limiting smotthing capacitor inrush current
Miguel A
I'm planning to build a simetric variable regulated power supply,
0 to -+14V / 2A. I bought 2 used 15000 uF capacitors that I intent to
use as smothing capacitors. It's nothing professional just for
my bench.
My questin is: in face of such a large capacity value should I
include some form of capacitor current limiting when the supply
is switched on? I though of using a small lamp which as a
resistance has a positive temperature coefficient but I don't
know if that is the best solution. If I don't include any form
of current limiting cound'n the initial current damage the diodes
and the capacitors? Won't the transformer secondary being an inductor
contribut to limit the inrush current?
thanks in advance for any sugestions.
Miguel_A
John Popelish
Both primary and secondary resistance and the leakage inductance
between them will help to limit the peak inrush current. You can
estimate the peak current pretty closely, by shorting the secondary,
and driving the primary with a scaled down voltage (say, 1/10th of
normal) while measuring the primary current. Scale this reading back
up by the inverse of the ratio that you scaled the voltage down, and
you have a good measure of the actual peak primary current. This can
also be used to estimate the peak secondary current, by multiplying by
the inverse of the primary to secondary voltage ratio. Silicon diodes
with large peak current capability are very easy to find, but you may
have to use a slow blow fuse on the primary to keep it from popping.
If this test results in a larger inrush than you think is acceptable,
then you think about adding extra components.
The lamp is a bad idea, because it offers the least resistance when it
is cold, right when the inrush is happening, and then, when the supply
is up and running, it heats up, and robs maximum voltage. Negative
temperature coefficient thermistors are normally used for this sort of
thing, but I have also used a small inductor in the primary (which
also helps RF filtering and improves power factor) to lower the worst
case inrush. The thermistor would have to have to be rated for the
peak voltage that can be impressed across it, and not overheat at the
worst case full load current and survive a shorted diode till the fuse
blows. Thermistors also have the weakness that if the supply is
switched off an back on as soon as the capacitors are discharged, but
before the thermistor cools off, they allow a larger than normal
inrush. But they help enough that their use is very common.
A very rough estimation based on your 28 volts @ 2 amp output implies
about 1 amp RMS line current (@ 120 VAC), so you would need a
thermistor rated for at least 150 volts peak and 1 amp continuous
current when hot. Digikey sells a unit (KC013L) that has a cold
resistance of 50 ohms (about 3 amps peak inrush possible) and a rated
continuous current of 1.6 amps with a hot resistance (@ 1.6 amps) of
0.75 ohms which might work. But at 1 amp RMS it will produce about a
watt of heat in your supply box and drop you effective line voltage by
a couple volts (because of the big peak currents at the top of the
voltage wave), so take that into account. They have several other
slightly different choices, also. The larger diameter units have
bigger cold to hot resistance ratios, but take longer to cool off
before the next start is protected.
--
John Popelish
Rich Grise
I built a bench supply using a scrounged 24VCT 5A transformer and,
I think, about a 55,000 uF capacitor (single voltage supply.) I
used a pretty large diode, I think 25 amps, and didn't do anything
about inrush current. I believe, as John has mentioned, that the
transformer's leakage inductance and winding resistance will do
a fine job of limiting inrush current. I haven't seen a lot of
hobbyist-level supply projects, but of those I have seen, none
of them did anything special about inrush current.
Also, as John mentioned, a slow-blow fuse on the primary sounds
like a good idea.
Good Luck!
Rich
john jardine
The beauty about transformers, is that their own inductive
characteristics dont really enter into a circuits operation. They just
'transform'. In your case the transformer will simply 'transform' the
electrical effects occurring on the secondary side, into exactly the
same effects on the primary side but with a turns ratio scaling
effect.
You're probably as well just allowing the supply to run barefoot at
start up. Yes, the caps are real dustbins and need to fill with a few
joules during the first couple of half cycles of the incoming power.
Let the transformer winding resistances and the dynamic diode bridge
forward resistances do the current limiting work.
The alternative is say 5 ohms of external series resistance to limit
the peak cap charging current but this then brings with it other
problems of resistor heat dissapation and voltage losses under normal
use. Alternatively the external resistor could be switched out of
circuit after say a second but then this is getting even more
complicated.
