1000 VA 230 VAC 50 Hz - How Much Inrush Current?

[dannybeckett]

Hi guys,

I'm attempting to connect a chunky 1000VA torroidal to mains, however I'm annihilating fuses like crazy and tripping the ring main RCBO. At first I thought I had wired things wrong, or even damaged the transformer, but could I be vaporising fuses due to the inrush current this thing can draw?

The suspect in question is here : https://uk.farnell.com/multicomp/mcta1000-55/transformer-toroidal-2-x-55v-1000va/dp/9533192

Any help in terms of worst case inrush would be great, and perhaps some ideas of how to tame it if it is indeed inrush that's blowing circuit breakers and fuses. The RCBO is 20A and it's blown 10A fuses, which does make me question if I've done something fundamentally wrong.

[Benta]

I don't think you've done anything wrong.
In my experience, any transformer (especially toroid) over 300 VA will blow your fuses intermittently.
You'll need to think about a soft-start mechanism, which isn't easy.
The goal is to switch on the transformer at peak mains voltage, or a little bit later. That will avoid the inrush current.

[AlfBaz]

You'll need to think about a soft-start mechanism, which isn't easy.

Current limiting resistor on the primary which is shorted out by a relay after 1 to 2 seconds would be the easiest option

[dannybeckett]

Yes I think you're right. I've been doing a little reading on the subect and have just discovered how powerful inrush current can be for larger transformers. This is a good article on the matter:

https://iosrjournals.org/iosr-jeee/Papers/Conf.17017/Volume-2/8.%2059-63.pdf?id=7590

Plenty of vidoes on YouTube illustrating the point too:



In reality, I have 2x 1000 VA transformers in parallel I'm connecting to line voltage so no wonder these fuses were getting utterly oblitorated. I'll throw together some arduino + relay thing that will close a relay contact at max amplitude (accounting for any mechanical relay delay) and see if that does the trick.

Thanks for the confirmation   

[SeanB]

Arduino plus relay is really overkill, just use a simpler method, a small transformer, to say 24VAC, a bridge rectifier, and then a series resistor in the positive output of the rectifier. Then a 1000uF to 2200uF 63V electrolytic (can be any one, 85C is fine here with only 100/120Hz ripple and almost no current in the capacitor) and then a 24VDC relay, that will close to short out a resistor in the feed to the large transformers. Resistor can be any value from around 4R7 to 22R, 10W, with a 130C single shot thermal fuse, held to the body of the resistor with some kapton tape. Thermal fuse on the incoming side before the relay contacts, and the NO contacts of the relay short out the fuse. Relay contacts rated for operating current of the input side, so for a 1kVA transformer on 120VAC mains a 10A contact rating, though a 16A relay is likely to be the same size and price, and available off the shelf in industrial suppliers. Will just last longer, or outlive the equipment.

Series resistor for the relay has to be related to coil resistance, generally around a third to half the resistance, and adequate power dissipation as well.  This will take perhaps a half second to close after power is applied, well enough time to limit both initial magnetising current of the toroid, and also charging of the likely massive capacitors on the secondary side. Adjust capacitor value till you get to somewhere between a half second and a second for the relay contacts to close.

Thermal fuse in case the relay contacts do not close and the resistor starts to overheat. All the hot running parts mounted either spaced off the PCB, or mounted using tag strips so they are in free air and not close to anything flammable.

Did that, though I used series resistors to directly rectified mains, as I had some 48V relays with high resistance coils, and used 400V electrolytics instead. No need for a catch diode across the relay coil either, there is no back EMF on it at all, because of the capacitor. Done on a 2kVA variable voltage transformer, that had an annoying ability to trip the thermal breakers used in the supply, despite them being 32A breakers, and not supplying a ring main.

[dannybeckett]

Well I already have the bits to make the Arduino inrush-avoider, and the circuit + code is actually really simple. 5V USB adapter, arduino, relay, and a mains voltage measurement circuit. Something like the attached image I think will work fine.

Oddly, Microchip app note AN954 "Transformerless Power Supplies: Resistive and Capacitive" have all of the current resisting stuff in the neutral line. Any idea why this is?

https://ww1.microchip.com/downloads/en/appnotes/00954a.pdf

[Kleinstein]

The circuit as shown is odd. 100 Mohms in parallel to 10 K and the zener do not make any sense.

Why use an extra µC, if just an RC is good enough to produce the delay of some 200-500 ms. A µC is only more complicatged and likely still needs an extra driver for the power relay. An extra driver transistor for the relay could be used with a smaller cap for the delay.

The relay can often be driven from the main transformer as well. A capacitive droper supply for the relay is also be possible, but still needs the large filter cap.

[dannybeckett]

Allow me to elaborate:

• 100M is the simulated input impedance of Arduino
• 1M + 47nF limits incoming current
• 10k forms divider with 1M for usable signal
• Zener clamps voltage seen @ micro input between -0.7V and +4.7V (ish)

I prefer phase-accurate switching as you elimirate inrush current by addressing the root cause, rather than simply handling the fallout using big resitors / caps / additional transformer / rectifiers / thermal fuse etc etc. Attached image is refined version, 260 uA draw @ 50 Hz. THe idea with C1 is to protect against short circuits but unsure if it provides any real benefit. C2 could be removed if C1 is not fitted.

