Electronics > Projects, Designs, and Technical Stuff
AC Active Soft Starter for Inductive Load with less energy dissipation
<< < (4/7) > >>
BravoV:

--- Quote from: duak on December 31, 2018, 09:42:34 pm ---BravoV, just thinking out loud about the snubbers.  I don't know if they are needed but here are the situations they could be helpful in.
--- End quote ---

Really appreciate your response and thought here, thank you.  :-+



--- Quote from: duak on December 31, 2018, 09:42:34 pm ---Case 1: the triac stops conducting when its anode current drops below the holding current (a few mA?).  I expect this to happen some degrees after the zero voltage crossing since current lags voltage in an inductor and it will tend to keep the triac conducting.  Since there is little current flowing thru the transformer & relay coil, there should be a minimal inductive kick voltage generated.
--- End quote ---

At the on state (switched ON), once the relay is fired, will the triac off instantly ?




--- Quote from: duak on December 31, 2018, 09:42:34 pm ---Case 2: the line switch opens at a significant transformer current.  Initially, the inductive stored energy starts and maintains an arc between the switch contacts.  At some points the current either drops enough or the gap opens enough to cool the arc and the switch opens.  Snubber B has been absorbing some of the energy but now it has to absorb it all.  Snubber A doesn't do much because either the triac is on or the relay is.
--- End quote ---

So the arc on the relay contacts will still happen even there is a U3 MOV (390V) across it ?




--- Quote from: duak on December 31, 2018, 09:42:34 pm ---Case 3: a differential voltage spike comes down the line between the conductors when the triac is off.  Snubber A handles some of the current and, if it happened at the right time, the triac could be triggered.
--- End quote ---

Ok, I'm abit lost, so there is a chance that this circuit won't turn off ?



--- Quote from: duak on December 31, 2018, 09:42:34 pm ---Will you breadboard this circuit vs go to a PCB right away?  If the latter, there is a small chance that the relay will not pick up in time and it will chatter.  This isn't good for AC relays especially because their coils can burn out.  Good thing is, AC relays pick up quite quickly.  If it doesn't you might have to make provisions for a DC relay driven from a rectifier and capacitor.

Best wishes for a Safe & Happy 2019
--- End quote ---

Happy new year 2019 to you too.

I haven't decide it, cause I haven't receive the relay, but for final it will be on PCB, but for early prototyping, I'm thinking of dead bug style (yes, I'm aware this is dangerous voltage), it will be build with alot of precautions and shielded (shrink wrapped) at exposed metal parts.

Again, thanks.
duak:
G'day BravoV, some answers for you

Q: At the on state (switched ON), once the relay is fired, will the triac off instantly ?

  Yes, when the relay contacts close they take the current from the triac and it goes non-conducting.  If the contacts open with current flowing thru them, say during a bounce, an arc will be struck between them.  The voltage across the arc shouldn't be more than 20 V so the triac is not likely to be retriggered and even if it is triggered, the contacts will close again shortly.

Q: So the arc on the relay contacts will still happen even there is a U3 MOV (390V) across it ?

  I was referring to the switch contacts opening, not the relay contacts as the relay contacts should never open with current thru them.  The voltage needed to strike an arc between contacts that are opening is on the order of a few volts and a MOV is practically useless.  The voltage is so low because when contacts are opening electrons can jump the gap when it's still much less fractions of an mm.  If enough current is flowing the electrons vaporize the contact metals making them conductive ions and forming an arc.    An RC snubber is much better because if sized properly it can carry some of the current and extinguish the arc.

Q3: Ok, I'm abit lost, so there is a chance that this circuit won't turn off ?

 No, it will shut off because the input is AC and the instantaneous current will sooner or later drop to zero after the mains switch is opened.  The the arcs extinguish and the voltage isn't high enough to re-establish them.  Good thing this isn't a DC system where arc can and does go on, and on, and on...

 I'm referring to the case where the switch happens to be on but the triac hasn't yet been triggered.  If a voltage spike comes down the line at just the right time, the triac could be triggered too soon and orginal problem occurs, ie. the transformer core saturates and the fuse blows or the breaker trips.   This shouldn't happen often but since you are using this in an electrical lab, you could very well have voltage spikes on the mains that a snubber could help with.

  If you find this still happens, then a slightly more complicated trigger circuit will be needed.   I would exchange D1 & R1 then add provisions for a capacitor between their junction and the cathode of U1 (right hand side).  This will form an RC network to attenuate any spikes from the line.  I'd probably remove snubber A.   R1 would then be something like 470, 1/2 W and the new capacitor 100n, 630 V.   I'd also move U3's right hand terminal to the net connecting pin 4 of the power switch. 

Hope this is a bit clearer,

Take care,
MrAl:

--- Quote from: BravoV on December 23, 2018, 07:55:03 am ---Aware of various AC soft starter designs, but most are using resistance at the initial inrush moment and then later by passed by relay switch, or even much more simpler design using NTC to limit the inrush.

