SSR Timer

[RyanG]

A while ago I posted on here because I was working on a timed SSR based on a 555. I am trying to shrink the circuit so that I can fit it onto a board that fits inside the AC motors I use my original design on. The original ones are wired outside the motor. Does this look like it will work? I thought of it today, tested it with a lower voltage with some modified values, and it worked. I would just hook it up and try it but 9VAC is much safer than 120VAC. In the picture I only included the "timer" part of the circuit and optotriac. I did think about dropping the optotriac because isolation is not relevant in this, but that raises the current required through the transistor when the triac is off significantly. I'm sure I overlooked something stupid, so take it easy on me.

[Ian.M]

R3 will dissipate nearly 3W and R2 adds another 0.2W so when the motor's off with no airflow it will cook, and even if R3 has an adequate rating, the electrolytics life will be greatly shortened.

Look at capacitor dropper non-isolated PSUs to reduce the dissipation by an order of magnitude.

[JS]

  With a capacitive divider as Ian recommended you can even get the 555 still there, probably your circuit was too big due to a linear PSU, but you say you don't need insulation (still be safe, at least while testing) if you don't have a mains insulation transformer you could power your circuit from two back to back 9V transformers and get the mains insulation with somewhat improved safety. Then you can remove the insulation when you install the circuit outside the children's reach.

  As you don't need too much current you could use a capacitive divider to lower to 9V, from there rectify and regulate to say 5V and use the usual logic design. Other option is an off the shelf SMPS, small and cheap enough to save you quite some trouble.

JS

[RyanG]

Is the attached circuit better? Also attached is a picture of the board size it would go on. The only thing I would remove from that board in the end would be the long connector at the top left. The other connectors would be for the wires in the motor and the mains. I put the voltage regulator on there for scale. Also, all of those parts on the board except the holes would have to be in the same locations with my circuitry around it.

[Ian.M]

Probably, though the 560nF class X capacitor is likely to be inconveniently large if you have limited clearance for parts.   Have you simmed it in SPICE yet?

Personally I'd prefer something that switches with a 'snap' action - it may be worth looking at two transistor monostable circuits.   Also, another argument for two transistors is more gain would be preferable so you could use a much higher value resistor and smaller capacitor (both physically and in value) for the same time constant.

What's the actual application?  Is it replacing a centrifugal starter switch?
A photo of the housing that board goes into would be helpful to see what vertical space there is for parts.

[RyanG]

Yes, it replaces the centrifugal switch. My original idea with the 555 was to save the start windings of the motors if the overload switch failed. I have had that happen and I end up with an unusable motor with completely burned windings. The 555 shuts the start windings off whether the motor starts or not. There are other versions of a similar switch that say they use current sensing, which won't protect the winding if the rotor is locked and the overload switch fails. The overload fails because they are thermal breakers. I did simulate the circuit, and it seems to work. I will look at changing the circuit to use 2 transistors. I can get a picture of the housing when I go to work in a little bit.

[RyanG]

Actually found this on google image search. This is the board in the other motors we use, but the space and housing is the same.

[Ian.M]

That doesn't look bad at all.  There's enough space to fit the dropper cap against the back wall next to terminal 1 opposite the cable entry.   

To save PCB real-estate and increase the functionality, it may be worth considering using a MCU.  Done right, you'll be able to scrap all the timing components except the MCU and the MOC3043 to drive a sensitive gate TRIAC directly. The one I've got in mind is the PIC12HV1615, which has an on-chip 5V, 50mA shunt regulator so is capable of running from  a  rectified capacitive dropper with no further regulation.   Its got enough processing power to be able to do the timing, current sensing via its ADC and zero crossing detection via its comparator module, it has a PWM module so can burst fire the TRIAC to get around the limited gate drive current available from the cap dropper PSU, and is available in packages as small as 4x4mm DFN.   Code it in C using Microchip's XC8 compiler in FREE mode (which is actually licensed for free commercial use!),  and once you go to production, buy the PICs from MicrochipDirect, factory programmed with your code and drop-shipped to your assembler, so from a production viewpoint its effectively become a dedicated motor start controller IC.

If you want to put a TRIAC in series with the main winding as well, it would be possible to generate non-overlapping firing pulses on a second PWM output pin by setting up the PWM module in active high half-bridge mode wioth a nominal 50% duty cycle then using its dead time control to reduce them to a much narrower pulse width.  Once it had started, it would then transition to single output mode with a low duty cycle to keep the main winding TRIAC on while letting the starter winding TRIAC turn off.  If a locked rotor fault was detected, *BOTH* TRIACs could be cut off till the power has been cycled.

[NiHaoMike]

HVAC parts stores sell fairly cheap "hard start kits" that replace starting relays and capacitors. That can be a ready made option.

Many years ago, a friend of mine came up with a compressor starting circuit that used the voltage drop over a shunt resistor to activate a triac that connects the starting capacitor. The idea being that with the rotor stationary, the current draw will be very high. If the loss in the shunt resistor would be unreasonable for your motor, a current transformer can be used.

[RyanG]

I have thought of using an MCU before. I have some ATTinys here that I could use with a 5v zener. That doesn't take up much more space, and I already have them. With the MOC3043 only needing 5ma what would be the best value cap for the dropper? I think 220nf would be too low, but I am not completely sure. I have some 220nf, 330nf, and 470nf X2 caps I took off some other stuff. I don't need a TRIAC on the main winding, but could be something I could do in the future to completely protect the windings.

[Ian.M]

If you are pulse driving the TRIAC or Optotriac with a low enough duty cycle pulse train, you could probably run on as little as 1mA through the cap dropper.   220nf at 60Hz 100V low line voltage would give you over 7mA if full wave rectified or 3.5mA if half wave rectified in the Zener, with a second ordinary diode preventing the reservoir cap discharging back into the Zener on negative half cycles.

If you cant do a continuous train of >10mA pulses at lets say 31.5KHz with that, then you are either *VERY* new to ATtinys and AVRs in general, or aren't competent to code your way out of a paper bag!

[NiHaoMike]

Why overcomplicate it? The triac circuit my friend came up with is powered by the current through the main winding. If the current is high enough to signal a stopped motor, the voltage drop across the shunt triggers the triac.

[RyanG]

I want to stay away from current sense circuits because it doesn't protect the start winding if the motor fails to start. If a mechanical problem happens and the rotor cannot spin then using current sense would make the motor continuously draw massive amounts of current. If the thermal overload doesn't trip then it results in a fried start winding. I have had this happen, and it is the reason for me doing the 555 circuit before and wanting to do it with some sort of timer in general.

I'm sure I can code the ATtinys to do whatever. I am significantly better at programming than I am at electronics.

[NiHaoMike]

If you just want a timed start (HVAC techs generally hate those!), get a PTC motor starter.