Reading Solar Power "Generation" in the dark with the circuit switched off

I think a word is missing somewhere, but you shouldn’t feel anything when touching isolated metal. If you do, then there’s some leakage current that you’re feeling. It shouldn’t be dangerous, the a.c. adapter, the c.t. and the 5 V USB power supply all provide isolation between the mains and the emonPi, but there will be a small capacitance that could carry a very small current.

Your 10 A of PV will give you 10 A * 50 mA ÷ 100 A = 5 mA out The SCT006 will give you 10 A × 25 mA ÷ 20 A = 12.5 mA exactly. So the correction is 5 ÷ 12.5 = 0.4 (N.B. you can only multiply :wink: )

You’re correct about the missing word. That should have read
“Something else: If the power is off to the emonpi,…”

Thanks. Will try earthing the case, and will try the alternate CT when it arrives next week.

Thanks again for your help.


We have a winner!

Earth wire connected. Tingling feeling when running finger over case, gone. Solar “generation” now (it’s dark, nothing being generated) is -4, i.e. the inverter is consuming power. NB That’s raw unprocessed reading (without the adjustments that I put in as mentioned above)


I checked my emonPi that’s using just one c.t. on a meter tail (line), a.c. adapter, 5 V USB adapter and optical pulse, and using a DVM with an input impedance of > 10 MΩ, I measured approx. 1 V from case to earth.

If you can feel a tingling, that would suggest a current of up to around a milliamp or two flowing through your body, and that implies a voltage much greater than the 1 V I measured.

My voltmeter, (UNI-T from Maplin) claims impedance on “10MW” (sic - presumably they mean 10MΩ!) Accuracy is claimed to be 0.01V in the range I used. I put a fresh battery into the meter, but I’m not 100% confident on how accurate the meter is.

I disconnected the earth, then measured a voltage between the case & an earth. It was never stable & fluctuating between 0.01 and 0.1V.

Touching the case and the “tingle” was there. The Solar read approx +20W generation (should have been zero to -5W).

I tried my alternate 5.25V power supply. I didn’t measure the voltage between the earth and the case, but the tingle was not present with the alternate supply, and the Solar generation read -4W. This was without the extra earth cable.

I tried the supplied Power brick without the AC sensing power brick, and the “tingle” was still there.

This tells me that the problems I am experiencing are caused by the supplied 5V 2A Power supply. Do you think this qualifies as faulty? If not, can I use the other power supply, a Raspberry Pi Power supply (5.25V 2A), or similar Raspberry Pi Power supply (5.25V 2.5 or 3A) or is 5.25 too much voltage? Obviously I’d have to get a adaptor to get the right end on the cable.

If desired, I can return the supplied power supply to be checked, (providing I can run it on the 5.25V for a while).

Any thoughts?


from The Shop?

Probably, it’s high frequency from the SMPS, and that’s a good reason why your meter couldn’t see it, because from memory it won’t be good above a few hundred Hertz.

Let’s see what @Gwil, @TrystanLea or @glyn.hudson have to say, but it does sound wrong to me.

The Pi should be OK at 5.25 V, the regulator on the front end “Shield” is a “MC1700” on the circuit diagram (I assume that should be MCP1700) which is good to 6.0 V, so you should be OK at 5.25 V.

Yes, from

Will wait until next week and return to work for comments from above.

@Robert.Wall, sincere thanks for spending so much time talking to me, an unknown noob, about this. :slight_smile:

Can I just point out that we were all unknown noobs once upon a time. :wink:

I’ve done some measuring. Firstly, I can’t measure anything on the case - the anodising on the extrusion is making a good job of insulating it. But it wouldn’t take much to cut through the anodising. I could measure on the screw heads - so I did. Running without the a.c. adapter, I see this:

1.08 V rms and 4.92 V p-p, at 50 Hz.

Measuring on the USB connector shell, I get a bit more:

And with the a.c. adapter, back on the screw heads, I get:

And with the c.t. nowhere near a mains cable, it’s reading 0 or 1 W.

The shop is back on Monday, one of them might visit the forum before then, so we’ll see what’s said. But my money is going to be on a faulty 5 V USB power supply. There shouldn’t be so much leakage current that you can feel it, and that it’s coupling into the input to show as ~ 20 W corroborates that.

