Possibly the end of shootthrough in SSTC's
Teodor Zafiroski, Thu Mar 26 2015, 07:36PMAnybody who has ever tried to build or actually built an SSTC, knows that half bridge and full bridge designs are prone to shoottrough. This problem becomes rather expensive when you use more expensive transistors. This problem can be solved using dead time between the turn off of one MOSFET and the turn on of the other. This forces the engineer to implement some sort of a circuit that ensures a long enough dead time to effectively protect the switching elements.
   Many gate driver chips have built in dead-time, but when using devices with slow switching speed, this dead time may not be sufficient.
   At a time, I was thinking about solving this problem for good, sort of a universal short-circuit protection for SSTC's using half and full wave inverter topologies. I came up with this (simplified) design...
Here is how the two interwound coils would look in real life (again simplified).
Here's a schematic representation of the coil's magnetic field when the current flows through the red coil:
Here's a schematic representation of the coil's magnetic field when the current flows through the blue coil:
This could be implemented in full bridge topology using a system like in this schematic.
In this case the only path the current could take is through a coil.
So, my question is:
-Could the improved half bridge schematic work, and if it could, what would happen if the two switching elements were on simultaneously?
-Could the improved full bridge schematic work, and if it could, what would happen if the four switching elements were on simultaneously?
-What an effect could the current flowing through the both coils (in both improvements) have on the rest of the schematic?
Al comments are welcome.
Re: Possibly the end of shootthrough in SSTC's
Wolfram, Thu Mar 26 2015, 07:59PM
This could work in a zero-current switched application like an SSTC. In cases where you don't have ZCS, the MOSFETs will have to avalanche any energy stored in the leakage inductance between L1 and L2 on every switching cycle, giving huge losses, I think. It could be an idea to add some diodes to the full bridge one, to clamp the two primaries between the supply rails during dead-time, basically making it like two magnetically coupled two-switch forward converters. This will return any energy that's stored in the leakage inductance back to the DC bus capacitor.
In the case where both switches on the same side are on at the same time, there will be something like shoot-through, but the dI/dt will be limited by the leakage inductance between the primaries.
I haven't really had much trouble with shoot-through myself, even when using GDTs. For most of my projects lately, I've used the Silabs SI823x series of isolated gate drivers, they provide a wide range of dead-time, set by a single resistor. I haven't used them in an SSTC yet, but it's only a question of time.
Wolfram, Thu Mar 26 2015, 07:59PM
This could work in a zero-current switched application like an SSTC. In cases where you don't have ZCS, the MOSFETs will have to avalanche any energy stored in the leakage inductance between L1 and L2 on every switching cycle, giving huge losses, I think. It could be an idea to add some diodes to the full bridge one, to clamp the two primaries between the supply rails during dead-time, basically making it like two magnetically coupled two-switch forward converters. This will return any energy that's stored in the leakage inductance back to the DC bus capacitor.
In the case where both switches on the same side are on at the same time, there will be something like shoot-through, but the dI/dt will be limited by the leakage inductance between the primaries.
I haven't really had much trouble with shoot-through myself, even when using GDTs. For most of my projects lately, I've used the Silabs SI823x series of isolated gate drivers, they provide a wide range of dead-time, set by a single resistor. I haven't used them in an SSTC yet, but it's only a question of time.
Re: Possibly the end of shootthrough in SSTC's
Teodor Zafiroski, Fri Mar 27 2015, 07:49PM
Teodor Zafiroski, Fri Mar 27 2015, 07:49PM
Wolfram wrote ...
In the case where both switches on the same side are on at the same time, there will be something like shoot-through, but the dI/dt will be limited by the leakage inductance between the primaries.
I had a wild thought that the opposing magnetic fields created by the two coils would resist further flow of current. So, back to the drawing board...In the case where both switches on the same side are on at the same time, there will be something like shoot-through, but the dI/dt will be limited by the leakage inductance between the primaries.
Re: Possibly the end of shootthrough in SSTC's
omegalabs, Sat Mar 28 2015, 12:00PM
Check out skory's mini tesla from 2005: Link
Separated primaries can work really well.
omegalabs, Sat Mar 28 2015, 12:00PM
Check out skory's mini tesla from 2005: Link
Separated primaries can work really well.
Re: Possibly the end of shootthrough in SSTC's
Ash Small, Mon Mar 30 2015, 04:52PM
Doesn't a full bridge become redundant with this setup?
Surely you only need one switch in each leg. You can replace the high side switches with diodes to block back EMF or whatever it's called, can't you?
You only need a full bridge when using a single coil, that's the point of it, isn't it?
Or have I missed something?
Ash Small, Mon Mar 30 2015, 04:52PM
Doesn't a full bridge become redundant with this setup?
Surely you only need one switch in each leg. You can replace the high side switches with diodes to block back EMF or whatever it's called, can't you?
You only need a full bridge when using a single coil, that's the point of it, isn't it?
Or have I missed something?
Re: Possibly the end of shootthrough in SSTC's
Teodor Zafiroski, Mon Mar 30 2015, 05:24PM
Yeah, I didn't thought much about that... If we simplify it, Q1 and Q2 do nothing, we can remove them, and we are left with two coils fired alternately, a design already known in the HV community. I've seen it somewhere, just don't know where...
As for the replacement with diodes, I don't think they are even needed.
And no, you haven"t missed anything...
Teodor Zafiroski, Mon Mar 30 2015, 05:24PM
Yeah, I didn't thought much about that... If we simplify it, Q1 and Q2 do nothing, we can remove them, and we are left with two coils fired alternately, a design already known in the HV community. I've seen it somewhere, just don't know where...
As for the replacement with diodes, I don't think they are even needed.
And no, you haven"t missed anything...
Re: Possibly the end of shootthrough in SSTC's
Ash Small, Mon Mar 30 2015, 06:03PM
I think 'Centre Tapped Coil' is the word you are looking for
Ash Small, Mon Mar 30 2015, 06:03PM
I think 'Centre Tapped Coil' is the word you are looking for
Re: Possibly the end of shootthrough in SSTC's
hen918, Mon Mar 30 2015, 07:47PM
They are wound in bifilar. the Q1 / Q3 side might as well be shorted together. Then you just end up with a half or full bridge driving an inductor.
hen918, Mon Mar 30 2015, 07:47PM
They are wound in bifilar. the Q1 / Q3 side might as well be shorted together. Then you just end up with a half or full bridge driving an inductor.
Re: Possibly the end of shootthrough in SSTC's
Ash Small, Mon Mar 30 2015, 09:25PM
as anyone yet built a DRSSTC using a centre tapped, bifilar coil?
Ive been trying towork out a schematic in my head, I think I've got it, but not put it into CAD yet.
Ash Small, Mon Mar 30 2015, 09:25PM
as anyone yet built a DRSSTC using a centre tapped, bifilar coil?
Ive been trying towork out a schematic in my head, I think I've got it, but not put it into CAD yet.
Print this page