Interupting xenon flash current ?

[Don Klipstein]

In a typical xenon strobe, roughly how much of the cap electrical
energy becomes light?

I mostly find something like 40 lumens per watt overall luminous
efficacy.

In a really good case, 75-80% of the energy going intio the flashlamp
becomes radiation. Most of this looks like blackbody radiation, but some
goes into a cluster of major near-infrared lines around 820-1,000 nm. The
percentage going into those infrared lines is greater when plasma
temperature is lower, the flashtube is narrower and the xenon pressure is
lower.

Maximum luminous efficacy of a blackbody is about 95 lumens per watt, at
about 6600 Kelvin. Take 75-80% of that and maybe less with lower energy
strobes, take a bit to some off for those infrared lines, take a little
more off for the plasma temperature being other than 6600 K at least most
of the duration of the flash, take a little more off for losses in series
resistance, and I think one would do well to get half that 95 lumen/watt
figure.

I somewhat remember 40 or maybe 45 lumens/watt from an old catalog from
one of EG&G's divisions. I do suspect that with enough work into this,
one could get 50 or 60 - along with light that is on the bluish side and a
more expensive and bulkier capacitor (or array thereof) and a bulky
expensive low-resistance inductor to shape the current pulse, and only at
a higher flash energy in a larger (and more expensive) flashtube.

- Don Klipstein ([email protected])

 

[Mike]

It's probably laminated steel, with a lot of winding capacitance, too
slow for this measurement; but you could test it.

John

John,
Thanks for the CT info. I tried the CT that I found and it seems to work ok.
At 1:200 I used a 2ohm 1/4w metal film resistor (all I have) and it shows a 140us long 470A pulse.
I posted a photo of the waveform on abse.


Mike
When truth is absent politics will fill the gap.

 

[Glen Walpert]

John,
Thanks for the CT info. I tried the CT that I found and it seems to work ok.
At 1:200 I used a 2ohm 1/4w metal film resistor (all I have) and it shows a 140us long 470A pulse.
I posted a photo of the waveform on abse.

I think your CT is probably under responding to the actual current;
your shunt measurement was more believable IMO.

I tried a ferrite core CT wound approximately to John Larkin's
suggestion, and it says my 600 volt 70 uF flash with 4.5 uH 0.1 ohm
ballast delivers 1800 amps peak to the tube. The current amplitude
and shape are very close to what the theoretical discharge of the RLC
circuit should be. Trace posted on ABSE.

(My lamp is 12 mm OD, 10 mm ID, 12 mm between the tips of about 8 mm
dia tungsten tipped copper electrodes, and uses a 20 kV trigger.)

Glen

 

[Glen Walpert]

On Fri, 07 Dec 2007 08:49:00 -0800, John Larkin

You don't need an igbt, and they're slow anyhow. Just pick a capacitor
size, and arrange the parasitics, to dump the energy in less than a
microsecond.

I finally attempted this to see how fast I could get with parts on
hand. With two parallel 1 uF 1000 VDC oil/paper capacitors mounted on
the flash head with short leads I got a 250 amp peak pulse approx 7 us
duration, starting about 3 us after the trigger. At half amplitude it
is 3.5 us wide. I expect that lower inductance capacitors would not
help with the slow 3 us turn on time but might shorten the pulse width
a bit.

Incidentally 10 us is the exposure time claimed for Edgerton's famous
football photo (the first widely published high speed strobe photo):
http://www.cycleback.com/edgerton.htm
But the information on this site does not agree with Edgerton's
description of how the photo was taken, from an interview he gave to
SPIE (IIRC) just after he retired. The photo was taken outside, not
in his lab, and his description of hauling a big HV power supply,
capacitor and low pressure air filled arc tube out to the field,
manually triggering the flash by closing a switch directly in the HV
ckt as the ball was kicked, and finding out that they hit the timing
right on one of the photos only after they were developed, was quite
entertaining.

I don't know how fast the xenon will quit making light after the
current is gone.

