Powerful discrete GDT Driver
Reaching, Sat Sept 09 2006, 09:27PMwithin the last few days i experimented a lot with some useful plugins for sstcs and drsstcs and i really dont like stacking lots of ucc driver ics together to drive a fullbridge in cw so i decided to build a discrete driver instead and came up with a really good design less expensive and much more powerful than 2 pairs of ucc drivers
heres the shematic
then i tried to build a stand alone drsstc driver without using a single ic by adding an "enable" to the driver,. the funniest is, this thing works wondeful, not tested on a real drsstc but from the waveforms it could work
heres a pic showing the interrupted signal measured on one secondary winding of an 1:2:2 gdt voltage is 10V/div and frequency is 120khz
the transistor preamp can take any ttl signal to drive the bridge of an p/n channel mosfet, . for a simple cw feedback sstc its possible to add a feedback transformer on the preamp , the preamp and the mosfet bridge act like a schmitt trigger and converts a sinusodial input to a nice squarewave to drive the gdt. maybe someone is interested
Re: Powerful discrete GDT Driver
Marko, Sat Sept 09 2006, 09:48PM
I experimented vith various kinds of 'discrete' drives and decided that it is hard to beat power and speed of UCC's. Mosfets you use even if they are small and relatively fast again need something better than simple bipolar totem pole to acheive full-speed turnon.
I even used 7667's to drive mosfets directly and it still wasn't good.
You didn't state most imoprtant tihngs, what gate capacitance are you driving, and what are exact fall and rise times you get? You will need to zoom the waveform a bit to see that clearly.
I somehow doubt you can beat power of UCCs that way, exscept it may be a solution for bricks etc. at lower frequencies..
Marko, Sat Sept 09 2006, 09:48PM
I experimented vith various kinds of 'discrete' drives and decided that it is hard to beat power and speed of UCC's. Mosfets you use even if they are small and relatively fast again need something better than simple bipolar totem pole to acheive full-speed turnon.
I even used 7667's to drive mosfets directly and it still wasn't good.
You didn't state most imoprtant tihngs, what gate capacitance are you driving, and what are exact fall and rise times you get? You will need to zoom the waveform a bit to see that clearly.
I somehow doubt you can beat power of UCCs that way, exscept it may be a solution for bricks etc. at lower frequencies..
Re: Powerful discrete GDT Driver
Reaching, Sat Sept 09 2006, 10:35PM
i build a testfullbridge with the to247 40n60s and a single gdt driven with 120khz cw and the output signal is the best ive ever seen on a fullbridge. the driver draws around 0,7A at 20volts, how much uccs do you want to stack together to drive such a bridge?. tommorow i can take a few pics, but i doubt my scope is good enough to measure the fall and rise times. for testing i connected a toshiba brick igbt with 40nF gate capacitance and with this setup i could drive it up to 100khz with a gdt without much trouble, . without a gdt ala high side switching i could turn it up to 120khz.
Reaching, Sat Sept 09 2006, 10:35PM
i build a testfullbridge with the to247 40n60s and a single gdt driven with 120khz cw and the output signal is the best ive ever seen on a fullbridge. the driver draws around 0,7A at 20volts, how much uccs do you want to stack together to drive such a bridge?. tommorow i can take a few pics, but i doubt my scope is good enough to measure the fall and rise times. for testing i connected a toshiba brick igbt with 40nF gate capacitance and with this setup i could drive it up to 100khz with a gdt without much trouble, . without a gdt ala high side switching i could turn it up to 120khz.
Re: Powerful discrete GDT Driver
Marko, Sat Sept 09 2006, 10:43PM
You probably don't need much better than 10Mhz scope for such measurments, I have some old 20Mhz bucket and it proven perfect for everything yet
Such drivers are usually good for big bricks and slow stuff with large gate capacitances, so your mosfet slowness won't affect overall speed for much.
Anyway I would still replace the totem pole stage with at least 7667 driver
Maybe you can play with paralelling, maybe bipolars and mosfet's together ala UCC's output stage, etc..
Marko, Sat Sept 09 2006, 10:43PM
You probably don't need much better than 10Mhz scope for such measurments, I have some old 20Mhz bucket and it proven perfect for everything yet
Such drivers are usually good for big bricks and slow stuff with large gate capacitances, so your mosfet slowness won't affect overall speed for much.
Anyway I would still replace the totem pole stage with at least 7667 driver
Maybe you can play with paralelling, maybe bipolars and mosfet's together ala UCC's output stage, etc..
Re: Powerful discrete GDT Driver
HV Enthusiast, Sun Sept 10 2006, 02:54AM
Looks good, although unless you are using your discrete driver to power some of those super large CM300 or CM600 (or similar) bricks, you really don't need something as complicated as that.
For example, i use just single UCC2732x type MOSFET drivers for driving in large IGBTs such as the Fairchild 40N60's, and i don't parallel them either for a half-bridge. Even when using the small surface mount versions of these chips, they don't even get warm.
HV Enthusiast, Sun Sept 10 2006, 02:54AM
Looks good, although unless you are using your discrete driver to power some of those super large CM300 or CM600 (or similar) bricks, you really don't need something as complicated as that.
For example, i use just single UCC2732x type MOSFET drivers for driving in large IGBTs such as the Fairchild 40N60's, and i don't parallel them either for a half-bridge. Even when using the small surface mount versions of these chips, they don't even get warm.
Re: Powerful discrete GDT Driver
Reaching, Sun Sept 10 2006, 10:42AM
i wanted to use the discrete driver for a huge sstc i´ll build and that thing must work in cw so i think this discrete driver is the right choice. cause i use this bipolar transistor preamp i can use a optocoupler to connect the HF signal and i only need a single insulated supply voltage. and i have some huge fuji igbts left with 38nf gate capacitance and some toshiba and eupec igbt bricks. maybe i can use the circuit for a simple feedback controlled smps with a feedback winding driving the bipolar preamp, many useful applications for this thing
Reaching, Sun Sept 10 2006, 10:42AM
i wanted to use the discrete driver for a huge sstc i´ll build and that thing must work in cw so i think this discrete driver is the right choice. cause i use this bipolar transistor preamp i can use a optocoupler to connect the HF signal and i only need a single insulated supply voltage. and i have some huge fuji igbts left with 38nf gate capacitance and some toshiba and eupec igbt bricks. maybe i can use the circuit for a simple feedback controlled smps with a feedback winding driving the bipolar preamp, many useful applications for this thing
Print this page