controlled HV-capacitor charging

36 views
Skip to first unread message

Heindorf

unread,
Oct 10, 2003, 10:07:12 AM10/10/03
to
Hello
What's a simple way to charge an electrolytic capacitor >1mF from a DC
source of 200 to 500VDc in a controllable manner with the subsequent
inhibition of this start up circuit to enable normal operation?
I thought of a step down chopper with external MOSFET+PWM with soft
start. But that requires an additional supply for the PWM-IC.
A search in the internet yielded components like the TOPSwitch family by
Power Integrations "TOP250....".Are there more powerful components
available or a better/simpler solution?
Thanks for all replies.
Rolf

John Fields

unread,
Oct 10, 2003, 12:00:40 PM10/10/03
to

---
You could charge the cap through a resistor in series with the NO
contacts of a relay, then when the cap is fully charged let the relay
de-energize, which will connect the cap to the NC contact, to which
you'll connect your load.

To do this you could connect a 2 resistor voltage divider from the NO
contact to GND, and run the junction of the two resistors to the + input
of a comparator. On the - input of the comparator place a reference
voltage such that when the cap charges up to the voltage you want, the
output of the voltage divider will be more positive than the reference,
forcing the comparator's output high. Use the comparator's high-going
output to reset an R-S latch which will turn off a transistor driving
the relay coil, which will now demagnetize and allow the common contact,
to which the cap is connected, to swing over to the NC contact.

In order to star the cap charging, use a SPST NO switch to set the
latch. This will turn on the relay driver, connecting the HV supply to
the cap through the charging resistor.

--
John Fields

Harry Dellamano

unread,
Oct 10, 2003, 1:53:35 PM10/10/03
to

"John Fields" <jfi...@austininstruments.com> wrote in message
news:r3kdov0lprrb055o7...@4ax.com...

> On Fri, 10 Oct 2003 16:07:12 +0200, Heindorf <hein...@aeras.de> wrote:
>
> >Hello
> >What's a simple way to charge an electrolytic capacitor >1mF from a DC
> >source of 200 to 500VDc in a controllable manner with the subsequent
> >inhibition of this start up circuit to enable normal operation?
> >I thought of a step down chopper with external MOSFET+PWM with soft
> >start. But that requires an additional supply for the PWM-IC.
> >A search in the internet yielded components like the TOPSwitch family by
> >Power Integrations "TOP250....".Are there more powerful components
> >available or a better/simpler solution?
> >Thanks for all replies.
> >Rolf
>
> ---
> You could charge the cap through a resistor in series with the NO
> contacts of a relay, then when the cap is fully charged let the relay
> de-energize, which will connect the cap to the NC contact, to which
> you'll connect your load.
>

John, you are embarrassing the whole group.
The Op could use a one stage, PWM boost circuit. Control the "on time" of
the switch with peak current from the input and "off time" proportional to
output (stored) voltage. This simplifies stability and gives fast, efficient
charging with least parts count. To soft start the circuit place a resistor
in series with the output capacitor to ground, short out when the output
voltages charges to the input, them start PWM boost circuit. First 200V of
charge is free.

regards
harry

CBarn24050

unread,
Oct 10, 2003, 3:39:26 PM10/10/03
to
well harry your solution isn't much better! There are thermistors around
specifically for this purpose, they use then in motor inverters. 1 part 2 leads
what could be simpler. rswww.com part no 210-673 is an example

John Fields

unread,
Oct 10, 2003, 3:41:19 PM10/10/03
to
On Fri, 10 Oct 2003 17:53:35 GMT, "Harry Dellamano"
<har...@tdsystems.org> wrote:


> John, you are embarrassing the whole group.

---
Harry,

As soon as you can show me where it says you've been elected the group's
spokesman, then maybe I'll think you have something worth listening to.

'till then, dear boy... Well, you know... go fuck yourself :-)

--
John Fields

Harry Dellamano

unread,
Oct 10, 2003, 6:04:55 PM10/10/03
to

"CBarn24050" <cbarn...@aol.com> wrote in message
news:20031010153926...@mb-m19.aol.com...

Hey Barn, just goes to show you that there are a lot of ways to skin a cat.

regards
harry

Yzordderex

unread,
Oct 10, 2003, 9:04:24 PM10/10/03
to
"Harry Dellamano" <har...@tdsystems.org> wrote in message news:<zmChb.770$ko4...@nwrddc02.gnilink.net>...

