Tunde.
I never got around to changing a few tracks, as it works so well as it was, I did not bother to change it.
The highside spwm heat sinks do get hotter than the low side for the reasons Ross eluded too, but not enough to get my attention.
So laziness won through. It is incredibly successful as it is.
Yes, the current limiter shuts down the 8010, and the 2110. I wanted instant shut down if it was invoked. Remember it is set up in the 12-14kw area on my units, so if it blows, I wanted it to happen very quickly, rather than rely on the 8010 software.
It is important you don't allow the 2110 to come back on on it's own for any reason, other than a soft start where it is already on before the software starts to turn on.
You don't want any circumstance where there can be an errant pulse width into a "cold" torroid. The back emf is incredibly high.
Yes I am a lazy slob. I did mean to change to spwm on the low sides only.
As Ross has eluded to, it makes sense to switch the 50hz square wave on the high side, as it makes the switching constraints easier to deal with.
The high side requires driving the fet gate-source with a voltage that will exceed the 48v bus.... by 12v or so. We need this, as when the high side fet is turned on, the drain and source are at the same potential, with RDS on in the milliohms level... but we still need the gate to be 12v above the source to stay turned on hard in the conduction range for the fet to stay in the switched range, rather than the linear part of the conduction curve.
The fet is a very high impedance input ( insulated gate) but with a rather decent capacitance. This means that we need only very small current to switch it on, if we don't need to do it quickly..... but the slower we switch, the more losses we suffer in the linear section of the curve. We want to get past this from 0 to 12v instantly if we can. This means charging the gate up very very fast, which necessarily means decent drive current, at a potential higher than the DC bus.... so driving it is not as simple as the low side driver, where the 12 signal is 30 or more volts less than the DC bus, and so easy to get .
If we are switching very quickly at high frequency, we have a lot of transitions to deal with , and our switching losses will come from here... so it makes sense to switch the high side at the lowest frequency, and the highest duty cycle... ie we switch once in a 100th of a second... and thats it, rather than once in a 20 thousandth of a second.. we will suffer many times the losses with the 20khz signal, and so it makes sense to do this at "normal"voltages on the low side of things.
I get the 12v above the bus from a voltage pump arrangement on the high side driver, and the less it has to do the better..
As it turned out, it works much better than I imagined, and so I have not bothered to do the simple changes that I should have done to keep the high side switching at the slow speeds.. So one high side switches at low fundamental frequency, and one at high spwm frequency.
You can change that on your design if you like..
......oztules