Serious transformer problem

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I had rewound the secondaries on two 950 W transformers a few months back, both identically. I just tried measuring the inductances of their windings, and to my surprise one of the transformers measures more than twice as high as the other one... WTF
When connected to the mains, both transformers seem to work fine and put out the same voltages on their secondaries, draw the same current from mains with secondaries open, and drop about the same voltage on the secondaries when a heavy load is attached.
Yet they measure so hugely differently, by a factor of at least two, and when I feed the primaries from my scope's 1 kHz square wave source, they load it differently.
I don't get it, how to troubleshoot this. None of the four secondaries could be shorted, because otherwise a winding wouldn't put out the right voltage. The foil electrostatic shield between the windings I'm pretty sure is not shorted, and the two ends are separated so there's no complete turn; it would probably overheat if it was shorted. I don't know what else it could be. I haven't touched the primaries when I was rebuilding the transformers. Both transformers also have a copper flux band outside the core (parallel to the windings), as well as magnetic shielding alloy along the core's perimeter, but all is the same here as well. I'm completely clueless.
 
If they measure the same current when connected to the mains, they must have the same inductance at the mains frequency.


Sy, if there are varying gaps, I think that would affect the unloaded line draw as well. Since he said identical draw, I suspect a frequency dependence.

What freq is your inductance meter?

John
 
Gap? The core laminations are interleaved, E and I swapped each layer. There is at most a difference of 1 (of about 100) in the number of laminations in each.

Each has four identical windings. I measured current with a shorted secondary, and 150 W bulb as ballast on the primary from mains. Primary and secondary currents were same for both within 5 %.

The multimeter has two ranges and measures with 1 kHz on one and 200 Hz on the other. So for the primaries I get (open secondaries): transformer one: mH range: 146.1 mH, H range: 0.142 H; transformer two: mH range: 284.4 mH, H range: 0.265 H. On the scope, the 1 kHz square wave is loaded as sketched in the attachment, each color is one of the transformers.

[edit] jneutron, the frequency dependence makes sense for 1 kHz, but not the 200 Hz the meter uses for its H range.
 

Attachments

  • square.png
    square.png
    899 bytes · Views: 357
Nixie said:
Gap? The core laminations are interleaved, E and I swapped each layer. There is at most a difference of 1 (of about 100) in the number of laminations in each.

Each has four identical windings. I measured current with a shorted secondary, and 150 W bulb as ballast on the primary from mains. Primary and secondary currents were same for both within 5 %.

The multimeter has two ranges and measures with 1 kHz on one and 200 Hz on the other. So for the primaries I get (open secondaries): transformer one: mH range: 146.1 mH, H range: 0.142 H; transformer two: mH range: 284.4 mH, H range: 0.265 H. On the scope, the 1 kHz square wave is loaded as sketched in the attachment, each color is one of the transformers.

Ah..ok..the meter changes freq to pick up resolution..I believe it is going to 1k for the lower inductance range. So, tranny 1 seems stable with freq, tranny 2 drops inductance as freq goes down.


Your initial statement:

Nixie said:
When connected to the mains, both transformers seem to work fine and put out the same voltages on their secondaries, draw the same current from mains with secondaries open, and drop about the same voltage on the secondaries when a heavy load is attached.
From your test description, you didn't measure the unloaded current of each tranny. Try measuring it unloaded..what is the current draw when the secondaries are open. This will measure the current of the primary due to inductive reactance plus that due to eddy losses in the core.

John
 
I did: "draw the same current from mains with secondaries open"

I don't know if directionality of the windings can make this frequency dependent difference. I'm not sure all are wound in the same direction (the secondaries are not on top of one another but split-bobbin type arrangement, all on top of the primary and its foil shield). But I guess it shouldn't, as I hadn't connected them to each other.
 
Nixie said:
[edit] jneutron, the frequency dependence makes sense for 1 kHz, but not the 200 Hz the meter uses for its H range.

Agreed. Your whole problem isn't making sense, that's what is so intriguing about it..I love it..;)


Nixie said:
I did: "draw the same current from mains with secondaries open"

If the inductance at 50(60) hz is different, they hafta draw different currents. Your statement of "within 5%" seems to indicate that difference. I can't believe that the inductance would plummet to identical at line freq..it just hasta be diff. Measure the line current directly, secondaries open, with no series load on the primary.

Nixie said:
I don't know if directionality of the windings can make this frequency dependent difference. I'm not sure all are wound in the same direction (the secondaries are not on top of one another but split-bobbin type arrangement, all on top of the primary and its foil shield). But I guess it shouldn't, as I hadn't connected them to each other.

Winding directionality will have no consequence on the inductance.

Cheers, John..

ps..this problem is fun, thanks.
 
Nixie,

Your gap occurs between the E's & I's, in each layer, where they meet. The reluctance of your magnetic circuit is primarily a function of this small gap. It may be. 0.004" on tranformer and 0.006" on the other.

Measure the inductance of the secondaries... you will see the same ratio as the primaries.

The question is: does the transformer with the most inductance saturate at full power?



