 Ideas for alternative circuits for DC powered SGTCs
|
|
|
monokel
|
|
Wed Sept 29 2010, 05:24AM
|
|
|
Registered Member #2981
Joined: Thu Jul 08 2010, 01:47PM
Location: Germany
Posts: 35
|
I have two alternative circuits for DC powered SGTCs.
This circuit combines a rotary for charging and a triggered static spark gap in the primary oscillating circuit so that the current through the inductor is zero when the static SG fires; so quenching is no problem. R1 and C1 should be dimensioned so that the triggering voltage (10 kV) is not reached before all the energy of the inductor L1 has been transferred to the cap C1. Thus the current through L1 is zero and it doesn't rise while the SG is conducting.
One cycle is as follows: The rotary resp. switch is conducting and C1 gets charged via L1. Short time before the current through L1 gets zero the switch opens and the diode gets conducting. Short time later the current is zero. At this point of time C1 has twice the supply voltage (20 kV); this is not enough for the static spark gap to fire. Next, C2 has been charged via R1 up to 10 kV; this is enough for the triggering spark. Thus the whole spark gap gets conducting and the oscillation begins. When C1 has been discharged and the rotary conducts again, the cycle begins from scratch.
One advantage of this circuit is that there is less current stress in the rotary and that "arc trailing" will probably not occure. Disadavntages over a rotary beeing in the primary oscillating circuit are: The arc in the rotary consumes power as well as the static gap. The apparatus may be more difficult to build, ...
This circuit makes use of Terry Fritz's SIDAC IGBT spark gap modules (SISG). If you want to build a high power SGTC with them you may want to use rectified three phase current; this has the advantage of a high power factor compared to a one phase rectified DC. A primary oscillating circuit with a SISG cannot simply be connected to an (ideal) DC source via an inductor; otherwise there would be very large currents.
I tried to design a triggering mechanism by disturbing the equal voltage distribution between the SISG modules.
Power IGBT modules like the FUJI 1MBI600PX-140-01 have quiet large parasitic capacitances between emitter and collector (some nF); I added them in my schematic. Thus the resistors for equalizing the voltage distributions must be lower than those in Terry Fritz's schematic.
The circuit works as follows: C1 is charged via L1 until the current through L1 is zero. The voltage across C1 is now twice the supply voltage (10 kV). Until this point of time all SIDACS and IGBTs are nonconducting. This voltage of C1 isn't big enough for the SISGs to ignite. A short time later C2 has been charged via R1 up to 3000 V. Then the series connection #1 of SIDACs gets conducting. This causes the voltage of the series connection #2 of SIDACs to increase so than it also conducts. Next, the series connection #3 of SIDACs and the series connection of SISGs #1 and then #2 get conducting. Thus the oscillation begins. After the cap C1 has been discharged and the SISGs have got nonconducting the cycle begins from scratch.
However, while the SISGs are conducting the current through the inductor rises again. The ON time of the SISG must not be too large so that this current doesn't get too high. Remember that the ON time of a SISG can be arbitrarily adjusted.
Terry Fritz writes is his PDF paper that at higher BPS rates the streamers don't get longer. Maybe this problem doesn't occure with my circuit because the current through the inductor doesn't get unnecessarily high. With my circuit, a BPS rate of a bit more than 600 BPS and a power of more than 6 kW will be reached.
I would be interested in your opinion about these circuits.
|
|
Back to top
|
|
Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob
|
|
Powered by e107 Forum System
|