Judging from the availability of a flash with a 250 ns pulse, it must
quit pretty fast, but I did not find the lifetimes of the excited
states of xenon in a quick search, only a bunch of papers of
questionable use for sale.

The igbt's are needed in cameras where the scene is unknown and
unmetered before the shot. A sensor integrates bounced-back light
during the exposure and shuts off the flash when it's had enough.
That's especially useful for real film with a slow mechanical shutter.

For "studio" setups, there are simpler ways to control exposure.

Right, although all good studio strobe lighting equipment allows for
adjustable light output, preset by the photographer rather than
depending on a photosensor, but using similar methods to turn the
flash off before the capacitor is fully discharged.

In my case this is just a pleasant diversion from real work, and as
such the requirements are rather arbitrary.

There are several options for speeding things up. I must reject the
options of using a tube with lower pressure, a different gas or
different geometry due to parts on hand limitations, leaving only the
most entertaining option: use of a simmer current to keep the tube
conducting while "off" (actually only almost off). Then either a
switched DC supply or a pulse transformer should be able to get
current through the tube very quickly, and the trigger can be
dispensed with entirely if the simmer current source is of high enough
voltage, like perhaps a HeNe laser power supply connected through a
ballast resistor, fitting the parts on hand requirement. Alas the
fast 1000 volt 2000 amp mosfets are not found in my junk box, so I
will have to try it driven by a pulse transformer.

The simmer current technique is used in the huge arc tubes which
switch the main beam amplifier mirrors at the LLNL National Ignition
Facility, rather quickly IIRC though I can't find the specs right now.

But by the time I get around to trying it this thread will be long
forgotten. Time to get back to real work again for a while .

Glen

 

[Mike]

I finally attempted this to see how fast I could get with parts on
hand. With two parallel 1 uF 1000 VDC oil/paper capacitors mounted on
the flash head with short leads I got a 250 amp peak pulse approx 7 us
duration, starting about 3 us after the trigger. At half amplitude it
is 3.5 us wide. I expect that lower inductance capacitors would not
help with the slow 3 us turn on time but might shorten the pulse width
a bit.

Incidentally 10 us is the exposure time claimed for Edgerton's famous
football photo (the first widely published high speed strobe photo):
http://www.cycleback.com/edgerton.htm
But the information on this site does not agree with Edgerton's
description of how the photo was taken, from an interview he gave to
SPIE (IIRC) just after he retired. The photo was taken outside, not
in his lab, and his description of hauling a big HV power supply,
capacitor and low pressure air filled arc tube out to the field,
manually triggering the flash by closing a switch directly in the HV
ckt as the ball was kicked, and finding out that they hit the timing
right on one of the photos only after they were developed, was quite
entertaining.


Judging from the availability of a flash with a 250 ns pulse, it must
quit pretty fast, but I did not find the lifetimes of the excited
states of xenon in a quick search, only a bunch of papers of
questionable use for sale.


Right, although all good studio strobe lighting equipment allows for
adjustable light output, preset by the photographer rather than
depending on a photosensor, but using similar methods to turn the
flash off before the capacitor is fully discharged.

In my case this is just a pleasant diversion from real work, and as
such the requirements are rather arbitrary.

There are several options for speeding things up. I must reject the
options of using a tube with lower pressure, a different gas or
different geometry due to parts on hand limitations, leaving only the
most entertaining option: use of a simmer current to keep the tube
conducting while "off" (actually only almost off). Then either a
switched DC supply or a pulse transformer should be able to get
current through the tube very quickly, and the trigger can be
dispensed with entirely if the simmer current source is of high enough
voltage, like perhaps a HeNe laser power supply connected through a
ballast resistor, fitting the parts on hand requirement. Alas the
fast 1000 volt 2000 amp mosfets are not found in my junk box, so I
will have to try it driven by a pulse transformer.

The simmer current technique is used in the huge arc tubes which
switch the main beam amplifier mirrors at the LLNL National Ignition
Facility, rather quickly IIRC though I can't find the specs right now.

