loading

Square wave DC transformer

If I was to feed a dc square wave into a transformer would I get a dc square wave as an output?

sort by: active | newest | oldest
AndyGadget7 years ago
No, you'd get positive and negative going pulses on the secondary. 
The transforming action happens at the edges and is related to the rise / fall times of the pulses on the primary.  When the input is DC (i.e. after each edge) the secondary output drops to zero.
What he said.  You'd probably get a messy AC at the output, which you'd have to filter and clean up before you could do anything useful with.
 . . . unless you're actually after a high voltage output, in which case you want to get lots of current into that primary as quickly as possible.

The output would probably look something like THIS. As Jeff says - messy.

If you're trying to level shift a square wave, using switching transistors may be a better way of doing it.
Nice find!
workinprogress37 (author)  AndyGadget7 years ago
Thanks for the info, this was a random idea I had and I was curious to know what would would happen.
prs092104 years ago
I realize that this is an old post, but in the event that someone comes across this by way of google--as I have--I think that they should know that the answers posted here are not necessarily true. I have actually connected a "modified sine wave" inverter to a 1:1 isolation transformer and measured the secondary voltage with an oscilloscope. MFW inverters output a square wave that goes from +170V to 0V to -170V to 0V... The secondary voltage with and without a load was a perfect duplicate of the primary. Now.. Even with a Bachelor's of Science in electromechanical engineering, I can only SPECULATE as to why this is happening. Before conducting the experiment, I had also thought (like those above) that you would essentially get an AC coupled signal on the secondary. That is, only the rising/falling edges would get through. I believe that at lower frequencies, this wouldn't work. My idea is that when you switch DC into the primary, the magnetic field grows at some rate. The changing magnetic field across the secondary induces a voltage. I think that in the short time that the square wave spends in one state, the magnetic field never gets to finish growing before changing directions. I believe this could be described as staying out of saturation for the core..

All that being said, a lot of inverters simply switch a primary coil at 12V through an H-bridge (allowing +/-12V) in a transformer with a 1:14.2 ratio (producing 170Vpeak).

Thanks prs09210 for posting your experiment. I came across this subject by googled it. I think your experiment result fits with the simulation picture that some one else had posted before. Here is what I think: depends on how long is the pulse on primary coil is and how much the Henry of the primary coil of the transformer is that we will have square wave or a short bust of voltage on the secondary coil. I.e. for a certain value of transformer, if the primary voltage stays high for long enough that when primary current reach near DC value then voltage on secondary start to fall, as a result, the voltage on secondary just looks like a burst of voltage compare to long width of the primary voltage, so the secondary is not a replica of the primary voltage. But imagine that if we turn off the primary voltage very soon (a very short time of high, during that time the primary current is linear) then the secondary voltage will stay at constant DC during this time, then falling at the same time as primary voltage; hence we have a replica of the primary on the secondary. The same thing can be achieved by using another transformer with much more turns on the primary coil instead of narrowing down the width of the square wave. The simulation picture posted by previous person showed this trend. So for transformers, the low frequency component is of more important I.e. any low frequency transformers can be used for RF transformer (if cost is not a factor), but the reverse is not true.

Please chime in.

Just a thought, to tidy up messy pulses, perhaps putting them into another transformer, might tidy them up a little bit more towards a sinewave?
Probably some sort of capacitance to help with the fall of wave so it is slower to fall rather than blocky, then putting that into the second transformer. As the rise and fall has a lot to do with the induction. And generally speaking, like in the old points system in a car, the bigger effect is when the magnetic field collapses creating the best spark, as there is no for want of a better word, "resistance" (not electrical), when the field collapses. But there is a push against the building up magnetic field on the rise voltage, which is an inductive reactance effect creating that resistance.

But if you want to get any current out of it, it goes without saying that the square wave will have to power a power transistor circuit.