Just went down and measured the primary inductance of a 600W toroidal transformer, unloaded, which was 0.63H. With your formula XL = 2*Pi*F*L that gives an impedance of around 210 ohms. So, it appears that even at twice the nominal current, on my 230V mains, we're looking at a current of no more than 2 Amps.
I also verified that when you disconnect the transformer from the rectifiers, there is no switch-on thump or any other noise. However, when you reconnect the rectifiers, there definitely is a switch-on 'groan'. These findings are fully in line with my measurements I posted earlier.
jd
I also verified that when you disconnect the transformer from the rectifiers, there is no switch-on thump or any other noise. However, when you reconnect the rectifiers, there definitely is a switch-on 'groan'. These findings are fully in line with my measurements I posted earlier.
jd
I'd expect magnetizing current to be even lower than that. For example, Rod Elliot measured a 500VA transformer on 240V 50Hz mains here
Mains DC and Transformers
and found it to have no more than 16mA RMS magnetizing current as long as there was no DC on the mains and that's about what I'd expect.
Rectifier conduction seems to be causing some confusion here. Output current does not affect the flux in the core, at least not directly, because the magnetic fields from load current in the secondary and the "reflected" current in the primary cancel! 😱
Flux is at its peak at the voltage zero crossing because
e = Npri * dΦ/dt (Faraday's law of induction)
where e = upri - ipri*Rpri with the standard sign convention.
Mains DC and Transformers
and found it to have no more than 16mA RMS magnetizing current as long as there was no DC on the mains and that's about what I'd expect.
Rectifier conduction seems to be causing some confusion here. Output current does not affect the flux in the core, at least not directly, because the magnetic fields from load current in the secondary and the "reflected" current in the primary cancel! 😱
Flux is at its peak at the voltage zero crossing because
e = Npri * dΦ/dt (Faraday's law of induction)
where e = upri - ipri*Rpri with the standard sign convention.
OK, but is my reasoning correct? Calculating the impedance (with an unloaded xformer) and using that to calculate the max inrush current?
jd
No, it is not. In the link in post #22 there are several interesting documents. Some of them are only in German, but you can probably unnderstand that language. Especially the pdf under 'Transformer Phsyic rules' should be interesting for your quest, because it explains very well what happens at transformer start.
The BH loop is almost square for a typical toroid transformer core so it doesn't look too much like an ideal inductor at all. 🙂 If you look at that Rod Elliot page you see that the magnetizing current looks more like a square wave than a sine! 😱 Not exactly what you'd expect from an inductor...
The maximum inrush current is only limited by mains impedance, transformer primary resistance and that there is still a little inductance after saturation. It can get pretty high!
The maximum inrush current is only limited by mains impedance, transformer primary resistance and that there is still a little inductance after saturation. It can get pretty high!
Yes I've read that paper. But why don't I measure it then?? I measured the current across a 1 ohms series resistor. I briefly checked with 0.1 ohms to see if the resistor changed the rules, but the result with 0.1 ohms was the same, only more noise so I went back to 1 ohms.
I understand the physics, read the papers, but reality (as measured) doesn't comply. Frustrating!
jd
What is it that does not seem to comply?
In the measurements of post #32 it looks like there is no transformer-created inrush. Also, magnetizing current will be invisible with that current scale. 🙂
In the measurements of post #32 it looks like there is no transformer-created inrush. Also, magnetizing current will be invisible with that current scale. 🙂
Well, for one thing, that tiny inrush current will not cause the switch-on 'groan'. The load will.
Why should we worry about the transformer inrush due to remanent magnitism when I don't see any?
jd
Because there is enough evidence that transformers blow fuses when they are switched on whether loaded with capacitors or not. Toroids are worse than EI core transformers with air-gap. There is also sufficient evidence for that. The fuses don't blow each and every time though, only when the remanent magnetism's polarity coincides with the voltage polarity at power on.
During your tests did you make sure that the transformer was switched on at the same polarity as the remanent magnetism to see the worst case inrush current? You could use the method described in that paper in reverse. Magnetize the transformer with DC pulses in one direction and then switch the power on with the same polarity.
Did you try what happens when you don't switch off at zero crossing?
What transformer did you use? Did you try a different transformer?
What SSR did you use? Some SSRs come with built-in current-limiters.
That may be the reason, I always switched off at zero crossing. But at any rate, any triac would always switch off at zero crossing, independent of when the gate drive is removed, right?
I'll see if I can do the test you described.
jd
The magnetizing current of a toroidal transformer may not be high enough to keep a typical triac conducting. Because of that continuous triggering (DC or picket fence) is often needed.
By the way, Bryston uses triac soft start in some of their models. Schematics are available online and might provide inspiration to those who are interested in making a triac soft start.
By the way, Bryston uses triac soft start in some of their models. Schematics are available online and might provide inspiration to those who are interested in making a triac soft start.
Last edited:
- Status
- Not open for further replies.