much secondary impedance should I be shooting for??
So, I tweaked the secondary geometry,
Now, with a larger impedance and a smaller topload!! :P
-
I was a bit surprised to hear, that Steves best results were obtained with a larger topload or equivalently, a secondary MMC. Usually the secondary Q (under arc load) should not be too large. My guess for an optimal Q would be around 1/k. An estimate for your last design suggestion would put it above 10. That depends much on the input power you can provide, so this is very uncertain.
You can get away with a high secondary Q, if you run in upper pole mode. I believe Steve used that and for your choice of a secondary that mode of operation seems necessary.
-
I was a bit surprised to hear, that Steves best results were obtained with a larger topload or equivalently, a secondary MMC. Usually the secondary Q (under arc load) should not be too large. My guess for an optimal Q would be around 1/k. An estimate for your last design suggestion would put it above 10. That depends much on the input power you can provide, so this is very uncertain.
You can get away with a high secondary Q, if you run in upper pole mode. I believe Steve used that and for your choice of a secondary that mode of operation seems necessary.
So, will the higher Q be a problem??
If yes, How do I reduce it??
-
Secondary Q is the ratio of arc resistance to secondary impedance. Arc resistance drops with higher power, but you might have constraints on that ;)
You can also increase secondary impedance by more inductance. But that implies a smaller toroid. And that again means larger frequency changes due to arc capacitance.
AFAIK most QCW designs run on the upper pole as e.g. mentioned in Loneoceans web pages. Upper pole operation works with much larger secondary Q than lower pole operation. I'm not sure how Loneoceans did this. My opinion was, that you need a PLL driver for that.
-
Hi Uspring correct me if I am wrong, high Q gives more output from the tesla by sharpening the point of resonance. it will also make the frequency splitting grater. It will also make the tesla harder to tune so arc capacitance will de tune easier. if arc capacitance is taken in to consideration shouldn't that give optimal performance. I remember somewhere that the upper pole has more voltage amplitude than the lower pole? this topic has been discussed on 4hv a couple of times before.
TDAF I found this
https://www.allaboutcircuits.com/https://www.allaboutcircuits.com/textbook/alternating-current/chpt-6/q-and-bandwidth-resonant-circuit
https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=6971783
https://electronics.stackexchange.com/questions/115933/why-does-frequency-splitting-happen-when-two-resonant-circuits-are-magnetically
http://4hv.org/e107_plugins/forum/forum_viewtopic.php?7499
-
AFAIK most QCW designs run on the upper pole as e.g. mentioned in Loneoceans web pages. Upper pole operation works with much larger secondary Q than lower pole operation. I'm not sure how Loneoceans did this. My opinion was, that you need a PLL driver for that.
In the QCW coil I am building the upper pole is significantly stronger than the lower and it oscillates there with either PLL or zero-crossing driver. Both primary and secondary are tuned a little over 300kHz in the graph below which shows both poles with the upper being the stronger of the two (lower impedance, though a little hard to see without log Y axis sorry):
-
@Teravolt:
The frequency, the DRSSTC runs on, is determined by the zero current switching condition on the primary tank. The frequencies at which this happens never really match the secondary resonance frequency. If you're outside the resonance, it is often better not to have a too large Q.
@Hydron:
From the diagram you show, it looks like the lower pole is at 225kHz and the upper pole at 360kHz. The Z peak at 245kHz corresponds AFAIK to the secondary resonance. I'm wondering about your statement of secondary res being around 300kHz. It looks like primary fres could be about 300kHz as you state.
Upper pole operation is no surprise under this condition, since during the initial phase of primary rampup there is little current in the secondary, which could affect the primary. So it will start at the primary res frequency. When secondary current picks up, the coupling takes effect and the operating frequency moves up to the upper pole. I believe, that startup at the primary fres is generally the case unless special precautions are taken in order to start at a different frequency (e.g. a PLL).
But: Upper pole operation has an advantage over lower pole operation only if the primary fres is below secondary fres. And it should stay that way even under arc loading. I've illustrated this here: https://highvoltageforum.net/index.php?topic=78.msg624#msg624
If primary fres gets over secondary fres, the upper pole, at which the coil is running, will be pushed away by the coupling from secondary fres. That will impede power transfer to the secondary. Have a look at the paragraph "Tuning Tricks" in Loneoceans QCW 1.5 writeup. http://www.loneoceans.com/labs/qcw15/
-
I should have included a couple of other measurements in my post as well, namely the primary (alone) and secondary+topload (alone) sweeps corresponding to the configuration in the previous graph. The primary is maybe not as far below the secondary frequency as is ideal, but I ran out of copper tubing about a turn earlier than expected when constructing it.
I've included a pic of the coil as-measured. It has changed a little since these sweeps (mostly due do a new topload), but I can do some new ones if it is of interest, this time with a log axis for impedance as well.
-
Now for the inverter IGBTS
I have narrowed it down to two
IRGP50B60
And fgh60n60smd
Now irgp has very low rose and fall times but at the same time costs twice of 60n60
So, is the lower losses worth it?
Also, could yall suggest something better IGBT if any
-
Price, low thermal impedance and past success all suggest going for the 60N60 (or the equivalent FGA60N65SMD if cheaper). FGY75N60SMD is also an option for slightly higher current (I have a thread about some very cheap ones I got from aliexpress).
-
Price, low thermal impedance and past success all suggest going for the 60N60 (or the equivalent FGA60N65SMD if cheaper). FGY75N60SMD is also an option for slightly higher current (I have a thread about some very cheap ones I got from aliexpress).
hmm...i see
-
@Hydron:
Your measurements all look fine for themselves. I just can't make them to fit to the coupled system measurement. Possibly its the new topload or some influence of the primary (capacitive or whatever) on your secondary alone mesurement.
-
OK, yeah that is a possibility. I'll make some more measurements and document them better - those were from a few years back.
-
Using the formulas off of Richie burnett''s website
My coil will have the upper pole at around 500-520khz
Doesn't that seem too high?
How will the coil ever get around to achieving upper pole operation?
Also, how can one use a startup oscillator with a ud2.7
I'm looking forward to using Steve ward''s tuning trick for QCWs
-
So, for the MMC
Will CBB capacitors work?
I'm thinking of something like 100 caps 10 in series 10 in parallel to give about 10nf of capacitance
https://www.ebay.com/itm/10Pcs-CBB-Capacitor-100V-101-683-0-1nF-68nF-101J-683J-Pitch-5mm/192143864676?hash=item2cbcaab364:m:miP5zeN54Fp-RSicta7nMyQ (https://www.ebay.com/itm/10Pcs-CBB-Capacitor-100V-101-683-0-1nF-68nF-101J-683J-Pitch-5mm/192143864676?hash=item2cbcaab364:m:miP5zeN54Fp-RSicta7nMyQ)
Something like these capacitors...
GUESS WHO JUST FOUND SOME DATASHEETS!!!!
looking at the datasheets even though the peak current ratings are for <100khz
the RMS ratings look quite sufficient
what do you guys think??