Limits of using 12V with HV Mazzilli driver ? / High Voltage / Forums

Forums

4hv.org :: Forums :: High Voltage

« Previous topic | Next topic »

Limits of using 12V with HV Mazzilli driver ?

Move Thread

LAN_403

Hel

Wed Mar 27 2013, 03:45PM

Registered Member #11213 Joined: Fri Mar 15 2013, 05:35PM
Location: Italy
Posts: 3

Hello everybody! I'm new here, so I hope you will be kind and patient...

To date I gained quite some experience and knowledge with electronics, but however
I'm not an electronic engineer, and I have everyday something new to learn and to
discover. All my experiments are aimed to learn and to understand things.

SHORT QUESTION:
I'm now trying to understand which and where the intrinsec limits are of using
just 12 V to power HV supplies for high voltages AND high powers. It's becoming
a kind of challenge to me: partly because I need to build a portable HV supply
for scientific experiments outside (no weapons of course !!!), and having just
one 12 V battery to carry around is for me simpler than having two or three;
partly because, as I wrote, I like understanding things, and I can't feel good until
I haven't got to the point !
As output, indicatively about 30 KV AC with a power in the range of 50-100 W would
be fine for me.

DETAILED QUESTION:
As most people did, I firstly tried the simple single transistor flyback driver, then I
built one Royer with two 3055s driving one AC (not rectified) flyback, a scheme
which I tried to optimize the best I could with the aid of my scope and lots of
patience. I ended up with a design using the resonant topology and the single (not
center-tapped) feedback winding pushing-pulling the transistors bases directly, each
one with a 100 Ohm pull up resistor and fast protection diodes between collectors and emitters. With 12V input, a 680 nF resonant capacitor, 2+2 turns primary and 2 feedback
turns, it produces a nice sinewave with arcs starting at about 10mm (I think about 10 KV)
and which can light a 15 watt incandescent bulb at full brightness. I couldn't do anything
better with all my efforts to improve voltage AND power or just voltage alone.
Then I was bored of the inefficiency of BJTs. To drive them hard on, large base
currents are needed, and the heat you perhaps save in switching is however lost
thru the base current and the limiting resistors. So I decided to switch to Mosfets
and I built my first Mazzilli driver (following the standard schematic everyone knows).

Well, I must say I felt a bit deluded

I expected it allowed me to pump more
voltage and power after fine tuning, but it didn't. Now, let's discuss the details of it.

As I specified, I aim at driving it at 12 V, no more.
My supplies aren't definitely a problem, since I'm using lead-acid batteries or
a 400W PSU (for bench tests), and I'm using large and short wiring which doesn't seem to
heat up.

For the Mosfets I used two IRF540N - my local shop has not the IRFP250, but read on:
they have quite good specs the same: 100V Vdss is more than enough at 12V
(at any frequency, no more than about +-40 V peak are circulating in my tank), also
the RdsOn 44 mOhm looks very fine, and Id 33A could theoretically be enough for
me. They always ran quite cold with just a little heatsink.

I had to lower the gate charging resistors from the 470 Ohms suggested to 150 Ohms
to have a nicer and steep square gate signal - perhaps I could decrease them even
further, to have a steeper drive at higher frequencies - what do you think?

As discharging diodes I used two BYV29-400 which I had around and which should be ok
(400V, 9A, 60ns).

For gate protection I placed two 13V zeners I had around. But I think that upto
20 V supply voltage (not my case though) I could even avoid them (Vgs is Â±20V for
the IRF540N)

As output transformers I tried many different flybacks (rectified and not,
airgapped and not, even a couple of identical capacitive discharge coils put in
series)... but in every case I really can't manage to pump out more than perhaps
20 W of power

That being always with a primary of 5+5 turns.

I tried with different toroids in the power line, ranging from some 60 ÂµH to even
1 mH and thick wire, I tried both a ferrite and a green toroid from a PC PSU, what they
affect is the shape of the tank current/voltage waveform, but definitely not (visibly) the
performance. I of course set it to always obtain a nice sinewave.

The resonating capacitors I use are all MPP or MKP, they don't heat up substantially.
In general I can lower the capacity to obtain higher frequency, higher voltage
and lower power at the output, or the opposite. But to some extent. In general, it
looks like I can't go below some 100nF, I get parasitic oscillations otherwise, but
even at the highest frequency I can reach, the output voltage never exceeds 10-15 KV
(at eyesight...). Conversely, if I increase the capacity, the output voltage
decreases but the power increases. I can light a 25W incandescent bulb at almost
full brightness but to do that I need to put at least 2 ÂµF - and I don't know if
increasing it further can help so much. The best I could achieve up to now is by
dareing with a primary of as low as 2+2 turns and 2 ÂµF capacitor: this way I can light
a 25W bulb at almost full brightness with an arc which starts at some 7-8 mm.
Morale: it looks definitely we ended up with plain transformer rules here, where
we have sinewave currents and voltages, and the step-up factor is given by pri/sec
ratio... how much am I right here ?

