Thank you.
I don’t have a valve like that but was just interested in how it worked, you never know, it might be useful information in the future.
But if there’s no current (because microswitch tripped) then the CT will sense no current?
Perhaps a voltage presence detector on the switched live would be more universal?
2 Likes
Exactly my thought too.
It seems to me that another microswitch (or an optical/proximity/magnetic equivalent) to give a positive indication of the valve position is needed.
That is why I asked, wouldn’t a second CT clamp on the other switched live trigger that the DHW cycle had finished?
I don’t know the heat pump monitor software, all the indications are that there’s a simple go/no go test for current that says the valve is drawing current therefore its position must be this. That’s an assumption (that the valve is spring return and it hasn’t stuck) and it’s fallen down and can’t apply to the valve we’re talking about.
If you sense the other live, then you’re dependent on history to know the valve position at any one time, it costs another current input and the logic is doable but starting to get complicated. You could conceivably put the other live through the c.t. backwards and then sense which current pulse the c.t. saw (assuming you’ve got the voltage to determine the phase), but it’s still complicated. By far the most reliable is a mechanical indicator of the valve position.
1 Like
Thank you for all the discussion. Just forced a DHW cycle (with both live and the only switched live to the 3-way valve running through the c.t.). Here’s a screen grab.
Hi Matt. On this actuator there is only one switched live.
Hi Robert. I guess that the same difficulty exists for the Heat Pump too. The HP cannot assume that the valve is in the same position that it last knew. Rather than just issue a change state command the HP must give a definitive command for the valve to move to the presently desired position.
Perhaps for each position the valve actuator has different coded voltage pulse timings that have been coded into the HP controller?
The test would be to check when what is actually live, and maybe take the switched live off and see what happens. It could well be that one winding puts a low torque (but enough to move the rotor) in one direction, and the switched live puts twice as much in the opposite direction - or something like that. I’ve come across that sort of thing before. If it is like this, then the HP just feeds it what it wants, and the valve moves to the desired position when power is put on. With no power, presumably it stays put.
I’ve now made some measurements of the voltages across the actuator terminals when the valve is steady-state in its two positions, HW and CH.
These were taken with my simple multimeter and might be slightly innacurate, but perhaps they are enough of an indication of what is occuring.
There are three terminals on the actuator, Brown, Blue and Black and the voltages are AC
I’ve had the c.t. aound just the brown, just the blue, and both brown and black. But none of those positions are giving any useful indications on the feed graph at emoncms, but the valve is being moved on command by the heat pump.
Any hints what I might try next? I could maybe test my 10m c.t. extension cable and emoncms logic by putting the c.t. around the immersion heater supply and then switching that on?
.
CT sensors sense current flowing
On dumb “traditional Honeywell” motorised valves the power is on ALL the time and the motor “stalls” against the stop. A spring moves them back.
(which wastes a material amount of electricity; but is idiot simple to manufacture)
A CT sensor will sense this.
On more sophisticated valves they mtor up to a switch then stop; and need power applying the other way to motor them.back the other way where they stop.
There is voltage on the wires in this scenario, but no current, so a CT sensor won’t sense anything.
If you want to measure some current you need a load. Connect something in parallel with the valve so that there is a small load when it is in the DHW position. The CT can sense this.
Or open energy monitor provide a “voltage sensing” module instead of a “current sensing” module. (an optoisolated input of some kind; or if you were being super simple about it you could put in an isolated 230>5V DC power supply that takes power from the DHW valve switched live then a couple of resistors to drop that onto a pin of the Pi and read that)
1 Like
And that could be a capacitor, so it would draw current but almost no real power. (It need not be large, you can multiply the current with multiple turns through the c.t)
Or maybe even easier and simpler, the old “a.c. adapter” with the output voltage divided down to 0.3 V and disguised as the c.t. 
(Though this could cost the same order of power as a spring-return valve.)
1 Like
Thank you Marco & Robert.
Please could you propose a suitable component for the capacitor solution. I suppose I would connect this between Blue and Black terminals?