Discovered this design -> Active Softstart for Welders



The designer claimed its suitable for inductive load, in this case it was designed for welder.

Quote :

Working principle
It exploits the fact that current in transformer is 90° out of phase of the voltage, which leads to conclusion that best time to switch on the transformer is in voltage peak of mains power line (which was quite counter-intuitive for me at the first time). This circuit detects the peak using 300V bi-directional transil (TVS diode) which is then used to trigger snubberless TRIAC which switches on the power transformer of welder. Once this happens, there is normally open relay with AC coil connected in parallel to transformer which shorts out whole active circuit and therefore prevents any excesive heating caused by high currents through semiconductor components (= No heatsinking needed!). To make circuit more robust and endurable there are two more components to prevent possible damage to semiconductor elements. In series with TVS there is resistor to limit current through TVS and to TRIAC's gate. Also there is 390V varistor across the whole switching circuit which prevents voltage spikes from transformer to kill the TRIAC and partially helps to supress arcing in relay contacts.


What interest me are these features :

- Much less dissipation, especially for load that getting turned on an off often in short period like a drill.
- Less delay or instantaneous power delivery at power on, cmiiw.


Question, is this design decent ?

- Do I need to adjust the 300V TVS (Bi-Diretional) for the triggering, to somewhat lower abit as my mains is 220V while its designed for 230V ?
- What is a good enough wattage for the R1 resistor ?

Appreciate any comment or feedback regarding this circuit.

TIA

--- End quote ---

Hello again,

I took a look at the "peak turn on" feature and it looks like it works.  Here's why.

If we normalize the time functions for everything except time for both possible ways of doing it (turn on at peak vs turn on at zero) we get two different functions:
1.  [turn on at peak] sin(t)
2.  [turn on at zero] 1-cos(t)

They have interesting characteristics.
#1 starts at zero then ramps up to some peak value which here is 1.  So it's a regular sine current.
#2 also starts at zero, but then ramps up to a higher peak value which here is 2.  It too is a sine wave, but now has an offset of 1 so it actually peaks at 2.  That's twice as high as #1.

Now in a real circuit with some small resistance as well as the inductance, #2 will eventually damp out to a regular sinusoid.  The speed at which this happens will depend on the value of both the resistance and the inductance and may actually limit that '2' to a lower value, but it will most likely be higher than #1 anyway.  This means that it makes sense to switch at the peak rather than at zero.

Of course this is quite easy to test too.  If you change the circuit a little you can get it to switch at zero, but beware the peak could go as high as 2 times normal if the exponential part of the response is slow.

I think that is what you were most interested in but if you have any questions no problem.

BravoV:

--- Quote from: duak on January 02, 2019, 08:39:03 pm ---Q: At the on state (switched ON), once the relay is fired, will the triac off instantly ?

  Yes, when the relay contacts close they take the current from the triac and it goes non-conducting.  If the contacts open with current flowing thru them, say during a bounce, an arc will be struck between them.  The voltage across the arc shouldn't be more than 20 V so the triac is not likely to be retriggered and even if it is triggered, the contacts will close again shortly.
--- End quote ---

So this is no different than ordinary contacts bounce right ?



--- Quote from: duak on January 02, 2019, 08:39:03 pm ---An RC snubber is much better because if sized properly it can carry some of the current and extinguish the arc.
--- End quote ---

Just assume this will be used only at the transformer I mentioned before, appreciate any help on finding the right RC values for the snubber, as hobbyist, tried read many documents from the net and I'm still lost & confused.  :'(



--- Quote from: duak on January 02, 2019, 08:39:03 pm ---If you find this still happens, then a slightly more complicated trigger circuit will be needed.   I would exchange D1 & R1 then add provisions for a capacitor between their junction and the cathode of U1 (right hand side).  This will form an RC network to attenuate any spikes from the line.  I'd probably remove snubber A.   R1 would then be something like 470, 1/2 W and the new capacitor 100n, 630 V.   I'd also move U3's right hand terminal to the net connecting pin 4 of the power switch.
--- End quote ---

Point me any error if any, sketched your propose change below.

Once all components gathered, I will start prototyping the original circuit 1st, and then try yours too as its easy to change and see the results.

Thanks for your assistance.

BravoV:

--- Quote from: MrAl on January 02, 2019, 11:09:14 pm ---
...

Of course this is quite easy to test too.  If you change the circuit a little you can get it to switch at zero, but beware the peak could go as high as 2 times normal if the exponential part of the response is slow.

I think that is what you were most interested in but if you have any questions no problem.

--- End quote ---

Hi MrAl, thanks for the reply.

What kind of "little change" are you talking about ? Modification for resistance type of load ?

I'm curious, and yes, definitely I will ask more.  :P

Navigation
Message Index
Next page
Previous page
There was an error while thanking
Thanking...

Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod