Modifying the Peacefair PZEM-004T AC Comms Module (with schematic)

The PZEM004T module is a mains powered module that measures voltage, current, power factor, apparent power and frequency. It has no display but provides these values over a serial interface. In this video, I am going to modify it to measure from 0 volts AC onwards and in a second step to take its operating power from the serial port. This makes using the module much simpler.
There will be a follow-on video, about accuracy and using the software I wrote to use the module.
Note: R18 in the schematic should be 0.5 Ohm not 0.05 Ohm
00:00 the PZEM-004T
01:38 specs
02:52 serial protocol
03:52 wiring diagram
04:49 current transformer
05:16 interface cable
06:23 the PCB
07:49 V9881D chip
09:13 schematic & function
13:45 modification options
14:59 measure from 0V
19:36 ..and power from serial
25:07 modified PCB complete
The x1 / x10 feature is explained here: • Improving the sensitiv...
The schematics and software is here: github.com/TheHWcave/Peacefai...
You can support the channel by becoming a patron. www.patreon.com/user?u=45604138
As a patron you get early access to all my videos as well as exclusive content.

[@bgable7707]

WOW, when you put it that way, it all appears clear and possible to accomplish! Option 1 for sure, two after I get the original and opt 1 working. THANK YOU, for the schematic, pictures of the wire and cut traces and detailed explanation!!!

[@Danfranschwan2]

Great Video and clear instructions!

[@BjornV78]

Nice and clear video, thanks for the effort. Grtz

[@TheHWcave]

Thanks, much appreciated

[@bertoid]

I was going to make some mods to enable operation at low voltage, but I wouldn't have thought of using a DC-DC converter as you've done, an excellent solution! I'm currently designing my implementation based on all your mods, but with a few variations. Thanks for making the schematic available, it's a great help. One small error: The cap across R16 is labelled as C1, but should be C12.

[@MegaJmanjman]

Great video! Just built a powermonitor using this Module and a nodemcu.

[@WaldekDudzinski]

Dobra robota 👍💪

[@andrey_pav]

nice work! best instructions!

[@BradCagle]

Great, thanks for the schematic. I wanted connect this directly to raspberry pi UART pins, but could not because of the 5v vs 3.3v levels. Now looking at your schematic I see I can just provide 3.3v to pin 4 on connector P2 instead of 5v.

[@abizarlakdawalla4519]

Was inspired by your video to buy a version with LCD for a variable AC power supply I am working on. The new PZEM 020 turned out to have a more sophisticated powersupply - but your general approach would work. I went with using an external DC supply instead ... have posted a video on my channel if folks are interested

[@kissingfrogs]

nicely done, thanks for sharing. Was considering powering a ESP8266 from the 7133 but datsheet seems to suggest it is a 3.3V 30mA reg which is under powered. Perhaps It could be powered via a separate 5v reg from the 12v rail. I think R? 100 could handle an extra ~50mA (~1/4watt dissipation) I like the idea of using the ESP8266 wifi for complete isolation and easy remote access.

[@t1d100]

These are now being offered on Amazon/USA in a white, panel mount case. They reference the PZEM-004T module, but they are not sold by PeaceFair, IIUC.

[@TheHWcave]

Interesting. I wonder if the circuitry is still the same. Maybe someone who bought one will tell?

[@ariel18012]

muy muy bueno amigo, excelente explicación, se agradece el esfuerzo para hacer el video, seria interesante hacer algo asi con un CS5463 de cirrus logic.....?

[@TheHWcave]

Sorry, no plans for doing something with the CS5463 in the near future

[@morganabraham1971]

This is a great video! One thing we are looking to do is increase the sample rate of the PZEM meter to track motors in realtime and identify current spikes/stalls faster than required by a draditional energy meter to provide close to realtime feedback loops for control systems and to prevent damage to the fuse from locked rotor conditions. From the video and initial documentation, it seems that the fastest sample rate is every 0.5 seconds. This seems a bit slow - also considering that we are not interested in measuring frequency (only current). Is there a way to increase the sample rate on the IC in the PZEM (ideally to >10x/sec), and do you know which registers/software functions are needed to do so if that's feasible. Thanks!

[@TheHWcave]

Thanks for the comments, unfortunately, I am not aware that the reporting rate can be changed. All chip registers are in the comprehensive datasheet for the V98xx family. While the hardware part of the chip samples energy values at kHz speed (as far as I remember the datasheet), the firmware that has been programmed into its Flash memory to convert the data into MODBUS format, calculate the checksum and send it on the serial port just doesn't work any faster. Since these chips were developed for use in electricity meters, there is an encryption-bit for the flash memory. If it it has been set by the developer, it would be effective since the in the PZEM-004T the mode pin is pulled high, enabling metering mode. It should be possible to force the pin low to get debug mode which I think turns off encryption and connect an in-circuit programmer to read, disassemble and possibly change the firmware. I have not investigated this myself but if you do get anywhere I would be interested in your results.