> "John Fields" <jfi...@austininstruments.com> wrote in message
> news:r3kdov0lprrb055o7...@4ax.com...
> > On Fri, 10 Oct 2003 16:07:12 +0200, Heindorf <hein...@aeras.de> wrote:

snip----


>> John, you are embarrassing the whole group.
> The Op could use a one stage, PWM boost circuit. Control the "on time" of
> the switch with peak current from the input and "off time" proportional to
> output (stored) voltage. This simplifies stability and gives fast, efficient
> charging with least parts count. To soft start the circuit place a resistor
> in series with the output capacitor to ground, short out when the output
> voltages charges to the input, them start PWM boost circuit. First 200V of
> charge is free.
>
> regards
> harry

What are we actually trying to do here? Just soft start a bus cap?
Boost 200v to 500v? I suspect it's just limit the charging current.
Maybe look at NTC thermister. NTC good for about 20 or 30 amps of DC
current. If you don't like that idea, just use a big diode bridge and
forget about the soft charge. If you insist on complexity use a
topswitch to run a relay. Perhaps a dc resistor to operate a relay in
parallel with charge resistor. There will be some hystersis as far as
open close points, but you could probably figure out how to get around
that with a 2 pole relay.

Next time try to be a little more specific about the application. Your
more likely to get meaningful answers

best regards,
Bob

Yzordderex

unread,
Oct 11, 2003, 8:37:47 AM10/11/03
to
yzord...@verizon.net (Yzordderex) wrote in message news:<a69e22f6.03101...@posting.google.com>...

> "Harry Dellamano" <har...@tdsystems.org> wrote in message news:<zmChb.770$ko4...@nwrddc02.gnilink.net>...
> > "John Fields" <jfi...@austininstruments.com> wrote in message
> > news:r3kdov0lprrb055o7...@4ax.com...
> > > On Fri, 10 Oct 2003 16:07:12 +0200, Heindorf <hein...@aeras.de> wrote:
>
> snip----

> Perhaps a dc resistor to operate a relay in
> parallel with charge resistor.

That was supposed to be a resistor to operate a DC relay:)

best regards,
Bob

Fritz Schlunder

unread,
Oct 11, 2003, 5:41:06 PM10/11/03
to

"Yzordderex" <yzord...@verizon.net> wrote in message

>
> What are we actually trying to do here? Just soft start a bus cap?
> Boost 200v to 500v? I suspect it's just limit the charging current.
> Maybe look at NTC thermister. NTC good for about 20 or 30 amps of DC
> current. If you don't like that idea, just use a big diode bridge and
> forget about the soft charge.


Err... In this case some sort of inrush limiting is almost certainly
mandatory. The OP's wording is slightly enigmous when he says, "from a DC
source of 200 to 500VDc". My interpretation is he has a DC bus voltage that
can be anywhere in the range of 200V to 500V. He has a 1000uF or larger
capacitor which he wants to charge up from zero to this 200V to 500V range
(depending upon what the bus voltage is when it is powered up).

So at the very lowest range 1000uF up to 200V we have an energy storage of
0.5CV^2. In this case:

0.5*0.001*200^2=20 Joules

At the higher range (but still assume 1000uF even though the OP said >1mF)
of 500V we have:

0.5*0.001*500^2=125 Joules


Now assuming we don't have some sort of switch mode constant current source
or other charging method using inductors, the total energy that must be
dissipated somewhere while charging the capacitor from 0V to 200-500V is 20J
or 125J respectively. That is way too much energy at way too high a
potential difference to avoid using inrush limiting.

If the DC bus voltage is stiff enough, the effect of charging the capacitor
instantly from it will be the same as discharging the capacitor by short
circuiting it while charged to 200V-500V.

Try short circuiting a 1000uF capacitor charged to 500V and you will
instantly understand why inrush limiting is required. To do this reasonably
safely wear sunglasses and use earplugs. The resulting bang is outright
deafening, and it will do severe contact damage (likely spewing small molton
balls of metal outwards). If you do this with a mechanical switch the
contacts will be eroded to the point of uselessness in probably the very
first closing cycle.

If the capacitor charging current must flow through diode(s), it doesn't
matter how large they are (well maybe a hockey puck size one might handle
it) they will be destroyed if they instantly dissipate 125J. How much
inrush limiting is really necessary depends upon how stiff the 200V-500V
supply is. If it is very stiff (plain mains sent through a bridge rectifier
for instance) you will need an inrush limiter that can handle the full 125J.
If the supply is something like the secondary of a transformer sent through
a bridge recitfier, then the leakage inductances and winding resistances
will dissipate/limit peak amplitude of that of that 125J and your inrush
limiter need not be anywhere near as big.


