zero cross detection

[otrotabi]

I would like to use an M171/M172 to fire a SSR to implement phase angle speed control for a small fan. We have been using something similar for a long time and most of the external rotor fans available on the market can be driven this way to control fan speed. The SSR part of the circuit is something like this:

The controller we use has a microcontroller to do this job and I would like to replace it´s functionality with the PLC.

This article here explains how to do it and claims that it is possible. Rather than using the mains I could generate an input signal to the PLC based on a 220 to 24 V transformer which we normally use to power the PLC.

Has anyone tried something like this ? Any comments on what to look for in the setup will be appreciated.

[DVarnerMinn]

How much of a delay time can you afford from zero crossing, to detecting it in the PLC, to acting upon it?

Micro Processors are typically much faster to react to inputs depending on how it is programmed.

PLC's you may have delays due to filtering the input, PLC scan time, and updating the outputs.

You would need to detect the zero crossing and then use a Fast Trigger event in the PLC.

Does this help at all?

[otrotabi]

Thanks for your feedback. 

For our frequency (50 hz), a complete half cycle is 10 msec. From my experiences with microprocessors, if I fire the triac gate once the curve starts going down (lets´s say when RMS voltage is 50% or less) the motor may start to stall (depending on the fan we are using). I try to avoid RMS voltages less than 120V usually (as read on a standard digital meter, I have not tried reading the actual voltage or current curve with a scope).

So my complete window would be 5 msec from detecting the zero cross to firing the triac gate in a usefull way), prefererably less than that.

What do you think ? Would it be possible ? What is a fast trigger event ?

[DVarnerMinn · Accepted solution]

I work with many different PLCs so it may help with a few screen shots on how I would go about this.

In the Project window under tasks right click Timed and select select Task Configuration.

The Default Timed scan time is 100ms.

Change Set Period from No to Yes and enter a time period. Let's say 0.1ms for now.

Click OK and you see that the minimum task time is 20ms.

This means you might need to use the Low Priority Background task which means you might not hit zero crossing every time if we can get inputs and outputs to be fast enough.

It looks like the M172 PLCs have 2 Fast Digital Inputs DI1 and DI2.

I am assuming that you would use an M172 model with 2 SSR outputs. Will these work for you? I don't see more details listed in the specs after a quick look.

So, if you can get something to give you the zero crossing pulse, the inputs should be able to handle it.

I background task MIGHT be fast enough BUT since it is low priority there may be several things that can delay it. 

Timers have a resolution of 1ms. This means you would need another way to time your delays assuming everything else is fast enough.

My opinion right now is that you would need to find someone with more inner knowledge of the M172 than I have to determine if it is possible.

The M172 was most likely designed for applications that did not require such speeds.

There are other models of PLCs that might be better suited for this application but at a higher cost.

[otrotabi]

Thank you for your very detailed information. Yes, it looks like timing issues will be hard to overcome with this PLC. A higher cost PLC is not something we are considering. I figure we can modify our original microprocessor PCB to let the PLC handle the sensor reading and PID strategy and let the microprocessor handle zero cross and convert an analog output from the PLC to a delay to fire the triac.

Once again, I appreciate your very in-depth analysis. Regards.