Power transformers test method ?

[kagliostro]

Often I have to do with old alone PT

sometime there are indicated Voltage and Current ratings

sometime nothing is known

assuming to know exact voltages of the windings the problem every one has is to know the currents

some time ago there was a thread about to know spec of transformers in rapport with the weight and wire section (if is possible to see the thickness of the windings)

the question is:

Is possible to build an adequate variable resistor load to be connected to the winding we want to test

then connect this load and a precise voltmeter and an ammeter to the winding under test

proceeding with increase the load till the nominal voltage of the winding decreases

at that time look to the current in order to establish the current rating of the winding under test

do you think this is feasible and can give reliable result ?

Kagliostro

[eleventeen]

You can do all of those things you suggested, but I am not so sure you could get a "maximum rating" with all that much precision.

You are proposing what, fire up the transformer with a variac and a voltmeter and an ammeter and a dummy load? Fine so far.

Let's say you are going to put a 2 amp dummy load on your 5 VAC winding to simulate a 5Y3 or 5AR4 rectifier = 2.5 ohms @ 25+ watts. (15.6 watts actual)

Let's say you are going to put a 6 * 300 ma = 1.8 amp (we'll use 2 amp) dummy load on your 6.3 winding for 6 qty 12A_7 = 3.15 ohms @ 20 watts (12.6 watts actual)

And let's figure on 120 ma of B+ for 2 qty 6L6 and the preamp tubes. You'd dial up the variac until the heater volts are right, then read the HV winding. Let's say it's 380-0-380. Figure on a GZ34 rectifier and B+ should be about 465 per a Pro Reverb schematic. Need a 3875 ohm 100 watt resistor (55.8 watts actual)

Everything is fine so far, but here's where you start to run into limitations.

How hot do you want to allow this to get?

Could you simulate the idea that in use, the tranny is semi-enclosed in a cabinet with only fair ventilation by putting it into a cardboard box?

I'm not entirely sure you could detect a drop in voltage (meaning...the tranny can't supply any higher current) using (only) a +15% overvoltage variac.

You'd need a fair assortment of power resistors  that you could build up into dummy loads of various sizes.

And you'd have to take into consideration that the thing might be 20 years old....or 30 years old....

And in the end, you'd have to de-rate your findings by at least 10%, probably more like 20%, to be safe.

So you could get a pretty good idea, but "precision"?...Kind of doubt it.

[sluckey]

A PT has internal resistance in the windings. This will cause the winding voltage to smoothly drop as you smoothly increase the load current. The winding voltage will not just sit at some useful level and then suddenly drop when you hit some magical current limit (not talking about saturable reactors). I think after you collect a page full of data, you'll still have to make an educated guess based on weight, wire size, and how hot the PT gets. Then you'll still have to smoke test it!   

[eleventeen]

Sluckey, I think you stated what I did, more succinctly.

The issue of course is that if you want to "smoothly" increase current you're gonna have to have a pretty fat rheostat (100 watt) or some ability to switch in more and more load on the B+ (really, what you're interested in after you've handled the heater demand) which is a little more tedious than cranking a variac.

It's doable. But by the time you derate it for heat and for age...the word "precision" doesn't exactly apply.

I guess this is why precision current sinks are made and sold.

[sluckey]

Sluckey, I think you stated what I did, more succinctly.

Huh? I've never been accused of that before!     

[Tone Junkie]

There you go using big words I dont understand. 

[PRR]

Except in odd cases, the "rating" is to give an internal temperature which just barely gives the required service life.

So you could power and load-up a hundred PTs and leave them running for many thousand hours. If they run hot and most fail "early", get new PTs and try a lower load. If they are not hot and most of them are fine after many thousand hours, get new PTs and run a heavier load.

You can measure the internal temperature from the copper resistance. Not easy. And you do not know exactly what kind of insulation was used. Plain shellac won't take heat and voltage for long. Some modern plastics can run very hot and live a long time.

Go for "sag". A small, 30W, PT may sag 20%. Over 100W, 5% sag may be the safe limit. So if you have a BIG hunk of iron, 300V no load, load it until it drops to 285V. Leave it in a safe place and let it cook 20 minutes, an hour, all day. Does it seem "too hot"?

[kagliostro]

Many thanks to all

Your consideration and councils clearly make to understand that such method will not be a reliable method

Go for "sag". A small, 30W, PT may sag 20%. Over 100W, 5% sag may be the safe limit. So if you have a BIG hunk of iron, 300V no load, load it until it drops to 285V. Leave it in a safe place and let it cook 20 minutes, an hour, all day. Does it seem "too hot"?

As always PRR has a good practical council

only one further question

sometime ago there was a thread were was explained how to use the weight of the PT and the wire thikness (for this I've a reference table that PRR posted some time ago) to have some ideas about PT ratings, unfortunately I lost that link, can you give me it ?

Thanks

Kagliostro