Inrush current/"Bounce back" voltage 120/12v landscape transformer

[gar]

150326-1742 EDT

grasfulls:

Your package arrived today. I have done a limited test on one of the Heath units.

This model uses a phase shift type of output switch. It can provide two lamp light levels with an incandescent bulb. Probably no other type than incandescent should be used. Full brightness to a 100 bulb has about 16 to 17 V drop across the output solid-state switch. If I measure bulb voltage with 123 V source the bulb reads about 107 V, and in dimmed mode about 42 V. There is probably no damage to this one sensor. I will do more tests later.

There are numerous reasons why this motion sensor probably failed in your application.

.

 

[grasfulls]

Failures

Failures

Gar:
I was not expecting you to test them but that is great!
Yes, I saw the setting but it allowed for it to be set at full bright versus dual level, so I did that. I can totally see that all of its functionality is probably through no real relay, but solid state....so yes, I learned a lesson there. Full bright is not via a contact.
Have fun with that massive bags of capacitors... Well, massive to me anyway.

 

[gar]

150326-2035 EDT

grasfulls:

I am interested in trying find out why you had a problem. Since I have not tried the other sensors yet I don't know whether some may have shorted, or if the problem has to do with an inductive load on the SSR.

.

 

[grasfulls]

make one that works

make one that works

gar:

Given the leaning toward LEDS...maybe you could make a bulletproof PC/MS.

 

[gar]

150327-1110 EDT

grasfulls:

What is a PC/MS?

All the following relates to a working Heath/Zenith motion sensor with an SSR output relay.

A correction to a statement I made yesterday.

Full brightness to a 100 bulb has about 16 to 17 V drop across the output solid-state switch. If I measure bulb voltage with 123 V source the bulb reads about 107 V, and in dimmed mode about 42 V.

I don't really have "a 16 to 17 V drop", but rather --- an apparent 16 to 17 V drop --- across the switch. What occurs is the full instantaneous source voltage drops across the Triac until the Triac is triggered, then the drop is about 1 V. On average it looks like about said 16 to 17 V.

I have run another experiment with a 20,000 load resistor instead of the 100 W bulb. Now the voltage is non-zero, but not large. This tells me two things. A 20,000 ohm load at the time of trigger is insufficient to exceed the Triac holding current, and that the Triac is being triggered with a short pulse, and not a sustained signal to the Triac gate.

Next I placed an unloaded transformer on the motion sensor. The SSR functioned with this load, but it is clear there is a DC component to the Triac output because of the unbalance between the positive and negative half cycles. This I detected without using a DC meter by the transformer sound. After application of power to the transformer its audio hum gradually increased indicating increased core saturation. This test should not be run very long.

More later.

.

 

[gar]

150328-1955 EDT

Continuing with the same Heath/Zenith unit.

I previously mentioned that one needs to know and understand how sources and loads work individually and together. This is well illustrated in the application of this motion sensor.

The output switch of the sensor is essentially a phase shift dimmer in both its high brightness mode and its more dimmed mode. Further turn-on for each half cycle is by a short pulse to the Triac gate. This means the load has to draw sufficient current at the time of the trigger pulse, and for the remainder of the half cycle to exceed the holding current of the Triac. Otherwise there is only conduction to the load during the duration of the trigger pulse.

A transformer as the load requires that any DC current component in the input voltage be very small to prevent excessive transformer core saturation. Extremely important is that any dimmer on the input side of a transformer produce very little DC current. Unless a phase shift dimmer has been designed to work on the input side of a transformer, then it must not be used for this type of application. Thus, if a dimmer does not state that it is designed for this purpose, then assume that it should not be used.

A unloaded transformer I tested with direct 120 AC supply, no motion sensor, read about 0.25 A AC. This same transformer when connected to the Heath output quickly went to over 10 A no load input current. This was because of the DC component of the voltage. This quickly burned out the 1 ohm 1W carbon resistor I was using for current measurement, 100 W or more into a 1 W resistor.

Next I tried my small P&B AC relay as the load. Alone it won't pull-in. Does not produce sufficient holding current. Next added a parallel 0.1 ufd capacitor. Bad result, some hum, no pull-in.

Using a 5 k ohm resistor in parallel with the relay coil the relay pulled in, and relay was quiet, but voltage was not good. Changing to 2.5 k and voltage was about 116 V from a 123 V source. The 2.5 k provided enough holding current for both half cycles, and damping of the inductive current lag.

.

 

[gar]

150328-2206 EDT

grasfulls:

None of your four motion sensors show a short between black (hot input) and red (hot output) using the diode/continuity position on an older Fluke 27.

I have connected two of the motion sensors in parallel, and loaded with one 100 W incandescent or one 10 W Cree 120 V LED bulb. The "OR" functionality of this circuit worked fine. Nothing stayed on continuously.

I see no explanation of why you had the continuous on problem. With the transformer load it is possible there was overheating in of the TRiac from excessive current from transformer saturation, but neither of the two sensors that I have now powered show any sign of damage. I wonder why there was no transformer damage. Could show up later. Would depend upon what was the input current.

