EMRFD Message Archive 2878

Message Date From Subject
2878 2009-04-07 12:15:50 ha5rxz Transformer Slow Start Circuit
Dear EMRFD Group

I need your help in settling an argument. Last night at the radio club we were talking about slow start circuits for the primary (230v) side of toroidal transformers. The general idea is a 50 ohm resistor in series which is shorted out after a fraction of a second by a pair of relay contacts.

What wattage should the resistor be?

Some of the members argued that the resistor should be a high wattage device on a heatsink (say about fifty watts) because of the power involved. Other members thought that a lower wattage resistor could be used because it would only be under load for a fraction of a second before the relay contacts closed, perhaps a ten watt resistor could be used.

All comments regarding this problem are welcome but please remember that we have 230v electricity here in Europe so please scale any formulae accordingly.

HA5RXZ
2879 2009-04-07 13:14:33 Art Re: Transformer Slow Start Circuit
You need to specify the peak load current in order to answer the question.

The limiting resistors are generally not needed unless there is a
heating element (such as in the heater of a vacuum tube or a large
bank of conventional light bulbs). Another case might be when a large
bank of capacitors need to be charged initially. In both cases, the
initial current draw is significant, especially in the case of the
heating element which starts are near zero ohms resistance at room
temperature and decreases significantly once it is warm(er).

However, you didn't state whether either of these types of loads were
present, so it's not even possible to know whether current limiting
is needed or not.

In either case however, you will need to supply the current vs time
data for load you intend to power up.

Enjoy.

Art



>Dear EMRFD Group
>
>I need your help in settling an argument. Last night at the radio
>club we were talking about slow start circuits for the primary
>(230v) side of toroidal transformers. The general idea is a 50 ohm
>resistor in series which is shorted out after a fraction of a second
>by a pair of relay contacts.
>
>What wattage should the resistor be?
>
>Some of the members argued that the resistor should be a high
>wattage device on a heatsink (say about fifty watts) because of the
>power involved. Other members thought that a lower wattage resistor
>could be used because it would only be under load for a fraction of
>a second before the relay contacts closed, perhaps a ten watt
>resistor could be used.
>
>All comments regarding this problem are welcome but please remember
>that we have 230v electricity here in Europe so please scale any
>formulae accordingly.
>
>HA5RXZ
2880 2009-04-07 14:08:46 John Levreault Re: Transformer Slow Start Circuit
There are also devices known as "inrush current limiters". They are
basically thermistors with a very large negative temperature coefficient
(NTC). They possess a high resistance when cold, but the resistance
drops rapidly as they heat up. They can handle very large inrush surges
and are very effective and useful for your application. However, you
need to specify the quiescent load current to properly select a device.

They're made by companies like Ametherm, AVX, Cantherm, Murata, etc.,
and are readily available in a variety of values here in the US from
Digikey.

de John NB1I


ha5rxz wrote:

>Dear EMRFD Group
>
>I need your help in settling an argument. Last night at the radio club we were talking about slow start circuits for the primary (230v) side of toroidal transformers. The general idea is a 50 ohm resistor in series which is shorted out after a fraction of a second by a pair of relay contacts.
>
>What wattage should the resistor be?
>
>Some of the members argued that the resistor should be a high wattage device on a heatsink (say about fifty watts) because of the power involved. Other members thought that a lower wattage resistor could be used because it would only be under load for a fraction of a second before the relay contacts closed, perhaps a ten watt resistor could be used.
>
>All comments regarding this problem are welcome but please remember that we have 230v electricity here in Europe so please scale any formulae accordingly.
>
>HA5RXZ
>
>
>
>
>
>------------------------------------
>
>Yahoo! Groups Links
>
>
>
>
>------------------------------------------------------------------------
>
>
>No virus found in this incoming message.
>Checked by AVG.
>Version: 8.0.100 / Virus Database: 270.11.45/2045 - Release Date: 4/7/2009 6:41 AM
>
>
2881 2009-04-07 19:36:11 Paul Anderson Re: Transformer Slow Start Circuit
2882 2009-04-07 21:34:37 Don Hackler Re: Transformer Slow Start Circuit
Depending on the resistor and load ratings, the resistors get hot or
burn up. On high power transmitters, this scheme is used to step-
start the plate supply. Sometimes they interlock the system to shut
it down if the shunt relay doesn't kick in quickly enough.


2883 2009-04-08 03:19:14 Ian White GM3SEK Re: Transformer Slow Start Circuit
Paul Anderson wrote:
>
>
2886 2009-04-08 05:39:04 Nick Kennedy Re: Transformer Slow Start Circuit
It's interesting to consider where the inrush comes from. Saturation
is generally blamed, but why does it occur? Residual magnetism is
given as a reason and it's probably a contributor, but the point of
energization in the voltage sine wave is a major factor. No residual
magnetism is needed for a doubling of normal flux during start-up.

