Rebuilding a PCB mount transformer

[electronwaster]

(Crosspost from the shoutbox, thought this might be the best place to ask after all.)

I am fixing a TDK ipod stereo thingy for someone, and a thermal fuse inside a PCB mount transformer has blown. I have swapped it out for a physically larger one - one with the same temperature specs, but higher current specs - all I could find.

Trouble is, I couldn't get all the EI laminations back in (got 21 of 24 back in), since the bobbin was a bit bulgy. What will be the result of this? Lesser efficiency, and therefore more heat, and so even more likelihood that the fuse will blow in the future? Hope not.

(At the very least I'd like to be able to run it for a while and measure the input/output turns ratio, so I can get a replacement.)

Cheers,
electronwaster

Edit: It passed a small smoke test. Connected to 18V rms input, gave 0.764V rms out, which is a ratio of 23.5, so for 240V mains, that would be 10.25V rms unloaded. So - I'm guessing a 9V AC transformer? Does that make sense?

[Psi]

Transformers are very resilient.

If you run it at full load for a while and it doesn't get hot then it should be fine.

If you plan to leave it will missing laminations..
Modifying mains parts and then giving it to other people isn't the best plan.
They will likely miss any sign of imminent failure that you might pick up on. Such as detecting a slight smell of electric smoke in the air before going to bed.


Edit, if your only trying to get the old transformer working so you can measure its specs for a new one, then yes, what your doing is fine.

[electronwaster]

Hi Psi,

That is a very good point! I can't believe I didn't think about it like that, all I had my mind fixed on was whether the device would continue working, not what could happen if it stopped working.. I will get a replacement transformer.

I have a feeling that I'll have to get a slightly larger transformer from jaycar or similar and connect it to the PCB with flying wires, but that is no problem.

From the information in my previous post, does it seem like it would be a 9V transformer to you? Is there a way to know for sure without melting the thing? I have no current/voltage/power specs at all. there is nothing on the PCB silkscreen, and I have spoken to TDK technical support twice - they have no information other than to sell me a new unit.

thanks,
electronwaster

[Psi]

Yeah, it's a bit different when it's a repair for someone else. Ya have to consider safety aspects
that you wouldn't consider if it was just for yourself.

The missing laminations shouldn't effect the no-load input/output ratio for testing.

9V is quite a common transformer value but having only 18V test input may not give you a good reading on the ratio.

Have you checked the PCB markings, sometimes they will have the voltages written next to where wires connect.

If you feel comfortable working with mains you could connect it up on the bench and then carefully measure the output.
Assume the output maybe at 230v until you confirm that it isn't with the meter.
Use an RCD/GDI if you have one, and have someone else present in the room while your testing.

[electronwaster]

Hi Psi,

Yes -  I'm very comfortable with mains, have set up ring mains, lighting circuits, fuseboards, etc. It's the small actual electronics stuff where I have not much idea.

The secondary winding measures 9.50V when connected to mains (through a resettable fuse), so I think I'm probably reasonably safe replacing it with a 9V transformer.

There were no PCB markings at all, other than something like T1 or TR1. My wife is near jaycar, so I just this minute asked her to pick up a transformer. 30 minutes and it will be here!

cheers,
electronwaster

Edit: I just put it all back together with the old transformer, for 3 reasons;

* Does it work? Yes
* What's the output voltage under a normal amount of load? 9.odd volts, less than 9.5
* How warm does the primary winding get under load? Currently testing this. Will leave it at 75% power for 15 minutes or so, just to give me a rough idea. The thermal fuse was a 125C one, and it's at 28C after 5 minutes..

[ivan747]

Tell us how it goes!

[electronwaster]

It's such a small transformer (low thermal mass), I don't think it needs testing for hours on end like big power transformers.

It's at 35C or so after 10 minutes of triple j on nearly max volume, I can't believe much current goes through this thing, and I don't think it will get to 125C.

The primary winding wire was like hair. It was the smallest thing I've ever soldered - and that was with my big clunking 40W iron .

[electronwaster]

Here is the old (rebuilt) and new transformer. The new one was $7.95 from jaycar (http://jaycar.com.au/productView.asp?ID=MM2017&form=CAT2&SUBCATID=1002#12). I'm sure I could have got it for $2 delivered on ebay, but oh well.



It's working nicely. The previous transformer eventually got to about 62C, which is fine, but I didn't want to take any chances.

The new transformer is rated for 1.35VA, which is very low, but it's physically bigger than the old one, so I am assuming it is rated for more power, I have no idea what the old one is rated for, of course. It's about 3mm taller, and now only JUST fits in the box. I had to cut the feet off and glue it to the board too, which I don't like doing, but I don't know what else I could have done.

It's been running for 5 minutes now, and still feels cold to the touch, so I think that's a success!

thanks for your help,
electronwaster

[Psi]

The previous transformer eventually got to about 62C, which is fine, but I didn't want to take any chances.

Small transformers like that should never get to 62degC.  It may still work fine at that temp but it's unusually hot.
Most of the time transformers sit between cold and very warm.

A shorted turn may have been the reason for 62degC. In which case its good you replaced it

[electronwaster]

Okay - fair enough - I was naïvely employing the "hot to the touch?" rule, which is OK for semiconductors. I could hold my finger on there just about, so it passed that test, but now I know better. It'll go in the bin as soon as I've shown it to the owners of the hifi unit to explain what went wrong. I spent all that time dismantling it and renewing the fuse too! Ah well, it's good as new again now.

Cheers,
electronwaster

[Psi]

Yeah, An overall temp of 62C isn't hot enough to cause damage, but it is unusual.
So you have to wonder if it really is 62C through and through, or 300C+ at one tiny point inside.