Small DRSSTC CT problems / Tesla Coils / Forums

Forums

4hv.org :: Forums :: Tesla Coils

« Previous topic | Next topic »

Small DRSSTC CT problems

Move Thread

LAN_403

Extreme Electronics

Thu Nov 26 2009, 05:52PM

Registered Member #74 Joined: Thu Feb 09 2006, 09:17AM
Location: Nottingham UK
Posts: 99

All,
Whilst building my last small DRSSTC I found a problem with using a
Current transformer as a primary current phase sense with the DRSSTC at
low power. The Inductance of the CT can become a substantial part of the
total inductance of the primary circuit. This causes two problems. Loss
of power in the primary and tuning is difficult as the CT / sense
circuit's inductance is not linear WRT current.

I have been looking into making another coil which will operate at
much higher frequencies (1 Mhz or so) so the problem is much worse.

Doing some simulations of the primary the biggest factor in
minimizing the problems is the coupling in the CT. With a high coupling
the impedance of the CT's secondary is reflected back to the primary and
almost removes the affect of the inductance of the CT.

At present I have been simulating in LTSpice using a 1:100 turn CT wound on an
FT37-43 toroid (Al=350) on the basis that the smaller the ferrite, the
smaller the inductance in the first place.

My question is how to wind a CT to maximize the coupling, assuming that
the primary will have a single, or at most two turns.

or

Would a rogowski coil be a better option as its influence on the primary
circuit would be much less (assuming I can get enough output to extract
a usable signal)

Any Comments welcome...

Derek

GeordieBoy

Thu Nov 26 2009, 06:30PM

Registered Member #1232 Joined: Wed Jan 16 2008, 10:53PM
Location: Doon tha Toon!
Posts: 881

An interesting problem!

I would get the highest-perm Mn/Zn core that you can find, and then design the CT to have the absolute tightest coupling possible. (If the pri is one single turn, then space the sec turns evenly around the core in a single layer. If the pri is more than one turn then wind it evenly spread over the top of the sec around the core.)

A single pass through the CT really shouldn't add that much inductance to the conductor _PROVIDED THAT_ the output of the CT is terminated into an "almost short-circuit".

As you said the CT just reflects it's burden impedance back to the primary side conductor that passes through it, (scaled by the turns-ratio squared.) This reflected burden impedance appears in parallel with the CT's own magnetising inductance (the one that's causing you all the problems!) So on that basis, if you make the burdern impedance extremely small, then the impedance refelected back and to the primary side will be even smaller and should adequately shunt the magnetising inductance of the CT.

The key thing here is making sure the impedance of the burden on the CT's secondary is _ALWAYS_ extremely low. If you can terminate the CT into a 1R resistor then that is good because it will always be 1 ohm for small signals and for large. If you can terminate the CT into an active current sense amplifier (TIA) then this is even better as it approximates a short-circuit to the CT's secondary.

Problems are likely to occur where the CT is terminated into something involving diodes! For very small signals these diodes will not conduct and the CT secondary is effectively open-circuited! This would cause the CT's magnetising inductance to appear at the primary side in all it's glory :( As the signal level increases the diodes will gradually conduct and the effective burden impedance will fall, gradually making the primary side look less and less inductive.

So if you must use diodes then try to minimise the voltage drop by using schottky's and ideally only one of them instead of two. You want the maximum voltage on the secondary of the CT to be clamped at as low a voltage as possible. (Whatever voltage develops across the secondary of a CT sets up a magnetising current and ultimately leads to "droop" in the sensed current measurement anway, so it is undesirable.)

I hope that helps!

-Richie,

Nicko

Thu Nov 26 2009, 08:48PM

Registered Member #1334 Joined: Tue Feb 19 2008, 04:37PM
Location: Nr. London, UK
Posts: 615

Could you use a small burden as suggested but an active rectifier instead of passive schottkys?

Extreme Electronics

Fri Nov 27 2009, 12:06AM

Registered Member #74 Joined: Thu Feb 09 2006, 09:17AM
Location: Nottingham UK
Posts: 99

Thanks All,

I hadn't thought of using a different core material, Ill look into that and using a TIA, Any pointers in using them would be appreciated.

My idea to get the best out of a core was to have a small'ish burden resistor and a high speed op-amp to get the signal back to logic levels.

If I have at least a 1:100 ratio and use a very small core to hopefully to keep the coupling high.

As the purpose of the diodes is purely to clip the resulting signal I should be able to remove them completely or put them after the opamp.

Of course the problem with op-amps is that the circuit may become more susceptible to noise as the CT will probably never have more than 100A to sense, so the voltages involved are very small (1V=>1Ohm max) and of course the local environment may be rather electrically noisy (with any luck !)

I had thought of using hall effect chips, but again the signal strength is very small and most hall effect chips have a roll off around 500khz.

Derek

Extreme Electronics

Tue Dec 01 2009, 10:51PM

Registered Member #74 Joined: Thu Feb 09 2006, 09:17AM
Location: Nottingham UK
Posts: 99

Well, half the problem is solved.

I tried a small core with 100T and there was a marked difference, the inductance visibly dropped, but the output from the CT was much smaller, probably due to the smaller Al value of the core. I put this down to the better coupling of the smaller core (e.g. the CT primary and secondary are physically closer together) and the smaller Al value of the core. I deliberately kept the wiring of the coil EXACTLY the same using both cores so I could make the comparison.

So I rewound the core with 200T of 0.125mm wire, which left a smallish hole for the primary wire (2.5mm Dia). I'm still using a pair of 5V clamping diodes and a TTL gate, but the results are hugely better, I don't have any low power tuning issues and my primary Cap (at resonance) has gone from a ~0.001uF to a 0.047uF. (JavaTc says 0.07uF but JavsTc isn't very accurate with very small coils.)

Picture of final coil with a Â£1 coin for comparison

Is it just me do you get the urge to cut all the way around a fully wound toroid and make a pom-pom??

I need to see how this performs at higher powers.

I doubt I will go for an op-amp solution anymore as all the immediate issues are solved. I may need an op-amp if I bother to put in some form of OCD, but again, the solution could just be loads of turns and a small core.

Thanks for all of the comments.
Derek

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