Smart Panel Meter - can it be modified to read down to zero volts? [SOLVED]

[enut11]

I am in the process of building my second Electronic DC Load based on a Chinese 150w/72v/10A kit. I was looking at using the following smart panel meter, mainly as a convenient way of keeping an eye on watts during testing (Model PZEM-031, DC6.5-100v, 0-20A).
https://www.ebay.com.au/itm/DC-20A-100V-LCD-Digital-Volt-Voltage-Watt-Current-Power-Meter-Ammeter-Voltmeter/281766296611?hash=item419a947c23:g:2iEAAOSw9N1VwjN7

The problem with this type of meter is that it only reads down to 6.5v minimum as it draws its own power from the circuit. A bit limiting as an electronic load monitor.

Other, simpler, v/A panel meters can be powered separately to allow them to read to 0v.

Has anyone modded one of these smart panel meters with a separate power supply to allow it to read down to zero volts?
enut11

[Zero999]

No doubt it's possible but why? The correct solution is to buy something suitable in the first place.

[sokoloff]

Take it apart and have a look.

If you don't want to do that, for less than a 10-spot, you can get something like this:
https://www.aliexpress.com/item/7-in-1-Multi-function-Electrical-Parameter-Meter-Voltage-Amp-Current-Power-Energy-Tester/32870900520.html

[Seekonk]

They all seem to have an internal jumper to supply  power from external. Since not digging into the unit wasn't your first inclination, I assume you are not up to that.

[enut11]

Thanks @sokoloff. The Aliexpress unit would work down to 0v.

@Seekonk, I might try the eBay unit as it has a clearer display. Will pull it apart and have a look.
enut11

[enut11]

Smart Panel Meter arrived today. Physically a bit larger than a normal panel meter at 90x50mm.

Inside it is highly integrated and, on first inspection, does not appear easy to mod for independent supply for the display.

On the front, next to the display panel there is the 20A shunt on one side and a control button on the other.

The positive terminals (+DC IN and + LOAD) are connected together. The 2 negative terminals connect to the current shunt.
enut11

[NiHaoMike]

It looks like modding might be as simple as soldering a wire to the diode near the top? Need clearer picture to tell.

[enut11]

Yes, M7 appears to be an input protection diode.

As received, the meter is powered by the DUT and reads volts, amps and computes watts. To get this meter to read less than 6.5v I need to find the IC that reads the voltage and isolate this input from the power supply to the rest of the meter circuit.

At the moment I can pass a current through shunt R002 (display side of PCB) while powering the meter from another supply but the voltage, and hence the watts, is wrong. It is the meter supply voltage not the DUT voltage. Hope this makes sense.
enut11

[enut11]

These pics show the chip IDs.

[Gyro]

I would think that Z1 or the package next to it will be the regulator. Whatever's feeding the 470uF 6.3V cap will be the giveaway.

[Zero999]

I Googled the part numbers for the ICs and found the SD8103, a Chinese ADC, is the part you'll be interested in, especially pins 4 to 11: the analogue inputs. There will be a potential divider connecting one of the inputs to the supply voltage, that will have to be reconnected to whatever voltage you want to measure. It will also need to be powered from an isolated supply, from the voltage you need to measure.
http://tucnak.nagano.cz/wiki/images/c/ca/Sd8103.pdf
http://www.sdicmicro.com/application.html

[enut11]

Problem solved!
Thanks to @NiHaoMike for the M7 clue and to @Zero999 for the ADC clue.
Some cut and connect mods were necessary.
Will post details and pictures soon on what I did to get down to 0v.
happy enut11

[enut11]

OK, with help from Forum members, got this smart panel meter to work down to zero volts.
Thanks to those who contributed to this project.

The first thing I did was find the input to the ADC (U1) that reads the input voltage. This voltage goes to a divider R10/R11. To prove the point I injected 0.25v at the junction of R10/R11 via a 10K resistor and this changed the readout. I then set about isolating this input which was connected to one of the screw terminals.

The two positive terminals are connected together via a solder pad on the display side of the meter. I removed the solder, cut the pad in 2 with a blade, then re soldered the terminals. One terminal was now open and the other connected to both the input voltage divider and the rest of the circuit via diode M7. I removed M7 and replaced it with a through-hole component connected to the "open" positive  screw terminal.

That was it really. Simple looking back on it 

[enut11]

And, to show that it works, here is a pic of the meter reading less than 6.5v, the previous as-purchased minimum. Watts is correct to the display resolution (this is a 20A meter).