Well, what should I do to get an higher output voltage possibly with even a
higher power, let's say, of 50 or 100 W, without increasing the supply voltage ?
It's out of question that my supplies can be a limit. They can actually power a 100 W
12V lamp without any substantial drop ! Here it's just a matter of transforming voltage.
But, even searching around the Internet, it always looks that people wanting more power,
or more voltage, from their HV supplies tend to increase the supply from 12 to 24, 36 or
even 50 V! - Something that seems applying to Tesla coils also...
Which is the intrinsic limiting factor of using just 12 V ??
Should I really use/find/wind a secondary with more turns than usual flybacks have,
to get more output voltage at the power established by the capacitor ? And just,
what about power ? Have I actually to increase brutally the capacitance to, say,
10 or more ÂµF to reach 50 or 100 W, or is there any limit ?

A suspect then came to me: is it perhaps a matter of secondary resonance ?

Perhaps I should really tune the primary tank to match the frequency of secondary,
so that all energy can be transfered to it without reactive current, pretty like in
a Tesla coil ? Well, if it's so, I see it hard, since this secondary resonance
(I tried) depends heavily on load and even just on how the transformer is
placed with respect to ground and other subtle, unpredictable details...
Also, how current and voltage are intended to be in the tank ? In or out of phase ?
I suppose out... unfortunately I tried to measure and show both on my scope but
that's definitely not so easy, since I have not a differential probe. Leaving the entire
circuit floating (battery powered) I used my two channels scope with a shunt resistor,
one probe connected across the shunt, the other to the other tank's leg, then
I inverted one of the channels (I assume this should be the proper way...) - this
way I can monitor tank voltage and current but I suspect it's not a reliable method,
since I noticed different phase offsets depending on the shunt used, perhaps it's
interfering with the circuit around causing some little mess...

Thank you in advance for any constructive comment or suggestion...

Steve Conner

Wed Mar 27 2013, 04:20PM

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

The transformer turns ratio limits the voltage, and the leakage inductance between primary and secondary of the transformer limits the maximum output current.

The problem is that as you reduce the number of primary turns, the leakage inductance increases, because the primary gets physically smaller, leaving more area for leakage flux to pass around it.

You can redesign the transformer. To get high output power at 12V you need a winding of only 1+1 or 2+2 turns with extremely low leakage inductance, and you can get this by tricks such as primary on the same core limb as the secondary, copper foil instead of wire etc. You also need very low leakage inductance between the two half-primaries. Look at the transformers inside car audio amps for inspiration.

Alternatively you can try to tune the leakage inductance out by resonating it with the various capacitances in the system.

Hel

Wed Mar 27 2013, 07:37PM

Registered Member #11213 Joined: Fri Mar 15 2013, 05:35PM
Location: Italy
Posts: 3

Steve Conner wrote ...

The transformer turns ratio limits the voltage, and the leakage inductance between primary and secondary of the transformer limits the maximum output current.

Steve, thanks for the quick and very explanatory reply !
Really I should have theorized something similar, but I simply haven't yet the necessary skill.

Steve Conner wrote ...

You can redesign the transformer. To get high output power at 12V you need a winding of only 1+1 or 2+2 turns with extremely low leakage inductance, and you can get this by tricks such as primary on the same core limb as the secondary, copper foil instead of wire etc. You also need very low leakage inductance between the two half-primaries. Look at the transformers inside car audio amps for inspiration.
Alternatively you can try to tune the leakage inductance out by resonating it with the various capacitances in the system.

I have not access now to such a car audio trafo but yep, I think I got the concept.
I will eventually try to omptimize my flyback, the flat coil seems a good idea.
But I would like also to try to resonate my flyback first, as you suggested (and as I partially
anticipated in my first msg).

The problem is, as I already wrote, that I can't so easily play with the resonating frequency,
because by lowering the capacitance to some 120-100 nF I start getting parasitic oscillations
(read: a tiny initial spark, then the system starts running crazily to some very high frequency
and I suddently turn it off for fear that some components can be damaged).
What do you think is, in general, the cause of such behaviour ? Maybe decreasing the gate
driving resistors furtherly could help here ?

Also - unless my shunt isn't going mad or I messed up with something - I measure a current
of about 22 A circulating in the tank with no load, which increases to 75 A with the little
15W bulb connected (1 ÂµF cap, 2+2 turns) ! Does it sound to you as a sensate amount ?
Nonetheless my Mosfets stay quite cold, and are rated for just 33A (110A pulsed)... how is
that ?
Thanks...

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