Hi Steve,

Very sorry about all this (as well as wishing you a Happy New Year!).

Would it be possible for you to post a photo of the power supply and cable you received from the shop?

As a first step, perhaps we can send you a replacement power supply and cable next week? Please PM with your order details or email the shop at [email protected].



These are the photos I PMed to @Gwil

I was using the 1.2 A Ideal HK-U-050A120-CP

OK, well the replacement PSU and cable arrived earlier this week. I tried them as soon as I could, but haven’t had a chance to report. Also arrived this week are a Micro USB socket to mini USB plug (so I could try more power supplies), and the alternate CT sensor, described above.

As I said before, my voltmeter doesn’t seem accurate enough to measure much voltage between the case and an earth. My earth is a gas pipe that has an earth wire from the circuit breaker connected to it.

With the the original PSU plugged in, as documented, without an earth, there is a “tingle” when I touch the case. Also, I can wiggle the power cable where it enters the emonpi and this causes the display to flicker slightly. However, I don’t think this flicker is significant.

With the replacement PSU and cable that arrived this week, unfortunately, exactly the same happens.

As well as those two PSUs, I tried 3 others.
A genuine Raspberry Pi 5V 2A. The “tingle” when using this was less, but still there.

A 5.25V 2A unofficial Raspberry Pi PSU from ModMyPi with integrated cable.
This has as much “tingle” as both emoncms PSUs.

A 5.25V 2A unofficial Raspberry Pi PSU from ModMyPi with detachable cable. This is the ONLY PSU that does not generate a “Tingle” without an earth.

Both 5.25V PSUs make the display run brighter than with the 5v PSU.

I wondered if it was the CT sensors or the temperature sensor that were causing the tingle. I disconnected the CTs from their mains cables, and from the Emonpi, and disconnected the temperature sensor, leaving only the original 5Vdc official PSU and the AC sensor input plugged in. The tingle was still there.

So, I’m now trying the alternate CT sensor (SCT006). Unfortunately, I’ve done this slightly too late in the day to measure generation, but I’ve put “x 0.4” into my input to see if that gives me roughly the right readings (as described in a previous post)

To get correct PV readings, and no tingle from the case, I either have to use the a specific 5.25V modmypi PSU, or add an earthing cable to the emoncms case.

I’ve attached pictures of my setup. PV Meter and CT sensor, out of picture top right. Mains meter just visible right of picture. “Fuse box” top left. Gas meter bottom left.

I’m a bit reticent about leaving an earth cable connected without some specific advice. As you can see from the pictures, the main earthing cable from the fuse box is connected to the gas inlet pipe. I have connected a cable to a redundant earth cable connection slightly above it. The other end of the cable goes to one of the small screws that connect the black plastic mounts to the emoncms case. Does this cable have to be a specific size or rating? Is it safe to connect it as described/pictured? Where is the best place to connect it to the Emonpi - as current, via the display’s screws, or via the screws on the edges of the case?

I think I’d be happier running the emoncms from the 5.25V PSU that doesn’t produce a “tingle”, but there’s a snag as it won’t quite fit in the mains socket as there is something next to the mains socket stops it fitting & would require a multiplug or extension lead. Is it safe to run the emonpi from the 5.25v 2a PSU?

@Gwil I can return the replacement PSU and cable as it doens’t seem to make a difference?

So 3qns 1) about earthing 2) 5.25v OK 3) return the replacement PSU?

Any comments gratefully received.



Your gas pipe should have earth bonding, but you must not use it as an earth for anything else. Your main earth should be provided as part of your electricity supply. If not, and like mine it is an older house that has not been upgraded, you can use a water pipe for your main earth as long as there is not an electrically insulated section in it, which would make it ineffective.
The best place to pick up an earth is either your main earth block (if you have one), or the back of the socket outlet that you’re using.

That’s a good question. I would use a ring crimp (or solder tag) under the head of one of the screws fastening the end plates, because the anodising on the aluminium makes quite a good job of providing an insulating layer. On the emonPi that I checked, I don’t have continuity between the end screws and the display screws - it might be necessary to scrape the anodising off the ‘runner’ edges of the display panel to provide continuity. (I’ll try that.)

It should be, as that’s the voltage of the official Raspberry Pi power supply, as I wrote earlier.