A few comments on use:
1) Meter draws ~1mA from 8v supply. This goes up to ~3mA with the back-light on. Makes it suitable to power from 2 LiPo cells. The display will tell you when the battery needs to be charged (it will go out). Back-light on/off is controlled by the Set button.
2) Visibility: The LCD is meant to be used at eye level or above, eg on a shelf. The display is dim below eye level. Probably something to do with the LCD polarising filter??
3) The meter has user set low and high voltage warnings - the display blinks when these are exceeded. I set mine to 0v and 30v. Occasionally, the low voltage warning blinks the display at switch on. A simple press of the Set button restores the meter to normal operation
enut11

[Zero999]

Good, I'm glad you've got it working. Thanks for posting an update.

Another interesting modification would be to move the current sensing, from the negative to positive side, but it would be much more complex: it would involve adding another op-amp.

[enut11]

I am still thinking about how to make the display more viewable at a lower angle. At the moment this panel meter needs to be viewed straight on or from slightly below. When is sitting vertically on a bench and viewed from slightly above, the characters are quite dim.
I read somewhere that the viewing bias of an LCD screen can be changed somewhat by adjusting the contrast. So, what I need to find out now is which pins control contrast for this particular display. There is not much to identify it on the PCB unfortunately.
enut11

[Seekonk]

Some of these are almost impossible to read. Leftmost pin looking at display is common #1.  #2 is 3V3.  Connect a pot from 1 to 2 with wiper going to pin 3 through a 1.2K resistor.  I used a 10K pot and that resistor because it was what I had handy. Far from optimum but it worked.  Pin #3 is in the 1.3V range.

[enut11]

Hi @Seekonk. I tried your suggestion but did not get enough contrast effect to make it worthwhile. Thanks anyway.
enut11

[kathanas]

Hi enut11, and thank you very much for your post and modification!
Ive tried to follow your example but didnt work for me.
Could you please tell me what diode your proposed? Ive tried with 5404 and 1n4004 unsuccessfully.
I put the voltage limits to 0.00V and 38 V.

As you understood I am a new in this field, and if you could send a procedure I will very much appreciate.
 Thank you in advance

Kostas Athanasiadis
(from Greece)!

[enut11]

Hi @ kathanas
That was almost 3 years ago! Time flies, and it looks like they changed the PCB, near the inputs at least.
The diodes you are using are OK.

I see that you connected the red wire from the diode cathode to the second PCB pad? Have you tried to connect it to the first pad, the one closer to the input terminals? Can you monitor the meter current via a separate 12v power supply with voltage and current meters?
enut11

[kathanas]

Thank you for your prompt reply dear @enut11.
Actually I connected only to the first one pad.
Now, I will try to connect the red wire to the first pad, the one closer to the input as you suggested.
I will inform you, for the results!

Kostas!

[maq67]

http://tucnak.nagano.cz/wiki/%C4%8C%C3%ADnsk%C3%A9_m%C4%9B%C5%99idlo_PAEM-031

[ozarchie]

Thanks enut, a very useful project that helped me.
The latest version is a PZEM-031 V4.0 which uses a different micro but the display and current sense are the same.
I am using it as a solar panel monitor and separating the meter power supply from the solar positive was essential.
Unfortunately for solar systems there are many sources of power to the batteries, so the load grounds are common. The only way to measure the load current on the low side is to connect the load- to ground and let the meter- 'float' below system ground. The load current voltage generated across the resistor is then negative. Fortunately, your circuit shows that you can simply swap the inputs to the I2C device and get a result.
I am currently using an ESP32 to spy on the SPI command/data to the display controller and publish the results to homeassistant using ESPHOMEs Custom Control. I will add a link when finished.

[warf135]

I wouldn't normally reply to an old post, but there is an easier way to power this unit without having to modify anything...

On the opposite end of the board to the regular connections, there is a small white rectangle with two unused connections (marked with a W), these are power and ground...

Just feed 5v into these and the meter is very happy!

Found your post while looking for a contrast mod for this thing, it looks great straight on, but terrible from above.

[Kim]

Thanks for that, I came looking for the exact same issue, and feeding 5V into the W terminal worked fine, except for one gotcha:

The calibration procedure for this meter is to power it up with a 20V supply and 1A load, while this W terminal is pulled to ground.  I discovered this because I'm using it to monitor the DC input of a dual power supply for a cycle race timing system, with the 5V supply to the W terminal derived from the AC PSU so the meter flashes its undervoltage alarm when a battery isn't connected, rather than just appearing blank.  (Similarly, I have an audible alarm on the AC side to warn of generator failure.)

It turns out that my RFID reader and Ethernet switch is close enough to a 1A load to put it into calibration mode when powered up from a ~20V source

Resolved by feeding the W terminal via a diode, so the 5V regulator can't act as a pull-down when unpowered.  The voltage drop doesn't appear to be a problem (indeed, the display contrast was slightly worse when fed with the full 5V).