UPS Isolation Transformer???

[jjasilli]

ISSUE: Any suggestions on how to safely test this puppy to see which wires are which?

I salvaged a transformer from a defunct UPS power supply.  (It's PCB had a burnt out trace.)   The tranny is fist-sized and ways over 3 lbs.  I'm confused by the wire leads --  one side: red, white & black leads, all 18 AWG;  the other side:  red, brown & yellow leads, all 14 AWG.

It's only possible purpose would be as an isolation transformer.  I would like to out it into service in this capacity.  But how to wire it?

Thanks!

[kagliostro]

Have you tested which is the transformation ratio ?

Attached an image that can be of some interest

K

[jjasilli]

Well it's not an isolation tranny.  Set my variac to 10 VAC for input into this tranny under test.  Measured all possible two lead combinations on both sides.  Did not try putting leads in series or parallel.  First with the 18 AWG as primary; secondary output was 0; 0.2; or 0.5 VAC.  Next with the 14 AWG side as primary, secondary output was 0; 30; or 140 VAC.

Don't know what to do with it.

[kagliostro]

There is something I can't understand

the transformation ratio can't be different from one side to the other

if 10v on 14 AWG winding gives 30v and 140v on the 18 AWG winding

the transformation ratio is x3 and x14

then

10v on the 18 AWG (on the winding where you previously measured 30v) must give 10v/3 = 0.33v on the 14 AWG winding

10v on the 18 AWG (on the winding where you previously measured 140v) must give 10v/14 = 0.71v on the 14 AWG winding

----

if you connect an oscope to the 18 AWG winding and your variac to the 14 AWG winding, you can increase the voltage till you see on the oscope an image like that on the low part of the image I posted

so you know you have reached ~1.5 times the voltage to which the transformer is rated, divide this voltage by 3 and multiply for 2 and this is (near) the voltage that is rated that winding on the transformer

and, yes, I don't think that is an isolation transformer, I think that the electronic circuit of the UPS put a square wave (at an unknown frequency - 60Hz ??) on one side of it to have the rated voltage on the other side (120v ? - Idon't know which is the output voltage of your UPS)

Which was the battery voltage ? 12v or two 12v in series for 24v ? this can help you to understand better the transformer

K

[John]

It should be a step down transformer to charge the battery in the UPS... but I thought those were 12V batteries. Can't figure how that ratio would give you that, unless it's a 240V tranny, which would then make the ratio closer. Seems odd the primary wires would  be 18 ga. though.

[jjasilli]

It'a all a bit odd to me too.  However the predicted .33 and .71 V is pretty close to the .2 & .5 V actually measured. 

If it's a step-down transformer, it's purpose would probably be to operate active SS components on the PCB.  Would this require a 3 amp tranny.

[PRR]

A UPS with a battery?

There is NO need for isolation. The computer runs right off the line and the battery is in a sealed box, can float at line potential. (Things are different in very-large UPS racks where batteries can be changed "hot".)

The transformer does two things:

Steps 120V down to about 12V at low current to constantly charge the battery.

Steps ~~~12V AC (from battery and inverter) up to ~~~120V at high current to run the computer for a short time in a blackout.

The ratios must change from standby to Run mode. We need 120V down to more than 12V to charge. In Run, the battery sags below 10V but the output must stay near 120V.

I have tried to re-purpose these things and decided the designers were much too clever for me. Between Best and APC, brutal competition drove the price of UPSes _VERY_ low, which means they are full of penny-shaving tricks. Add to that the many different ways inverters can drive transformers, and the infinite variety of Run waveshapes possible.

For more fun: small UPSes do not deliver sine-waves in Run. One I had made pure square waves at 150V peaks. This runs electronics just fine (similar to the 168V peak of a 120V sine) but makes a lamp Really Bright because it feels 150V. Most now use a 3-step wave which has peak/RMS ratio more similar to a sine, but is far from a sine. There are also 5-step waves, and whatever waves heavily filtered to near-sine, but you won't see these.

> you have reached ~1.5 times the voltage to which the transformer is rated

UPSes have a special case. Charge can last forever, but the Run time is known exactly and is very short. Office-size UPSes give you much less than 20 minutes of run time, often less than 10 minutes. It can NOT run longer because you only paid for so-much battery! Also most UPSes will only run their full 7 minutes once or twice in the 5-year life of the battery, and most USPes are discarded when the battery gets old (most offices don't have a screwdriver; a whole new UPS costs very little more than the battery).

If the original wall-wiring were known.... but I have tried to trace this through the PCB through filters and relays and it is a real puzzle. We know that two leads go to the 120V/240V wall, for Charge mode, and this is safe "forever". However there may be step-up leads at higher voltage for Run mode, not good for more than a few minutes at the higher voltage.

And it is very likely it has NO isolation. The only use I can think of is charging a battery in a sealed box or running 12V lamps from wall-power.

> It's PCB had a burnt out trace.

That can be patched. However that is not the problem, and if you do a large-wire patch you will probably have a bigger problem. Most likely a diode or an inverter transistor failed short. The entire battery capacity flowed into the short in an instant. While thee may be fuses, an inverter has too many ways to go wrong to cover every possibility The PCB trace widths are designed so they will blow fairly quickly when stuff goes wrong. Throw down a patch of house-wire, it will hang-on until something else goes boom.

[John]

and decided the designers were much too clever for me.

In that case, I'm done even thinking about it! 

[jjasilli]

Thanks PRR.  I had vague suspicions in that regard; your response really fleshes that out.

Also there is some continuity between the opposite leads.

cheesy applications,  Could I suggest determining if the xfrmer has a manufacturers and model number on it.  If you are lucky, you might be able to find technical data on the part.   

Tried that.  It's a huge Chinese company supplying other big manufacturers.  No data on the website.

Looks like 3 lbs of iron is off to the re-cycling bin.