Step down transformer

asif

I have made a step down transformer of 220v input and 4 taps of 20v. The sectional core area is 60 sq cm .
It starts heating up after afew minutes even without any loading.
Spec are
158 turns of 17 guage awg
15 turns of each tap of 10 guage awg
Input current is 3amp and output current is 15amp.
Can any one help me to resolve the issue

Heater.

Do you really have a core that is 8cm by 8cm? That is a huge chunk of metal.

I would not dare advise on winding a mains transformer. Except to say "Don't do that". Sounds like a reasonable way to commit suicide if you have no idea what you are doing. Or worse still kill someone else.

But out of curiosity:

158 and 15 turns does not sounds anything like enough. Surely every mains transformer I have ever seen has more.

What is that core made of? Is it a laminated core?

Do you have a picture of this beast?

Have your read this: http://ludens.cl/Electron/trafos/trafos.html

davejames

@asif,

Can't answer your question directly, only provide a link with some information that may help:
http://www.electronics-tutorials.ws/transformer/transformer-basics.html

LoopyByteloose

It sounds like a traditional iron core transformer.

3amps @ 220VAC input is roughly 660 watts. Of course it is going to get warm, even when unloaded. The iron core created eddy currents that waste some of the power.

If you are getting 20VAC on each tap and the transformer doesn't burn up if left on for 48 hours, the ratios of the windings are about right.

Be sure to put a 3amp fuse on the 220VAC mains while testing.

Ideally you would put 660 watts in and get 660 watts out.
20Volts x 15 amps = 300 watts... less that 50% efficiency.

Could it be that each 20Volt tap will only provide 3amps, and 360 watts is wasted as heat? Yes, that is possible. If you are going to use this transformer, it should be in a metal enclosure with vents for hot air. 360 watts is a lot of heat.

Do you require AC or DC power at 20V? You could get something that runs cooler and better in a switching power supply to provide you 15amps near 18VDC... after you convert the 20VAC to DC with bridge rectifiers, you are going to have only 18VDC or so. And you still need some big filter capacitors that are expensive and difficult to buy.

Heater.

Potentially dangerous advice from Loopy there. I seriously recommend ignoring it.

LoopyByteloose

Yep, if this is being built from scratch, forget building it. But if it is just an old junker, it worked once and likely can work again.... and waste a lot of electricity.

It could be a nice transformer for building a 40watt BJT stereo power amp, but it will also warm your home in the cooler months.

It certainly demonstrates all the reasons that high watt iron core transformers are obsolete. They run hot, they are heavy. When we do need the really big ones for power distribution, they are liquid cooled.

Heater.

Those big power distribution transformers are in the high 90 something percent efficient. They have to be or they would melt down immediately on power up. It helps that they are working at high voltages to reduce resistive losses.

From my memories of building big audio power amps certainly the transformers were big and heavy but most of the heat was being generated in the output transistors. The bulk of the transformer is all to do with the magnetic coupling and avoiding saturation. A black art I know little about.

RickB

I would estimate that your core is going into saturation. 15 turn secondaries for 20 V @ 15A implies significant primary flux. Ok for short duty cycle usage like a microwave oven or spot welder. No load primary current should be a small percentage of full load primary current. As long as the laminations are still properly insulated from each other, a properly designed home brew transformer should only be slightly warm to the touch after hours at no load.

However, If you have bolted mounting brackets to it with uninsulated bolts/screws and no insulating washers under the screw/bolt heads and nuts, you will have created shorted turns in your otherwise good transformer.

When I test unknown transformers with a variac, I watch the rate of rise of input current with no load. Its very easy to see the current rise dramatically at the beginning of core saturation. Try that and note the voltage.

asif

Sory gentlemen the sectional area (solid) is 40.23
These are the detail specifications.
The core i have used is;
EI 250 sections laminates 2.5 '' central leg widh
Cross sectional area ,solid 40.323 sqcm
Volume ,solid 1516.5 cu cm
Length of megnetic path 38.1 cm
Calculation have been made by core geometry approach.
One more thing i would like to ask if the voltage input are near 220v it heats up quickly but if input voltages are between 180-210, it takes more than hour to start heating up.
And how saturation factor is controlled.

LoopyByteloose

So you are saying this is actually a design from scratch power transformer.......

My guess is that core saturation would be resolved by a bigger core... but that is just a guess. There is a tendency to be optimistic and build too small for a first try. That is how dangerous devices end up on the market.

These days, a toroid core would be used, not iron laminates. And I suspect that nobody would bother with something as high as 660watts.

RickB may have access good design information, I don't.

