User-agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:24.0) Gecko/20100101 Thunderbird/24.2.0
Em 31/01/2014 01:21, romsigurthr@xxxxxxxxx escreveu:
It turns out my formula for determining maximum cycles was off by a factor of 2 for a half-bridge - meaning that it would apply for a full bridge. So I can actually run for twice the burst length that I thought without exceeding the primary capacitor voltage rating.
Vcap = (1/2 * Vbus) * ((2N)-1) for a half bridge, ignoring losses and transfers, where N = number of half cycles.
The tuning that I use produces the first output voltage rise after the
end of a full beat of the primary current. The current envelope is
sinusoidal, not straight. When streamers appear the beat doesn't
complete, but the current peaks remain approximately constant around the
maximum value.
That is quite a nifty calculator program you have there! A bit all over the place as far as the GUI goes but I think I managed to figure it out with enough experimentation. I put my coil’s parameters in and came up with about 16 cycles burst length before VCa (primary cap’s voltage) gets too close to the rated voltage for comfort. The difference in projected primary circuit amplitude between your calculator and my formulae comes down to the fact that my calculations don't take into effect transferred power or loading or detuning. Mine aren’t meant to be an accurate model so much as a worst-case-scenario design guideline. After all, figuring out the necessary voltage rating for primary capacitor isn’t as straight forward for a DRSSTC as it is for a SGTC. Here’s a screenshot of my results: http://imageshack.com/a/img42/8379/a73d.jpg
A straight rise appears when there is no secondary coil. It's really the
worst case. I still have to include streamer loading in the simulator.
That is quite strange behavior with the 1MHz noise, but good to hear that ferrite beads clear it up.
The beads become quite warm. A common-mode choke at the power input also
helps to eliminate the oscillation.
There’s significant explosion/fire hazard in overvolting the capacitor in a resonant system, do be careful!
My mmc is assembled in an acrylic board inserted in a PVC tube. I hope
that the tube can contain an explosion. So far I have not exploded or
damaged any capacitor.
About your (half) bridge; are you using Schottky isolation diodes and external freewheeling diodes to take the body diodes out of the circuit? I’ve found that they make tremendous differences in the longevity of the bridge and its ability to tolerate adverse conditions. While my site is intended for the SSTC, the bridge design and info there is solid for DRSSTCs and Induction Heaters as well, so I invite you to take a look. http://sigurthrenterprises.com/the-usstcc.html If you would like a GDT core suggestion one can also be found there in the BOM file for my USSTCC board. I’ve used this core to drive four FDL100N50F mosfets (about 60nF of total gate capacitance!).
I am using ultrafast diodes in parallel with the mosfets, but nothing in
series. Curiously, in a first assembly I forgot to mount the diodes. The
system operated well, but at the energy return after each burst there
was a delay between the inverted current and the bridge output voltage.
I then mounted the diodes expecting a smaller delay, but the waveforms
didn't change in nothing significant. Maybe the intrinsic diodes of the
mosfets are fast enough, or my diodes are too slow.
I installed new mosfets to replace the others that were damaged, but the
new waveforms are not as good as they were. There are high-frequen cy
oscillations at the transitions, specially in the energy return. Anyway,
new full power tests will be done only after I make a better GDT. I am
using a different from usual GDT, with a center-tapped primary. Trying
to use one of several toroids scavenged from various sources that I have
accumulated, measuring their characteristics.