I still don't understand why Microchip have put all their current limiting stuff in the neutral line...

[Kleinstein]

Phase correct switching is a possiblitiy. However one than also needs to remember what happend on turn off. The right time to turn the transfromer on depends on exacty how it was shut down. Turn on at zero crossing without caring about the past has a 50% chance to hit the worst case.

With the capacitive droper the capacitor and additional sereis resistors can be at different positions. Splitting it to both sides is a reminder that the power is no isolated. Often it does not matter, but sometimes it makes a difference at what side the parts are.

[dannybeckett]

Actually, turning on at zero crossing will give you 100% chance of worse case

As Benta mentioned, the idea is to switch the transformer on when mains voltage is at a peak (or just after) to eliminate inrush. The following goes into why:

https://iosrjournals.org/iosr-jeee/Papers/Conf.17017/Volume-2/8.%2059-63.pdf?id=7590

[Kleinstein]

Part of the problem, why the toroidial transformers are effected so much is that they have very little effective air gap and a high remanance magnetization. So quite often the magentic field does not start at zero, but can be up to about half the saturation value. The remanent field helps when starting with one polarity and makes the other polarity worse. With a remanent field at 50% the start up at the voltage peak would be about as good or bad as starting at the better polarity with zero crossing. Starting at the voltage peak is good when the core starts demagnetized (e.g. about OK with an EI core).

[Benta]

So quite often the magentic field does not start at zero, but can be up to about half the saturation value.

Source?
Must be some Chinese pig-iron toroid.

[dannybeckett]

Fascinating, I wasn't aware of remanent magnetization! Thanks for the knowledge. I wonder if it's easy to measure...
 
https://www.belfuse.com/resources/otherdocument/signaltransformer/other-ST-signal-transformer-transformer-basics.pdf

Some useful transformer basics in this doc ^

[Benta]

I wasn't aware of remanent magnetization! Thanks for the knowledge. I wonder if it's easy to measure...

Not easy. And it's not really significant if you buy quality transformers. Kleinstein's "50%" is wildly exaggerated. It might be 10% if you're unlucky.
Unless, as said, the core is Chinese pig-iron.

Nice document you've found.

[jonpaul]

Hi very odd, we are using  up to 5 KVA isolation both EI and toroïdal never seen this.
Signal Transformers EI and Germany toroïdal

 inrush current may be from cores near saturated, check the BH curve, ask manufacturer exact specs, peak B, core area, materials.

Your putting a band-aid on a cancer....fixing a much deeper problem.

If it is junk from China, suggesting to  get a high quality EU or USA product.

Jon

[themadhippy]

, we are using  up to 5 KVA isolation both EI and toroïdal never seen this

But i bet they aint protected by a fast acting protective  device,if the 20A device protecting the ring final is a type B it  only takes 0.1 seconds at 100A for  it to go pop.

RCCD

Hope you mean RCBO as an rccd is just the old name for an rcd as in residual current circuit device and has no overload protection,if it is an rcd/rccd thats tripping your looking up the wrong tree for a solution

[Kleinstein]

Magnetic remanence is a problem with high quality toroidial transformers too. The good materials have quite some remanence too, though there seem to be also steel grades optimized for less remanence (likely with downsides in other parameters). It is not so easy to find a good reference - with the graphs often shown it is not really clear if this is real or schematic only.
One example ref I have found is this, where it is about measuring the remanence. The get a range up to 30 / 50% of the actual used magnetization some some 25-40% of the saturation magnetization as there highest measured point.

https://pdfs.semanticscholar.org/e243/dcfd93b3cf368457f06043844cf65367b225.pdf

How high the remanence magentization can also be effected by the ciruit the transformer is used in (load, capacitance) an the moment when the voltage is turned off. So the remanence can vary a bit random.

EI cores usually have enough air gap to reduce the remanence quite a bit. In addition EI cores often use non grain oriented materials and a relatively low magntization. So they have more headroom to saturation.
Torodidial transformers on the other side often use grain oriented steel and this allows to get closer to saturation in normal operation. This is one may point, why the torroid form can get higher power density. As a downside it makes the inrush current from saturation more of a problem. It would be the ones with cheap non oriented material that are less effected by the inrush current problem.

So getting a problem with inrush current is not a sign of poor quality, but more an indication of the use of the better grain oriented matirial and a relatively high magnetization.

[Benta]

Hi very odd, we are using  up to 5 KVA isolation both EI and toroïdal never seen this.

The main problem today is, that transformers are "starved" on material, meaning iron and copper. Everyone wants "Aliexpress" prices and get corresponding products.
The cores operate within a hair's breath of saturation, and the copper operates at too high temperatures. That makes them smaller and cheaper for the same quoted "power".