If the transformer is energized at a zero crossing, a worst case peak
in flux occurs. That's because you integrate an entire half cycel of
voltage before reversing. A means of energizing at the peak of the
sine wave could eliminate the problem, but that would be a lot more
complicated than the limiting circuit.

No one here has tried to calculate the required wattage and I'm not
sure it's possible. It might have to be done empirically. And how
long does it take for the offset flux, the DC component, to decay
away? That's probably another one for the lab. I'm guessing a few
tens of cycles ...

In the world of large power transformers, protective relaying can be
fooled by the inrush caused by saturation and trip as though there
were an internal fault. Relay designers addressed this by adding a
"harmonic restraint" to the relay. Saturation produces harmonic
currents as the transformer becomes non-linear, and these currents are
detected and "restrain" the force on the tripping contact.

73--Nick, WA5BDU
2890 2009-04-09 01:25:18 ha5rxz Re: Transformer Slow Start Circuit (More Information)
Thank you for the posts so far. The possibility of fitting NTC devices to limit inrush current is interesting but we cannot use them in this case because there are two transformers and we intend to use relays for sequencing as well as current limiting. The relay sequence completes in a quarter of a second

We have some thermal fuses in the shack junk box that trip at 70C. These can be fitted in series with the resistors in case of relay failure.

I went down to the club shack this morning and measured the input current of the transformers. Both of them are toroid units with one used for the tube heaters and the other for the high voltage supply. The heater transformer consumes 4.1 amps @ 230 volts and the HT transformer consumes 3.3 amps, the startup current is a LOT higher and enough to make the lights flicker when we throw the switch, this is why we need to fit a slow start circuit.

All of this work is to upgrade the club linear amplifier.

HA5RXZ
2904 2009-04-10 10:23:12 Bill Linn Re: Transformer Slow Start Circuit (More Information)
2910 2009-04-11 02:00:00 ha5rxz Re: Transformer Slow Start Circuit (More Information)
This looks like another way of controlling the switch on delay, we use relay logic because of the high RF environment but this website uses analog circuits for the delay. The result is the same but I regret that it does not answer the original question, do we need to use large power resistors or can we get away with smaller devices as they will only be passing current for a fracti
2911 2009-04-11 08:21:08 michael taylor Re: Transformer Slow Start Circuit (More Information)
2912 2009-04-11 09:36:08 Gary Johnson Re: Transformer Slow Start Circuit (More Information)
Here's my view of the problem, based on slow-start circuits for 50 kW high-voltage e-beam supplies. The reasons for the surge are 1) charging the capacitors and, to a much lesser extent, 2) initial magnetization of the transformer. We use wirewound resistors and calculate the value from two fundamental requirements. First, what peak line current do you wish to tolerate? That's simply peak_current = 1.4*rms_line_voltage / R. We usually set that at no more than 5X the circuit rating because circuit breakers are tolerant of this kind of overload. So are slo-blow fuses.

The other way to look at it is energy delivered. You need to deliver 0.5 * C * V^2 Joules to charge the caps. Since a Joule is a Watt*sec, you can roughly estimate the charge time from your inrush resistor value and line voltage. Since the power supply isn't a dead short the whole time it's charging up, let's say that 1/2 of the line voltage appears across the inrush resistor, on the average. So the average charging power is P = (0.5* rms_line_voltage)^2 / R. Charging time is then roughly the number of Joules to charge the caps divided by that power. You can also simulate all this in Spice, or just fiddle with some numbers until something reasonable appears. Keep the charge time very short, maybe 50 ms is about right.

As for power dissipation in the resistor: First, use a very simple time delay relay, for reliability. Based on your calculation above, you can figure a rough delay time. It should be very short, like 0.1 sec. Second, use wirewound resistors. They have excellent ENERGY dissipation capability. The problem here is not average power, it is short-term (pulsed) energy absorption. They also are made of non-flammable materials (metal and ceramic) and so in case of long-term overload are not nearly as nasty as vaporizing a bunch of nasty phenolic material. Here is a nice datasheet with graphs of peak power vs. pulse time on p. 28: <http://www.vishay.com/docs/28730/acseries.pdf>. This data can be used to estimate the resistor power rating since you know the peak power you will deliver and for how long.

Third, if you want to be completely safe, size the resistor so that it will tolerate the failure of the time-delay relay. This may be hard to calculate because behavior of the load on the power supply may depend upon the delivered DC voltage. We used a second set of contacts on the relay to disable the application of the load and turn
2916 2009-04-12 08:55:08 ha5rxz Re: Transformer Slow Start Circuit (More Information)
"Did that answer the question?

Gary, WB9JPS"

Yes, thank you. The link to the Vishay document was also quite useful although I must confess that I needed some help with the mathematics.

HA5RXZ


2917 2009-04-12 09:35:54 Stephen Wandling Re: Transformer Slow Start Circuit (More Information)
I have been following this in a general fashion and so this may have
been covered and I missed it.