NTC thermistors are the easiest way to provide inrush limiting, although one
that can handle 125J or more is going to be awefully large and very uncommon
(maybe non-existant). As such the more practical method employed at these
levels might be to use a decent sized power resistor (maybe a 15W+ wirewound
at around 4.7R) in parallel with a relay. The relay is energized by some
form of delay circuit, preferrably one that monitors the voltage on both
sides of the 4.7R power resistor and turns on only when it approaches a
small enough value. The relay should be rapidly turned off when the device
is powered down.

If this is for a high power xenon stobe or similar application then you
really aught to use a switch mode constant current source for best
efficiency. That is more complicated. Perhaps an understatement.


JeffM

unread,
Oct 11, 2003, 11:20:25 PM10/11/03
to
Don't sugar-coat it John. Tell us how you really feel.

Yzordderex

unread,
Oct 12, 2003, 1:55:41 AM10/12/03
to
"Fritz Schlunder" <m...@privacy.net> wrote in message news:<bm9t6e$kkbmc$1...@ID-203926.news.uni-berlin.de>...

> "Yzordderex" <yzord...@verizon.net> wrote in message
> >
> > What are we actually trying to do here? Just soft start a bus cap?
> > Boost 200v to 500v? I suspect it's just limit the charging current.
> > Maybe look at NTC thermister. NTC good for about 20 or 30 amps of DC
> > current. If you don't like that idea, just use a big diode bridge and
> > forget about the soft charge.
>
>
> Err... In this case some sort of inrush limiting is almost certainly
> mandatory. The OP's wording is slightly enigmous when he says, "from a DC
> source of 200 to 500VDc".

Yes, enigmous, like 90% of all posts here on this group.


My interpretation is he has a DC bus voltage that
> can be anywhere in the range of 200V to 500V. He has a 1000uF or larger
> capacitor which he wants to charge up from zero to this 200V to 500V range
> (depending upon what the bus voltage is when it is powered up).

Perhaps your interpretation is more accurate than mine. I suspected
he was charging off some type of AC line with a diode bridge.


>
> So at the very lowest range 1000uF up to 200V we have an energy storage of
> 0.5CV^2. In this case:
>
> 0.5*0.001*200^2=20 Joules
>
> At the higher range (but still assume 1000uF even though the OP said >1mF)
> of 500V we have:
>
> 0.5*0.001*500^2=125 Joules
>
>
> Now assuming we don't have some sort of switch mode constant current source
> or other charging method using inductors, the total energy that must be
> dissipated somewhere while charging the capacitor from 0V to 200-500V is 20J
> or 125J respectively. That is way too much energy at way too high a
> potential difference to avoid using inrush limiting.

Nope, inrush limiter not always needed. If we further assume (Gosh I
love assuming... The things I can assume) that he is going to use a
diode bridge, then we can also assume that he is going to be charging
caps off of AC line. There is always some impedance associated with
line. 115v lines are for the most part very soft. The complex Z is
easily calculated by knowing a few of the transformer parameters. If
these not known, it can be easily deduced by measuring voltage loaded
and unloaded. Do some calculations - you might be surprised at how
much inductance there is in a 20kva 230v single phase transformer you
might see on a phone pole outside your house.


>
> If the DC bus voltage is stiff enough, the effect of charging the capacitor
> instantly from it will be the same as discharging the capacitor by short
> circuiting it while charged to 200V-500V.

Agreed!


>
> Try short circuiting a 1000uF capacitor charged to 500V and you will
> instantly understand why inrush limiting is required. To do this reasonably
> safely wear sunglasses and use earplugs. The resulting bang is outright
> deafening, and it will do severe contact damage (likely spewing small molton
> balls of metal outwards). If you do this with a mechanical switch the
> contacts will be eroded to the point of uselessness in probably the very
> first closing cycle.

Ah, you're using the wrong kind of mechanical switch. Get hold of a
big piece of aluminum. Something sturdy with some mass - say
12"x12"x1" This is 1/2 the switch. Now get hold of a screwdriver and
grind the tip to almost a point - sorta round. Next wrap about 20
turns of 12 guage wire tightly around the screwdriver and tape it up
good with some sticky black tape.
Connect a similar wire to the sink using a crimp and tapped hole. Now
you got some kinda switch! To operate just jam sharpened screwdriver
into sink. This will discharge them caps in a hurry. It may take you
a few trys to get a nice clean hit. Look at it with a scope and you
will see no bounce too. It's like firing an SCR. Energy is
dissapated in cap ESR and wiring according to ratios of resistance.


