Transformer soft start with Triac and uC

First of all, you need to start at the peak or falling of the peak wave form. Personally, I would sync lock a few cycles to make sure we have a clean wave form and not static from contact closure.. This should always be running in the Uc so if you do receive noise interruptions on the line, you can at least allow it to keep using the same timing cycle until your verification loop expires, in which case you should stop any further triggers on the Triacs. In my last circuit i used 12 cycles before I determine I had a good signal. This has to match your line system. 50hz = 10ms, 8.3ms for US. etc.

Have 2 timing algorithms operating, one to sync lock with the service coming in, and the second to phase delay ~ 7 ms before sending a trigger to the triac. This assumes you're on a 50hz system..

You have a counter in there where you can decrement the delay shift back to lets say 1 or 2 ms to then get the primary to turn on at just above the base or maybe keep it on?

Remember to turn off the output when ever you see the input reference drop while soft starting, other wise, you'll be kicking in the next cycle before you can complete the soft start incase you want to have a ramp up effect.

I suppose you can soft start on a single cycle once you have confirmed you have proper line reference and are locked in but it might be a plus to allow the connected load on the secondary side to see a few cycles come up soft. This may not be desired and that is your decision. That being the case, then sync lock the line for a solid reference then perform a initial ~ 7ms pulse. (that's 7 ms after the base of a reference coming in gets at least 5% or more in level.

Using a bridge type circuit into a compression scheme to get a rather square wave from your line voltage should give you what you need as a reference for you to work with. Make sure you pass this through a resistor or optical coupling so you don't get a phase shift issue, other wise, you'll have more problems to content with.

P.S. Yes, I've made things like this before with lots of power delivery on 3 phase systems.

Thanks to you all(Phil A. too),for spending time helping me describing detailed experiences,or even just for reading my post. I read carefully every answer . If i would solve i will post the adopted solution,hoping it could be of some interest to other newbies as i am; Please forgive my English.

Have a good Sunday.

Diego Milan,Italy

There are special NTCs for this purpose.

"blisca" Inscribed thus:

Yes that is a problem with them, they take a while to cool down before reuse.

Would a saturatable reactor in series work. DC from the rectified transformer output could be used to control the ramp up.

We had that problem with ultrasound machines, back in the days when digital circuits were extreme guzzlers. NTCs didn't work in situations where power could flicker. Nothing bad happened in our machines but breakers popped and in a hospital that's a real nuisance.

You have to add a hold-off circuit that, for example, says "enough already" after the x-th start attempt and puts in a pause interval so nothing gets hot. If you want to get by with the smallest possible resistor you need to set the number of starts per interval to one but that may not work well in areas with shaky power distribution, like this one possibly is:

from

Around 1980, during my disco music "career", I was designing boom boxes, and all kinds of dancing and dimming lights (frequency and amplitude sensitive).

To dim neon, I used integer-cycle switching, plus a DC safety sensor. Never croaked a neon transformer :-) ...Jim Thompson

What he needs is something simple like this.

Jamie.

The trend these days when driving transformers is to go reverse phase control. In which the output is turned on immediately and off later in the cycle. This requires a true linear switching device, rather than a triac or scr.

Advantages are ridiculously less RFI (you switch slowly...) and better control of transformers and fluorescents.

LSI/CSI has parts.

I suppose that means an IGBT or power MOSFET bridge. That would be an option, and the device could probably be fairly small if used in conjunction with an electromechanical relay as I saw in some "hybrid" configurations on the Crydom website. The problem with solid state controls is the pesky forward diode drop or On resistance that make for rather high power dissipation during extended power on conditions. Electromechanical contactors for high voltage, high current, and inductive loads must often be made larger than required for just current carrying capacity and breakdown voltage because of switching transients and arc control. If you can make initial connection to power and final disconnection by means of a solid state control, the contactor need only handle the current and be able to provide safe voltage breakdown for open contacts.

Motor control is now moving toward inexpensive small three phase induction types driven by cheap PWM modules, even for items as small as food processors and blenders, and especially for things like washing machines.

Paul