[@hnnn5914]

Hello sir, if I need to use this module with ct 50/5 or 100/5 or more than this, how I can do?

[@mikepaul6688]

great video. do you know if an adum6000 could be used as the isolated dc-dc converter? they are smaller and could be soldered to an soic to dip adapter.

[@TheHWcave]

I had a look at the datasheet (rev D) and it should be ok but check the limits on isolation voltage and recommend PCB layout

[@tobienmarc]

Hello Thanks for the great explanation. I plan to convert 6 modules. Since I unfortunately do not get the R05P12S I have another question. Couldn't I run the 6 modules with their own small 12 V supply supplierend by on source? E.g. a HLK-PM12 for all 6 modules? All 6 modules are running on the same phase for my heating system.

[@TheHWcave]

Yes, that should be ok. I assume you would power the HLK-PM12 from a different AC source than the one you are measuring.

[@mosix76]

Hi, Suddenly my device no measures the current, but still measures the voltage correctly. I tested the CT with the tester, and it works. Thanks to the scheme you published, I was able to identify the elements related to the CT. I measured on PCB the values of R18, R11, C9, C11, C10, and R12: the resistance values are correct, and the capacitors are not in cc. Is it possible that the V9881D doesn't work? If so, why measures the tension?

[@TheHWcave]

A strange fault indeed. If you look at the datasheet of the V9881D, it has separate differential inputs and ADCs for 2 current and 1 voltage channel, so it is quite possible that voltage still works. If you are very careful, you may try to measure if there is a voltage between IAP and IAN when current is flowing through the transformer. That would prove conclusively if the fault is in the chip or in the circuitry before it. The analog inputs can only handle -0.3V to +5V. This means a huge current spike could possibly overload and damage the input. The same is true if for some reason R18 is missing or not connected to one of the CT input. A CT must be always loaded or it will produce damaging voltages. If the chip''s current input is damaged, I wonder if one could use the unused second current input but that also depends on the firmware being able to use it.

[@andremedeiros7136]

Hello. I want to connect the PZEM to a regular ESP32 via the RX2 and TX2 ports. Is a 5v to 3.3v voltage converter needed?

[@TheHWcave]

Don't connect directly to the V9881 because that is really very unsafe. That part is directly connected to mains. Use the opto-coupler that is on the board. Just feed it with 3.3V instead of 5V on the USB side if you want to avoid a voltage conversion. The LEDs should still light to make the opto-coupler work,. You can check by looking at the LEDs that are onboard. If they still blink when receiving and transmitting, then the optocouplers should as well.

[@eduardofranca5805]

This is a great video. I would like to put a ESP01 inside to the Pzem004. Yes, to install in the same original box and connect the ESP in the TTL using 3.3/5v and gnd from pzem. Is it possible? How?

[@TheHWcave]

I have no experience with the ESP01 but it sounds doable, sorry that is all I can say

[@indrapurnama8659]

Hi, what value of capacitor is used in c1-c8?

[@roolon6604]

English socket, German accent -Mods I`ve already done:) Same way. Shunt transformer with smaller hole/not demountable is more accurate. It`s better to use DC-DC converter 5v to 5v and recalculate divider R1 and R2. HT7133 is low drop linear converter, so 1,7V is enough.Converting 5V to 12V and then 12V to 3V3 it is a waste of energy. Thx

[@TheHWcave]

I agree that it's more efficient to skip the 12V stage and go to 5V directly after modifying the divider as you did.

[@D9ID9I]

How about to design a proper board so others could just upload files to some sort of jlpcb, order them and then re-solder missing parts?

[@sebydocky5080]

Great video.... As far I understood of the V9881D datasheet & the PZEM-004T V3, it's not possible to retrieve the sign of the current,voltage.... For photovoltaîc application would have been great

[@TheHWcave]

No, unfortunately it does not distinguish between flow-direction in the current transformer. Sign is maybe a bit misleading, because its AC, it is actually a phase shift. But I think there are modules available that can do this.

[@tuiastro]

A really nice customization. BTW, any chance to make the PZEM have bi-directional capability.

[@TheHWcave]

I must admit I don't know. I have not checked the market for some time and with solar now a lot more common, it would make sense to have a bi-directional module.

[@fibranijevidra]

Fantastic video. Thank you! I see you winded the wire multiple times trough the CT clamp. Why? I would expect that mess the measurement, but I am ibviously wrong.

