Testing the current limit of a transformer

[Govmnt_Lacky]

What is the easiest way to test the output current limit of an AC to DC transformer?

[Perrow]

http://dl.dropbox.com/u/20165528/Power_Transformer-98.exe

Measure some values from your transformer and input them and get estimates of current limit and such. Tried to find a reference for it online but couldn't. Did find the download in my downloaded files folder and uploaded it to dropbox, couldn't find a licence that said I couldn't 

[Govmnt_Lacky]

Thanks Pele.

Cant download it now but I will look at it later this evening 

[R.G.]

The current limit of any transformer is ultimately the temperature rise compared to the temperature rating of its internal insulation. A transformer will work fine at temperatures hot enough to fry eggs on the case. Iron loses its magnetic abilities at something like 700F and copper works as a conductor up til it starts to melt at over twice that. It's the insulation that fails.

For non-permanent failures, the resistance of the wires cause the output to sag. You can often measure the wire resistance with an ohmmeter and make a few calculations.

My normal objections to special purpose calculators that one does not understand the math behind apply. These things keep you from ever learning what's really going on.

All that's quite apart from the current limit of any DC output. Transformers produce AC outputs. This has to be rectified and filtered and sometimes regulated to get DC. The DC side may either current limit or die before the transformer.

The only good way to test this without knowing what's inside the box is to go ahead and load it.  The only good way to do this *non-destructively* if you don't know the design is to pulse-load it. You set up a circuit to load it to varying currents and turn on the currents while watching the output voltage. When the voltage sags during the loading, you make a judgement about whether that's destructive enough. Short pulse loads are over and done with before heat can destroy the circuit.

This can sometimes be done by hand if you are good at turning on a load and reading a meter and then getting off the load before smoke comes out. Automated testing of semiconductors for overcurrent is done this way, but with microsecond long pulses.

[Perrow]

I totally understand that it's better to understand the calculations, but for ballpark figures (i.e. power a Fuzz Face or a bunch of tube heaters) I can live with the uncertainty of plugging some numbers into a calculator.

ps. Thanks (again) R.G. for all the knowledge you keep pouring into us.

[merlinb]

By far the easiest way is to compare the size of the transformer with one of the same size and known VA rating. Most off-the-shelf transformers come in the same VA ranges; 6VA, 12VA, 24VA, etc.

[PRR]

> compare the size of the transformer with

+1

Pulse testing is too much work. For "small" (under 100VA) iron, especially in rectifier-cap DC service, sag is usually the big problem.

Transformer sag is a direct linear function of current (it is a resistance).

In small iron the rating is often 10% sag. In very-small iron, often 20%.

Eyeball the iron to estimate a VA (just as Merlin sez) then aim low. You think it may be 12V 12VA (1A)? Then figure a 0.1A load. 12V/0.1mA is 120 ohms. Power is 1.2 Watts. A couple 270 ohm 1/2W in parallel is a suitable short-time load.

Say that you measure 14.0V no-load and 13.5V with 135 ohm (0.1A) load. 13.5.14.0= 0.964, or 96.4%. You have 3.6% sag. You can probably go to 20% sag. So you can suck 20%/3.6% or 5.5 times more current. 5.5*0.1 is 0.55A probably-safe load.

This is still a guess. The "20%" is just typical. If you MUST know the exact rating.... you can't. You *can* put a heavy test load on it for an hour or two, and see how hot it gets.

Anyway, temperature and life are related. You can run much hotter than nominal, but life goes down. So was it rated for 20,000 hour life or 200 hour life? Do you need 20,000 hours? I've had rigs with 100,000 hours in service, and others that never did 10 hours.

What really matters is the inside "hot spot" temp, and you can get a little closer by comparing winding resistance cool and hot then consulting copper tempco. But you still don't know if it is wound with 200 deg C mica-glass insulation or 35 deg C flour-and-lint junk insulation. Aside from knowing all the data, the method is to get a large batch of identical transformers, work them in very heavy overload, note failure times, and use Arrhenius to extrapolate a lower temperature which gives useful life.

[Perrow]

I made a wiki page of this thread, hope nobody minds.

http://rumbust.net/Transformer+Estimations

[R.G.]

The overall size of a transformer is a good rough and ready estimate of its power capabilities. It's pretty good as long as the assumptions of (1) the same lowest frequency and (2) the same insulation system are held up.

With AC power line stuff, the lowest frequency is a slam dunk. As I mentioned and PRR reiterated, the ultimate issue is when the insulation gives up, and you can't really tell that without heating it up or guessing.

I agree that pulse testing is too much work to do right. But he did say "test the output current limit". 

