Optical sensor reverse reading

djh · 2019-02-23T13:02:59.387Z

I have an optical sensor on my main import meter. I’ve just noticed that when I am exporting power from my PV installation, a graph of the optical sensor goes up not down, i.e. indicating that I am importing power from the mains. See the graph below.

Screenshot-2019-02-23%20Emoncms%20-%20graph%2061010×1207 101 KB

The import power is a log of a CT on the meter tail. The optical sensor is on the meter itself. You can see the optical graph increasing whilst the power is greater than zero, but it also increases when the power goes negative!

(I was exporting from just after twelve because my hot water tank just reached its limit temperature, before that the solar power was being diverted)

Now I know the meter itself shows rEd when I’m exporting, so what’s going on? Is this normal, or a problem with my meter, or a problem with my emonTx installation or a bug or what? It makes the optical sensor record useless. I also have a power to kWh log of the CT data which does decrease when I’m exporting, but it isn’t as accurate as the optical sensor.

Robert.Wall · 2019-02-23T15:02:11.806Z

djh:

so what’s going on? Is this normal, or a problem with my meter, or a problem with my emonTx installation or a bug or what? It makes the optical sensor record useless.

It is a problem with the design or configuration of your meter. I’d ask the question: How can a single pulse of light indicate the direction of power? It cannot.

This is why most of us here caution against the use of the optical sensor except as a means of calibrating your emonTx.

The optical sensor - or more correctly, the optical pulse output on your meter, is only useful when the energy flow is unidirectional. When the energy flow reverses, most meter LEDs either show nothing or remain lit continuously.

Yours seems unusual in that it appears to indicate the same irrespective of direction. The more important question is, does it have separate energy registers for forward and reverse? Because if it doesn’t, and the (single?) register still increments when you are exporting, it means that you are paying to supply your energy company.

djh · 2019-02-23T15:24:14.823Z

Robert.Wall:

It is a problem with the design or configuration of your meter

Thanks for the clarification.

This is why most of us here caution against the use of the optical sensor except as a means of calibrating your emonTx.

Ah, I haven’t seen that caution. Is it useful to read it? Where is it? I’d been blithely assuming that the optical sensor was the best source since it [should] precisely match the meter reading, and since there’s a special feed type to make it extra reliable. Apparently I’ll need to think again.

The optical sensor - or more correctly, the optical pulse output on your meter, is only useful when the energy flow is unidirectional. When the energy flow reverses, most meter LEDs either show nothing or remain lit continuously.

Hmm, I can’t see the optical pulse output itself, because of the optical sensor stuck over it, but the green LED on the sensor appears to be permanently on and the meter says rEd. (It’s a Landis+Gyr 5246C)

So I don’t understand why the readings keep going up if the LED is permanently on.

Robert.Wall · 2019-02-23T15:39:27.764Z

djh:

So I don’t understand why the readings keep going up if the LED is permanently on.

They shouldn’t be. That sounds like a fault in the logic. Which sketch are you using?

(But if the LED goes off for more than 100 ms, it will count as a pulse.)

djh · 2019-02-23T15:41:20.098Z

Today seems to be the first day we’ve actually exported any significant solar energy since I installed the logging. So at least I’ve caught the issue early.

I’ve written to Landis+Gyr to ask for a technical manual and explained the issue to see if they understand what’s going on.

djh · 2019-02-23T15:44:03.895Z

Robert.Wall:

They shouldn’t be. That sounds like a fault in the logic. Which sketch are you using?

(But if the LED goes off for more than 100 ms, it will count as a pulse.)

I’m using whatever sketch was installed when I bought the emonTx. Presumably the green LED on the sensor mirrors the state of the red optical output LED on the meter? I’d think I would have noticed a 100 ms occultation. I’ll go and have another look, and perhaps take the sensor off the meter for a few seconds to be sure.

Robert.Wall · 2019-02-23T15:52:50.084Z

djh:

Presumably the green LED on the sensor mirrors the state of the red optical output LED on the meter?

Yes it does - with some caveats, the LED state and the output state can be different. There was a discussion about that a while ago.

djh · 2019-02-23T15:53:06.644Z

I’ve just been for another look. The green sensor LED looked to be permanently lit. The red optical output was mostly off and blinking occasionally whilst displaying rEd in the LCD.

edit: just seen your reply. What are the caveats?

Robert.Wall · 2019-02-23T15:56:19.993Z

I said - the LED indication and the output state can be different. A search should find the discussion.

