Hi,

I have bought some isolation transformers:


Howerver, I think them could be defective in some way. This is the output at the secondary:

A kind of crooked, isn't it?

With a rectification bridge I have this:

Definitely something is wrong.

I thought I could have a grounding problem with my scope, so I tried an old transformer from an old modem that I used long time ago:

Clean wave.

And through a bridge:

Flawless.

I tried a couple of them with identical behavior.
I wonder if my old transformer could have something to make the wave cleaner, or just my new transformers are defective.
What do you think?

 

Are your measurements done with or without a load? If with, how much current? Non-linear behavior could result from core saturation if the current is too high. Not sure what the effect would be be if the current is too low, i.e. no load.

I would try the measurements again with a load of about 25mA, half the rating.

Bob

 

Are your measurements done with or without a load?

Without load, in both cases.

I would try the measurements again with a load of about 25mA, half the rating.

Let's say 30mA:

Still a bit crooked.

However, through a bridge, signal looks more as it should:

Slightly leaning, though.

Well, there is something to do with load.
However, I am still not convinced of the quality of those transformers.

 

I think the transformers are not defective. You are running up against the physics of transformers.

Put a 1000μF 50V reservoir capacitor on the output with 25mA load and have a look again at the output.
Note that if you are planning to use the transformer with a bridge rectifier output you have to derate the current output to 50% of specified current.

 

can you share a picture of your test setup in use please

Can you also share a picture of the input waveform at the same time ,

 

Put a 1000μF 50V reservoir capacitor on the output with 25mA load and have a look again at the output.

In series, I understand. That makes about 3Ω at 50 Hz.
I don't have one of these on hand at this moment. I will try later.
Thanks for your advice.

 

Non-linear behavior could result from core saturation if the current is too high.

Core saturation occurs because of the input voltage being too high and/or the input frequency being too low.
High load current does not cause saturation.

 

In series, I understand. That makes about 3Ω at 50 Hz.
I don't have one of these on hand at this moment. I will try later.
Thanks for your advice.

The capacitor does not have to be exact value. 100-1000μF in parallel with the load.

 

You appear to be overdriving the transformer as the scope says the output of your new transformer is 26-30V while you old transformer is 16.6V (as would be expected from a 12V transformer).

How many terminals does the new transformer have?
Perhaps it's a 115V transformer mismarked as a 230V.

 

Core saturation occurs because of the input voltage being too high and/or the input frequency being too low.
High load current does not cause saturation.

I was thinking the same but it has been too long since I had my head in this game, I wasn't sure.

I noticed the transformers have 8 pins. I assume there is some jumpers on the back? Can these be configured for full(series)/half(parallel) input voltage and full(series)/half(parallel) output voltage like a typical control transformer? If so, how are they configured and what is your input voltage?

 

The capacitor does not have to be exact value. 100-1000μF in parallel with the load.

View attachment 230371

That would smooth the signal; I would get a flat response.
However, what I need is a full wave rectification, without smoothing.

 

There is some information on these transformers at the link below.
They are described as 'safety isolation transformers'. Can this construction affect the output waveform.
https://www.poweruc.pl/products/pe-series-pcb-safety-isolation-transformer-pe2012-m-110v-0-6va

 

How many terminals does the new transformer have?

This is the back of one of those:

The transformer has 4 terminals (2 primary + 2 secondary).
Although, you can see there are 8 terminals, half of them hidden by... silicone?

You appear to be overdriving the transformer as the scope says the output of your new transformer is 26-30V while you old transformer is 16.6V (as would be expected from a 12V transformer).

Yes, around 17Vp is I was expecting.
However, manufacturer says the secondary for this "12V" transformer is 20V under lo load:

Source

I'm still confused with this, because what I really wanted are those ~17Vp, and now I have ~30Vp, almost double.
However, is my fault for not checking the product data sheet. Well, is just a transformer, not a microprocessor, I thought I could not be wrong with something so simple.

Nonetheless, apart from my fault on choosing voltage, what I don't understand is why that crooked signal.

 

They are 0.6VA - transformers that small are generally wound with rather too few turns, and run in saturation using the resistance of the primary to limit the current.
I once bought some with the intention of using them as measurement transformers, and the output is not proportional to the input, even without a load.

 

can you share a picture of your test setup in use please

My test setup cannot be more simple:


The transformer is welded to a prototype board with some terminals and a bridge. I put an additional diode and capacitor for testing purposes.
The load is just a 1KΩ resistor.
The probes are connected to the resistor's terminals.
My first test was with the transformer alone, with no board, nothing more. The transformer, some cables and the scope.

Can you also share a picture of the input waveform at the same time ,

It is mains at 240V... it is scary.
Well, I could build a voltage divisor.

 

Your earth on the scope probe is terrible,
Wire a resistor straight across the output of the transformer, get rid of that test board,
then clip the scope direct across that resistor,

nothing else connected.

Can you also post a picture of the underside of the transformer soldered to that vero board .

 

Why no grid lines on the scope traces--can't read time or voltage like that?

What is the resistance of the windings?

 

Let's remember that the rating is RMS volts, not peak voltage. The 20V no load rating is a RMS value, with peak being close to 28V. If your line voltage is a bit high, that would explain the 30V peaks you are seeing.
Also, the 12V RMS rating is at a load of 50mA. Anything lower in current load will cause the voltage to read higher. Use a proper RMS voltmeter to read the voltages, or get your scope to report the RMS values, not peak.

 

The odd shape to the waveform out of the transformer is odd harmonic distortion that results because of a combination of the magnetizing current driving the core and an IR or IZ drop in the primary. These safety transformers are usually "impedance limited" so they cannot supply large currents because of the large impedance between the primary input and the secondary output.

I ordered a dozen of them once without realizing how poor the performance would seem. I still have all of them unused.

 

Your earth on the scope probe is terrible,
Wire a resistor straight across the output of the transformer, get rid of that test board,
then clip the scope direct across that resistor,

nothing else connected.

Let's take a new transformer (I'm unlikely use them for any other purpose, as DickChappels):

Everything soldered.
4x10K resistors in parallel (2K5 total) because only one 1/4W is too little and got warm after a few seconds.
Probe directly attached to the resistors.


Same result: crooked wave.

Why no grid lines on the scope traces--can't read time or voltage like that?

Grid lines are there. Cannot see them?
Maybe is a matter of monitor calibration?