But I am with Heater... what you are doing requires a lot of expertise and experience in order to be done safely. It would be best to source a transformer of modern design from a reliable and reputable manufacturer.

pjv

The rule of thumb for primary windings is one volt per turn. But I find that to be a little high. Half a volt per turn has worked for me.

The symptoms you just described is a classic case of encroaching on saturation. More turns will take you away from saturation. Other than practiacl physical size and copper cost consideration, more turns is better.

The ratio between primary and secondary of course gives you the secondary voltage.

Cheers,

Peter (pjv)

LoopyByteloose

Well, the OP has 157 turns of 17 gage wire. So that is in the ball park of 110 to 220 turns. Could it be that the gage of the wire is too small and generating heat?

Laminated iron cores are also a source of heat, and much depends on the quality of their construction. There is a lot of variation and a traditional factory would have a lot that would not past a final inspection. Some actually hum and create noise.

The newer toroid cores eliminate those problems, but are a bit more work to wind.

pjv

Loopy,

158 turns for a 220 volt input is asking for exactly the saturation problem being exhibited. More turns will solve that. Again, assuming that no phantom shorts such as uninsulated bolts or shorting laminations exist.

What is required is that there is sufficient inductance in the (unloaded) primary to prevent non linear current flow.

Cheers,

Peter (pjv)

Heater.

I may be missing a point here. Is this really about "saturation"? Bear with me:

My crude understanding is that a core material, like these laminations, increases the inductance of the coil wrapped around it (Can I ignore any coupling to the secondary for now?)

Further, that this nice increase in inductance only works until a certain point, when the currents get big enough and the fields strong enough we run out of, shall we say "magnetic head room". The inductance drops, currents increase, resistive losses increase, everything heats up, efficiency drops. I have seen this on scope traces from my dumb Smile switched mode PSU experiments.

Well, in the no load situation, ideally there is no current going anywhere. Practically we would like it to be very small. That means the turns and core material have to be such as to provide enough inductance on the primary to ensure a small current at the input frequency.

Bottom line is, if this thing is saturating in the no load, low current, situation it is never going to work under load.

The core described sounds huge. So the only thing left is more turns on the primary. Of course do check the core is assembled correctly.

LoopyByteloose

Ummm. pjv is usually spot on, one of the best eggheads on Parallax Forums. It seems he does quite well running his own electronics manufacturing.

I suspect his point of view and experience comes from having to either buy or make quite a few power transformers. He generally won't settle for second best.

pjv

Ignore the secondary for now. Voltage applied to the primary will want to drive current into it. If there is insufficient primary inductance by means of insufficient turns, the current will become large. Think of zero turns equals zero reactance equals a short circuit. As the current becomes large, the core will saturate, using up all the "headroom", and stopping any increase in reactance as the current rises.

So, the primary, with sufficient turns generates flux linkages into the core, and can be thought of as the "motor". The secondary(s) can be thought of as the " generator", and will counteract the primary's linkages, bringing the net toward zero.

Cheers,

Peter (pjv)

kwinn

Laminated core transformers have typical volts per turn (or turns per volt) ratio that can be determined experimentally by winding a few turns of wire on a known good 220V transformer and measuring the output voltage it produces. This was how I modified 6.3V filament transformers to add a winding for +- 15V outputs for op amp power supplies way back when. I no longer recall if it was 2V per turn or 2 turns per volt, but it was in that neighborhood. Once the ratio has been determined the current requirement can be used to determine the wire gauge.

Heater.

pjv,

Thanks. I think you are confirming my view of things. With no load we want current to be small. To get small current we need sufficient inductance. That means more turns on the primary. We should be far away from core saturation at this point else we will not be able to handle any load.

Hence my question about this being a saturation problem. In your extreme case of no turns, there is no field and no saturation. Surely the core is not really in the picture any more. There is however a huge current and resulting resistive heating losses.

pjv

Heater,

The "no turns" scenario is of course only a conceptual extreme, and not realizable.

The driven winding (primary) generates flux linkages into the core, and the reactance is, among other things, a function of the magnetization (think ampere-turns) present in that core. As the magnetization approaches saturation, the increase in magnetization due to increasing current becomes reduced, and the "back emf" effect of all the turns is likewise reduced. Thus there is less induced voltage to counter the driving voltage, and a nonlinear relationship ensues. This causes more current to flow, and that generates heat.

Hope my explanations are providing some clarity.

Cheers,

Peter (pjv)

Heater.

pjv,

Hope my explanations are providing some clarity.

No. But I'm prepared to believe you

As a practical matter, what test can we do to see if a core is saturating or not?

