Minimum detectable pulse width

[jrainey]

I am looking into using an EZFLORA to control my lawn sprinklers. I would like to use a rain sensor to help conserve water, by recording rainfall, and delaying watering based on the amount of rainfall sensed. The particular sensor I am considering is the Hydreon Optical Sensor. According to its specs, it can be programmed to generate a 50 millisecond pulse for a given amount of rainfall detected. I am thinking of using this in conjunction with an IOlinc, counting the pulses in the isy99, and adjust the watering schedule accordingly.

 

Two questions:

1) Has anyone used anything like this?

 

2) Can the IOLInc and ISY99 combo be depended on to detect every 50ms pulse without fail?

 

Thanks in advance

 

Jack Rainey

[ELA]

I tested one unit that met your requirement but since Smartlabs does not specify a minimum pulse width there are no guarantees.

 

See this thread:

http://forum.universal-devices.com/viewtopic.php?f=27&t=7994&p=62438&hilit=pulse+width#p62438

[jrainey]

Thanks for the info ELA.

 

I waited until Monday to call Smartlabs. The answer I got was that none of the senior technicians available had any idea as to what the minimum detectable pulse width is.

 

My comment on this is that in the area of control systems, these are engineering specifications that should be published. Hopefully this will be passed on to the engineering group.

[ELA]

Hello jrainey,

I have found Smartlabs to be seriously lacking in details / specifications in lots of areas.

Sometimes I wonder if it is not intentional. In this case I cannot see why.

I have gotten used to setting up test scenarios to determine these variables the hard way. Without having a large cross section of units to test you still do not end up with a true specification.

 

I read about the sensor you are using and it was interesting. I am assuming you using the tipping bucket scenario?

If the bucket size is large enough, and the the IOlinc you purchase missed some pulses you could consider adding a pulse stretcher circuit. Hopefully it would not be required.

 

 

Best of luck to you

[Brian H]

I never checked the minimum pulse width needed for an I/OLinc but know that the Sensor LED can glow and the Sensor Input is not tripped.

Seems the LED is part of the input electronics and I found the Sensor Input needed to be pulled below 1.0 volts for mine to trip. The Sensor LED can glow and the input not be pulled below 1.0 volts.

[ELA]

Thanks for that added input Brian,

If it has an opto-coupled, input with the status LED in series on either side of the coupler, that would be expected behavior when running the opto-coupler in its linear region ( using a resistor as opposed to a dry contact closure).

 

It should not affect the minimum pulse time unless the input is very slow rising.

 

When I performed my testing I switched the input with a steadily decreasing pulse width and after each pulse application I monitored to be sure Insteon commands were being generated.

[Brian H]

No opto coupler at all. It actually goes through series resistors, pull up resistors and an LED with current limiting resistor to +5 Volts from the Sensor Input to a pin on the controller IC.

[jrainey]

Thanks for the suggestion ELA.

 

I actually had to do exactly that in a previous job. That is what tweaked me to ask the question.

 

A conveyor scale emitted a 50ms pulse for each ton that passed over the scale. This was adequate to increment the electromechanical totalizer mounted in the control panel, but would not reliably increment a counter in a control system that had a 250ms scan time. Interim solution was install a pulse extender to extend the pulses to 300ms. Unfortunately, the device was only good for about 10 million cycles - not so good when production was on the order of 16 millions tons per year. In the end, the solution was to replace the scale with a more modern version that had a programmable pulse width.

 

So far, I am only researching the feasibility of what I would like to do, haven't bought anything as of yet. The main goal is to integrate the irrigation system into my home automation system. My current Rainbird controller does the job, but can't be integrated, and is nowhere near as flexible program wise as I would like it to be.

 

Another issue is that weatherbug is not very useful to me. The nearest weather station is at our airport, which is located on the river, about 1800 ft lower in elevation than my home.

 

So far, everything looks good if I can get on of the Hydreon sensors into Canada without a lot of hassle.

 

Will keep the board posted on my progress should I decide to go ahead.

 

Thanks

 

Jack Rainey

[bmercier]

jrainey,

 

I'm looking into doing the exact same thing. I was to try out a rain collector with a tipping bucket, until I saw this post.

 

This Hydreon Optical Sensor seems to be a very interesting solution. Have you purchased it? Does it work with the IOlink?

 

I'm not sure if I would hook it up to an IOlinc, or to an unused input of my ELK. I would guess that the ELK can detect a 50ms pulse. But I'm a bit worried that the Elk integration with ISY adds overhead and causes problems if the counting pace is too high, compared to an IOlinc. Anyone knows?

 

Thanks,

 

Benoit.

[bmercier]

I would guess that the ELK can detect a 50ms pulse.

 

Actually, I checked my Elk documentation, and yes, it could detect it. I have to set the zone for a "Fast Loop response", and configure this to lower than 50ms (Can be as low as 20ms).

 

But still, I'm not sure it's a good idea to use an Elk input for counting. I don't know how the integration works between the Elk and ISY, and there may be too much overhead. Any input on this appreciated.

 

Thanks,

 

Benoit.

[jrainey]

To those who may have be interested in this topic, I am pleased to say that I have successfully integrated a Hydreon sensor into my system. A little embarrassed though that I have waited so long to post. My apologies.

 

The system components include the Hydreon sensor an EzFlora, and 8-button KeypadLinc an IOLinc and a 24 volt power supply.

 

The Hydreon is set up to emit a pulse with each 0.2 mm of rainfall detected. The IOLinc detects each pulse and triggers a counter program. The pulse counter increments a millimeter counter by one on the fifth pulse, and resets to zero.

 

Every midnight, a counter is incremented by 1. When tis counter exceeds a preset value, an irrigation cycle is run, and a program is run to reset the day counter and rainfall counter to zero. The reset program is also run whenever the rainfall counter reaches 10 (or 1 cm), postponing the next irrigation cycle.

 

The Keypad linc provides a control interface. The first 6 buttons provide individual on/off control for each of the six zones in the system. The seventh button manually runs a cycle. The eighth button toggles the operating mode between auto and manual.

 

Installation was completed and tested late last summer. The system has been running well so far in this, the first full season of operation. Timing is great, since this is also the first season that my water usage will be metered and I will be charged on consumption.

 

So far we have received about 60 mm of rainfall, and the system has yet to run an irrigation cycle.

 

The sensor is working well, and, as a result, so is the system.

 

Thanks for the feedback.

 

Jack Rainey