With knowledge of how the sensor output switch works it is possible to judge the kind of load circuit that is possible to use with the sensor.

I believe with this type of sensor that a shunt resistor across an AC relay coil is a good way to get a contact type relay output to handle high power loads, transformers, and other types of loads.

.

 

[grasfulls]

No Problems

No Problems

Gar:
wow...

I am not sure we had an always on problem. We first had one that stopped turning on the transformer (about 8 feet away). Then I tried two others, they too would not turn it on, then the one about 100 feet away failed. Then I got the relay, it chattered at a very high rate with multiple sensors, capacitor nor not. Then the RAB sensors worked great, I just left the capacitor in place.

I think you can see an electrician may never have the ability or the wherewithal to do an analysis such as you have performed. You have also seen that some just want the easy way out.

It would be great if manufacturers would provide a sheet that shows loads, to include transformers with secondary loads, and then tell us what sensor they have that WILL work, or properly tell us what we can do to make one work.

I would love to be able to take items, figure out what one can do to use them and then proceed. The tech support people I spoke with had no idea they could clearly relate, or perhaps they had no good ideas so poorly told me whatever sounded good.

I will take the advice "If it is not listed as a controlled item, do not use it for that". I like using a relay ad separating the actual load from the sensor, I am hoping that actually makes the system a little more robust. The client is happy, you have a hodge-podge of sensors (I found one more I threw in the van) and .1uf capacitors, and I have learned a lot. Thank you for the relentless pursuit of answers. I AM going to read up on triacs and their gate, AND the DC component...I do not even want to ask about that yet, I want to try and find some information first.

 

[gar]

150329-0838 EDT

grasfulls:

At one point you had mentioned something about a partially on state. That apparently led me to think that a failure state was that circuit did not turn off. My error in making that interpretation.

The manufacturers need to provide adequate information on how their product works and how it can be applied. This also means that technical support needs to truly understand how their product works and how it can be applied.

In the Heath/Zenith instructions there are two statements of significance, but not adequate.
1. Not for use with fluorescent lights. But no other NOTs.
2. Electrical load --- up to 500 W maximum incandescent.

There is no indication that the output switch is a solid-state type vs a mechanical contact type, and/or a discussion of the significance of each type. Your RAB sensors have a mechanical contact output.

I don't believe your external relay with the added parallel 0.1 ufd capacitor would have worked on the Heath/Zenith sensor, but probably would have worked with a parallel 2500 ohm 10 W resistor (actual dissipation about 6 W). I believe the reason for the 0.1 ufd capacitor with the RAB sensors is to snub transient voltage when the relay is turned off.

Relay lifetime, typical might be:

Mechanical --- 1,000,000 no load, 100,000 full load.
Solid-state -- Very many cycles at full load. Multi-billions.

At 50 times per day 100,000 cycles occurs at about 2000 days, or 5.5 years.

A mechanical contact relay is going to work well with many different loads without complication.

.

 

[grasfulls]

yuck on lifetime

yuck on lifetime

gar:
The remote relay or the relay withing the sensor? Or both?

Seeing the what i perceive to be as a very inadequate life compared to solid state, making a solid state device function properly seems best. However, if I still need a relay due to the actual load of the trans and LEDs on secondary, there is the need to let the client know about the potential expected lifespan. I am not a get paid and forget about it electrician. Full disclosure is a life-saver.

I am not sure I have ever seen a sensor rated for MLV loads, I always construed incandescent as encompassing that, what a mistake.

Though a shared gate (driveway, not triac), I am hoping the activations at night are more around 20 or less....

 

[gar]

150329-1520 EDT

grasfulls:

The relay life figures are ballpark values from memory. So I went to look for some references.

I tried GE on the GE RR7. I have seen numbers in the past, but could not find anything now. Google to GE RR7 is about worthless. The GE website search is also worthless. The figures from my previous post were ballpark GE values.

Then I tried Potter & Brumfield. They are now TE. Could not find numbers here either.

Also note that for numbers to mean anything the test conditions must be specified.

My kitchen lights are switched by RR relays, as are all other lights in the house. The main lights are two twin 8' Slimlines, about 250 W. These are probably switched on and off 4 times per day. For 48 years this amounts to about 70,000 cycles. This is light loading for this relay. I have had some RR relays fail. Probably because they were the plug-in type and that applied mechanical stress that distorted the mechanical structure for the snap blade.

A properly designed Triac circuit used as an on-off switch in an AC application should be able to serve almost any application that a mechanical contact switch can. This means the Triac needs to have a constant minimum current flow to the Triac gate.

The differences in operation will be:

1. The mechanical contact will have a voltage drop in the millivolt region. The Triac in the 1 to 2 V range. The mechanical has lower losses associated with the contact.

2. There will be a very slight perturbation in the current and voltage waveforms very close the current zero crossing for the Triac, and none for the mechanical contact.

3. Average DC current is zero for the mechanical contact and very small for the continuously driven Triac.

4. If operated at maximum rating the cycle life of the the mechanical contact may be around 100,000, and the Triac near infinity.

.