Sizing an Isolation Transformer for a DC Power Supply

[aemla]

TL;DR Summary

To properly size an isolation transformer suppling DC power supply, should the input or output power be used?

Good afternoon,

How should an upstream isolation transformer be sized for a dc power supply?
For example, if we have a power supply with the following specs:
Input Voltage: 120 VAC
Input Current: 2 Amps
Inrush Current: 22 Amps
Efficiency: 90%
Power Factor: 0.55
Output Voltage: 24 VDC
Output Current: 4 Amps
(Spec for 1606-XLP95E power supply)

Should the Isolation transformer be sized for the input or output power of the DC power supply?
I've seen both answers in different blogs. What I haven't seen is a good explanation for either case.

On one side, DC power supply has an input rating implying that it is what it needs to operate properly.
On the other hand, power supply output is much less to drive it's loads. It would also make sense to size the transformer based on the output power with a correction for efficiency. In this case, it would be P=V(out)*I(out)+Efficiency*V(out)*I(out).

This leads me to another question: What happens to the reactive power? Does it need to be accounted for when sizing a transformer?
Because a dc power supply contains capacitors to smooth out the dc waveform after the bridge rectifier, does the reactive power reflect fact to the supply? Does this also mean that when sizing an isolation transformer for an ac motor (inductive load, therefore, reactive power is being used and not reflected back to the line), the reactive power needs to be accounted for?

 

[Baluncore]

The 1606-XLP95E power supply appears to be a switching supply, with direct rectification of the input voltage, maybe with PF control.

The DC output comes from an efficient buck regulator. Such a power supply must be internally isolated by the high frequency transformer in the buck converter regulator.

So why do you need a heavy, low-frequency, isolation transformer, between the 120V AC supply and the 1606-XLP95E power supply input?

 

[aemla]

The 1606-XLP95E power supply appears to be a switching supply, with direct rectification of the input voltage, maybe with PF control.

The DC output comes from an efficient buck regulator. Such a power supply must be internally isolated by the high frequency transformer in the buck converter regulator.

So why do you need a heavy, low-frequency, isolation transformer, between the 120V AC supply and the 1606-XLP95E power supply input?

If actual supply voltage is 480 vac where a step down transformer is needed.

 

[DaveE]

I would size it for the worst case transformer load in normal operation as determined by the maximum PS specs. If you always have a smaller load than the PS capability, then buy a smaller PS. This has as much to do with failure modes and safety standards as what normally occurs, IMO.

I wouldn't worry about the reactive power unless you have a particularly obnoxious transformer load. As @Baluncore said, if it's a SMPS, it'll be OK. If it's an old school rectifier and capacitor, the PF isn't as bad as you might think at full load where you do the worst case analysis. OK, wait, yes, it's bad. But not as bad as you might think. It's very sensitive to the source impedance. Ask if you have this sort of problem and I'll give you my favorite, sort of ancient, reference (or run your spice program on it). Basically it's about the RMS current in the transformer load.

 

[Baluncore]

If actual supply voltage is 480 vac where a step down transformer is needed.

Analyse for step-down, rather than an isolation transformer.
A step-down transformer will handle and limit the short inrush without problems. The input current is the vector sum of the in-phase and quadrature currents.
Input: 120 V * 2 A = 240 VA. Output: 24 V * 4 A = 96 W. That 2 amp input includes the circulating energy, quadrature current, PF and efficiency. The 2 amp specification indicates how thick the supply wires must be, not to get hot during normal operation.

The secondary of the step-down transformer needs to be rated for 2 amps, at 120 volt.
One cycle of 22 amp inrush current will flow when the supply is turned on. The circuit breaker must be rated to survive that inrush. Consider a slower thermal breaker, rather than a fast magnetic breaker.

 

[aemla]

Thank you for the explanation, that makes complete sense.

 

[Dullard]

I agree with everything here, but 'Properly' is a loaded term. There are some practical considerations that haven't been mentioned:

Do you expect to operate near 'full power?'
Depending on the efficiency/rating of the transformer, it may be very hot when operating. That also means it will stink for a while when first operated. Consider how many times you want to explain that 'it's supposed to do that.'

Is the application weight or cost-sensitive?
In this case, a 50% oversized transformer will cost almost nothing. It's unusual to get into trouble for bringing a gun to a knife fight.

There are 480V DC Power Supplies. Depending on the circumstances, sometimes the 'best' transformer is no transformer.

 

[aemla]

I agree with everything here, but 'Properly' is a loaded term. There are some practical considerations that haven't been mentioned:

Do you expect to operate near 'full power?'
Depending on the efficiency/rating of the transformer, it may be very hot when operating. That also means it will stink for a while when first operated. Consider how many times you want to explain that 'it's supposed to do that.'

Is the application weight or cost-sensitive?
In this case, a 50% oversized transformer will cost almost nothing. It's unusual to get into trouble for bringing a gun to a knife fight.

There are 480V DC Power Supplies. Depending on the circumstances, sometimes the 'best' transformer is no transformer.

Perhaps I should of said "more accurate" out of the two choices.

For this example, yes, power supply and transformer will operate near full power.

Weight, size, and cost are always considered in real applications. For this example, I ignore those factors.

I've seen 480 VDC power supplies but because this transformer feeds other 120 VAC circuits, I went with 120 VAC power supply.
I wonder if NEC says anything about this specific topic.

 

[rbelli1]

The secondary of the step-down transformer needs to be rated for 2 amps, at 120 volt.

In this case, a 50% oversized transformer will cost almost nothing. It's unusual to get into trouble for bringing a gun to a knife fight.

This is 360VA plus whatever 120VAC other loads you need.

If your other 120VAC loads will be large compared to 360VA then

because this transformer feeds other 120 VAC circuits, I went with 120 VAC power supply.

seems like a reasonable path to take as 120VAC power supplies are easier to source than 480VAC ones and you need a big transformer anyway.

If the other circuits are small compared to the 360VA then you are bringing a heavy and expensive transformer to the party when it is not necessary.

BoB