[@TheHWcave]

It increases sensitivity for low currents. I have used the same trick in another power meter, maybe this video helps to understand it better: kzbin.info/www/bejne/hpOcd6Nrf5yfiqs

[@JohnMiller-cm5ne]

Great video. It looks like you have an error on the option 2 schematic. You seem to show connecting the DC to DC converter to +5 and TX, not +5 and ground. I know it's done correctly in the video. Again, great job on the video, keep up the good work.

[@TheHWcave]

Hi John, well spotted! I fixed the error and uploaded the new schematic to my Github page..

[@CarloCaddeo]

I really love your analysis and modification. Do you know if it is possible to use the module without connecting directly to AC? Connected only to CT to measure AC amps

[@TheHWcave]

The original module? No, because it needs to get its power from the AC power connection. After the modification? I have not tried it but it should work just fine.

[@CarloCaddeo]

@@TheHWcave yes of course After you're modification. Can you try It?

[@TheHWcave]

Just did, yes it works just fine.

[@CarloCaddeo]

@@TheHWcave thanks a lot. In this case I can Power the v9881d directly from USB cable?

[@TheHWcave]

@@CarloCaddeo You mean without DC-DC converter? No, because the circuitry of the chip is meant for 12V and its power management senses that. If you would never connect it to mains AC voltage, you could connect 12V directly to the input of the onboard regulator but you would still need a 5V supply to power the optocouplers, Better get a cheap 5V to 12V DC converter from Ebay. If you never use the AC volts, it does not need to be isolating.

[@Mr_ToR]

Awesome video. I have a question. Can I replace the 3.3volt 7133 Low Power LDO with a beefier regulator? I would like to use the VDD5 3.3V line to simultaneously power an ESP8266 chip, so I won't have to additionally supply an external 5V and I can simply use the lines going into opto-isolators directly with the ESP8266 chip without even having to power the isolators. It would be a self contained device. However the problem is that the 3.3v regulator, the 7133 Low Power LDO can supply at most 30mA. The ESP8266 can draw up to 320mA @3.3V. Can you recommend me something to replace it with? Would there be another component that I would have to replace as well? Also most importantly would this hinder the PZEM measurements in some way? Thanks a lot.

[@TheHWcave]

The 7133 or a replacement is not isolated from mains. In your suggestion the opto-isolaters would be useless, because either side would be connected to mains. Everything in your circuit, displays, buttons, interfaces would be potentially deadly to touch. Please don't do this. Alternatives. If you don't want to measure from zero volt, you can keep the original circuit and connect in parallel to the measurement connection a normal isolated power supply that feeds both the opto-isolators (you still need them!) with 5V and your new circuit with 3.3V. Make sure you connect the power supply so that its current doesn't go through the current transformer, otherwise you always measure your own circuit consumption in addition to the load

[@Mr_ToR]

@@TheHWcave The is a misunderstanding. The ESP8266 is a WiFi chip and there will not be any displays buttons or anything else other than the chip it self. Think of it as if it was already inside the V9881D. Imagine this chip in place of the LCD driver part of the V9801S version for example, except without any outside connection. No outside connection so no isolator needed. I'm eliminating any outside connection so I don't need the opto-isolators.

[@TheHWcave]

@@Mr_ToR Ok, fair enough, but the existing power supply using cap dropper an Z-diode is not able to support this. It would need to burn 60 to 70W . Best to use separate isolated transformer to get say 6V AC which you then rectify and feed into a standard 3.3 V regulator. I would not use a SMPS for that because you have to connect the negative DC back to mains and with (dodgy) SMPS you never know how well isolated the DC output is. If the SMPS leaks L to neg. DC , you effectively short-circuit mains.

[@Mr_ToR]

@@TheHWcave Great, thank you so much for this information. Yes, that zener diode and the half bridge rectifier. Thank you for pointing that out. I have to get myself more familiar with that kind of non-common power supplies. The main issue is that I already have 10 of these PZEM devices throught the house. All interfaced to that ESP8266 WiFi chip with the serial connection where both the chip and the optoisolars are powered by cheap 5V usb power adaptors.With the wiring I'm using PZEM measures the mains load as well as the it's own consumption and the 5V adapter load as well. The 5 V adapters consume about 1~1.5Watts. However they add up in quantitiy. I have 5 more PEM devices I have not installed yet. I was hoping to find a more efficient and easier wiring method. 15 x 1.5 W makes ~+20Watts.

[@kamilsobolak3316]

@@Mr_ToR hey! you touch good point, how your story end up? I do also look for simplification of ESP8622 power up.