But by the time I get around to trying it this thread will be long
forgotten. Time to get back to real work again for a while .

I'll be watching for further developmens.

Glen

I have changed to pretty small tube I scrounged from an old Polaroid camera because
the large tube was just too slow turning on. Even after removing all current measurment
components, shortening the leads, and increasing the charge voltage the light pulse,
as measured with a fairly slow photo transistor, took a little over 10us to peak. The
small tube using 15uf charged to 575V measures 4.5us to peak light output and the total
light pulse length is 60us. I'll have more accurate measurments when the PIN photo diodes
I ordered arrive. In the mean time I think I'll see if I can blow up an IGBT by trying to
open the dischagre path with it. The one I have is rated at 300A for 1ms and I'm hoping
it can withstand the current since it only has to conduct for 10us worse case.
The current shouldn't be nearly as high as yours with the smaller caps and higher
impedance tube. I'll lower the charge voltage to the minimum for the tube to flash and
increase it if the IGBT survives.

Mike


Television is a medium because it is seldom well done.

 

[Don Klipstein]

<snip>

(responding to someone else who wrote this)

I have a Photogenic Powerlight 1500SL, and it appears to me to work
differently.

The main energy storage capacitor is charged to a variable voltage. It
appears to me that this big capacitor is in parallel (except for see
below) with a smaller capacitor that is charged to the full 500 volts or
so.

A diode separates the big capacitor from being truly in parellel with
the small capacitor (or this is the way things here appear to me to be).
The small capacitor is in parallel with the flashtube. The small
capacitor gets the arc in the flashtube going once the trigger pulse
(kilovolts applied to the trigger lead) ionizes the flashtube.
Once the small capacitor's voltage drops below that of the large
capacitor, the large capacitor starts conducting through that diode. Of
course, this has to be a hefty diode!

This scheme is used because, with a lower flash intensity setting, the
large capacitor is only charged to a voltage that can sustain the arc once
it gets good and rolling, but not enough voltage to get enough current
through the thin streamers initiated by the trigger pulse to get them to
"blow up" into an arc.

It is easier to terminate the charging of a big capacitor at some
selectable voltage than to interrupt the current being discharged from
such a capacitor through a big flashtube.

One side effect of this flash output adjusting scheme: The color
temperature of the flash varies slightly with flash output. At minimum,
my unit is closer to 5,000 Kelvin, and at maximum it is closer to 6,000
Kelvin.

One other thing: If you turn down the flash setting and use the flash
before the big capacitor bleeds down to the new corresponding voltage
setting, the next flash will be brighter than you have the unit set for.

- Don Klipstein ([email protected])

 

[Glen Walpert]

<snip>

I'll be watching for further developmens.

OK, I'll post any results I get.

I have changed to pretty small tube I scrounged from an old Polaroid camera because
the large tube was just too slow turning on. Even after removing all current measurment
components, shortening the leads, and increasing the charge voltage the light pulse,
as measured with a fairly slow photo transistor, took a little over 10us to peak. The
small tube using 15uf charged to 575V measures 4.5us to peak light output and the total
light pulse length is 60us. I'll have more accurate measurments when the PIN photo diodes
I ordered arrive. In the mean time I think I'll see if I can blow up an IGBT by trying to
open the dischagre path with it. The one I have is rated at 300A for 1ms and I'm hoping
it can withstand the current since it only has to conduct for 10us worse case.
The current shouldn't be nearly as high as yours with the smaller caps and higher
impedance tube. I'll lower the charge voltage to the minimum for the tube to flash and
increase it if the IGBT survives.

Mike

For short accurately timed flashes it looks like pseudosimmer could be
the best mode of operation for small tubes. If you haven't already,
get dts_flashcatalog.pdf from http://optoelectronics.perkinelmer.com/
and look at the pseudosimmer circuit. (I searched their site for
"flash catalog" to find this.)

 

[Glen Walpert]

(responding to someone else who wrote this)


I have a Photogenic Powerlight 1500SL, and it appears to me to work
differently.