I can’t, without knowing what is the minimum current that the heat pump can detect via the c.t. you’re using. @glyn.hudson - can you specify this please?
I’d still add low value series resistor to limit the inrush current because we don’t know the switching arrangement inside the heat pump.
@dmajwool I’ve just had a look at your data. The DWH diverter valve CT seems to be reporting power readings totally unrelated to the diverter valve.
Are you sure the CT is clipped around the correct conductors? I.e Live and Switched Live (usually brown and black) conductors feeding the diverter valve and the CT is plugged into CT1 on the emonPi2. Could you post a photo?
What’s the make and model of your diverter valve? Some diverter valves only draw current when moving in position. We have an alternative solution, using an optoisolator to detect the presence of 240V for these valves. However, we’re not currently registering any current which looks like it’s the diverter valve, therefore this indicates the CT could be in the incorrect position.
As an alternative to a switched live spring return valve, I would use a Sintesi full bore diverter valve, which has a nice high Kvs and doesn’t slam shut like spring return valves. They can be configured as either 2 or 3-point and seems to trigger the OEM in DHW mode with the CT on the switched live (to the internal relay) when used with 2-point control (They also have volt free and 240V live end switches). I don’t have a ESBE 2-point actuator on hand to test, but would of thought it would of been the same for them? - we only use the ESBE 3-point actuators for mixing valves and not for heat pump DHW/HTG diverters, but you could control a 3-point actuator via a relay, with the CT on the switched live to the coil to give it some constant load for the CT in DHW mode. The ESBE can also be supplied with a volt free end switch for other options.
I agree. There is a lot of random fluctuation in the readings.
I have had the ct clipped around 1. Just the live. 2. The live and the switched live. 3. The Neutral. All of these combinations give the same results on the emoncms data. I wondered whether this might be because my c.t. extension cable is faulty and the c.t. isnt connected to the pi at all, so I clipped the c.t. around the immersion supply (which is adjacent in the same interconnection box) and switched it on.
So this suggests the wiring is OK for registering a 3kW load, but maybe there is some interference when reporting smaller loads? Presently the c.t. is still around the immersion supply. But the long-term trace (and short-term activity when the actuator is moving) on emoncms is no different when the c.t. is clipped around the live & switched live.
Maybe it’s just not practical to eavesdrop on these actuators with a c.t.?
I wonder if something a lot more physical that looks at the dial on the actuator and connects to a local emonTX might be worth looking at?
Perhaps something like this Photo Interrupter Sensor | The Pi Hut could be mounted on the actuator and an appropriately placed lolipop stick attached to the dial?
Just to confirm - have you tried the CT on just the switched live (black) wire, with the CT load arrow pointing towards the actuator?
Hi Tony,
Thanks for the thoughts proposing different fundamentals. If I’m understanding you correctly, in my case the valve hardware is already installed and is working and sadly we can’t get into changing it at this stage.
I have previously tried the c.t. on just the switched live, but as per the instructions haven’t cared about the polarity (however, it was with the arrow pointing away from the actuator). Just now I have moved the c.t. and now it is as you suggest - on the black wire with the arrow pointing towards the actuator.
I’ll report back after the next DHW cycle.
Thank you 
As mentioned in my earlier post I haven’t used/tested this with the ESBE 2-point actuator, but expect similar results to the Sintesi 2-point, as the OEM feed is looking for a non-zero value, but maybe there is a minimum threshold? Also worth noting is the OEM refreshes the feed every 30 seconds, I think, so the output maybe a little out of sync with the valve.
Glyn didn’t answer my question about the minimum detectable current (nor the c.t. rating), but noting the valve is 6 VA, the current is 25 mA give or take. If you have the smallest c.t. from The Shop, this is rated at 20 A, so you’re looking for 0.125% of full load current. I’ve seen split-core 333 mV c.t’s down to 5 A rated current, this would give you 4 times the sensitivity (as would passing the wire 4 times through your existing c.t, if you can loop the wire trough the c.t. many times, this will multiply the current it sees).