Do I have a defective PZEM 004T? Voltage and Current readings show incorrect values

RendyIrawan · July 9, 2021, 6:56pm

I recently bought a PZEM 004T V3 (split CT) to use with an ESP8266 running ESPHome. I managed to get it up and running, hooked the split CT to the live wire. Reading seems to be accurate. The problem is what happens after I use it after a while (around 36 hours). The sensor stopped giving out readings.

After a bit of tinkering, I managed to get it up and running again. And after that, I noticed that the watt values are so high. Upon further investigation, it turns out that the ampere values are too high. Because I’m installing this at an electricity panel, I uninstalled it so I can troubleshoot this from my desk.

About a week later, I started working on this again. I hooked it up to an ESP8266, and it now shows correct readings! I wanted to be sure, so I left it running overnight with my laptop charger connected to it. The next day when I saw the readings, it was way higher again. My 40W MacBook Pro charger shows a 150W consumption. This is of course an incorrect value. So I checked the history graph. I forgot to take a screenshot, but the graph basically shows that the ampere values slowly drifted higher and higher even though the laptop charger obviously draws a fairly constant amount of power draw. Since the amount of drift in the value always changes, I couldn’t even calibrate it on ESPHome. Also, when testing with a load that has a constant power draw like a light bulb, the ampere values isn’t constant (it should be constant).

Another thing, the voltage value is also incorrect. One time, 215V shows as 205V, and another time 215V shows as 180V. It’s not linear and always changes so I also can’t calibrate it on ESPHome either. This combined with the faulty current values makes measuring the power (watts) impossible to do, which is why I bought this thing in the first place.

So then I bought another PZEM 004T, this time it’s the 10A model that doesn’t require a CT, and after running it several days the readings came out consistent.

Now, the PZEM (the one with the split CT) that I first bought even shows ampere readings show a value that is almost 25 times bigger than it is supposed to be.

So, did I receive a defective PZEM 004T? Or is this a common issue that could be fixed?

li_gangyi · January 29, 2022, 10:29pm

I experienced the same issue and initially thought it was the 1M resistor that was part of the divider network, changed that to no avail (the original also measures good on my DMM, however I was unsure what it was doing when exposed to HV and warmed up), put the original 1M back on (R17).

Next I started blowing on the board (covers exposed) and found that the readings would drop back to regular when blown on. To localize the cooling, I used a small cotton swab to apply alcohol around suspect areas.

Cooling the Vreg, a 7133, made the readings ok for a few secs everytime my swab touched it, measuring it’s output, it was low at 2.0+ V and dropped further as it heated up. Even setting it aside and letting it cool all the way, it never got anywhere close to the output voltage it was supposed to be at (3.3v), input voltage was right around 12V.

I have yet to replace the voltage regulator, I suspect the heat over time has killed it, no mention of thermal shutdown features is included in the datasheet from Holtek. Will report back again.

li_gangyi · February 1, 2022, 5:54pm

Quick update to this, I removed the 7133 and jury rigged an external Vreg (7805 and 3 diodes in series is what I scrounged up), ran hookup leads to this regulator thru the existing zener+series capacitor + resistor + diode config (where the Vin, Gnd & Vout of the original 7133 went) and I can report that accuracy went back to normal (at least for the VAC reading, did not hook up the current transformer to test). Am waiting on parts to properly fix this thing up (I ended up ordering a bunch of 7533s).

Not exactly sure why/how the original Vreg died. Will see how this replacement fares.

li_gangyi · February 17, 2022, 7:48pm

Final update, I’ve replaced the bad 7133 with a 7533, measured the voltage output to be 3.29V and confirmed everything works as normal again. I suspect my unit came with a defective regulator from the factory as I’ve measured the power line voltage higher than normal (about 5% higher), with this change it has gotten to within 1%.

RendyIrawan · February 18, 2022, 10:33am

That’s great news @li_gangyi! I’m glad to hear you managed to figure it out. Thank you so much for sharing your findings here. I’ll be sure to check the 7133 on my pzem as well and see if swapping it out fixes my incorrect readings too.

ronalehm · October 17, 2023, 10:56pm

Solucione el problema como lo indico aqui: https://www.youtube.com/watch?v=qRsjsenvlJA, basicamente la alimentación del integrado V9881D, Que se alimenta del 7133, debe suministrar 3.3VDC en el VDD5, y el Voltaje de seguimiento en VDCIN debe ser 1.5VDC como minimo para que el CI funcione correctamente.

Con esto suministre a la entrada PIN 2 (VIN - 7133) un voltaje 10.5VDC para que pueda funcionar con normalidad.
cualquier duda me escriben: [email protected]

Slds
Ronald