:)
 
Nixie said:
...
The multimeter has two ranges and measures with 1 kHz on one and 200 Hz on the other. So for the primaries I get (open secondaries): transformer one: mH range: 146.1 mH, H range: 0.142 H; transformer two: mH range: 284.4 mH, H range: 0.265 H. On the scope, the 1 kHz square wave is loaded as sketched in the attachment, each color is one of the transformers.

[edit] jneutron, the frequency dependence makes sense for 1 kHz, but not the 200 Hz the meter uses for its H range.

Hi Nixie,

Just my view: don't you think a primary induction of merely 146.1 / 284.4 mH (!) is a bit low for a fat @ss 950VA tranny? It should be in the range of multiple Henrys. I just measured a 500VA toroidal that I have here and it's about 14H :eek:.
Maybe your measurement is wrong?

Regards,

Ray.
 
6h5c said:
is a bit low
Well sure, the cores are designed for 60 Hz, not 200 or 1000. An EI choke I wound and I'm certain is at least 5 H measures less than a tenth of that at the meter's frequencies. Yet high frequency cored chokes with known inductance give the expected number at measurement, so I know the meter is working fine. The point is, something's gotta be wrong if I get such a huge difference even if at frequencies that are much above the mains.

poobah said:
Your gap occurs between the E's & I's
Yes, and they are likely all different, since I assembled the thing by hand after rewinding. I hammered a bit on each transformer after assembly to minimize the gap, but there's probably on average up to a tenth of a millimeter variation. Still, due to variation over hundred laminations, it's likely to average out to very similar gaps on each transformer.

Measure the inductance of the secondaries... you will see the same ratio as the primaries.
Yes. But again, can I be sure it's the core? It still bugs me the small possibility the electrostatic shield aluminum foil between the primaries and secondaries may have its ends shorting each other, though I just poked it around with a small screwdriver from the side where the ends overlap around a paper insulator, and didn't make any difference. I can't see deeper towards the middle of the bobbin though. Partial shorts like that usually decrease output voltage, but since it's aluminum foil and it's just one turn, who knows.

The question is: does the transformer with the most inductance saturate at full power?
Don't you mean with the least inductance?
The sine wave when powered is the same on both transformers, both primaries and secondaries, so I guess neither one does, whether there's a load or not (actually more load should decrease the chance of saturation as it increases resistive losses, no?)
However, the transformer with lower inductance buzzes more. Not the windings, but the magnetic shielding material wrapped around the core (perpendicular to the windings. However, if the core was saturating, it would probably saturate the thin wrap as well, and distort the sine waves on the scope. Might be close however and that could cause the buzzing, or more leakage flux is causing the buzzing. I know it's not the windings since when I squeeze them through the copper flux band I feel no vibration. When I touch the shielding, it's vibrating.
The primary is multitap. Buzzing is loudest with the tap for the highest secondary voltage (least primary turns), as makes sense; I'll probably avoid that one; luckily the middle tap gives me enough voltage, though there is still some buzzing -- which is likely related to the difference in measurement.
 
Nixie said:
Well sure, the cores are designed for 60 Hz, not 200 or 1000. An EI choke I wound and I'm certain is at least 5 H measures less than a tenth of that at the meter's frequencies. Yet high frequency cored chokes with known inductance give the expected number at measurement, so I know the meter is working fine. The point is, something's gotta be wrong if I get such a huge difference even if at frequencies that are much above the mains.

Nixie,

I know the inductance is depending on the measuring frequency. But still the inductance seems low. What I meant is: could your meter just not be able to handle this large inductance?

Regards,

Ray.
 
The one buzzing more is the one with the lower measured inductance. Could something be causing more leakage perhaps? Maybe a bolt shorting between laminations, though that could at most affect a few layers. I guess I could unscrew the bolts and test again; the nail polish should still hold things together.
 
It could be gapping, but like I said, I treated both transformers the same, so though various layers vary, the variations would average out similarly for both transformers. It's doesn't seem likely this is the issue, especially because the difference is more than a factor of two. I'd expect small variation in gapping to have a lesser effect.
Each layer is slightly offset from the next, but the shift on the E side corresponds to the one on the I side at the time I built it. The nail polish is very strong and I doubt it could have shifted over time, despite the tight squeeze of the core around the bobbin.

I'll try this now: put it in the vice and squish the Es with the Is, see if it makes any difference in the reading.
 
Makes less than 10% difference when I put on all my weight, instead of the 100% I need. Pulling out the bolts made no difference at all. Sigh...

I gotta get some sleep now, but then I'll remove the magnetic shielding and put more pressure on, see if that does it.

I don't understand with such a huge difference why measurements of current etc. when powered are so close to each other.
 
I wonder if the magnetic shield is shorting between laminations, since I didn't spread the silicone very well. I guess that would be the best case scenario, since if the gapping is a problem, I'd probably have to rebuild the core, unless I could somehow soften the nail polish without dissolving the windings' enamel.
Much worse is if something's wrong with the windings, especially the primary. But given that the ratios seems OK, I doubt it, unless weak insulation spots behave worse at high frequencies. Worst is if it's the electrostatic shield, as I'd have to completely unwind the secondaries.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.