I will need to study this paper.
I jumped to the conclusion and there I seem to understand that the unknown remanent flux swamps the switching angle minimising of the start up peak.
If one does not know the remanent flux state then my conclusion is that changing the switching angle to a predetermined fixed angle does not achieve a consistent minimising of the peak start up current. In other words don't bother, just let the bypassable added resistance reduce the peak current to an acceptable value.
the decision to use a soft start circuit will depend on:
1, size and type of your power traffo...
2, total capacity of your main filter caps...
3, current interrupting capacity of your mains power switch...
inrush is not by traffo alone but in combination with your filter bank,
You do not need to know the premagnetization state of the toroid.
One idea is to bring the toroid back to a controlled and therefore known premagnetized state with some small pulses before full swicthing on.
your description implies to me that you do end up knowing the remanent flux to allow the angle adjustment to become effective.
Or in other words:
you do need to know the remanent flux.
Last edited:
Yes - for every transformer type ...
Have a look at following example http://www.emeko.eu/uploads/media/HWA-Beschreibung-e.pdf page 4:
uses 5 - 6 pulses starting at e.g. 130 degree angle and afterwards full on with positive half wave. The start of pulse at e.g. 130 degree and the switch off at e.g. 200 degree has to be tested/set for every transformer type (EI, toroid ...) and size.
To be accurate, one still does not know the remnant flux. It depends on the load at the instant that the transformer is switched off.
What the EMEKO procedure does is determine how their circuit will deal with an unknown remnant flux, dependant on the transformer parameters.
Best wishes
David
Really?
I asked if this was measured because inductance seemed low.
1.25 H is only 392j ohms at 50 Hz.
So 0.56 A at 230 V, or 135 VA.
This seems excessive, does your transformer really draw this much under no load?
Best wishes
David
I asked if this was measured because inductance seemed low.
1.25 H is only 392j ohms at 50 Hz.
So 0.56 A at 230 V, or 135 VA.
This seems excessive, does your transformer really draw this much under no load?
Best wishes
David
For a toroid the only thing that will limit the inrush current
at power on is its DC impedance wich is on the order of a few
ohms for a 500-600VA transformer , hence why i used a 50R
limiting resistor that allow for 6.2A maximum with a 230V main.
Through a few tests it appeared as a better option than using
values often seen of 10-33R wich will be more fragile at same
power capability given the huge peak currents.
at power on is its DC impedance wich is on the order of a few
ohms for a 500-600VA transformer , hence why i used a 50R
limiting resistor that allow for 6.2A maximum with a 230V main.
Through a few tests it appeared as a better option than using
values often seen of 10-33R wich will be more fragile at same
power capability given the huge peak currents.
4KVA / 117V = 34A. Ordinary split-phase house wiring might not like that much juice through 12ga Romex..... Click a breaker??
I have used the photo-triac/triac combo to a series resistor with a relay to short when the slow start time has passed for a 1.4KVA toroid. I even used an auxillary winding I added on the toroid to power the relay coil.🙂 I plan to use this method again but try the solid state relay that Bonsai uses instead of the mechanical relay.
Looking at that app note gives a better understanding of why switching a torroid on at the zero voltage crossing is a bad thing because at that point the magnetizing current is minimal and the primary current peaks as a result. You gave to switch the primary at the max rate of change of B.
For big transformers, a triac is a viable way to go IMV- use a micro to switch the triac on at vmax and then off at vzero crossing. Do that a few times to charge the cap bank up before finally turning on for good.
For up to 2KW, I have had no probs with using resistive in rush limiting ( see my e-amp write up)
For big transformers, a triac is a viable way to go IMV- use a micro to switch the triac on at vmax and then off at vzero crossing. Do that a few times to charge the cap bank up before finally turning on for good.
For up to 2KW, I have had no probs with using resistive in rush limiting ( see my e-amp write up)
Read this interesting post from another forum. The poster notes that the best firing angle seems to be circa 70 degrees and not the 90 I alluded to above. Also note the comments about de magnetizing the core. Seems quite a bit more complex than I initially assumed. Would make a good project to research this up and write some C code.