>
> If the capacitor charging current must flow through diode(s), it doesn't
> matter how large they are (well maybe a hockey puck size one might handle
> it) they will be destroyed if they instantly dissipate 125J. How much
> inrush limiting is really necessary depends upon how stiff the 200V-500V
> supply is. If it is very stiff (plain mains sent through a bridge rectifier
> for instance) you will need an inrush limiter that can handle the full 125J.
> If the supply is something like the secondary of a transformer sent through
> a bridge recitfier, then the leakage inductances and winding resistances
> will dissipate/limit peak amplitude of that of that 125J and your inrush
> limiter need not be anywhere near as big.

Wall power is brought to you by way of transformer in your
neighborhood. Transformer with impedance. See my previous ramblings.

An interesting note is that I once designed a braking unit for a 10hp
480v ac motor drive. Drive had to be protected against fire if brake
transistor shorted. I used a normally closed temperature switch to
keep a small sensitive gate scr off. When switch opened up, little
scr fired. Little SCR turned on 10 amp darlington which fired a 90
amp 1200v SCR module. End result is that 1000uf cap discharged into
SCR. Same event also cleared JJS50 fuses in input line of drive.
Kerchunk, and about 7000 amps of anode current. SCR survived this
beating! So I've already tried discharging a 1000uf cap charged to
680vdc.


>
>
> NTC thermistors are the easiest way to provide inrush limiting, although one
> that can handle 125J or more is going to be awefully large and very uncommon
> (maybe non-existant). As such the more practical method employed at these
> levels might be to use a decent sized power resistor (maybe a 15W+ wirewound
> at around 4.7R) in parallel with a relay. The relay is energized by some
> form of delay circuit, preferrably one that monitors the voltage on both
> sides of the 4.7R power resistor and turns on only when it approaches a
> small enough value. The relay should be rapidly turned off when the device
> is powered down.

Not so. 200 Joule thermisters easily found. Amtherm in LasVegas
produces em by the truckload. 200J unit can be had in 32mm diameter
package. Digi-key may even sell em. Resistors are ok, but you have
to be careful how they are wound. Can use a few turns of skinny wire
or many turns of fatter wire to get same resistance. Few turns of
skinny wire could be a prob due to I^2t considerations. They will
blow open like a fuse. NTC and relay are a better choice for my
particular application.


>
> If this is for a high power xenon stobe or similar application then you
> really aught to use a switch mode constant current source for best
> efficiency. That is more complicated. Perhaps an understatement.

Ok Fritz, I think we agree on most points. But I've said this before
and I'll say it again, and again, and again. The problem is most
people don't supply the facts needed to converge to a solution. See my
post - What do you really want, what are you asking for, or something
like that. I got a lot of mileage outta that post. I think everybody
had something to say.

I enjoy nothing better than to explain, teach, help young engineers.
(I'm not so very far away from retirement) I realize we're not
dealing in a face to face situation here and it's difficult to get the
info across at times. Hope I didn't scare off the OP. OP if you're
there come back and tell us more about your application.

regards,
Bob

John Fields

unread,
Oct 12, 2003, 9:35:41 AM10/12/03
to
On 11 Oct 2003 20:20:25 -0700, jef...@email.com (JeffM) wrote:

>Don't sugar-coat it John. Tell us how you really feel.

---
If you would include a little of the post you're replying to it would
make it _much_ easier for everyone to figure out what you're talking
about.

--
John Fields

Winfield Hill

unread,
Oct 12, 2003, 9:53:45 AM10/12/03
to
Bob, Yzordderex wrote...
>
> Fritz Schlunder wrote ...

>>
>> NTC thermistors are the easiest way to provide inrush limiting,
>> although one that can handle 125J or more is going to be awefully
>> large and very uncommon (maybe non-existant). As such the more
>> practical method employed at these levels might be to use a decent
>> sized power resistor (maybe a 15W+ wirewound at around 4.7R) in
>> parallel with a relay. The relay is energized by some form of
>> delay circuit, preferrably one that monitors the voltage on both
>> sides of the 4.7R power resistor and turns on only when it
>> approaches a small enough value. The relay should be rapidly
>> turned off when the device is powered down.
>
> Not so. 200 Joule thermisters easily found. Amtherm in LasVegas
> produces em by the truckload. 200J unit can be had in 32mm diameter
> package. Digi-key may even sell em. Resistors are ok, but you have
> to be careful how they are wound. Can use a few turns of skinny wire
> or many turns of fatter wire to get same resistance. Few turns of
> skinny wire could be a prob due to I^2t considerations. They will
> blow open like a fuse. NTC and relay are a better choice for my
> particular application.

