# Why did the transformer heat up when connected to a UPS?

I have an electromagnetic vibratory feeder with 2-phase input, 380 V, 0.35 kW.

We noticed that the feeder has stable performance when it is connected to a generator because it has constant frequency and voltage. The main power source has frequency range 48 to 51 Hz and voltage range 370 V to 388 V.

To solve the problem we used a UPS with a pure sine wave output (SCL TWIN PRO2 catalogue) with 3000 VA power. We connected it to a 2 kVA, 220 V to 380 V transformer.

When we run the UPS with transformer we had stable performance, but the transformer heated up to 105°C after 30 minutes.

What is the problem here?

• Wasn’t there an identical question to this one about one week ago, but without the diagram? Commented Jun 5, 2021 at 16:31
• Which side of the transformer did you connect the UPS to and what voltage do you measure? Commented Jun 5, 2021 at 16:32
• yes, then i received some notes to add diagram ad here it is. Commented Jun 5, 2021 at 16:33
• i connect 220V side to UPS and 380V side to feeder, i measured the voltage of course before connecting Commented Jun 5, 2021 at 16:35
• And what’s the output voltage from the UPS? Commented Jun 5, 2021 at 16:39

To solve the problem we used a UPS with a pure sine wave output

You don't mention any problem beside the one using a UPS.

Possible reasons of the issue:

• The transformer heats up due to iron core losses. The output of the UPS I guess is a high frequency PWM modulated sine wave. You would perhaps need a filter or at least two chokes.

• The output if the UPS is not a sine wave, rather more a square wave that saturates the iron core.

simulate this circuit – Schematic created using CircuitLab

There is no exact way of determination of this filter. You have to know the PWM frequency, first. Then you determine the resonant frequency to be at least 10x 50Hz and at least 10x < fsw. Determine the capacitor current to be no more than 5% of rated converter output current at fundamental frequency.

$$I_c=\dfrac{V}{X_c}$$ $$X_c=\dfrac{1}{\omega C}$$ $$I_c $$C<\dfrac{I_n\cdot 0.05}{220\cdot 314} < I_n\cdot 7.24\cdot 10^{-7}$$

Supposing In = 11.8A then, C<8.5 micro Farad.

Let's choose C=3.3uF, and suppose the PWM switching frequency is 4kHz, so the filter resonant frequency shall be 400Hz.

$$f_{res}=\dfrac{1}{2\pi\sqrt{LC}}$$

$$L=\dfrac{1}{C\cdot f^2_{res}\cdot 4\pi^2}$$

It comes 48mH (2 x 24mH) , which is quite large.

I would suggest you to find the switching frequency, then use only two small chokes of 1mH without a capacitor.

• when i used METER i get 1.7KH ( on output of UPS, input of transformer), and 4.5KH on the output of transformer. is this could be right values? what is the value of the filter then? are you recommend any method for make transformer has lower losses, like welding Slices and but E & I Slices like microwave transformer? Commented Jun 6, 2021 at 14:20
• could i use any method for cooling transformer like fan or oil tank? Commented Jun 6, 2021 at 18:16
• KH = kHz?? It could be a sign of 3rd harmonics which means that the transformer is saturated, i.e. primary voltage too high or bad build transformer. Usually cheap toroid transformers are cheating on specs. What kind you have? Other possibility is a ferro-resonance, in case you have a capacitor on the motor. In normal conditions, there should be no issues. Try to use a lamp or a heater as load and see if the behavior changes. Commented Jun 6, 2021 at 19:28