The soft-starter is one way of getting around this.

In a couple of projects, I've gone a different way: Using 400 V primary transformers for 230 V input.
You'll need to specify the secondary voltages differently, of course, and you'll need to downgrade your power expectations. A 1 kVA would then be specified as 800 VA instead. This due to the thinner winding wires (but lower core losses).
This takes the transformer away from being very close to saturation. I've not had any of those 400 V types blow the fuses yet.

[dannybeckett]

Yes I meant RCBO, I'll change the text in my original post. Also attached an image of it for clarity. The fast blow fuses get annihilated along with the RCBO tripping.

I've started the phase detect firmware for the Arduino tonight, should be able to get something going in the next couple days. Once it's working, I'll do an experiment measuring inrush current vs switch phase angle on this big torroid. Very interested to see the results.

[dannybeckett]

One example ref I have found is this, where it is about measuring the remanence. The get a range up to 30 / 50% of the actual used magnetization some some 25-40% of the saturation magnetization as there highest measured point.

https://pdfs.semanticscholar.org/e243/dcfd93b3cf368457f06043844cf65367b225.pdf

Really interesting paper. It may end up being the case that the simple inrush limiting resistor / relay timer circuit is the way to go as we don't have to worry about any such residual flux.

The engineer in me wants to see if an arduino is capable of making those small signal LCR measurements to determine magnetising inductance and hence remenance. Knowing that will let us energise at the correct phase angle.

Alternatively, there might be an easy way to degauss the transformer before energising.

[dannybeckett]

Hi guys, made some interesting progress. I bizzarely found a complete soft-start circuit harvested from a microwave when rummaging through my old salvaged parts box! What are the chances I noticed 3 problems with the circuit though:

• The current limit resistor value was very low (15 R), worried this would still cause trips
• If you powered it on and off quickly, the relay remained closed (no inrush limiting)
• If the limiting resistor somehow blew or failed, the shorting relay would still close (no inrush limiting)

I bought a 40W halogen bulb (~90 R) and replaced the load, simulated the circuitry needed to mitigate the points above and then built an add-on board to the original soft start circuit I found. Tested and it works perfectly! Transformer gives me no inrush trouble at all now. Relay shuts down extremely quickly after loss of AC and won't engage if the halogen bulb has for some reason blown.

In parallel, I tried connecting mains to my arduino but it tripped my RCBO as live current found a path to earth via the cable shield. I've got a USB isolator on the way and so will resume exploring the phase-accurate switching method when that arrives.

But for now, see attached for the Tina-TI simulation, pics and graphs of the fancy inrush limiter as it stands today!

[Kleinstein]

15 Ohms is a reasonable size for the resistor to limit the inrush current. The current may still exceed the 16 A current rating for a short time, but that is Ok with the fuses. In addition to the resistor the transformer also has some impedance.  To trip the fuse need significant higher current (like > 50 A) and the inrush current may well reach that level when the core is saturated.

If the relay is power from behind the current limiting resistor, than with a blown resistor nothing happens, as the relay would never close.

A very short power down phase is a possible problem, but the relays are usually reasonable fast in turning off. It is responsibility of the user to avoid such short turn off phases. This is a larger problem with the NTC type solution that is sometimes used with lower power. Here the cool down may need 10s of seconds.

The USB isolators may not all be safe for mains work. To be on the safe side, I would test the circuit first with a lower AC voltage from behind a transformer.

[BravoV]

Alternative circuit, soft starter for inductive load which will turn on at the peak voltage, the snubbers (A or B) are optional.

[dannybeckett]

The current may still exceed the 16 A current rating for a short time, but that is Ok with the fuses.

I'm using fast blow fuses and my ring main is on a 20A breaker

If the relay is power from behind the current limiting resistor, than with a blown resistor nothing happens, as the relay would never close.

If there's a sizeable load attached to the xformer and the resistor isn't blown, this arrangement prevents the relay from closing like it should due to vdrop developed

A very short power down phase is a possible problem, but the relays are usually reasonable fast in turning off. It is responsibility of the user to avoid such short turn off phases.

The user should not have to count to ten before they re-energise a device to prevent blowing fuses. My home experiences short duration blackouts (cost of rural living)

The USB isolators may not all be safe for mains work. To be on the safe side, I would test the circuit first with a lower AC voltage from behind a transformer.

Good idea. The mains is being fed to my arduino via a 1M resistor

Alternative circuit, soft starter for inductive load which will turn on at the peak voltage, the snubbers (A or B) are optional.

Very nice, I like the simplicity. Ty for sharing

[Kleinstein]

Even a fast low fuse is OK with quite some current for a few short peaks. The peaks from saturating the core are usually short, like 1-5 ms once per mains period. Here the resistor is quite effective to limit the current. So with 120 V mains even 5 Ohms would be enough to limit the current enough. With a significant load on the secondary the 15 Ohms from the microwave may be too high.

With a transformer there is usually no need for a fast blow fuse.