Bu508d breaking down. / High Voltage / Forums

Forums

4hv.org :: Forums :: High Voltage

« Previous topic | Next topic »

Bu508d breaking down.

previous 1 2 3 next

Move Thread

LAN_403

Dalus

Sat Mar 06 2010, 03:10PM

Registered Member #639 Joined: Wed Apr 11 2007, 09:09PM
Location: The Netherlands, Herkenbosch
Posts: 512

My guess is that the 555 has way to little power to drive your big transistor. I'd switch to a mosfet and use a gate driver IC in between of your mosfet and 555.

hboy007

Sun Mar 07 2010, 12:07AM

Registered Member #1667 Joined: Sat Aug 30 2008, 09:57PM
Location:
Posts: 374

Dalus wrote ...

My guess is that the 555 has way to little power to drive your big transistor. I'd switch to a mosfet and use a gate driver IC in between of your mosfet and 555.

I hope Tiemes also thinks about using a gate driver IC, such as TC4426 or higher current versions, at least he relies what his intentions are.
I am not a friend of MOSFETs in high voltage applications. I really like the BU2508, it has great switching performance and I haven't managed to kill one for a long time now.

The high voltage supply is a homebrew solution, so why not buy a toroid core and wind a current transformer? The primary side can easily be driven by a medium power high voltage transistor (I have one power supply running with a reconstructed TV circuit with a MJE340 (TO-220, rated 300V for inductive spike tolerance)). The base driver side consists of a small RCD pulse forming network, you can look up the ciruits online in TV repair manuals.

Patrick

Tue Mar 09 2010, 02:23AM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

i think you shoulduse Uzzors2k or other examples from this site. I too am building a lifter device, however this forum is the place to find such a power supply schematic i would just use IGBT's and a clamp diode, like the fairchild HGTG series. in anycase the psu's on JLN's website are kind of flimsy, if you have the knowledge/skill i would build your own flyback as seen on uzzors et al. websites.

And which country do you come from?

-Patrick

Tiemes

Tue Mar 09 2010, 05:04PM

Registered Member #2718 Joined: Thu Mar 04 2010, 03:36PM
Location: Netherlands
Posts: 3

hboy007 wrote ...

I hope Tiemes also thinks about using a gate driver IC, such as TC4426 or higher current versions, at least he relies what his intentions are.
I am not a friend of MOSFETs in high voltage applications. I really like the BU2508, it has great switching performance and I haven't managed to kill one for a long time now.

So you are saying that don't have to change the BU508D but put a gate driver IC between the 555 and the bu508d?
@ patrick, I live in Holland.

hboy007

Tue Mar 09 2010, 08:15PM

Registered Member #1667 Joined: Sat Aug 30 2008, 09:57PM
Location:
Posts: 374

Tiemes, you seem to be a little confused now. you can
a) use a high voltage mosfet in combination with a suitable GATE driver such as the IR4426 or
b) use the BU2508AFI, BU508AF, BU508D, BU508DF or whatever you can buy in combination with a high current BASE driver.

In both cases you need to have good control over the device during turn-off. You want to obtain a situation where the voltage drop across the device is low until it is completely open-circuit. This isn't possible if you just connect an inductor to the FET/BJT.You want to keep the driving voltage and current within the device ratings, this applies to forward/backward current of the BJT as well as to the gate drive current of the MOSFET. The latter has some nF of gate capacitance, the BJT suffers from charge storage effects and base-collector coupling.
The next thing would be to control the inductive voltage spike. You can either hope that the transformers parasitics will work for you or add some capacity in parallel to the primary winding or in series (namely in parallel to the switching device). If the voltage rises too fast, you won't be able to turn your semiconductor off properly and it gets slapped by a voltage spike when half-open, exceeding the SOA by an order of magnitude. If the voltage rises too high, your poor little semiconductor will turn into a 0 Ohms resistor. You can reduce the voltage swing by adding capacitance but this will be severely punished by long oscillations and significant inrush current spikes.

I am not telling you to use the bipolar HV transistor, I just tell you that there are drawbacks everywhere and your decision should be motivated by your abilities to control all the parameters. If you don't want to handle the base driving current waveforms, go for the MOSFET, add some transient voltage spike diodes or a decent dissipative RCD snubber, a VDR or whatever appears applicable to you.
If you choose the BJT, get the collector emitter voltage rise under control and be sure to drive with enough (but no more than that) base current - and build a circuit that generates a negative voltage kick to turn it of as fast as the device ratings permit regarding base-emitter breakdown voltage and current figures.

uzzors2k

Tue Mar 09 2010, 10:22PM

Registered Member #95 Joined: Thu Feb 09 2006, 04:57PM
Location: Norway
Posts: 1308

A gate driver is utter overkill for driving a single mosfet at these frequencies and power levels. You can get away with a gate resistor in the 10s of ohms range and still have sharp rise and fall times.

Patrick

Wed Mar 10 2010, 01:30AM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

Uzzors might be right, the se/ne555's that i use, have an internal totem-pole type output to drive extewrnal transistors with high gate capacitence (Mosfets), or high gate charge (IGBTs). try a single transistor and an
o-scope if you have one. get rid of that electrolytic cap ,c6 that will be a troublemaker for sure.

also, if the people from poland are called "poles", why arent the people from holland called "holes"?

lol, just the ranting of a crazy american

EDIT: added this example, do this simple stuff first before doingany of the above which might confuse you.
and someone validate he TVS directions im dyslexic and iget those dam diodes backwords.

oh fuck me, i got the A and B backwords.

-Patrick

hboy007

Wed Mar 10 2010, 12:55PM

Registered Member #1667 Joined: Sat Aug 30 2008, 09:57PM
Location:
Posts: 374

Uzzors, I know you're right. In contrast to the bipolar timers, the LMC version will source only 10mA. With a gate driver, even if it's a pair of BC638 / BC639 or BC548/BC558 transistors, you can simply forget about gate driving for the rest of the day.

Patrick, why didn't you add a capacitor across the IGBT?

Patrick

Fri Mar 12 2010, 12:10AM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

oh yeah hboy007, a cap is a good idea. close to the IGBT. but this is not my preferred circuit, it seemed to be flimsy (killed itself, often) to me when i built it for my ionocraft.

And if you do use that c6 cap, it needs to be paper/film or similar, not an electrolytic, that chemical polarizability of the electrolyte is slow to ware off. and hence we use electrolytic mostly in dc and static bias applications. not for controlling complex signal current of high frequency.alsoi it doesnt have to 100uF it can be less, but i advise not using that c6 at all.

hboy007

Sat Mar 13 2010, 05:04PM

Registered Member #1667 Joined: Sat Aug 30 2008, 09:57PM
Location:
Posts: 374

Patrick my experience is a bit different. A small cap in the range of 220pF .. 1nF will modify the parasitic oscillation of the transformer enough to push the nasty spikes down below 1kV instead of 1.8kV. The key may be generating a rather soft turn-on waveform which in turn slows down the draining of the cap through the transistor. I guess it will be much harder to gain this level of control over IGBTs so the cap is noch suitable without a series resistor but now the cap evolves to a passive snubber, I wonder if that's a good idea.

I assume it will be easier for Tiemes to just clip the spikes with a TVS Diode, VDR or RCD snubber instead of controlling the net resonance parameters. I hope he can report his progress soon!

previous 1 2 3 next

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