Here:

Problem with Optical Pulse sensor emonTx

The optical sensor output is not a true TTL output as is claimed, it appears to be an over-driven analogue signal - in that is appears to be linear over a small range. You might be able to see this with a multimeter measuring the output voltage. Also, the operation of the green LED on the sensor isn’t a reliable indication that the emonTx has or hasn’t detected a pulse, due to that pseudo-analogue output again. One thing I do notice is, the single-phase emonTx sketch detects a pulse on the fall…

djh · 2019-02-23T16:02:16.083Z

Sorry. I don’t know what to search for. A search for ‘LED indication and the output state can be different’ returns this thread and two irrelevant ones.

Here’s another graph, BTW

Screenshot-2019-02-23%20Emoncms%20-%20graph%20131010×1408 154 KB

The yellow trace is effectively showing the export. The green is generation and the red shows where it is being absorbed by the diverter. The brown trace shows the optical sensor readings.

djh · 2019-02-23T16:03:21.016Z

Thanks for the link. I’ve read it now but don’t understand what it means.

Robert.Wall · 2019-02-23T17:06:54.664Z

I did some measurements - I’ve been searching but I can’t find them.

What it means - and I can’t remember which way round it was - is the emonTx and the LED use different decision voltages to determine how they see the state of the output voltage of the sensor. So the situation is the LED could be lit and the emonTx does not see a logic high, or the LED could be dark and it does. I think it was the former, but I can’t be sure.

djh · 2019-02-23T18:13:48.709Z

Robert.Wall:

What it means - and I can’t remember which way round it was - is the emonTx and the LED use different decision voltages to determine how they see the state of the output voltage of the sensor. So the situation is the LED could be lit and the emonTx does not see a logic high, or the LED could be dark and it does. I think it was the former, but I can’t be sure.

So the state of the green LED is entirely determined by the sensor unit itself? And the value state in the emonTx is somewhat ambiguous depending on exact voltage levels? Seems like a poor design of sensor, although that’s just a comment and not very helpful at that.

But a green LED that stays on when both the meter output is flashing in what looks like a normal fashion and the emonTx is observing those flashes sounds like a faulty sensor to me. Or am I missing something?

Then there’s a completely separate question of whether the meter should be flashing and whether it provides some way to discriminate. Otherwise I suppose the answer would be to modify the sketch to use the CT to determine the direction of flow being measured by the flashing.

Robert.Wall · 2019-02-23T19:21:17.373Z

djh:

Seems like a poor design of sensor, although that’s just a comment and not very helpful at that.

…but I’d agree that it’s accurate.

djh:

Then there’s a completely separate question of whether the meter should be flashing and whether it provides some way to discriminate.

That I cannot answer. And the meter manufacturer is almost certainly not going to answer, as I believe identical models can be configured differently according to the requirements of the owner. Who does the configuration, I don’t know. It might be the owner, in which case the manufacturer won’t even know.

djh:

Otherwise I suppose the answer would be to modify the sketch to use the CT to determine the direction of flow being measured by the flashing.

I guess that’s a possibility, but at very low powers, that might not be reliable. Though in that case, the error is likely to be small.
If the meter did as you suggested - occults when exporting, then provided that the m:s ratio never got close enough to 1 (from either direction) software should be able to discriminate. But I don’t think that is done. I think most meters show flashes when importing, and a steady indication when below “creep” or exporting.

Going back to this:

djh:

The green sensor LED looked to be permanently lit. The red optical output was mostly off and blinking occasionally whilst displaying rEd in the LCD.

have you tried shielding the sensor from ambient light? If I recall correctly, as I indicated above, I could get the green LED lit on ambient light. And the hook-and-loop fastener is by no means light-proof.

djh · 2019-02-23T19:31:32.691Z

Robert.Wall:

And the meter manufacturer is almost certainly not going to answer, as I believe identical models can be configured differently

Well, they should be able to tell me what the possibilities are, so I may be able to tell whether it is faulty or not. And with the information, I may be able to persuade my energy supplier to instruct somebody to configure it sensibly, if there is such a possibility.

I guess that’s a possibility, but at very low powers, that might not be reliable. Though in that case, the error is likely to be small.

Yes, exactly.

have you tried shielding the sensor from ambient light?

The meter is in a meter box, so it should certainly be shielded from light when in normal operation. When I observed it, I held my hand over it to shield it from ambient light.

Robert.Wall · 2019-02-23T20:41:28.043Z

A little more here:

LED pulse sensor problems Hardware

Try moving the sensor around very slightly on the meter, also rotate it on it’s axis (ie cable entry down, up,left and right etc) and/or angling it slightly whilst watching the LED. Since you have now observed this odd behavior I shall also share my findings (I didn’t want anyone thinking I’m even more nuts than they already did) I have noticed a strange situation, where I can move a sensor and position it to give anyone one of 3 outputs from the same pulsed light source, LED flash only, no …