[Amps] Testing filament xfmr

[Will Matney]

Really the best test is use a load bank so the secondary will pull the maximum rated current. The load bank of course is a large set of resistors. An air cooled transformer for our type work shouldn't get hot but only warm to the touch. Never hot enough that you couldn't leave your hand on it. If enough of the right amount of iron is used, copper loss is always the predominant one. This can be controlled by using large enough wire with a minimum of 750 circular mils per ampere for ICAS work. You can fudge on this a little for a low voltage transformer as long as the primary is wound using the 750 figure. The secondary, if it's the outside coil, and no more than two layers, can be a little less than 750. The reason being is it can cool off easy being the outer coil. For other secondaries though, using 750 is a must. For CCS, 1000 Cir. Mils per amp or up to 1200 needs to be used. Commercial door bell type transformer are designed to be ran kind of hot though and they use something like 500 cir mils per amp. They dont care about the life span on these nor if the secondaries waveform is a bit distorted. They run pretty warm without a load. Some of the other cheap mass-produced transformers kind of skate by too, so using them is not a real good idea for long term performance. You can tell by the resistance of each winding about how many watts it will dissipate by simply measuring the resistance while cold and after ones been under load for a while. If you dont have a load bank though, the test by shorting the coil is done.

Best,

Will

*********** REPLY SEPARATOR  ***********

On 8/11/05 at 2:53 PM Peter Chadwick wrote:

>David said:
>
>>I can't really see how you could determine the temperature rise too 
>>easily by this.
>If it doesn't get too hot on open circuit, the iron losses aren't too
>high. If it doesn't get too hot on short circuit, the copper losses aren't
>too high.
>Probability is that it won't get too hot in service. Rough and ready
>tests; most transformers seem to get hotter under load, suggesting that
>copper losses predominate.
>
>Time to worry is when the pitch or resin or whatever bubble
>out.......messier is the oil filled xfmer where the pressure leads to an
>oil leak....
>
>73
>
>Peter SM/G3RZP/P
>
>
>
>
>
>
>
>========================================
>Message Received: Aug 10 2005, 11:18 PM
>From: "David Kirkby" 
>To: g3rzp at g3rzp.wanadoo.co.uk
>Cc: StephenTetorka at cs.com, "amps at contesting.com" 
>Subject: Re: Testing filament xfmr
>
>Peter Chadwick wrote:
>> Simpler is to connect an ammeter across the secondary, feed the 
>> primary from a Variac and wind up the Variac until trhe desired filament 
>> current flows. then leave it for a time and see how hot it gets. 
>
>The classic short circuit test.
>
>> The 
>> probability is that most of the losses will be copper losses: 
>
>agreed
>
>> leave it 
>> on full primary with open circuit secondary to get an idea of how much 
>> iron loss there is by the temp rise.
>
>It's a long time since I did this, but the short and open circuit tests 
>give you copper and iron losses respectively. I think you can show that 
>theoretically copper_loss == iron_loss for maximum efficiency, but the 
>transformer would no doubt have been wound on a suitable former, rather 
>than one that is theoretically optimal.
>
>I can't really see how you could determine the temperature rise too 
>easily by this.
>
>It's not my area of expertise, but I would have thought a load resistor 
>a safer bet.
>
>
>-- 
>David Kirkby,
>G8WRB
>
>Please check out http://www.g8wrb.org/
>of if you live in Essex http://www.southminster-branch-line.org.uk/
>_______________________________________________
>Amps mailing list
>Amps at contesting.com
>http://lists.contesting.com/mailman/listinfo/amps