smaller coils (and a 12MHz target) / Tesla Coils / Forums

Forums

4hv.org :: Forums :: Tesla Coils

« Previous topic | Next topic »

smaller coils (and a 12MHz target)

first previous 6 7 8 9 next last

Move Thread

LAN_403

Avalanche

Mon Oct 02 2006, 12:50PM

Registered Member #103 Joined: Thu Feb 09 2006, 08:16PM
Location: Derby, UK
Posts: 845

I've never seen surface mount components mounted on the edge of the board

Wilson

Mon Oct 02 2006, 02:25PM

Registered Member #78 Joined: Thu Feb 09 2006, 11:27AM
Location: Sydney, Australia
Posts: 133

I guess functionality comes before aesthetics

teravolt

Mon Oct 02 2006, 07:40PM

Registered Member #195 Joined: Fri Feb 17 2006, 08:27PM
Location: Berkeley, ca.
Posts: 1111

does it work

Marko

Mon Oct 02 2006, 08:25PM

Registered Member #89 Joined: Thu Feb 09 2006, 02:40PM
Location: Zadar, Croatia
Posts: 3145

teravolt wrote ...

does it work

I guess it does but not as well as it could if he etched some board and placed those UCC's closer and nicer..
The second one looks somewhat better though..

...

Mon Oct 02 2006, 10:51PM

Registered Member #56 Joined: Thu Feb 09 2006, 05:02AM
Location: Southern Califorina, USA
Posts: 2445

hehe, good to see my 3oz copper pcb getting some use, since it is too thick to be chemically etched worth a crap

And since the ucc's seem to be able to drive the fet at 17mhz, you should have no problems getting the 12mhz coils of yours working?

Do you think it might help a little to put some small smd ceramics right across the power pins to ground?

ragnar

Sun Oct 08 2006, 09:35AM

Registered Member #63 Joined: Thu Feb 09 2006, 06:18AM
Location:
Posts: 1425

Those were SMD tantalums across the rails... the underside of the second style of board is the +ve plane. After drilling and countersinking the top/gnd side, I made feedthroughs, which you see connecting to pin1 and pin8 on the chip (and also the tantalums)

17MHz?
Well, that's the thing... the UCCs do "work" at 17MHz... but I discovered (it was no surprise) that they're not nearly reliable enough to be usable. Note that even with no load, after several minutes they get stinking hot, until they latch or go into runaway. Add a load, and it's hopeless after about a minute. The bipolars in the output stage just aren't fast enough.

Lower that to 13.56MHz, and the output waveforms are very shakey... I note specifically that only the UCCX7322 (non-inverting) is useful at this frequency. The output inverting UCCX7321 doesn't quite get to the supply voltage, and the output pulse is too slopey, and only occupies ~25-30% of the cycle!

After much twiddling and fidgeting, I was able to find an adequate solution.

The first pic is a breadboarded driver, @ 13.56MHz (RC osc.) into an IRF510.
The second pic is the veroified driver, @ 12MHz (crystal) into an IRF510.

I bias the gate at 6V, and the resulting swing is about +14 to -2V.

The IRF510 will provide the gatedrive for a larger FET.

Marko

Sun Oct 08 2006, 01:35PM

Registered Member #89 Joined: Thu Feb 09 2006, 02:40PM
Location: Zadar, Croatia
Posts: 3145

Yup, i'm afraid that for very high frequencies you'l have to goodbye the UCC's no matter your emotional devotions to them.

R. burnett used simple transistor push-pull to drive first class E stage, and it's not problematic at all since sinewave is fine for class E, and with resonant drive little power is required to get optimal drive.

Probably any small complementary pair would do.

Can you hook up the upper setup to an IRFP450 to see if it would work? =)

ragnar

Sun Oct 08 2006, 10:56PM

Registered Member #63 Joined: Thu Feb 09 2006, 06:18AM
Location:
Posts: 1425

wrote ...
Yup, i'm afraid that for very high frequencies you'l have to goodbye the UCC's no matter your emotional devotions to them.

Never!! My emotional attachment absolutely and overtly requires me to discover the secrets to getting these drivers to work at stupid frequencies three orders of magnitude higher than they were intended for. I've already disclosed one of them; that you may only use the noninverting driver, UCCX7322 if you want half-reasonable waveforms and a close-to-50% duty cycle output for a 50% duty cycle input.

wrote ...
R. burnett used simple transistor push-pull to drive first class E stage, and it's not problematic at all since...

At 4MHz with 12V drive, fine, yes. I don't have an easy way of getting a nice 12V drive to these little transistors at 13.56MHz. The CD4024B divider dies at 9.5MHz out (19MHz input).

wrote ...
with resonant drive little power is required to get optimal drive.

I think you'd be surprised... I'm probably looking into losing at least 10W into the gate, and I haven't even started di/dt- and dv/dt-ing it all over the place yet.

wrote ...
Probably any small complementary pair would do.

Trust me, been there, done that, I found it too sensitive, unreliable.

Can you hook up the upper setup to an IRFP450 to see if it would work? =)
The gate charge of an IRFP450 is 75nC. The gate charge of an IRF510 is 5nC. It is not possible to swing 2600nF at 13.56MHz with a single UCCX7322.

Usings TWO UCCs with an 11:3 stepdown transformer with a resonating secondary and 30Vp/p across the primary, I found it impossible to exceed ~6.5MHz, even when the UCCs were heatsinked.

And even if a single IRFP450 could be driven at 13.56MHz with a single UCCX7322, that takes all the fun out of it, so I'd do it another way anyway

Marko

Mon Oct 09 2006, 10:03AM

Registered Member #89 Joined: Thu Feb 09 2006, 02:40PM
Location: Zadar, Croatia
Posts: 3145

The gate charge of an IRFP450 is 75nC. The gate charge of an IRF510 is 5nC. It is not possible to swing 2600nF at 13.56MHz with a single UCCX7322.

Å¾

Oh, I meant to drive '450 with with IRFP510's output

10W isn't really too much power, when you think what would it look like with hard drive at that frequency

IIRC richie needed about 5W at gate for his coil at 4Mhz.

Steve Conner

Mon Oct 09 2006, 10:25AM

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

I was always amazed that Matt could get the UCC37322 to work at such high frequencies as he did. UCC's use a double output stage with MOSFETs and BJTs in parallel, and BJTs take a finite time to turn off, so I'd have expected huge amounts of shoot-through leading to overheating and destruction.

I'm thinking that he probably gets away with it as long as he draws little enough current that the MOSFET part of the output stage takes all the load and the BJTs don't conduct.

Firkragg: Even with resonant drive, you need to supply a fair amount of power to overcome losses in the MOSFET's gate resistance, since the gate electrode is made out of a poorly conducting crappy polysilicon. 10W of gate drive at 14MHz sounds in the ballpark, for an IRFP460 sized device, and I'd expect it to rise to 25-40W at 30MHz, at serious risk of burning out the gate structure.

It should still be considerably less than if you used hard drive and burnt 1/2*Cg*Vg^2 of energy at every cycle.

first previous 6 7 8 9 next last

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