A friend of mine has built a DIY square wave inverter, which for driving an 120V 8.3W pump. The pump was turned on an off at twice a second. He picked a transformer rated for 12VA. But the transformer was overheating. I wonder what could be causing that?
On the web, there are many DIY inverter projects with schematics. Unfortunately, they don’t show which transformer model to use.
The below link has fairly good amount of explanations of what causing the transformer over heating and about the power consideration. Paying a visit to the site will be helpful. Practical considerations : TRANSFORMERS
Dear Kender
Hi
Did your friend used , EMI filter?
BTW : your loss power , is due to the harmonics of your square wave. you can improve it with some simple filters.
Best Wishes
Goldsmith
In addition to the points raised by others, and in particular expanding on what Johansen said, AC motors usually draw a higher current than would be expected from the simple ohmic formula I = W/V. This is because the current is out of phase with the voltage. But the current is what heats up the transformer coils and the major contributor to transformer heating. Hence the need for a higher VA rating.
Could the pump be drawing extreme surge current? This is typical of AC motors when starting up, before they have reached speed. Also, AC motors are known to draw high current when bogged down.
Thanks for these comments. We may be underestimating the current consumption of the motor. So, we bought a cheap inverter rated for 65W continuous and 200W peak for 5 sec. That inverter powered our pump without overheating. Then we've looked at how it was designed. It's built around a KA7500B chip.
What was surprising for me was the compact transformer (or "xfrmr" for short). The photo below shows the 65W inverter (I drew a purple frame around its xfrmr) and a 12W xfrmr we were playing with. Inverter's xrfmr is made with a powdered iron or ferrite core (it doesn't have markings - no way to find the datasheet). I'd like to buy a transformer like that, but I haven't found one yet. If we can find a core, we can wind one ourselves, probably. But I'd prefer to buy one fully made.
The xfrmr we were playing with in the beginning has a laminated core, and it's heavier and bulkier. It's on the right side of the photo (click thumbnail to enlarge).
So many factors determine transformer behavior. It's mystifying.
The outfit that made the efficient inverter has someone who understands transformer specs and parameters, to do their research and development.
I have seen toroidal transformers mentioned in some discussions, as being a more efficient type. A donut shape is less easily broken when molding powdered iron/ferrite into transformer cores.