Peter Jakacki

I can't see where the mains frequency has been mentioned as surely this is important. The 240V mains here is 50Hz vs 60Hz and that makes a lot of difference to the transformer if it has only been designed to only operate at 60Hz. At 50Hz the primary impedance is less and this could saturate the core etc. You can run this 60Hz transformer at 50Hz but I think only about 83% of the rated voltage.

pjv

Heater,

The test you can do is to monitor the current in the unloaded primary while you slowly bring up the driving voltage with a variac. The current will rise very markedly as you encroach on the area of saturation.

Peter,

The mains frequency will not significantly change the secondary voltage as long as you stay away from the saturation zone. It is set strictly by the turns ratio and of course any IR drop due to current flowing in both windings.

Cheers,

Peter (pjv)

pjv

Peter,

I forgot to say that of course the mains frequency affects the impedance. I assumed the comparison would be with more or fewer turns at the same driving frequency.

Cheers,

Peter (pjv)

Peter Jakacki

pjv wrote: »

The mains frequency will not significantly change the secondary voltage as long as you stay away from the saturation zone. It is set strictly by the turns ratio and of course any IR drop due to current flowing in both windings.

I'm not talking about the secondary voltage, simply about the no-load impedance on the primary and saturation. As you mentioned and putting aside primary/secondary ratios, there does not appear to be enough turns on the primary. We know for instance that if we put 240V DC on the primary that the impedance will be the same as the resistance, extremely low (short). The same for very low frequency AC too. Now as we ramp up the frequency the impedance will rise for sure but the important thing is that even with no load current the core will go into saturation if the frequency is too low.

I'm no expert on transformers and it sounds like you know your stuff but I have designed switch-mode supplies and it seems to me that the OP's transformer may be just on the saturation threshold, especially if he had used 60Hz design rules for a 50Hz transformer.

pjv

Peter,

I think we are saying the same thing. In my first post I suggested insufficient turns on the primary as I was suspecting an encroachment on the saturation region.

What has not been said is that, on more detailed analysis of the (any linear) transformer, is the presence of self inductance in all windings. That is the inductance that only associates with the specific winding, and does not magnetically couple to the other windings. The (typically small) impedance generated by this effect will somewhat affect the voltage ratio.

It is however a minor component and not in the realm of the OPs question.

Cheers,

Peter (pjv)

asif

Design considerations are based upon 50Hz rules and input is of course 220v 50Hz. Laminates are Silicon M6X with operating flux density of 1.6tesla as per provided information by the manufacturer.
Turns of primary are calculated keeping in the view that max fulx density is 1.6T. A point difference in flux value will considerably increse or decrease the number of turns in primary. Though the flux value is provided by the manufacturer but i think there is a slight chance of variation in the value of flux persists or it may be a manufacturing fault.
I agree ... the number of turns should be increased due to strange core behavior and saturation.
But how many turns?
Would it be a hit and trial game? The greater number of turns will affect the current on load and impedence.

Heater.

pjv,

The test you can do is to monitor the current in the unloaded primary while you slowly bring up the driving voltage with a variac. The current will rise very markedly as you encroach on the area of saturation.

Thanks for the tip. I think this calls for some experimentation around here.

Next question:

Let's assume my transformer does not go into saturation under no load at the normal input voltage, as confirmed by the test you describe. Presumably it will do so when I start loading the secondary. Can I watch for a similar marked rise in primary current as I increase the secondary load current?

I'm guessing that when saturation occurs things go very non-linear and there is noticeable distortion of the sine wave input as seen on the output. Is this something we can see with a scope? Perhaps it's even possible to hear saturation occurring, due to the harmonics generated, given a suitable speaker connection?

Calls for more experiments...

pjv

Asif,

".......The greater number of turns will affect the current on load and impedence."....

Well yes, it will increase the primary's inductance, and hence impedance, but the secondary's current causes a magnetic flux of the opposite polarity that works against the primary's flux. So the total flux stays about the same.

For a particular secondary output one needs only keep the winding ratios set, and allow for some IR losses.

A transformer overloads mainly due to the current in its windings, and not due to excessive flux. Although there are always some smaller side effects due to non-ideal parameters.

An interesting experiment is to take two *identical * transformers, and hook the primaries in series and their secondaries in parallel. This will simulate doubling the turns.

Then apply the standard design voltage to the series primary pair. The unloaded current measured will be markedly lower than with a single transformer.

User Name

@asif:

I think you've hit upon your problem. 1.6 T is a little high. When designing transformers with tape-wound oriented-grain silicon steel cores, I've used 13,000 Gauss (or 1.3 T) as an upper limit. 11,000 is even better.

pjv

User Name,

Could I ask where you get tapewound steel cores?

Cheers,

Peter (pjv)

asif

From islamabad market.