Active clamp network - non-ZCS SSTC feedback method / Tesla Coils / Forums

Forums

4hv.org :: Forums :: Tesla Coils

« Previous topic | Next topic »

Active clamp network - non-ZCS SSTC feedback method

previous 1 2

Move Thread

LAN_403

Steve Conner

Tue Aug 04 2009, 10:05AM

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

LithiumLord wrote ...

it's pretty hard to get a CT to operate well on very low currents.

No it's not. My induction heater worked fine with the same CT feedback circuit on the output from the half bridge, which was about 8 amps of RF. Then again, maybe there is a phase lag, but I compensated it without noticing when tuning the PLL.

In any case, you can just lower the ratio of the CT. For secondary base current, you might want, say, a 1:1 with 10 primary turns and 10 secondary turns.

About the dissipation: Of course it's negligible compared to the power being fed to the coil. But the circuitry needed to dissipate it adds complexity and cost. You used 20 amp diodes, a MOSFET and a UCC gate driver chip, whereas I argue that you only need the diodes, and they don't need to be that big.

If the hysteresis on a 74HC14 is too much, use something with less hysteresis. Like a 74AC04, or even a gate driver chip. Or use two sets of two diodes to double the burden voltage.

I'm sure your circuit works fine, but you said you wanted a debate

LithiumLord

Tue Aug 04 2009, 03:36PM

Registered Member #1739 Joined: Fri Oct 03 2008, 10:05AM
Location: Moscow, Russia
Posts: 261

>you said you wanted a debate
Yeah, as debates are where the truth is born (double(?) semi-accurate xlation of some quote :P)

>8 amps of RF

Well, the goal was improving the non-DR coils, meaning we can not use the primary feedback. And even the peak secondary currents do never reach that high, while on the ringup they run like a few mA only - there's no good way to deal with those with a CT ;) Also there's no use for PLL in an interrupted coil. Indeed, first of all the PLL needs a few cycles to lock which is unacceptible with the constant frequency shifts there. PLL usage will require higher rated switches and still not allow for a fast ringup. Also the other fact - with the PLL you can set up a phase correction, but you are missing another problem here - the CT during the ringup sequence floats through different lag values, meaning it's usage with a CT is unacceptible but in CW devices. From this point, by the way, the controller/driver/mosfet lag is more acceptible.
Also one other point to be considered. Some people prefer using 100Hz-modulated supply, and the CT will perform poorly on the lower input voltages.

>For secondary base current, you might want, say, a 1:1 with 10 primary turns and 10 secondary turns.

And then we get a full current to be clamped, also with the fact the control electronics are grounded anyway so we can just add the couple of antiparallel diodes as a clamp just directly at the secondary base. However the long signal slopes will render us unable to use the histeresis (made a few low-voltage scope tests to find that out), making the noise issues possible. By the way, I also considered using serial diodes as a "Zener", but the bad side of this trick is that diodes are "too non-linear", and therefore the clamp amplitude will float from 0.5v/diode to 1.5v in some cases! Meaning the power dissipation will exceed the one of the active clamp (a drop in a sea anyway, but still). Also the signal shape will remain in not the best condition anyway.

>You used 20 amp diodes, a MOSFET and a UCC gate driver chip

Yep, but that was what I used ;) Eg the schematic presented is no option for a replication, it just depicts the actual device I assembled, with the parts I used there as I didn't want to go to a store but had the parts listed scattered around. In a construct to be used in someone's coil the list can be heavily modified. The usage of the 20A diodes is an overkill, a cheap 1A one is suitable for most coils. The MOSFET can be replaced with a cheap low-current one, and UCC can be replaced with a cheap low-current driver or even with a single Schmitt gate along with, probably, an output stage and a TTL FET. Even price-wise it will be cheaper then a fast comparator. However what is nice in this is the fact you can just add two more UCCs, a GDT and a halfbridge and immediately end up with a simple halfwave-fed SSTC that will perform well even without any startup generator.

Firnagzen

Wed Aug 05 2009, 02:22AM

Registered Member #567 Joined: Tue Mar 06 2007, 10:55AM
Location: Singapore
Posts: 147

LithiumLord wrote ...

As the counterexample form the last post I used a shunt of two antiparallel diodes as is clear reading down the thread ;) For topology comparison, tuning tricks, tests and design notes see the first post here ;)

Whoops, my bad.

previous 1 2

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