The main energy storage capacitor is charged to a variable voltage. It
appears to me that this big capacitor is in parallel (except for see
below) with a smaller capacitor that is charged to the full 500 volts or
so.

A diode separates the big capacitor from being truly in parellel with
the small capacitor (or this is the way things here appear to me to be).
The small capacitor is in parallel with the flashtube. The small
capacitor gets the arc in the flashtube going once the trigger pulse
(kilovolts applied to the trigger lead) ionizes the flashtube.
Once the small capacitor's voltage drops below that of the large
capacitor, the large capacitor starts conducting through that diode. Of
course, this has to be a hefty diode!

This scheme is used because, with a lower flash intensity setting, the
large capacitor is only charged to a voltage that can sustain the arc once
it gets good and rolling, but not enough voltage to get enough current
through the thin streamers initiated by the trigger pulse to get them to
"blow up" into an arc.

It is easier to terminate the charging of a big capacitor at some
selectable voltage than to interrupt the current being discharged from
such a capacitor through a big flashtube.

One side effect of this flash output adjusting scheme: The color
temperature of the flash varies slightly with flash output. At minimum,
my unit is closer to 5,000 Kelvin, and at maximum it is closer to 6,000
Kelvin.

One other thing: If you turn down the flash setting and use the flash
before the big capacitor bleeds down to the new corresponding voltage
setting, the next flash will be brighter than you have the unit set for.

- Don Klipstein ([email protected])

OK, I stand corrected on that. How about "at least one major brand of
studio flash lighting equipment uses interruption of lamp current to
control output" instead.

I recall perhaps 3 decades ago reading advertisements for studio flash
equipment that bragged about no color change due to no voltage
adjustment, with duration limiting instead. But I had completely
forgotten about the voltage adjustment method until you reminded me.

Glen

 

[Mike]

For short accurately timed flashes it looks like pseudosimmer could be
the best mode of operation for small tubes. If you haven't already,
get dts_flashcatalog.pdf from http://optoelectronics.perkinelmer.com/
and look at the pseudosimmer circuit. (I searched their site for
"flash catalog" to find this.)

Thanks, I had missed that document.

Btw, My igbt has survived a few flashes with a 550V charge, but with the gate constantly
turned on, I.E. tied to +12 thru a low value resistor. Now if it'll survive the turn off.
The pseudo simmer is interesting, but it may be a problem for some high speed photos since it
requires at least 100ms advance notice of when the event will occur.

Oh yeah, my igbt is only rated for 150A peak. I've been looking at too many data sheets I guess.

Mike
If there is no absolute truth then nothing can be known.

 

[[email protected]]

Hi,
I've been working on a high energy flash system for some time now
(about 4000 Joules) and just come across this forum. Like some others
here I've been having difficulty with interrupting the current to
terminate the discharge current. I had originally intended to run the
tubes at 1400 volts but I have had to scale that down to 700 volts.
I am currently using 24 IGBTs in parallel each rated at 60 Amps 900
Volts (will increase to 50 eventually to handle the full energy). The
flash tube draws some 900 amps once the curent stabilises after the
initial surge.

Unfortunately the IGBTS are not surviving (I have blown 3 already) and
I have had to limit the voltage further to 500 volts, reducing the
steady current to about 600 amps. I'm still trying to find out why
I'm getting the failures at 700 volts. The current sharing scheme is
admittedly quite primitive (or perhaps I should say non existent) but
it does not seem to be the problem and neither does the initial 5000
amp current surge. What does seem to be the problem is some very ugly
oscillation when the IGBT is turned off which probably means I need to
work more on the gate drive circuitry.

I've been reading further back in this forum and noted some discussion
on the initial surge of current when a flashtube is first triggered.
I have also found that phenomenon and with a very low resistance path
between the capacitor bank and the flash tube I was getting an
estimated 30 KA surge at 1400 volts, the duration of the surge being
about 30 to 40 microseconds. I had initially suspected an inductance
problem, as was also suggested here, but to eliminate that possibility
I tried using different shunts, ranging from a flat piece of lead to a
short length of 10mm2 copper cable and got similar results in each
case.