John
Stevie Blunder
relay that shorts out a (big) series resistor when the caps reach ~80% of
their normal voltage.
"Miguel A" <Migu...@sapo.pt> wrote in message
news:23ca9cd8.02070...@posting.google.com...
Miguel A
for the sugestions.
I will first follow John's indications and determine the
peak inrush current expected, by the process he described.
That will help me to chose the adquate rectifier diodes.
If the value is to high I'll probably include a small
inductor on the primary side.
cheers,
Miguel_A
John Popelish
Please come back with your measurements and any difficulties you found
making them. Thanks.
--
John Popelish
Joakim Asplund
>
> The beauty about transformers, is that their own inductive
> characteristics dont really enter into a circuits operation. They just
> 'transform'. In your case the transformer will simply 'transform' the
> electrical effects occurring on the secondary side, into exactly the
> same effects on the primary side but with a turns ratio scaling
> effect.
>
Umm... The transformer is an inductor with more windings on it. In an
ideal transformer the output voltage equals to the input voltage times
turns ratio, and current the other way around. But transformers are not
ideal. There is leakage inductance, wire resistance, stray fields, core
losses etc. thatmakes the input differ from the output.
--
Joakim Asplund
http://megajocke.cjb.net
john jardine
> ideal transformer the output voltage equals to the input voltage times
> turns ratio, and current the other way around. But transformers are not
> ideal. There is leakage inductance, wire resistance, stray fields, core
> losses etc. thatmakes the input differ from the output.
news:<3D2C5EC3...@hotmail.com>...
(input voltage -divided- by the "turns ratio" if the usual step down
:)
Joakim, I of course agree in what you say about the usual transformer
losses. I was coming in from the point of view of Miguel mentioning
the transformer inductance as possibly helping to limit the inrush
currents. As it's the electronics 'basics' newsgroup I made the
assumption he was actually thinking of the Henry's of inductance he
could have measured in the transformer windings and I could
understand what he was thinking with regard to it's possible 'rate of
current rise' limiting effects, (I've been there myself!).
I was though, trying to suggest, that at switch on, the transformer
via normal transformer action, 'reflects' the electrical state (the
secondary's real and imaginary vector situation) relating to the
discharged capacitors, into the primary winding and thus presents this
'discharged capacitor load', straight to the input power terminals.
The 'big discharged caps' as an electrical load is significantly
greater in magnitude and effect than those caused by the normal
transformer loss mechanisms, which in this case, the ones we want,
would be the lumped winding resistances and leakage inductances,
(which in turn, also appear as a primary 'load').
The biggest fuse blowing problem of all has not actually been
mentioned here and is the regular and quite normal occurrence of
residual magnetism in the transformer core material, causing core
saturation at the instant of switch on. Thousands of amps will flow
even if the secondary is open circuit. This is the cause of most fuse
blowing problems with PSU's and big hi-fi setups.
Regards
John.
Miguel A
> mentioned here and is the regular and quite normal occurrence of
> residual magnetism in the transformer core material, causing core
> saturation at the instant of switch on. Thousands of amps will flow
> even if the secondary is open circuit. This is the cause of most fuse
> blowing problems with PSU's and big hi-fi setups.
>
> Regards
> John.
I had already thought about that. If there was no residual magnetism
in the transformer core, probably this inrush current wouldn't be such
a problem because in the imediate instant after power was applied to
the primary, the transformer woudn't be nothing more than two coils
inte "air". It's as if the core wasn't there, and the current even in
a secondary short circuit situation would be very small. As the core
becomes more saturated, the magnetic flux increases, secondary current
increases but by that time some charge would be stored in the
capacitors and the short-circuit situation would be gone. The result
would be a "soft" power up in the secondary all as a result of the
magnetic "inercia" of the transformer core.
Did I understand correctly? This has allways been something that makes
sense to me but I've never been very sure of!
Regards,
Migel_A
Joakim Asplund
> (input voltage -divided- by the "turns ratio" if the usual step down
> :)
>
> I was though, trying to suggest, that at switch on, the transformer
> via normal transformer action, 'reflects' the electrical state (the
> secondary's real and imaginary vector situation) relating to the
> discharged capacitors, into the primary winding and thus presents this
> 'discharged capacitor load', straight to the input power terminals.
