Add "soft start" to Isolation transformer

[staze]

All,

I just picked up a 120V 15A isolation transformer for $1. Yes, $1. It's huge, and weighs a ton. Opening it up, there's no soft start, and I'm worried the power on inrush current will trip a breaker. It looks like I should be able to remove the "jumpers" on the terminal block (upper right of the overhead picture), and replace those with NTC Thermistors... but I'm unsure how to calculate the value needed. Or should I just throw in a couple high wattage resistors? Or am I crazy, and you really only have to worry about inrush with toroidal transformers?

Any info would be great.

[uncle_bob]

Hi

With no load on the secondary, a proper isolation transformer presents no load on the primary. There's no need for a "soft start" unless you do something silly. Even then it's no worse than plugging in whatever load you already have attached to it.

Bob

[staze]

Then are toroidal transformer based isolation transformers just that much worse? Because I know they generally have soft start (through relays or NTC thermistors) (the one I have certainly does). I thought there was a inrush current due to the magnetic "mass" of the core (sorry, can't think of the word I'm looking for there)...

Thanks!

[uncle_bob]

Then are toroidal transformer based isolation transformers just that much worse? Because I know they generally have soft start (through relays or NTC thermistors) (the one I have certainly does). I thought there was a inrush current due to the magnetic "mass" of the core (sorry, can't think of the word I'm looking for there)...

Thanks!

Hi

I suspect what you are looking at is a fuse. Transformers, regardless of the design do not have an inrush issue. Capacitors have inrush issues.... Some loads with capacitors in them have inrush issues.

Bob

[Ian.M]

So why do the experts contradict you?
http://sound.westhost.com/articles/inrush.htm#s4
http://www.ametherm.com/inrush-current/transformer-inrush-current.html

The inrush current of a E&I core transformer is typically much less than a similarly rated toroidal transformer, but the O.P. has one big-ass transformer so it may in fact be a problem.   One can get all fancy and try and measure the inrush current to determine if a soft start circuit will be needed, but why not simply see if it ruptures a T6.3A fuse if switched on and off repeatedly with no load?  If it does, its going to need a soft start circuit - either a NTC thermistor and a time delay relay to bypass it, or a series power resistor and a time delay relay, with another relay slaved to it so the output is disconnected until the resistor has been shorted out.

[uncle_bob]

So why do the experts contradict you?
http://sound.westhost.com/articles/inrush.htm#s4
http://www.ametherm.com/inrush-current/transformer-inrush-current.html

Hi

The first one seems to be *very* confused about motors vs transformers. Hint: they are not the same thing at all. The second one is talking about a transformer that has not been properly wound and they *really* would like you to buy their inrush gizmo.

Transformers have been used for over 100 years now. Go climb up a pole and take a look (carefully). Seen any inrush limiters? Nope. Tear open several thousand pieces of gear with transformer based supplies. See any inrush limiters? Nope. Take a look at the one you just bought. See an inrush limiter? Nope. Do they all work just fine? Yup. Have they been doing it for about a century or more ... yup.

Bob

[Ian.M]

*BIG* transformers on limited current capability supplies can have an inrush problem.  Big toroidal transformers are known for it.   Try fusing one with an ordinary F fuse, using the next value up from its rated current some day.   It wont last very long and you may even see the fuse wire glow red and sag momentarily at switch-on.

[IanB]

An explanation would be helpful.

The primary of a transformer with an open secondary looks exactly like a big inductor, meaning it will tend to resist any sudden changes in current. If you connect AC to it it will naturally present a high impedance and try to prevent a large current from flowing.

If there is to be an inrush current problem, something else must be at play. What might that be?

(In contrast, a transformer with a shorted secondary has a primary that also looks like a short circuit. For instance, smoothing capacitors on the [rectified] secondary side look exactly like a short circuit. So if you apply power to the primary of such a transformer you had better expect something interesting to happen.)

[Ian.M]

Residual magnetisation of the core is one issue.   If the supply is switched on at the beginning of a half cycle that will give the same flux direction, unless the transformer has a significant excess core area, its likely to saturate.

Our workshop Variac often tripped the breaker on the isolating transformer even if it was switched on with no load.  It certainly didn't have any capacitors.  The only reason you don't see problems more often is that most fuses and circuit breakers have a surge capability significantly in excess of their continuous rating.

[Zeranin]

Residual magnetisation of the core is one issue.   If the supply is switched on at the beginning of a half cycle that will give the same flux direction, unless the transformer has a significant excess core area, its likely to saturate.

Our workshop Variac often tripped the breaker on the isolating transformer even if it was switched on with no load.  It certainly didn't have any capacitors.  The only reason you don't see problems more often is that most fuses and circuit breakers have a surge capability significantly in excess of their continuous rating.

You are exactly correct, core remanence is one culprit, and the resultant inrush current can in some cases be very large indeed. Remanence is the magnetic property of steel, whereby it stays in a magnetized condition after removal of a magnetizing coil current.