I understand that there will be a large inrush current and I grasp why.
What I don't see is "What is the problem with the large inrush
current?" Is there potential damage to the charging circuit? Is there
risk of tripping the supply breaker? Is the limiting simply good
engineering practice?

Thanks

Stephen
VE7NSD

ha5rxz wrote:
> "Did that answer the question?
>
> Gary, WB9JPS"
>
> Yes, thank you. The link to the Vishay document was also quite useful although I must confess that I needed some help with the mathematics.
>
> HA5RXZ
>
>
>
2918 2009-04-12 10:24:18 Don Hackler Re: Transformer Slow Start Circuit (More Information)
On large DC supplies, the problem is keeping the inrush current from
tripping breakers or causing the line voltage to sag.
With large broadcast transmitters, the primary fuses on the power pole
outside could blow without the step-start circuits.
The beam supply for each output tube of a typical UHF TV transmitter
has a 480volt three-phase feed fused at 150 amps. Even with the step-
start, the lights dim when it's first turned on.

On a small DC supply, like in a ham transmitter, it's the same problem
on a smaller scale. The branch circuit running your ham gear might
not be up to handling initial surge. Also, If you are running on a
generator, the inrush current might be enough to stall the generator.

2919 2009-04-13 07:31:12 Ian White GM3SEK Re: Transformer Slow Start Circuit (More Information)
Don Hackler wrote:
>With large broadcast transmitters, the primary fuses on the power pole
>outside could blow without the step-start circuits.
>The beam supply for each output tube of a typical UHF TV transmitter
>has a 480volt three-phase feed fused at 150 amps. Even with the step-
>start, the lights dim when it's first turned on.
>
>On a small DC supply, like in a ham transmitter, it's the same problem
>on a smaller scale. The branch circuit running your ham gear might
>not be up to handling initial surge. Also, If you are running on a
>generator, the inrush current might be enough to stall the generator.
>

Also, if the fuse or breaker on the power supply is sized to handle an
uncontrolled switch-on surge, that leaves the PSU with much poorer
protection against faults occurring after startup. One of the greatest
advantages of a step start is that it allows the use of a faster,
lower-current fuse or breaker.

Coming back to the practical question about the size and type of surge
limiting resistor required, most amplifiers of the 1.5kW RF class can
manage with a 50W wirewound resistor. For example, with a 230V AC
supply, a 22 ohm resistor will limit the peak current to 10A.

The optimum combination of resistor value and step time will depend on
parameters like the size and winding resistances of the transformer, and
teh value of smoothing capacitance that needs to be charged. The timing
and the resistor value can be optimized to make the two current surges,
at switch-on and when the resistor is shorted, roughly equal though this
is not critical. With our kind of power supplies, it can often be done
well enough by judging the loudness of the two "thumps".

The step time rarely needs to be more than 1 second. This allows the
series resistor to be protected by a slow-blow fuse (eg a "4A T" rated
fuse for a peak current of 10A). This fuse will never blow in normal
operation, but it should protect the resistor if the timing circuit ever
"sticks".

Another important point about timing circuits is that they should be
re-triggerable. If the mains supply suffers any kind of glitch - and
this includes operator error on the main switch - the startup timer
should always deliver a new full-length timing pulse. Many timing
circuits won't do this.


--

73 from Ian GM3SEK
http://www.ifwtech.co.uk/g3sek
2921 2009-04-13 11:03:27 ha5rxz Re: Transformer Slow Start Circuit (More Information)
"I understand that there will be a large inrush current and I grasp why. What I don't see is "What is the problem with the large inrush current?" Is there potential damage to the charging circuit? Is there risk of tripping the supply breaker? Is the limiting simply good engineering practice?

Easy to answer. Here in Eastern Europe we do not have the luxury of heavy current mains supplies. When I lived in the UK our house had a 150 Amp supply but when I came back home to Hungary my apartment 230v supply was divided between three ten amp breakers. We face a similar problem in the radio club shack, our total supply is twenty amps @ 230v split between two breakers. Without a slow start circuit on the linear (and a similar circuit on the 12v rig supply) the breakers trip every time due to the switch on surge.

HA5RXZ
2926 2009-04-14 04:56:59 Harold E. Johnson Re: Transformer Slow Start Circuit (More Information)
In days of old..........The first Signal One, the CX-7, used a
Hypersil cored 550 Watt transformer and would blow the push button
on-off switch in about 3-5 on-off sequences. The switch was good for
10 Amps I believe. (120 Volt operation) Meaurement with a scope,
showed first cycle currents well over 55 Amps. Core magnetization and
capacitor charging currents. This from a transformer not much larger
than your fist. We finally cured the problem with a series Globar
resistor with the AC line. 110 Ohms cold and 6 Ohms hot and not a lot
bigger that a couple of quarters stacked together. Another way to skin
a cat, although you really shouldn't expect a lot of help from it of
you turn the rig off and then immediately turn it back on again.

W4ZCB