 

[[email protected]]

Hi,
I've been working on a high energy flash system for some time now
(about 4000 Joules) and just come across this forum. Like some
others
here I've been having difficulty with interrupting the current to
terminate the discharge current. I had originally intended to run the
tubes at 1400 volts but I have had to scale that down to 700 volts.
I am currently using 24 IGBTs in parallel each rated at 60 Amps 900
Volts (will increase to 50 eventually to handle the full energy).
The
flash tube draws some 900 amps once the curent stabilises after the
initial surge.
Unfortunately the IGBTS are not surviving (I have blown 3 already)
and
I have had to limit the voltage further to 500 volts, reducing the
steady current to about 600 amps. I'm still trying to find out why
I'm getting the failures at 700 volts. The current sharing scheme is
admittedly quite primitive (or perhaps I should say non existent) but
it does not seem to be the problem and neither does the initial 5000
amp current surge. What does seem to be the problem is some very
ugly
oscillation when the IGBT is turned off which probably means I need
to
work more on the gate drive circuitry.

I've been reading further back in this forum and noted some
discussion
on the initial surge of current when a flashtube is first triggered.
I have also found that phenomenon and with a very low resistance path
between the capacitor bank and the flash tube I was getting an
estimated 30 KA surge at 1400 volts, the duration of the surge being
about 30 to 40 microseconds. I had initially suspected an inductance
problem, as was also suggested here, but to eliminate that
possibility
I tried using different shunts, ranging from a flat piece of lead to
a
short length of 10mm2 copper cable and got similar results in each
case.



(PS: sorry if this post appears in duplicate. I had to post it several
times for it to appear - I'll check later and delete the surplus ones
if they actually appear)

 

[Spehro Pefhany]

Hi,
I've been working on a high energy flash system for some time now
(about 4000 Joules) and just come across this forum. Like some
others
here I've been having difficulty with interrupting the current to
terminate the discharge current. I had originally intended to run the
tubes at 1400 volts but I have had to scale that down to 700 volts.
I am currently using 24 IGBTs in parallel each rated at 60 Amps 900
Volts (will increase to 50 eventually to handle the full energy).
The
flash tube draws some 900 amps once the curent stabilises after the
initial surge.
Unfortunately the IGBTS are not surviving (I have blown 3 already)
and
I have had to limit the voltage further to 500 volts, reducing the
steady current to about 600 amps. I'm still trying to find out why
I'm getting the failures at 700 volts. The current sharing scheme is
admittedly quite primitive (or perhaps I should say non existent) but
it does not seem to be the problem and neither does the initial 5000
amp current surge. What does seem to be the problem is some very
ugly
oscillation when the IGBT is turned off which probably means I need
to
work more on the gate drive circuitry.

I've been reading further back in this forum and noted some
discussion
on the initial surge of current when a flashtube is first triggered.
I have also found that phenomenon and with a very low resistance path
between the capacitor bank and the flash tube I was getting an
estimated 30 KA surge at 1400 volts, the duration of the surge being
about 30 to 40 microseconds. I had initially suspected an inductance
problem, as was also suggested here, but to eliminate that
possibility
I tried using different shunts, ranging from a flat piece of lead to
a
short length of 10mm2 copper cable and got similar results in each
case.

Have you tried snubbers and softening the turn-off to reduce di/dt?

Best regards,
Spehro Pefhany

 

[[email protected]]

Have you tried snubbers and softening the turn-off to reduce di/dt?

Best regards,
Spehro Pefhany

Have tried but I probably do need to put some more effort into that.
All I tried was a 75nF capacitor in series with a 1 ohm resistor and
it is connected across the emitter-collector of the IGBT group. It was
there already though when I burnt out the IGBTs. It had made a
considerable improvement but not nearly enough it seems. Without it I
had a huge over/undershoot at cutoff - so much that it went offscale
on my scope (and even crashed the laptop sometimes ! ) even though I
was using a 100:1 probe.

