No load losses - normal vs reverse fed

[gar]

170118-1457 EST

This thread has digressed to various aspects of transformers. But the original post question was does the transformer core loss differ between forward and reverse feeding the transformer.

For a transformer with a high permeability core material, relatively little air gap in the magnetic circuit, the primary and secondary coils coupled to the same part of the magnetic circuit, and the same "voltage per turn" applied to one coil or the other, then the core loss is the same whether the power input is to the primary or the secondary.

The assumptions in this statement are pretty much valid for any normal power transformer of any size (at least covering this range 0.1 W to 10 MW) using conventional design techniques.

Core loss consists of the sum of of hysteresis and eddy current losses. It is not easy to separately measure these losses. Eddy losses are a result of induced voltage in the magnetic material and the electrical resistance of the magnetic material. Lammination thickness or particle size and material electrical resistivety determine this loss. Hysteresis loss is a function of the magnetic material characteristics and frequency of excitation.

.

 

[iwire]

You're reporting the samething I reported on a 600VA transformer, except even smaller. The GE whitepaper notes transformers over 3kVA are generally uncompensated. Sometimes the most difficult journey is implementing theory into reality.

And the something many of those in the field reported from the start of the thread.

No experiments needed.

 

[GoldDigger]

One possible very small difference is in the wire size of the windings as it affects copper losses at full load amps.
In a one way design the primary will be sized based in full load current plus magnetizing current.
The secondary is designed for full load current only.
The result is that the copper losses may be *slightly* higher when reverse fed.
At zero load, no big difference.

mobile

 

[Phil Corso]

170118-1457 EST But the original post question was does the transformer core loss differ between forward and reverse feeding the transformer. .

No... OP asked about "No-load" losses! Thus, all losses should be determined to an accuracy greater than used to date!

I do agree subsequent posts answers went way off into la-la land! But, then, that is not unusual!!

Phil

 

[gar]

170118-1527 EST

Phil Corso:

I am still on your shorted secondary question. The results I got were a surprise. The phase angle is only about 10 deg. So a low Xl/R ratio. A 175 VA transformer is more in the 50 deg range. More on this later, and also with a comparision to a 175 VA unit, and a 500 VA control transformer.

On your newest question I need to know more precisely what I am to measure. The 80 W isolation transformer is not a 1 to 1 ratio. Am I to apply the same absolute excitation voltage for both forward and reverse? This will not produce the same core flux level. Or do I adjust excitation to approximate the same flux level, meaning the same voltage per turn?

For the 80 W isolation transformer the core loss vs voltage to primary at 60 Hz is:

Volt --- Power
100.1 -- 3.6
109.6 -- 4.5
120.4 -- 5.7
130.1 -- 7.2

A 9.6% change in voltage from 100.1 to 109.6 V produced a 25% change in core loss. From 120 to 130 the voltage change was 8.1% and produced a 26% change in core loss. So it is seen that % change in core loss increases as % change in voltage increases. IR loss in the primary has been ignored at this level. At substantially higher input voltages magnitizing current can not be ignored.

.

 

[gar]

170118-1619 EST

Phil Corso:

You are correct that primary IR loss from input current needs to be considered in a high accuracy complete experiment. In my post numbered 93 the IR loss was about 0.038 W.

But it wouldn't have been much different in reverse feeding the transformer. In that direction it was about 0.048 W.

.

 

[Phil Corso]

Gar...

Exactly what is shown on the nameplate? Arethere parameters other than just V, Hz, and W?

Phil

 

[Phil Corso]

Gar...

Your experiment results (Post #93):

1) Vin = 110.1 V, Iin = 0.09 A, Pin = 4.5 W, VAin = 10.1 VA, PF = 0.41 --- Vout = 116.2 V.

2) Vin = 116.1 V, Iin = 0.09 A, Pin = 4.6 W, VAin = 11.0 VA, PF = 0.41 --- Vout = 110.1 V.

If 1) above is the step-up case, and 2) the step-down case, then you proved that there is a change in no-load loss!.

To increase accuracy have clamp-on jaws around, say, 5-loops of source input lead!

Good work! now strengthen your case with lamp-load!

Ps; it is an isolating Xfmr!

Phil

 

[gar]

170118-1942 EST

Phil Corso:

I do not have any instrument to make accurate power measurements. The instrument I used as I stated was a Kill-A-Watt EZ, power resolution of 0.1 W, but not necessarily that accurate even at low power levels. The Kill-A-Watt EZ is really a pretty good instrument for it's price and quality of manufacture. It has better power measurement accuracy than a TED system.

The implication of the OP was that there are persons that believe there is a major no-load power loss difference between exciting a transformer on the primary vs the secondary. If the core is excited to the same flux level for the two different directions, then my measurement is sufficiently accurate to show that there is no major difference. A 10% voltage change produces a much greater difference in the no-load loss than the 0.1 W shown in my post which has a flux level close to the same value in both directions. The 0.1 W difference could simply be a rounding error or other instrumentation error.

I am willing to run the loaded test, but I need to know how to run the test, and what the test is supposed to prove. A near full load test is not going to provide a good means to prove anything about a no-load loss test.

.

 

[gar]

170118-2135 EST

Phil Corso:

Your post numbered 107 I did not see until now.

The only nameplate information was 117 to 117 V, 50-60 CY, and 80 W. There are no other parameters.