And I suspect that the current limit of an AC to DC transformer, which is what he asked about, may be more limited by the DC parts than the AC parts. It is for all the commercial AC to DC transformers I've ever delved into. Simply the fact that it's being rectified and filtered to DC will up the RMS current in the transformer to 1.6-1.8 times the DC current out, so the DC-ness of the issue is not inconsequential.

Manual pulse loading (touch the resistor, read the meter, pull the resistor off) is about the only way to semi-nondestructively test a completely unknown DC circuit without hacking it open to find out what's inside.

But there are easier ways. In almost all cases, the device will have its current capability printed on it. It's been quite a while since I've seen one that doesn't.

Estimating transformer death is hard. Reading the label is easier. 

[Paul Marossy]

In almost all cases, the device will have its current capability printed on it. It's been quite a while since I've seen one that doesn't.

If it's a power transformer and it's new, usually it will be on there somewhere. It can be quite cryptic for NOS power transformers that you might get off ebay though. Been there, done that... 

[tubelectron]

Well...

I worked with transformer companies, and they usually consider for our kind of power transformer that the load limit is reached when the secondary voltage drop is 4% of the nominal design value of the transformer, with an ambient temperature of 40°C. Note : it's in France, not USA, but I could measure that this rule was usually correct...

A+!

[R.G.]

I worked with transformer companies, and they usually consider for our kind of power transformer that the load limit is reached when the secondary voltage drop is 4% of the nominal design value of the transformer, with an ambient temperature of 40°C. Note : it's in France, not USA, but I could measure that this rule was usually correct...

The larger the power rating of the transformer, the less it usually sags from no load to full load. Many very small transformers in the 1VA to 10VA range have sags of 10% or over, sometimes to the 20% extreme PRR mentioned.

The thing that drives this is the need for enough turns on the core to make the primary inductance big enough to keep the magnetizing current low. (This is another way of saying that you have to have enough turns to keep the flux density in the core from the primary voltage down to values that don't start saturating the core.)

The smaller the iron core area gets, the more turns are needed to make the same primary inductance. And on a smaller core, there is a smaller area to wind these primary turns, so the wire diameter gets very small indeed. So the wire resistance goes up. The added resistance means more sag.

[Govmnt_Lacky]

To shed a bit of light...

I picked up a couple NOS transformers used in the older MXR flangers and delays. I want to use them in a different design that I am currently collaborating on and would like to know if it would handle the current draw requirement of this circuit.

I know this circuit is eating at least 50mA because my testing with the 1054 pump failed and was sagging.

I suppose there is nothing to it but to do it! Fingers crossed! 

[Ronan]

Yep, do some testing, take note of the sag figures mentioned in previous posts, sag is the difference between the unloaded voltage and the loaded voltage. You can use resistor/s on the AC output of the transformer and use Ohm's law (which is usually only true for DC circuits, but works for AC if the load is purely resistive). Be careful of the power going through the resistor/s so the smoke does not escape.

I keep clear of charge pumps and SMPS's unless they are designed by some clever dude and known to work really well. In contrast, a mains transformer and a linear regulator is more bulky and can generate a bit of heat, but works well and is easy and cheap to construct.

[Kesh]

if these are unregulated PSUs then measuring ripple might be one indirect way of measuring the designer's intentions. ripple increases linearly with current. so when ripple become unacceptable, that's your current limit. figuring out what the designer thought was unacceptable may be an issue.

[duck_arse]

surely the included thermal fuse will "protect" the wire's insulation from smoke, and make heat-test measurements tricky/destructive. are these fuses now standard world-wide, or just in my local parts supplier's range of txs?

[Govmnt_Lacky]

Did the testing last night! 

Seems to power the circuit just fine and very quietly (at least, as quiet as the original  )

Collaborating on a layout and BOM for the forum. More to come hopefully soon! 

[R.G.]

surely the included thermal fuse will "protect" the wire's insulation from smoke, and make heat-test measurements tricky/destructive. are these fuses now standard world-wide, or just in my local parts supplier's range of txs?

Thermal fuses are a solution, not a standard.

The safety standard specifies that the transformer will not endanger the external user if any one failure occurs. The one usually used is a 'soft' short on the output which pulls the output down and causes an overcurrent in the primary and/or secondary. The way this is usually done is to open the primary side of the transformer.

In larger transformers, the primary AC fuse at the AC power inlet can be sized to open and provide this protection. In very small transformers, the wire resistance makes the difference between normal load and shorted output smaller, and it's hard to design a fuse current so that it will not trip on normal maximum load, but will protect against a short. So a thermal fuse inside the transformer directly detects the internal heating that is the result of an overload.

It's not required by the standards to use a thermal cutout, but it's a cheap, workable solution for small ones. So it gets used a lot. They're not required, just very common.