I probably need to do some more research into designing a better
snubber as the one I made was based more on what I had in the junk
box than any calculation. With the snubber I now have only some 30%
voltage overhsoot but the cutoff is plagued with ringing - some 50 to
200 cycles of about 400V p-p with a period of about 100nS.

My worry with lower resistance/higher capacitance in the snubber is
that I will cause problems on turn on when the IGBT will have to
discharge the snubber capacitor through just one ohm - not very kind
to the IGBT given the stresses it already faces with the flash tube. I
have been thinking about some snubbers incorporating diodes (for
different turn on/off behaviour) but it looks like the cost/size of
the diodes may be a problem.

What I was thinking could be the problem is that rate of change of
voltage rise on the collector causing feedback into the gate. If that
is the case the snubber will certainly help, however I have been
trying to approach the problem more from the gate side than the
Collector side.

By applying capacitance across the gate and emitter and playing around
with different gate resistors I managed to tame the ringing somewhat
such that I now have less than 20 ringing cyles with an amplitude of
about 200 V. However I am just trying random things at this point but
I cannot honestly say that I know what I'm doing.

I have an excel file with the trace of one of the latest tests and I
would post it here but I can't quite figure out how to do that.

 

[Robert Baer]

Have you tried snubbers and softening the turn-off to reduce di/dt?

Best regards,
Spehro Pefhany

About 40 years ago, SCRs, ignitrons (and some transistors) had I*I*T
ratings that were necessary to use for designs in impulse circuits.
Demand that data from the vendors.
Also try an SCR in place of a FET, but watch tat eye-squared-tee rating!

 

[[email protected]]

   About 40 years ago, SCRs, ignitrons (and some transistors) had I*I*T
ratings that were necessary to use for designs in impulse circuits.
   Demand that data from the vendors.
   Also try an SCR in place of a FET, but watch tat eye-squared-tee rating!- Hide quoted text -

- Show quoted text -

I*I*T seems not to be a problem. I have succesfully fire the tube many
times at 700 volts through the IGBTs, keeping thjem turned on long
enough for the capacitor to drain almost fully and it seems the IGBTS
can take that in their stride. The problems happen when I try to turn
off the IGBTs at an early stage while there is still a high voltage
and current. In fact all three failures have occured during test
firings with a duration between 50uS and 200uS. After such a discharge
the capacitor voltage would not have reduced appreciably. Whereas with
the longer duration pulses (which seem not to cause problems) the
voltage would have gone down to some 400 volts or lower.

I have actually even tried a stupid experiment using just one single
IGBT to handle the full current without switching it off and
surprisingly it survived. The transistor is rated at 60A continuous or
120A for 1 mS. I put some 700 amps for about 300uS through it and it
didn't compalin. (The duration was reduced by using less capacitance
rather than by turning off the IGBT). I estimate that is about 10
times the quoted I*I*T rating for this device.

That makes me think that it is what goes on during the actual
switching off that is killing the transistors, perhaps the VxI during
parts of the ringing when both the voltage across the transistor and
the current through it are high. ANother possibility I am thinking
about is overvoltage on the gate due to the high dv/dt at the
collector pulling up the gate through the gate-collector capacitance.

I had originally planned to use thyristors but it tuirned out that
they would be even more difficult to get to work properly. The two
major problems where how to switch off the thyristor and how to limit
the di/dt enough to avoid destruction of the thyristor on turn on.
The switching off problem is the most difficult though as one needs to
shunt current away from it (by another equally rated thyristor) and
then ensure that it actually stays off when it experiences the fast
voltage rise after cutoff. There used to exist so called 'gate
turnoff thyristors' which were a bit easier to use and were used in
many consumer camera flashes but it seems they no longer exist and in
any case seem toi have never been made with sufficiently high current
and voltage ratings for what I need.

Next thing I am thinking of trying is to get a 'softer' turnoff as
Spehro suggested, but the downside of that is potentially destructive
V*I*T due to the period of partial conduction if I overdo it. I'll
also need to learn a lot more about how to design a better snubber.