How much of an inrush you get on any particular switch on is purely a matter of chance. In the worst case, the transformer is switched off at the peak of the magnetizing current, with the result that the core remains in a substantially magnetized condition. That is a matter of chance. Then, in the worst case, the supply is switched on at the beginning of a half cycle, of phase such that the magnetizing current magnetizes the core still further, in the direction in which it is already magnetized. If you are unlucky enough for both these events to be worst case, then there is a fair chance the core will saturate, after which the inrush current is limited only by the winding resistance. After a few cycles, the remanent magnetization disappears, and all is well.

Toroidally wound xformers are generally much worse that EI laminated types, because they use a strip-wound core that has zero air gap in the magnetic path, whereas conventional EI lamination types do have some magnetic air gap because the laminations are crudely butted together. The air gap helps to resist remanent magnetization in the first place, and also reduces saturation when inrush currents do occur, limiting the magnitude of the inrush.

I have a 2kVa Variac, toroidally wound as Variacs are, and every 4th or thereabouts time I turn it on, it trips it's breaker, as per the above, and as Ian has also experienced.

You may or may not have a problem with your 'huge' conventional lamination xformer. Sounds like it might be conservatively designed and thus less likely to saturate, so with luck you will be OK.

[edpalmer42]

I have a 250V, 1.6KVA variable autotransformer (i.e. Variac).  Many years ago, I wondered how much current it drew at no load.  I carefully hooked it up with my bench DMM in series to measure the current.  When I turned on the power, LIGHT came out of my DMM!      Note that there were no gaps for the light to come out, but it did anyway!  I quickly unplugged everything and pulled the top off the DMM to let the smoke out.  But there was no smoke.  No leads that were missing a component.  No damage at all.    Eventually I saw that the 2 amp glass fuse had a perfect mirror finish on the inside of the glass.  That's where the light came from.  The initial current surge vaporized the fuse wire and deposited it in a perfect, mirror smooth layer on the inside of the fuse.   

So yes, a transformer with no load can generate a scary current surge.  However, Staze, since you're in the States, your circuit breaker is probably a thermal breaker, rather than a magnetic one.  If so, I'd be very surprised if the surge was large enough or long enough to blow the breaker.

If you decide, or testing shows, that you do need some form of soft start, you won't be able to use an NTC because the NTC required to suppress the starting surge would be totally wrong if you load the transformer with a 15A load.  You'd have to use power resistors with a time-delay relay to short them out.  Even then, you could find that they would interact badly with some loads.

For now, I'd recommend that you just run it as is and see what happens.

Ed

[staze]

Ah, interesting. Actually, this whole thread has been.

The confusion, I suppose, is that my 5.28A toroid isolation transformer does have two NTC thermistors for "soft start". That's it. No relays, etc.

So given the huge additional current capacity, those NTCs would have to be pretty large I guess. So hence the power resistors and a relay?

Interestingly, the unit isn't fused. Probably for this reason. :p

I'll give it a go.

Thanks for all the input. I don't really NEED a 15A one... But who could turn down a dollar. :p

[johansen]

Our workshop Variac often tripped the breaker on the isolating transformer even if it was switched on with no load.  It certainly didn't have any capacitors.  The only reason you don't see problems more often is that most fuses and circuit breakers have a surge capability significantly in excess of their continuous rating.

my experience is this is only a problem with relatively new breakers. i think they are now designed to be sensitive to the magnetic fields from the 100 amp inrush current on the first cycle.

[edpalmer42]

Ah, interesting. Actually, this whole thread has been.

The confusion, I suppose, is that my 5.28A toroid isolation transformer does have two NTC thermistors for "soft start". That's it. No relays, etc.

I'm actually surprised that a general-purpose isolation transformer would use NTCs at all.
 

So given the huge additional current capacity, those NTCs would have to be pretty large I guess. So hence the power resistors and a relay?

But if the NTCs were big enough to handle a 15A (~1875W) load, a 10W load, for example, might be too small to heat them up,  or it would take so long that it would affect the load.  If the transformer was built-in to a piece of equipment, the NTCs could be sized appropriately, but not for a general-purpose, variable-load isolation transformer.

Interestingly, the unit isn't fused. Probably for this reason. :p

I wondered about that.  Even a crappy power bar has a 15A breaker.  Is the switch just a switch or does it have a breaker built in?  By the way, for 15A, the line cord had better be 14 AWG.

Ed

[TerraHertz]

I'm surprised at all the posters insisting there's no inrush current problem with big transformers. Don't they have any big transformers themselves?

Even my measly 240V 2.2A  to 240V isolation transformer sounds like it's going to blow up if I happen to switch it on in the 'wrong' instant of the 50Hz cycle. Takes several complete mains cycles for the "BRRRMMMmmmmm" to fade away.

I had thought it was due to the magnetic field taking a while to build up in correct phase to the mains voltage. Bit simplistic I know. The explanation involving remanance magnetization is much more satisfying.