[@kemalaryapanduwicaksana8402]

How the process of this sensor can produce the power factor value?

[@TheHWcave]

The V98xx chip measures current and voltage continuously. It multiplies current with voltage to get active power. In a second part of the chip the current signal is first phase shifted by 90 degrees before being multiplied with voltage. That gives reactive power. The power factor is then calculated as the ratio between active and reactive power.

[@hobby-electronics-repair8655]

rectification using the forward conduction of the Zener is not "wasteful" as you state ... it is in fact necessary to avoid DC current flowing in that series cap that has the main purpose of reducing the voltage...if current would be only + halfwave, then series capacitor would then overcharge and would cause the circuit to malfunction

[@team69ch0tu9]

Hi, this is a really great Video. Can you tell me the value of the capacitors from the schematic? I need that information for a project I am working on. Thanks

[@TheHWcave]

Sadly these are all SMD caps with no readable marking. Since they are in-circuit, you can't just measure them reliably because the other components will influence the result. The only way to be sure is to de-solder them and then measure, but, sorry, I am not going to do that. There are some testers that claim to be able to measure capacitors in-circuit by using very low voltages, below where most(!) semiconductors start to conduct, and that may work in some cases but it wont make a difference for networks of capacitors and resistors as in this case.

[@nitinkumar29]

At our location, we have voltages fluctuation and it can fluctuate from 210V to 280V (higher voltages during night when there is less load especially in winters, local grid transformer get in flames sometimes under load so I have seen 110V twice in a year. yes, we have very poor grid supply). Therefore at home, we are using a voltage protection relay with a beefy contactor to limit supply to 200V-270V, thereafter RCCD is there along with the MCBs, furthermore all sensitive electronics and low powered electrical devices is powered through a phase out of 3 phase which has a servo transformer based voltage stabilization (230V output) which further powers an inverter with 24V battery bank for backup as we have power cut in our city as well. I am looking for a solution to measure 3 phase power supply between 200V to 270V so I brought 3 PZEM. I thought of going with Option 2 and powering all 3 PZEMs with a power supply using a old fashioned way of a locally available transformer with primary and secondary windings and bridge rectified it to get 12V using voltage regulator (no smps). And maybe using another transformer based supply to power ESP32 and RX, TX side to keep opto-coupler side isolated too. Though both transformer will be powered using same phase supply which comes from inverter and always on even during power cut so I can measure power cut state, length of power cut beside voltage and current monitoring per phase for further automation in the home. I'll change the resistor to 0.6 watts at input as suggested. Though I'm not sure if that will be safe in measuring 3 phase power supply or not, can you suggest if I can go with this plan?

[@buimanhha3741]

To bother you again. This time I have another project I want to ask you: When the current is greater than 100A(sometimes the current is about 120A), will the pzem explode or just measure it wrong? Thank you so much!

[@TheHWcave]

Good question. 120A instead of 100A will generate an overvoltage on the current input. I don't think that will cause any permanent damage. I was trying to test that by running a 2KW heater at 240VAC in x10 mode but that is still just 8.3A. I don't have anything handy that would bring it up to the >10A. Anyway, if you want to try it out, do maybe 15 turns in the CT and then a 2KW heater will produce 124A. It would be great if you could report what the chip reports in this case. I was also thinking how to make a x0.5 mode. If you measured the length of the wire that goes through the CT and used a 2nd piece of the exact same wire and length and run it in parallel but NOT through the CT, you would theoretically now have a 200A measurement range and you would need to half all current and watt values you read. How accurate that x0.5 range is, I can't say. I guess how equal the current splits between the two paths would not only need the exact same length and diameter but also depend on the contact resistance where they join together.

[@buimanhha3741]

Really thank you very much. I also thought of running 2 wires in parallel, but it is quite complicated because the machine capacity is very large and the power supply wire is very large. I will try to replace original CT with CT HCDT SCT24TS(200A, 2000:1, 10omh). Then the current and power values ​​are multiplied by 2 to get the actual value. When the project is completed, I will report back the results. Thank you very much!

[@awebuser5914]

Just wondering if the voltage input is relevant at all if you are just doing home power use testing? (you already know the voltage and the daily variance is minimal) Can you delete the hot/neutral input circuit completely and power via USB? This will make it easier to implement temporarily in a fusebox, etc.

[@TheHWcave]

the chip needs mains voltage and current to calculate and display power (simply put V * A = P ) but of course its more complicated in AC circuits...