Charging up the cap bank also adds to the issues to deal with as does the required snubbering . . . Mmmmm
Re: Transformer soft start with Triac and uC
Charging up the cap bank also adds to the issues to deal with as does the required snubbering . . . Mmmmm
Re: Transformer soft start with Triac and uC
I like the idea of using a triac to short out a NTC or a resistor. No real contacts to deal with and none of the phase angle, snubbering etc probs mentioned in the link above.
What app note?
This doesn't make sense to me, is it a typo?
Bad idea for a toroid. Potentially major inductive inrush, dependant on switch off conditions. Of course the resistor bank can be sized to handle it but it's a clumsy solution, reminds me of rheostat lamp dimmers. Smarter to adjust the phase of the switch-on point and use the transformer's own primary inductance to limit the current. This works nicely as the transformer VA increases whereas the resistor bank becomes more and more unwieldy.
Best wishes
David
Seems you learned of some of the problems as I typed.😉
Thanks for the link.
The OP in had a reasonable idea but messed it up.
Need to switch over to full power synchronously.
The 70 degrees recommendation in response is only of much use for that particular transformer.
Can still be excessive.
Last edited:
If you use a TRIAC cleverly then you shouldn't need resistors/NTC. That is the whole point of this thread.
But TRIAC is non-linear near zero volts. I asked above about possible EMI from this, you have any experience or data? Otherwise still need nasty contacts to short TRIAC after stabilization.
Best wishes
David
Last edited:
... of course not. The measured data can not be correct as the core will be re/demagnetized during measurement and the measurement frequency was 120Hz. (My Peaktech 2165 USB has 2 possible measurement frequency settings: 120Hz and 1kHz).
Have measured again two different toroid transformers:
230V 650VA / 2 x 48V 6.77A
1.2H / 470R @ 120Hz
2 x 85mH / 39R @120Hz
DC pri: 1.5R / sec: 2 x 0.16R
5200g
1.2H / 470R @ 120Hz
2 x 85mH / 39R @120Hz
DC pri: 1.5R / sec: 2 x 0.16R
5200g
230V 500VA / 2 x 50V 5A
pri: 3.5H / 1.5k @ 120Hz
sec: 2 x 400mH / 240R @ 120Hz
DC pri 2.5R / sec: 2 x 0.25R
3900g
Other coils kept open during measurement. These values differ depending on previous magnetization and differ if one measures them at 50/60Hz. pri: 3.5H / 1.5k @ 120Hz
sec: 2 x 400mH / 240R @ 120Hz
DC pri 2.5R / sec: 2 x 0.25R
3900g
I just used the measured values for a quick simulation test as I found no data tables about toroids inductive values.
Didn't think so😉 Just like to be polite.
Best way to measure would be just to run on 230V no-load and measure the current, wouldn't it?
Best wishes
David
The voltage and current are 90 degrees out of phase remember. When Vi is at max, current is passing through zero. So, switch on at Vmax is not a bad idea at all.
If you use a TRIAC cleverly then you shouldn't need resistors/NTC. That is the whole point of this thread.
But TRIAC is non-linear near zero volts. I asked above about possible EMI from this, you have any experience or data? Otherwise still need nasty contacts to short TRIAC after stabilization.
Best wishes
David
The app note posted a bit earlier in this thread makes it clear that this is no trifling task, and especially so when you have a cap bank attached to the secondary - hence my suggestion that an NTC or resistor that is subsequently shorted out is a sensible solution.
If you use a TRIAC cleverly then you shouldn't need resistors/NTC. That is the whole point of this thread.
But TRIAC is non-linear near zero volts. I asked above about possible EMI from this, you have any experience or data? Otherwise still need nasty contacts to short TRIAC after stabilization.
Best wishes
David
Burst fire the triac with HF from a gated oscillator set to be enabled after the required delay 😀
I claim zero credit for the idea... Eva demonstrated it about 6 or 7 years ago with what I thought was a truly inspired and simple circuit. Getting it to work in simulation was a right royal $*"!+&^$% though. I built a low voltage (as in running off a 30 volt transformer) breadboarded circuit feeding a dummy load resistor and it worked well. Also tried using a CMOS oscillator, and while that worked I just couldn't get it to run in LTspice... thats my fault though.
Attachments
- Status
- Not open for further replies.