> The 'big discharged caps' as an electrical load is significantly
> greater in magnitude and effect than those caused by the normal
> transformer loss mechanisms, which in this case, the ones we want,
> would be the lumped winding resistances and leakage inductances,
> (which in turn, also appear as a primary 'load').
The winding resistance in the secondary and primary and the leakage
inductance will limit the current a bit at switch on, so it won't be
exactly reflected.
> The biggest fuse blowing problem of all has not actually been
> mentioned here and is the regular and quite normal occurrence of
> residual magnetism in the transformer core material, causing core
> saturation at the instant of switch on. Thousands of amps will flow
> even if the secondary is open circuit. This is the cause of most fuse
> blowing problems with PSU's and big hi-fi setups.
Yes this problem is probably bigger. How about a transformer degaussing
circut? :-)
Joakim Asplund
>
> I had already thought about that. If there was no residual magnetism
> in the transformer core, probably this inrush current wouldn't be such
> a problem because in the imediate instant after power was applied to
> the primary, the transformer woudn't be nothing more than two coils
> inte "air".
the fuse would blow instantly.
> It's as if the core wasn't there, and the current even in
> a secondary short circuit situation would be very small.
As long as the core is not saturated it will work normally as a
transformer, so the short circuit on the secondary will be reflected to
the secondary, but the winding resistances and leakage inductance will
limit the current.
> As the core
> becomes more saturated, the magnetic flux increases, secondary current
> increases but by that time some charge would be stored in the
> capacitors and the short-circuit situation would be gone.
As the core becomes more saturated, it will draw a huge magnetisation
current causing a high voltage drop across the primary resistance, so
the output voltage will go down fast.
How it will get out of saturation:
If a flux is still in the core at start up from a previous turn off, the
core can saturate, which will cause the inductive part of the primary
to become almost a short, the only thing limiting current would be the
primary resistance and resistance in the power feed, switch and fuse.
This limiting will make the voltage over the inductive part lower in
that direction which causes a DC offset because it would not saturate in
the other direction, making the voltage higher. This will cause it to
eventually reset, but the current during this phase can be quite high
and blow the fuse if you are unlucky. And of course, the capacitor
charging on the secondary would make it draw even more current.
> The result
> would be a "soft" power up in the secondary all as a result of the
> magnetic "inercia" of the transformer core.
Actually, the core doesn't take much part in the transfer, but it is
needed because of without it the primary inductance would be way too
low, and the leakage inductance too big. Why a more powerful transformer
needs a bigger core is mostly due to that the windings will take more
space, and the bigger core also means that fewer turns can be used for
the same voltage.
> Did I understand correctly? This has allways been something that makes
> sense to me but I've never been very sure of!
>
Not exactly, but almost :-)
> Regards,
>
> Migel_A
John Popelish
You have got a few facts wrong. The transformer works as a
transformer, not because it holds a magnetic field, but because the
core can have a large change in its total magnetic field as voltage is
applied. This change in field is what produces the inductance that
holds back the magnetizing current. The magnetizing process runs
effectively in parallel to the transforming process that produces
voltage across the secondary winding.
The problem at start up occurs when the last half cycle of voltage
that was applied to the primary is in the same direction as the first
half cycle of applied voltage the next time the transformer is
powered. The last half cycle had its magnetizing current held back by
allowing the core to be strongly magnetized in one direction. Then,
when the transformer is again powered up, the first half cycle tries
to continue that magnetizing process in the same direction, and if the
core held onto most of the magnetic field from the last half cycle,
accumulates so much magnetic field that it saturates. So, for the
rest of that first half cycle, there is effectively no iron inside the
primary, so the effective inductance of the primary (that is in
parallel with the transformer process) gets very low, and passes
current that is limited mostly by the resistance of the primary. If
the first half cycle is in the opposite direction as the last half
cycle, no problem occurs.
If the core had no remnance (ability to hold a magnetic field
internally, when there was no current applied to the primary) this
would not happen, but silicon steel and most other practical core
materials will retain a significant fraction of their magnetic field
when the primary current goes to zero, unless the core has some air
gap in the flux path, and such a gap increases the magnetizing current
during normal operation (lowers the primary inductance that is in
parallel with the transformer process).
--
John Popelish