[@TheHWcave]

The problem with AC power is that current and voltage are usually not in phase and not sinusoidal either. This chip has the necessary signal processing (which is quite a lot and complex) on board to take care of all that. Simply "assuming" 120V and multiplying the current with that will give you wrong values. As far as I remember, the module will give current if voltage is zero (provided it is powered from some other source)

[@awebuser5914]

@@TheHWcave I have an Amprobe clamp-meter so I can verify the accuracy of the readings, so I'm not worried there. Most household loads are measured just for curiosity and power budgeting, so it's not like you need significant accuracy, as long as you're in the correct ballpark. Having the AC input completely deleted would make implementations far safer and easier to use.

[@TheHWcave]

@@awebuser5914 As I said, to accurately determine power in an AC circuit, the phase shift between current and voltage is essential. Measuring voltage separately from current is not going to work. Please take the time and read-up, for example the Wikipedia entry on AC Power en.wikipedia.org/wiki/AC_power

[@NeerajSharma-ko5wv]

I have an energy meter which has V9821 IC. I have gone through the datasheet for V98XX which covers both V9821 and V9881d ICs. As per the diagrams in the datasheet I've observed that in V9881d pins Rx and Tx used for UART communication are UART2 pins. which are named as P2.4 RXD2 and P2.5 TXD2 in the datasheet. and these are for IR communication because UART2 by default is set for IR communication. how pzem-004t module is able to provide this functionality to connect it serially instead of IR. Is it possible that both have different firmware? Could it be possible to run same command over IR communication to read the metering data?

[@TheHWcave]

Sorry, I have never looked into this in detail and can't answer this.

[@NeerajSharma-ko5wv]

hi TheHWcave, thanks for your reply.

[@malaparnass]

Just played around with this item. I've removed the 3.3V regulator and piggybacked a galvanic isolated DC DC Converter 5V -> 3.3C onto the Optocouplers. connect output to 1 ( gnd) and 2 (3.3V) of the regulator pads. What I have an issue with it on how to determine power flow direction. Anyway I connect the CT coil it just show _positive_ energy flow.

[@TheHWcave]

Yes, unfortunately the firmware doesn't seem to distinguish flow direction. It was probably written when nobody thought of solar installations.

[@georgiypanaetov9497]

Good afternoon, please tell me why you replaced the 1 MΩ resistor with two series resistors of 510 kΩ and 470 kΩ. Does this affect the measurement accuracy? I want to redo it according to your scheme, but use a phone charger. The only question is, if the phase is reversed at the input and the neutral wire will not damage the unit and power supply and the module itself? Thank you very much for your work.

[@TheHWcave]

I measured the 1 MegOhm resistor and it was only 980K. Since the unit was calibrated in the factory with this resistor, the replacement has to match it or it will be off. But you need to measure your R17 because your resistor may have a different value. So don't just use 510K and 470K like I did. You need to come up with your own substitution and please use resistors rated for 350V (or more). I would NOT use a phone charger. It has to be a properly isolated power source and phone chargers are notorious for not being very good in that department. The 12V input on the module is connected to mains. It would be far safer for you to leave the capacitor dropper in place (do only option #1) and power the module through its original pin on the terminal block. Swapping live and neutral has no influence on the original module or my modified one. But if you use a dodgy phone charger, it might cause damage.

[@georgiypanaetov9497]

​@@TheHWcave Thank you very much for your answer! I use this module to measure the voltage at the output of a car inverter - there is a rectangular sine wave and the capacitor power supply does not work well (the current-limiting resistor, the largest on the module, heats up). This resistor burns out during continuous operation.

[@georgiypanaetov9497]

@@TheHWcave But if the phone does not fit, can you use a classic linear power supply on a conventional transformer?

[@TheHWcave]

Yes, that is possible, but why go through the trouble of making a separate power supply? You can use the one that is already onboard of the PZEM-004T. If you need to measure voltage from zero, just do option 1 in my video, but power the PZEM-004T as normal. However, in your case I don't think you even need to measure from zero as the inverter will always put out the same voltage. You can just use the PZEM-004T without any modification

[@georgiypanaetov9497]

@@TheHWcave Thanks for the answer, I plan to use these modules for different purposes. It is very convenient when the module is powered separately from the source, this allows you to control the moments when the device is turned on and off.

[@danhhuy2312]

Great video ! Please tell me why Pzem can measure Frequency , i dont know this .Thank you so much !

[@TheHWcave]

You may want to use it with a generator where the frequency isn't quite as stable as from the power grid. So its nice that it allows you to check what frequency is being produced.