That's Ametherm in Carson City, NV. see http://www.ametherm.com/
They offer the bigAMP 32mm inrush-current limiters rated for 100 to
250J, and the MegaSurge series rated at 110 to 900J.

"Ametherm's bigAMP Inrush Current Limiters are specially designed
thermistors able to withstand up to 36 amperes of continuous current
and 260 joules of input energy."
http://www.ametherm.com/download/bigAMP1.pdf

and also this...

"Ametherm's MegaSurge Inrush Current Limiter is an economical and
space saving way to limit high inrush current. It is specially
designed to withstand up to 50 amperes of continuous current and
900 joules of input energy."
http://www.ametherm.com/download/MegaSurge%20%20DS%202003.pdf

The 900J rating is their MS35 0R550, a 35mm dia x 9mm thick 0.5-ohm
(25C) part, droping to 0.008 to 0.05 ohms warm, depending on current.

Ametherm has a wide range of part ratings to chose from. They say,
1) Use the maximum allowable surge current and Ohm's Law to determine
the least allowable resistance at turn on for your application.
2) Using the formula J = ½ C V², determine how much input energy
the thermistor will absorb when the device is turned on.
3) Determine the maximum steady state current that will flow through
the Inrush Current Limiter.
4) Select the bigAMP or MegaSurge Inrush Current Limiter that will
work for your application.

Obviously that's for capacitor-charging. But they also discuss
how to design transformer inrush current limiters:
http://www.ametherm.com/Inrush_Current/transformer_case_study.htm

Ametherm, Inc. mailto:sam...@ametherm.com
3111 N. Deer Run Rd #4
Carson City, NV 89701
(800) 808-2434, (775) 884-2434 or FAX (775) 884-0670

Thanks,
- Win

whill_at_picovolt-dot-com

Heindorf

unread,
Oct 13, 2003, 6:50:06 AM10/13/03
to

John Fields schrieb:

Hello John
Probably it's one of the simple solutions but I was actually going to avoid
any mechanical components. The problem is that a boost switcher is supplied
via the capacitor and the start-up circuit should disappear afterwards if
possible without series resistance.
Rolf

Heindorf

unread,
Oct 13, 2003, 6:53:54 AM10/13/03
to
Hi,
the main current path should have "zero"resistance after start-up, if possible.
Rolf


CBarn24050 schrieb:

Heindorf

unread,
Oct 13, 2003, 7:03:50 AM10/13/03
to
You're right, Fritz
the source is a fuel cell of 200 to500V and this voltage has to be boosted to
700V.
The boost circuit is using the cap on its input. Without using a relay I think
we can only employ 2 seperate circuits one charging the cap and "disappearing "
afterwards and the main current path is enabled subsequently.
Sorry for being not specific enough.
Rolf

Fritz Schlunder schrieb:

Yzordderex

unread,
Oct 13, 2003, 11:41:59 AM10/13/03
to
Heindorf <hein...@aeras.de> wrote in message news:<3F8A835E...@aeras.de>...
> John Fields schrieb said something too:


----snippety do dah here ----

> Hello John
> Probably it's one of the simple solutions but I was actually going to avoid
> any mechanical components. The problem is that a boost switcher is supplied
> via the capacitor and the start-up circuit should disappear afterwards if
> possible without series resistance.
> Rolf

SCR and large NTC inrush limiter work? Maybe have boost switcher
tertiary winding supply picket fence gate drive to SCR.

good luck,
Bob
N9NEO

edward....@gmail.com

unread,
Mar 28, 2016, 12:44:07 PM3/28/16
to
This old thread describes exactly our problem. The equipment was designed/built in the 2000s. Problem dealing with 1000uF 200V caps. Perhaps the OP was dealing with the same equipment.

On Saturday, October 11, 2003 at 2:41:06 PM UTC-7, Fritz Schlunder wrote:
> "Yzordderex" <yzord...@verizon.net> wrote in message
> >
> > What are we actually trying to do here? Just soft start a bus cap?
> > Boost 200v to 500v? I suspect it's just limit the charging current.

Yes, inrush current as well as overall current limit.

> > Maybe look at NTC thermister. NTC good for about 20 or 30 amps of DC
> > current. If you don't like that idea, just use a big diode bridge and
> > forget about the soft charge.

20A is way too big. Ideally, we don't want more than 2A.
How about a large resistor, says 10K, for a pre-power up switch, then power-up switch a few minutes later. It's a quick and dirty temporary solution until we have something better.
Reply all
Reply to author
Forward
0 new messages