From my earlier test it is clear that at an 80 W load on it would not put out 117 with 117 input. But it is in the ball park.

On physical size it is about the same core size, but about 3/4 the stack height of a Signal A41-175 which is a 175 VA transforner. No load loss of the A41 is less than the isolation transformer, but has about twice the VA capability.

.

 

[gar]

170118-2143 EST

Phil Corso:

Were you a member of either IRE or AIEE at the time of the formation of the IEEE?

.

 

[Phil Corso]

Gar,

All three! Plus a few others! Why?

Phil

 

[gar]

170118-2405 EST

You were then one of about 10,000 founding members of the IEEE.

.

 

[GoldDigger]

170118-2135 EST

Phil Corso:

Your post numbered 107 I did not see until now.

The only nameplate information was 117 to 117 V, 50-60 CY, and 80 W. There are no other parameters.

From my earlier test it is clear that at an 80 W load on it would not put out 117 with 117 input. But it is in the ball park.

On physical size it is about the same core size, but about 3/4 the stack height of a Signal A41-175 which is a 175 VA transforner. No load loss of the A41 is less than the isolation transformer, but has about twice the VA capability.

.

It sounds from that that the transformer is intended to be 1:1 with compensated windings.

mobile

 

[Electric-Light]

170118-2135 EST

Phil Corso:

Your post numbered 107 I did not see until now.

The only nameplate information was 117 to 117 V, 50-60 CY, and 80 W. There are no other parameters.

From my earlier test it is clear that at an 80 W load on it would not put out 117 with 117 input. But it is in the ball park.

On physical size it is about the same core size, but about 3/4 the stack height of a Signal A41-175 which is a 175 VA transforner. No load loss of the A41 is less than the isolation transformer, but has about twice the VA capability.

.

Your no load power factor is quite high, but the magnetizing current is low. A big transformer operating at near actual rated voltage has a PF of 0.0x at no load. The transformer would be near intended magnetizing current at 140v 60Hz (V-Hz proportionate to 117v 50Hz).

 

[Ingenieur]

PID

used in a feedback control loop as the controller
the feedback is the actual measured/controlled system parameter
input is the desired or set point
output is the actuator/action that influences or changes the measured/controlled parameter
error E (or deviation) = desired - actual, if at set point E = 0

P = Kp x E, output signal is Proportional to Error
I = Ki x integral (E dt), output signal is a summation of Error over a time interval;
D = Kd x dE/dt, the output is based on the rate of change of Error, if Error changes in a large step, output will do the same

total output = P + I + D
most processes required only P and I, but some that have step changes may require D

 

[gar]

170119-0919 EST

GoldDigger:

Yes.

I think it is reasonable that a transformer labeled as an isolation transformer, or one labeled as a filament transformer, and apparently one labeled as a control transformer should have an output voltage that is near its labeled voltage at full load. This means that one can not use an ideal turns ratio to achieve the desired results.


Electric-Light:

The 80 W isolation transformer has about 50% more iron per VA than the 175 W rectifier transformer. We do not know what the design criteria is for these transformers or what are the differences in the magnetic materials.

I have not worked up the differences between these transformers yet.


Ingenieur:

I believe your post was for a different thread.

.

 

[gar]

170119-1649 EST

Phil Corso:

Earlier you ask for core dimensions.

The 80 W isolation transformer has a stack height of 1.38" when actually measured as best I can estimate because there are thin sheetmetal covers.

My Signal Transformer A41-175, a 175 VA transformer, has the same core dimensions except stack height is 1.5". This core is completely open and easy to measure.

The core magnetic material is quite different between these two transformers based on viewing the no-load input current wave shapes. The A41 has a more more square loop material.

From a comparison based on percent of full-load the 80 W isolation transformer has a higher no-load loss.

.

 

[Electric-Light]

Just throwing down some ideas here. For discussions sake lets assume 1:1 turn ratio.

If one winding is wound over the other one, you can not have symmetric coils.

The outer layer needs more length to get the same turns because the inner winding increases the diameter. So energizing the outer layer as primary causes more I^2R loss since added coil length adds R. If it used a split bobbin so the coils don't overlap they could be symmetrical.

A square core with two identical bobbins on the outer limbs would also have symmetrical winding but the efficiency might not be as good as single direction optimized design.

 

[gar]

160119-1955 EST

Electric-Light:

If one winding is wound over the other one, you can not have symmetric coils.

The outer layer needs more length to get the same turns because the inner winding increases the diameter. So energizing the outer layer as primary causes more I^2R loss since added coil length adds R. If it used a split bobbin so the coils don't overlap they could be symmetrical.

Correct.

I believe IR losses from no load excitation current are small for typical power transformers.

As an illustration my Signal Transformer A41-175 type, 175 VA, has the following no-load values as measured with a Kill-A-Watt EZ:

Volt ---- Amp -- Pwr -- VA ---- PF -- Calculated -- IR loss as % of Input power loss
---------------- Loss ---------------- Pri IR

120.1 -- 0.21 -- 4.2 -- 25.6 -- 0.16 -- 0.071 --- 1.7%
110.7 -- 0.14 -- 3.3 -- 16.5 -- 0.21 -- 0.031 --- 0.8%

Note: Kill-A-Watt's calculated VA is slightly in error. While volt, Amp, and Power are measured by Kill-A-Watt.

Note that at 120 V input the no-load loss is 2.4% of rated VA.

.