 

[legg]

On Tue, 19 Feb 2008 11:09:54 -0800 (PST), [email protected] wrote:

Have tried but I probably do need to put some more effort into that.
All I tried was a 75nF capacitor in series with a 1 ohm resistor and
it is connected across the emitter-collector of the IGBT group. It was
there already though when I burnt out the IGBTs. It had made a
considerable improvement but not nearly enough it seems. Without it I
had a huge over/undershoot at cutoff - so much that it went offscale
on my scope (and even crashed the laptop sometimes ! ) even though I
was using a 100:1 probe.

I probably need to do some more research into designing a better
snubber as the one I made was based more on what I had in the junk
box than any calculation. With the snubber I now have only some 30%
voltage overhsoot but the cutoff is plagued with ringing - some 50 to
200 cycles of about 400V p-p with a period of about 100nS.

My worry with lower resistance/higher capacitance in the snubber is
that I will cause problems on turn on when the IGBT will have to
discharge the snubber capacitor through just one ohm - not very kind
to the IGBT given the stresses it already faces with the flash tube. I
have been thinking about some snubbers incorporating diodes (for
different turn on/off behaviour) but it looks like the cost/size of
the diodes may be a problem.

What I was thinking could be the problem is that rate of change of
voltage rise on the collector causing feedback into the gate. If that
is the case the snubber will certainly help, however I have been
trying to approach the problem more from the gate side than the
Collector side.

By applying capacitance across the gate and emitter and playing around
with different gate resistors I managed to tame the ringing somewhat
such that I now have less than 20 ringing cyles with an amplitude of
about 200 V. However I am just trying random things at this point but
I cannot honestly say that I know what I'm doing.

I have an excel file with the trace of one of the latest tests and I
would post it here but I can't quite figure out how to do that.

A current snubber or other RCD clamping configurations allow the
capacitor to absorb current during the voltage rise-time only.

If you know the amount of energy that you wish to deliver prior to
switching, this could be adjusted by separating the ignition source
from a ( lower voltage ) programmed energy source, allowing the tube
to self-extinguish on the discharging sources. It would also allow use
of SCRs.

You can't count on parallel IGBTs sharing without external assistance.
Perhaps you might consider a cascode HV bipolr/ LV mosfet.

I'd also invest in a current transformer to assist in monitoring.

RL

 

[Tim Williams]

Isn't the point of SCRs to shunt the flash tube? Obviously, through some
sort of energy damping medium, like an inductor between the cap and tube.
You only have to deal with the flyback energy later.

(Hmm, if you just let it swing, the inductor will run the capacitor's
voltage negative by however much, no complicated snubber or switching or
damping, the SCR only conducts peak current. That might be useful. Flip
polarity with a beefy relay and recharge the difference?)

Tim

--
Deep Fryer: A very philosophical monk.
Website @ http://webpages.charter.net/dawill/tmoranwms

About 40 years ago, SCRs, ignitrons (and some transistors) had I*I*T
ratings that were necessary to use for designs in impulse circuits.
Demand that data from the vendors.
Also try an SCR in place of a FET, but watch tat eye-squared-tee rating!- Hide quoted text -

- Show quoted text -

<snip>

I had originally planned to use thyristors but it tuirned out that
they would be even more difficult to get to work properly. The two
major problems where how to switch off the thyristor and how to limit
the di/dt enough to avoid destruction of the thyristor on turn on.
The switching off problem is the most difficult though as one needs to
shunt current away from it (by another equally rated thyristor) and
then ensure that it actually stays off when it experiences the fast
voltage rise after cutoff. There used to exist so called 'gate
turnoff thyristors' which were a bit easier to use and were used in
many consumer camera flashes but it seems they no longer exist and in
any case seem toi have never been made with sufficiently high current
and voltage ratings for what I need.

<snip>

 

[Tim Williams]

You can't count on parallel IGBTs sharing without external assistance.