[@jensschroder8214]

10:00 The zener diode ensures that the required 12 volts are generated in the positive half-wave and in the negative half-wave the zener diode works like an ordinary diode and charges the capacitor negatively. The capacitor must always see AC otherwise the capacitor power supply will not work. It is not possible to connect a diode in series with the capacitor, because then the capacitor will no longer be discharged. A full capacitor will stop the current flow. The zener diode can pass current in both directions. Once with 12 volts and the second direction with 0.6 volts. Therefore the capacitor sees AC and it works. But because 12 and 0.6 volts is an imbalance, the capacitor still needs a discharge resistor.

[@paulcameron7629]

Hello, I hope you can help me, I have this sensor and I want to know how to calibrate it, because when I connect it to the 220Vac network it is giving me readings of approximately 245Vac, but this does not happen with the current that it measures correctly, (compare it with a laboratory instrument). Thank you very much in advance and greetings from Perú

[@TheHWcave]

Which country are you in? In a lot of places it is 230V and in some 240V. The UK is 230V but at the moment I measure 237.8V. I think the tolerance is +10%/-6% so over 250V may be possible, all depending on daytime and network load and changing all the time. What I am saying is the 245V may be correct. You should check the voltage with a reasonable meter at the same time as the module to see how far the module is off. 2nd question: is this the original module or did you do the mod? In case of the latter, did you measure the removed 1M resistor because that needs to be matched as close as possible with the new resistors and in my case the original resistor was way lower than 1M. Since the module was calibrated in the factory with that resistor, the replacement resistor needs to be of the same value.

[@NetworkGeek280]

Is it possible to measure the current direction? This sensor outputs a positive value regardless of CT clamp orientation. It'd be nice to use this device to measure the mains import/export of a house with solar.

[@kirillodessa7668]

Можно ли просто подключить напряжение питания 3,3 В к разъему P2 (контакт 4)? Тогда не понадобится преобразователь уровня для подключения PZEM к ESP01.

[@TheHWcave]

Possibly, it all depends on the current drawn by the ESP-01. I had a look at the datasheet (never used one myself). It can draw up to 170 mA worst case. That would probably be too much. The lowest TX power draw was still 120mA. Quite a lot. Not sure if that would work

[@YogeshkumarGobinathan]

Fantastic video sir , I am able to follow the same Method for making Zero Voltage measurement and it is working Good . But sir, I want to measure 440vac and switch of a phase to protect .It is possible using PZEM004t? In a 3 phase supply system, when the Neutral is lost at the utility junction point, by some reasons, devices connected in one phase will get 440V thru the neutral inside the premises getting another phase via common neutral connection. Hence this requirement

[@rodv92]

I got in touch with Peacefair to order a sample rated up to 437V rms. That would allow to make line to line measurements possible and also line to neutral in case of a neutral fault in a 3P+N setup. Unfortunately, I hadn't seen this video at the time of the order so I could not send them the feedback. I wonder if they changed the layout and resistors for such a high voltage rating. i'll probably make a video of the test when i get the samples.

[@TheHWcave]

I am curious about the modifications that make it rated for 437Vrms. Please share when you receive it.

[@rodv92]

@@TheHWcave So i got the modified module. The safety yellow X1 capacitor C15 was changed to a brown 630V film capacitor. As for resistors, only R16 was changed to a 470 ohms, which gives a top measured voltage of only 427V instead of 437V i requested. I guess the engineer didn't have tailored SMD resistors at hand ;( Besides that the vango 9881D SOIC datasheet must have an error for the UP pin max measured voltage, it should be 200mV RMS not peak) with a 750 ohms R16 resistor on the unmodified voltage divider gives the proper max RMS voltage of 266,86V. I did not notice majors changes on the powering part resistors uphill of the zener diode. R15 is strictly the same. R14 the big thru resistor is the same wattage it seems (same form factor but is beige so it has been probably swapped to a carbon film resistor). As for the layout and traces I did not notice any change. I will power it up between two phases and wait for fumes (with a ceramic fuse upstream). I hope that the engineer modified the SOIC firmware for the new voltage rating at least. Will keep updated.

[@TheHWcave]

@@rodv92 Many thanks for providing us updates on that modified module. I am not surprised that they had to change C15 as the original was only rated for 250VAC. It acts a capacitor dropper to reduce the voltage for the Z-diode that creates the raw supply voltage. R15 is just there to discharge this capacitor safely so you don't get zapped if you touch the circuit after power was removed.

[@buimanhha3741]

Happy new year! I use 3 phase, 200V power supply . I can use this device to measure the power consumed between line1 and line2(two phases) . Thanks!

[@buimanhha3741]

Sorry, my English is not good. I want to ask you: can I use this device to measure the power between line1 and line2? ( two phases - 200v). Looking forward to your early response. Thank you so much!

[@MrArif009]

you have done a really amazing job. but the problem is dc to dc converter, it's so expensive. can I use a normal XL6009 boost converter?