Re Earth:
@Robert.Wall - thank you. I have removed my added “earth wire” to the emonpi.

Re alternative earth sources - my water inlet pipe is plastic, so that’s a no-go.
I am not going to run any wires into the fusebox or the back of the mains socket to pick up an earth, because I know that there are regulations, but I don’t know what they are, and I don’t know that any work I did would be safe or not. I think this is illustrated by my previously connecting a wire to the earth-bonding.

Indeed you did mention earlier that the 5.25V supply would be OK earlier. I initially misunderstood what you said & thought you were just referring to the Pi bit rather than the whole unit.

Again, sincere thanks for your help.


I wouldn’t have suggested that if it wasn’t safe. The rules (Building Regs) in the UK say that you can replace like for like, but you must get approval for extensions and ‘new’ circuits. In my book, that’s so that you don’t overload the existing installation. Adding an earth such as this cannot do that, the only thing you must be certain of is you maintain the integrity of the existing wiring. There’s a little-known regulation that says that you don’t have to adhere to the rest of the regulations when you can demonstrate that what you’re doing is no less safe than the approved way.

So you can flip the breaker off that feeds that socket, take it off the wall, feed your earth in from the back, put it into the terminal and make sure all 3 (or both if it’s a spur and not part of a ring) earth wires are under the screw and tight, and put it all back. Job done.

No, the power goes into the emonPi Shield, that’s good to 6.0 V as I wrote, and in parallel feeds the Pi, for which the 5.25 V supply is designed, via the GPIO pins. I think I was assuming that you knew the emonPi is a cut-down emonTx (called the emonPi Shield) and a Raspberry Pi in the same box.

Please don’t get me wrong - I didn’t mean to imply that you suggested something that wasn’t safe. I meant that I don’t know that I could do it safely. For instance, I didn’t know the regs you described or that that would be safe. Also, I have no idea what size (gauge) wire to use. That’s why I’ve gone for the 5.25V PSU instead.



I would have used 0.75 mm² multi-strand flexible. Rated at about 6 A, that should be well capable of blowing the fuse in the power supply, should there be a serious fault.
(Like, not that you want a 30 m roll.)

I’ve been looking carefully at the emonPi case. The anodising covers the metal end plate, the two extrusions and the front and back plates, leaving only the sawn ends of the extrusions and the ends and edges of the front and back plates. The screws on the metal end appear to break through the anodising, so that there’s continuity between the two extrusions via that end, but there’s no guaranteed contact with the front and rear plates. The only way I can see to provide a good connection to the front plate would be to trap a wire under one of the display nuts at the pushbutton end, and trap it under the acrylic end plate (that end because that’s the one to take off to get to the SD card). Then you can put the earth tag under one of the screws at that end too, or one of the screws for the fixing clips of course - any screw that cuts through the anodised layer is OK.

1 Like

I’m wondering whether or what EMI suppression components that power supply has. And whether Steve’s problem is leakage via that route.

I remember at work many years ago, we had laptops that were sold on to the customer as programming tools - these had suppression components rated for the American 120 V system (although the power supplies were ‘universal’ voltage - something like 85 V – 260 V) and those were apt to give a painful shock if you touched the plug soon after switching off and unplugging. Not quite the same situation as we have here, but possibly a related mechanism. Because the plug on the PSU has a plastic earth pin, any suppression capacitors must be between line and neutral only, and it’s just possible that the 5 V side could be connected to the centre of a capacitive voltage divider between L & N. I’m not sure how that reconciles with it being double insulated though.

Sorry for the delay in updates. I can confirm that using the Raspberry Pi power supply mentioned earlier, that the system appears to be working fine. There is still no “tingle” when touching the case and the PV monitor reads negative when not generating (or would do if I hadn’t suppressed negatives from the logging.)

Do you want me to return one of the official PSUs @Gwil ? I’m happy to.

Re earthing: It’s moot as the alternate PSU has fixed the problem, but, should I need to earth it, I’m still loathed to touch inside a plug-socket. Is it acceptable to wire up just an earth wire inside a mains plug’s earth pin and run that to the emonpi case?

The alternate CT connector works fine, provided that the relevant conversion factor is incorporated into the script.

But to confirm it’s one specific PSU, or adding an earth that makes the PV monitor read correctly when solar is not generating, and removes the “tingle” from the case.