[Pjotr]

What happens when you switch on a transformer? Or an iron cored inductor? At that moment the iron is not magnetised. Now when you switch at the zero crossing (zero voltage) you will get a DC component on the magnetisation current of the iron. This DC slowly fades out finally due to the transformer losses.  The net result is at switch on: Twice the peak induction as without DC. To cope with this you will need twice the amount of iron as is needed for normal steady operation. It is sometimes done for medical transformers but for normal transformers it will be a waste of material and money. In that case the core will go into saturation for the peak induction for a while. When that happens, no self induction any more, only the primary DC resistance to limit the current. With small transformers that is no problem, primary copper resistance will limit inrush current, but with larger ones the fuse will trip. It is the way inductors work.

[TheAmmoniacal]

I'm surprised at all the posters insisting there's no inrush current problem with big transformers. Don't they have any big transformers themselves?

I'm asking myself the exact same question, I've had high inrush current problems on all my large transformers, they would trip the circuit breaker for every cold start. The ferrite core isn't instantaneously magnetized!

I just throw an EPCOS NTC in series to fix the issue. Still not confident on which values to pick though, but they've all worked so far.

[Jeroen3]

A transformer will have at least 3x Inom inrush current. This will depend on the impedance of the source.
Inrushing transformers from an inverter or generator will cause low inrush, from mains this will cause high inrush. Very high if you turn it on at 90 or 270 degree.
Using a solid state relay can reduce it significantly.

I remember on this forum someone strapping simple water heaters in series to reduce the impedance of the source, then bypassing them when core magnetization it complete.
This generally takes only 10 periods.

[VK5RC]

I can vouch for the in-rush current issue, built a 28V 30A linear PSU (for some RF gear - so wanted to avoid SMPS). The in rush current did vary from time to time at switch on - in keeping with a phase type issue.
The only time I have worn safety glasses in addition to my corrective glasses was when I was testing that puppy, scared the daylights out of me when I had a tantalum in backwards!  O0  (had straight hair before I started)

[Zeranin]

What happens when you switch on a transformer? Or an iron cored inductor? At that moment the iron is not magnetised. Now when you switch at the zero crossing (zero voltage) you will get a DC component on the magnetisation current of the iron. This DC slowly fades out finally due to the transformer losses.  The net result is at switch on: Twice the peak induction as without DC. To cope with this you will need twice the amount of iron as is needed for normal steady operation. It is sometimes done for medical transformers but for normal transformers it will be a waste of material and money. In that case the core will go into saturation for the peak induction for a while. When that happens, no self induction any more, only the primary DC resistance to limit the current. With small transformers that is no problem, primary copper resistance will limit inrush current, but with larger ones the fuse will trip. It is the way inductors work.

Agree with all you say, but core remanence is another important player. Here is a paper analysing xformer inrush current in detail.

http://www.ee.ktu.lt/journal/2011/03/04__ISSN_1392-1215_Calculation%20and%20Analysis%20of%20Transformer%20Inrush%20Current%20Based%20on%20Parameters%20of%20Transformer%20and%20Operating%20Conditions.pdf

Take note of Figure7, showing that the initial remanent flux density has a very large effect on the inrush current. Even if switched on at zero remanent flux, there will be a large inrush current exactly as you describe, but if you get unlucky and combine this with a peak remanent flux in the wrong direction, then the brown matter really hits the fan.

[Pjotr]

The large inrush current is due to the iron core getting into saturation. If the core goes into saturation and how much depends on what phase the transformer is switched on. Switching on at peak voltage doesn't cause the inrush problem. Remanance magnetism add or substract to this but for normal silicon grain oriented transformer steel it is very small. In the end the inrush effect fades away due to the transformer losses. For that time you need to limit the current with an extra resistor if the primary resistance can't do.

[Zeranin]

The large inrush current is due to the iron core getting into saturation. If the core goes into saturation and how much depends on what phase the transformer is switched on. Switching on at peak voltage doesn't cause the inrush problem. Remanance magnetism add or substract to this but for normal silicon grain oriented transformer steel it is very small. In the end the inrush effect fades away due to the transformer losses. For that time you need to limit the current with an extra resistor if the primary resistance can't do.

So presumably you believe that the paper that I linked to is wrong, and the effect due to remanence is 'very small' ? Who are we to believe? Did you actually read the paper? Can you provide evidence for your claim that the effect of remanent magnetism is 'very small'. How small is 'very small' ?  What is your reference on this matter? I'm not saying you are wrong, but you will need to provide some evidence.

[Pjotr]

I didn't say the effect is small. I did say that remanence magnetism for transformer steel is usually small. Note that the area of the BH loop represents iron losses. You want to keep that as small as possible.

But finally it doesn't matter much. The iron will go into saturation now and then at switch on. For that situation, anyway you need to take provisions to limit the current in some way.

[timofonic]

I just to say I'm amazed by that big cat. Please put more photos of that great feline ?

[staze]

So starting the unit up resulted in a whole lot of nothing. As in, disappointment that it didn't trip a breaker, or make a bunch of noise, etc.

To answer the question, what I thought was just a switch is actually a breaker, a 3120-F321-P7T1-R01D, or on Digikey (http://www.digikey.com/product-detail/en/e-t-a/3120-F321-P7T1-W02D-10A/302-1219-ND/509821) a $40 part!

So, I won't worry about it tripping the breaker. Now I just have to find a place for this monstrosity. =D Might use it to hold up a shelf below my bench. =P