[@TheHWcave]

Sadly no. It needs to be a good(!) isolating DC-2-DC converter, or you have live mains voltage connected right into your PC & Arduino etc. It will probably kill your equipment and other stuff that may be connected including the user (you!)... Instead please just stick with option 1 (if you want to measure from 0V) and keep powering the module through its capacitor dropper.

[@MrArif009]

@@TheHWcave thanks. Please don't mind can I ask you a question Are you an Electrical engineer? I'm asking because your programming skill is also amazing. (I am an electrical engineer student but my programming skill is not as good as yours. How can I improve, is any suggestion)

[@TheHWcave]

@@MrArif009 I studied electrical engineering but even in those long gone days I already found programming very interesting. Initially I worked mostly as software engineer but for a long time I am now a Systems Engineer (designing complete systems) and do programming and hands-on electronics only as hobby. Don't worry too much about your programming skills, good hardware engineers are still in high demand, possibly even more than ever. One thing I would recommend is getting a popular Raspberry Pi or Arduino and a couple of breadboards to add peripherals, like LEDs, switches, sensors etc. There are plenty of tutorials out there how to program them. Don't worry if you struggle understanding a tutorial, its probably not you but the way it is explained. Just go to another tutorial, until you find one that makes sense to you. Start with simple projects like making a LED blink and then add things, like changing the pulse width or making the speed depend on temperature and such. As an electrical engineer you will understand the hardware which is a big advantage.

[@buimanhha3741]

Dear! I use 3 phase 200v power supply. And I want to ask you: Can I use this device to measure the power between line1 and line2? ( two phases - 200v). Looking forward to your early response. Thank you so much!

[@TheHWcave]

Not without extensive modifications. If the phase to neutral voltage is 200V then the phase to phase is 346V which is way too much for the original module (the big yellow capacitor can only handle 250V and the chip cannot deal with such high inputs either). You could replace the yellow cap with a capacitor of somewhat lower capacitance but that is rated for at least 400V (lower capacitance = higher Z because it needs to drop more voltage). Also the resistor divider at the voltage sense input of the chip needs to be changed to reduce the measured voltage. Probably easiest (but bad for resolution) is to divide by an additional factor of 10, then the voltage shown and the power values needs to be multiplied by 10 to get back the correct value. Considering all this, I recon it is easier (and safer!) to get a module that is meant for higher voltage measurements.

[@buimanhha3741]

First of all, thank you very much. Maybe because I didn't explain clearly, you misunderstood. Phase to phase voltage is only 200v. So can I use this device to measure power between 2 phases ? One more time thank you very much.

[@TheHWcave]

@@buimanhha3741 200V is no problem. I am using it with 240V. Of course this will need to be installed floating with respect to ground. As the USB output is opto-isolated it should be fine.

[@buimanhha3741]

@@TheHWcaveI am working for a Japanese factory. The phase to phase voltage in the Japanese power grid is close to 200v. Today, I installed the device at the factory. it runs well. Thank you so much !

[@raflyakhsani]

Excuse me, what if i plug the output pin in reverse?? L pin on neutral and N pin on load? thank you..

[@TheHWcave]

This is no problem for this module. It will work either way.

[@raflyakhsani]

@@TheHWcave Let me ask again sir, I have taken data from the sensor (Voltage, Current, Power, and Energy), then I communicated it via the MicroSD module. When I took the data, I tried to multiply V by I with the basic formula P=V.I, but the results did not match the data I had taken on the MicroSD. Can this be considered normal? I think that the sensor installation is appropriate and there is no interference whatsoever. Thank you in advance.. 🙏

[@TheHWcave]

@@raflyakhsani The module measures active power as well as RMS volts , current and the phase angle between them. On an AC circuit multiplying RMS volts with RMS current only gives you watts if the phase angle between volts and current sine waves is 0 (=power factor 1). At any other time you get apparent power (VA not Watts) The module does a lot of signal processing and number crunching to get this calculation right for all phase angles (power factors). The only time you will have a power factor of (nearly) 1 is with a pure ohmic load, i.e. an old-style incandescent light bulb. These days, nearly all loads have a power factor of less than 1, especially cheap and low-power devices like LED bulbs, chargers, etc. If I use a 6W LED bulb, for example I get 233V 0.043A, Pwr = 5.36W, power factor 0.54. If I multiply volt and current I get 10 VA but the active power (which you pay for) is just 5.36W .

[@ahmadmuhaimin2852]

great video, but can i change the current transformer?