Actually, you can. IR wrote an AN somewhere that describes the process.
At low current, it's true, but presumably, at these current levels, the
paralleled devices are each rated for enough that there is no problem. At
higher current levels, the resistive component of the IGBTs comes into
play, and current shares to an increasing degree. What you get is a
roughly constant difference in current between transistors, and as long as
this difference is less than the rated current of each device, you're fine.

Tim

 

[[email protected]]

A current snubber or other RCD clamping configurations allow the
capacitor to absorb current during the voltage rise-time only.

If you know the amount of energy that you wish to deliver prior to
switching, this could be adjusted by separating the ignition source
from a ( lower voltage ) programmed energy source, allowing the tube
to self-extinguish on the discharging sources. It would also allow use
of SCRs.

You can't count on parallel IGBTs sharing without external assistance.
Perhaps you might consider a cascode HV bipolr/ LV mosfet.

I'd also invest in a current transformer to assist in monitoring.

RL- Hide quoted text -

- Show quoted text -

The RCD snubber seems like a good solution but the problem is the cost
of the diodes needed. I have three high current diodes on the way in
the post and I may be able to spare one of them for that. However
even with that I am having some difficulty getting the values right so
that the snubber can be used multiple times, for instance to have
three 100uS pulses spaced at 100uS. The snubber capacitor finds
itself still charged on the second current disconnection and will not
have much effect. However that is probably something that can be
solved with the right values.

The capacitor self discharging is not an option because thje pulse
widths and spacing need to settable to arbitrary values controlled by
a pic from some 50 uS all the way up to 5mS, so it is not possible to
tailor component values to fit the timings required. Also, I need the
pulses to have a very sharp cutoff, something that cannot be achieved
if one depends on the capacitor voltage decaying to below the holding
voltage to cut off.

As for current sharing I am currently not doing anything except
paralleling the emitters and collectors of all 24 transistors on two
thick copper busbars. The gates are separated into four groups of six
to reduce to some extent the interaction between the transistors on
turn off. In future I intend to have a 0.25 ohm resistor on the
collector of each transistor to help distribute the current. It will
cause some 15 volts drop at maximum load, but that is a small price to
pay if it does its job. The plans are also for each gate to have its
own optically isolated driver to completely eliminate any interaction
between the gates of the various transistors.

 

[JosephKK]

Have tried but I probably do need to put some more effort into that.
All I tried was a 75nF capacitor in series with a 1 ohm resistor and
it is connected across the emitter-collector of the IGBT group. It was
there already though when I burnt out the IGBTs. It had made a
considerable improvement but not nearly enough it seems. Without it I
had a huge over/undershoot at cutoff - so much that it went offscale
on my scope (and even crashed the laptop sometimes ! ) even though I
was using a 100:1 probe.

I probably need to do some more research into designing a better
snubber as the one I made was based more on what I had in the junk
box than any calculation. With the snubber I now have only some 30%
voltage overhsoot but the cutoff is plagued with ringing - some 50 to
200 cycles of about 400V p-p with a period of about 100nS.

My worry with lower resistance/higher capacitance in the snubber is
that I will cause problems on turn on when the IGBT will have to
discharge the snubber capacitor through just one ohm - not very kind
to the IGBT given the stresses it already faces with the flash tube. I
have been thinking about some snubbers incorporating diodes (for
different turn on/off behaviour) but it looks like the cost/size of
the diodes may be a problem.

What I was thinking could be the problem is that rate of change of
voltage rise on the collector causing feedback into the gate. If that
is the case the snubber will certainly help, however I have been
trying to approach the problem more from the gate side than the
Collector side.

By applying capacitance across the gate and emitter and playing around
with different gate resistors I managed to tame the ringing somewhat
such that I now have less than 20 ringing cyles with an amplitude of
about 200 V. However I am just trying random things at this point but
I cannot honestly say that I know what I'm doing.

I have an excel file with the trace of one of the latest tests and I
would post it here but I can't quite figure out how to do that.

What are you doing to balance the currents between the IGBTs?