CA3500 Instability with GFCI load

I have a CA3500 ZWave (Intermatic receptacle) outlet that I am using to control my outdoor lights and exterior holiday light outlets.

The outdoor outlets are protected by a GFCI outlet on the first outlet and then there is a second outlet on the load side of the GFCI.

At first I had a situation where the CA3500 would just lock up in a some funky state needing to be power cycled to recover but only then to lock back up again very quickly. I swapped out the GFCI with another one (same vintage circa 1995 Leviton GFCI) and the same symptom. I then double checked the wiring to make sure I didn’t have some foul up on the wiring including keeping the load side neutral isolated from line side neutral. No problems. Then I went to store and bought a new GFCI circa 2013 Leviton. Hooked it in an no more lockup.

Fixed I thought. But now a different side effect. The CA3500 will randomly turn on!!! This isn’t in response to a ZWave command its just a spurious turn on…

So then I disconnect the GFCI from the switched side of the CA3500 and the outlet never cycles.

Other data points. There is no load on the GFCI. I have no lights or devices running this time of year. The CA3500 switched output feeds a pair of coach lights with 60W equivalent CFL bulbs and the line side of the GFCI.

So it seems the CA3500 is sensitive to some interference sent back from the GFCI. Maybe the load impedance of two CFL bulb ballasts in parallel with an unloaded GFCI creates some unstable condition as seen by the CA3500.

The mystery continues. I’m going to swap the CFL bulbs for incandescent and see if it makes a difference.

Anyone seen erratic behavior on CA3500’s? I have tried two different units so I suspect it is a design issue with the CA3500 and not simply a bad unit.

Most puzzling.

This is a great post! Somewhat complex description includes clear and (mostly) accurate wiring diagram (not important but CFL neutral goes nowhere) and a picture. :smiley: Now, you don’t go into detail about the modification of the C3500, but I know what you did.

My guess is that the Load Sense feature of the CA3500 is at play here. I would suspect the CFL as the primary cause, but your trials certainly do imply that the GFCI is in play. I don’t know if Load Sensing can be configured (turned off) on this outlet.

Plot thickens… I switched CFLs to 60 W incandescent and it hasn’t spuriously triggered yet.

I am starting to believe either or both the CFL ballasts and GFCI electronics are sending some noise back up to the CA3500 and it is not decoupled in the CA3500 design and is creating some type of havoc. It goes away with the incandescent load because they are such low impedance it kills the noise getting back to the CA3500.

As you can see I modified the CA3500 to send the switched line out a blue 14AWG wire. I’ll do a separate post on how I did that for others.

I did notice when I modified the CA3500 there is a light gauge wire on the switched output side of the CA3500 that goes to the internal electronics. I am not sure if that is part of a current sense loop or just a confirmation of the switching to ensure the micro-switch achieved closure. My suspicion is this is a coupling path for load noise.

I wonder how much of this automation stuff will be impacted by the switch to CFL’s as incandescent bulbs go away? Unintended consequences of designs that maybe didn’t consider complex electronic loads.

I’d have to say that, assuming I’m right about Load Sense, the receptacle is working as intended and designed. The Load Sense runs a trickle current(possibly the light gauge wire) and senses changes in the load, like a flickering CFL ballast/capacitor, and turns on the switch. The incandescents will provide steady current flow and won’t trip the Load Sense.

If the CA3500 Load Sense is not configurable, you’ll need to replace it with a relay that doesn’t have the Load Sense feature or, if it does, allows you to disable it in configuration.

I think/hope that CFL will be quickly supplanted by LED bulbs. However, in your scenario I would expect an LED to behave in the same way as the CFL.

Re: CA3500 random turn on. Due to Load Sense enabled.

Thanks to Z-Waver for some good advice.

I went into the device and set the configuration setting for variable 29 to zero. I reinstalled my CFL light bulbs and the CA3500 hasn’t toggled in a few hours.

I think the load sense must send a current that is expecting a constant load. Since there are electronics in CFLs and GFCI’s the assumption of a constant impedance is not always the case.

Anyway attached is a diagram of how I set the parameter in case this load sense feature is plaguing you too.

Thanks curiousB and Z-Waver.

I just received a 6 pack of these today and experienced the load issue. Now I don’t :smiley:

I have been working on this further. I am finding I cannot disable the load sensing feature. I set up a CA3500 on a bench with a simple table lamp and 60W incandescent bulb.

When I turn the lamp off the CA3500 turns on and stays on. Then when I switch the lamp back on it is lit.

So I know the load sense feature works but I want to disable it.

I tried parameter code 29 with many settings:

0 in default setting
0 in 1 byte dec
0 in 1 byte hex
00 in 2 byte dec
00 in 2 byte hex
0000 in 4 byte dec

then I tried with 1 thinking maybe the feature is inverted and a 1 is actually disabled. No luck.

A couple questions:

  1. When is the config parameter written to the device? Once you save the change in Vera or do you have to do some sort of configure node for the changes to take?
  2. When I set the device config variable to monitor only I get the following in VariablesSet 29,m, (under the advanced tab). Its as if the parameter is not returned due to a query.

So I don’t know if the device is even getting the parameter of if 29 is the proper parameter for a CA3500.

Any suggestions??

Usually Vera is smart enough to see that a node needs configuring when you press Vera’s Save button, but in some cases changes to device parameters may require that you click the “Configure node right now” button on the device’s Settings tab.

I’m not sure that the load sense feature can be disabled on the CA3500. I had wondered where you got parameter 29 from? I can find no official reference to any configurable parameters on this receptacle.

I think you will have better luck if you go with a receptacle that does not have load sensing capability. Many don’t have load sensing capability including the GE 45605 and the Intermatic HA01 and others.

The config parameter 29 was somewhere on this web site but I am not sure if was for Intermatic. Somewhere else it was said manufacturers generally try to use the parameters uniformly. All vague but that is all I had to go on.

My problem is I bought two six packs of the CA3500 on the ebay sale. ??? ::slight_smile:

I did take one apart and snipped the load side sense wire back to the electronics. I though if it never sees the load then it can’t tell if it changed. Anyway powered up fine and when I turned on via ZWave it turned on fine. But then it couldn’t be turned off. I think maybe the cpu uses a voltage detect on load to confirm the switch was properly turned on. With the wire snipped it didn’t get this confirmation so is stuck in some void state.

I believe the load sense works by injecting a small current down to the load (when the switch is in the off state). If it detects the current flow, then it assumes all is ok. If the user toggles the switch at the lamp, the load then becomes infinite and there is no current flow to the load, this is when the switch is powered on, the detection of zero load. Then when you switch the lamp back on it lights up. So to be clear it is the state transition from load to no load that triggers a load sense on event. So load sense is a bit of a misnomer, its really load not sensed that triggers the on condition.

I might try another experiment to tie the load side sense wire through a fixed resistor to neutral and see if a constant current then allows the device to toggle on/off. I am not hopeful but its an easy enough experiment.

Update: I tried rewiring the load sense line of the receptacle to a fixed resistor to see if that would work. It didn’t the receptacle must use the detection of line voltage on the load side of the switch as part of its switch proving sequence.

At this point my options are to put a relay after this device (so all the load sense circuit sees is a relay coil) or get a different type of Z Wave device. Too bad it would seem a trivial software change to disable the load sense feature but these guys exited the market.

If anyone knows of a way to switch off the load sense feature please chime in. The parameter 29 to zero doesn’t seem to do the trick for me.

Update 2,22761.0.html