[@TheHWcave]

You could but the unit has been designed an calibrated to use the one that comes with it. What you would need to do is characterise this transformer by measuring how many mA AC it produces when a known AC current passes through a conductor going through the core. If your replacement transformer has the same ratio, the unit should work correctly.

[@TechnoloJ1M]

My PZEM-004T shows 0.23A of current even I am not using any power in the wire where the sensor is attached. Even if I remove the CT, the current of the sensor does not return to 0, it just gives a random value of current from 0.20A to 0.50A.

[@TheHWcave]

Sounds as if the chip has a problem. While there is supposedly a calibration mode, I have no idea how to get into it or use it. If you still can, I would return the module.

[@elpeepo8844]

Would a MT3608 work instead of the R05P12S?

[@TheHWcave]

No. The reason I chose the R05P12S is that it isolated to 3.2KV. It protects you and the devices you have connected on USB from mains voltage. Most DC-2-DC converters including the MT3608 are not and you will blow up your USB device and get a serious shock if you would use it (or any similar non-isolated converter)

[@elpeepo8844]

@@TheHWcave Oh I see, guess I will have to stick to the original design then since there is no stock (aliexpress isn't helping either). Thank you.

[@TheHWcave]

@@elpeepo8844 I understand but I would not buy such a critical component from eBay or Aliexpress anyway. This thing is essential for your safety (and survival of your connected devices). Better get it from a proper electronics distributor (far less likely to get a fake)

[@Jihenguesmi]

can i use this module in proteus simulation?

[@TheHWcave]

No idea, never used Proteus. But I would use some other circuit because with this one all the smart parts are inside a chip. All you could simulate is the power supply part.

[@Mr_ToR]

Would it cause a problem to switch Live and Neutral inputs?

[@TheHWcave]

No, it would make no difference. You can still use this circuit in countries where main plugs are not polarized, e.g. the euro-plug can be inserted in to a mains socket either way around.

[@Mr_ToR]

@@TheHWcave 👍

[@Mr_ToR]

@@TheHWcave Thanks a lot for your quick answer.

[@Mr_ToR]

@@TheHWcave I would really appreciate it if you would take a look at my other question as well about replacing the regulator?

[@Dialbist]

Please help, what is resistor R18? 1206 005 Om?

[@TheHWcave]

It should be 0.5 Ohm. My schematic says 0.05 which is an error. The resistor is marked R500. I need to fix the schematic, thanks for pointing it out

[@regolitharakiri2751]

@@TheHWcave I accidentally mixed up the voltage input and the current input. There was a short circuit, but fortunately no one was hurt, only an ugly stain remained on the table. It turned out that this R18 resistor literally evaporated along with the adjacent tracks. The rest of the parts were not affected. However, one of the 100-ohm resistors through which the current signal goes to the microcircuit increased its resistance to 160 ohms and had to be replaced. Not knowing the value of R18, I started looking for a circuit and found it with you. After carefully rubbing off the soot and sprayed copper from the tracks (this is a conductive layer), and also taking precautions, I connected the module to the network with a 0.05 ohm resistor. To my surprise, the module still worked. The voltage showed correctly, but the current readings were much lower than the real ones. When, in addition to the diagram, I watched this video, I realized that the denomination was indicated in the diagram erroneously. Therefore, the readings were 10 times smaller. After replacing the resistor with 0.5 Ohm, everything returned to normal. Thus, if PZEM004 has a linear current error, then it can be easily corrected by connecting, for example, another one in parallel with the R18 resistor. By the way, in my opinion, its size is not 1206, but 0805.

[@regolitharakiri2751]

When I got low current readings, I reasonably thought that the analog input of the microcircuit was damaged and it might have to be replaced. These chips are commercially available and relatively inexpensive, and I was tempted to assemble my own device, because the circuit is simple and other parts are easily accessible. if the chip is already programmed, you only need to calibrate it. It's not easy, but I think it's possible. But if it still needs to be flashed, then the problem arises to find the right firmware. Most likely, you will not be able to copy it from the working module.

[@TheHWcave]

@@regolitharakiri2751 Wow, good story and good recovery. I added a note in the description of the video that R18 should be 0.5 Ohm not 0.05

[@abdelkarimkhelifi863]

I want to modify my pzem to read Voltage between two phase and also read frequency 200Hz Can you help me please ?

[@TheHWcave]

While it could measure the voltage between 2 phases provided it does not exceed 260V and while 200Hz are beyond its spec but my unit was still reasonably accurate at that frequency, I am not sure what you want to do with current?

[@abdelkarimkhelifi863]

@@TheHWcave I want to connect the pzem to terminal of a VFD without neutral The voltage between two phases is just 230V and frequency is 50Hz But when I tried to connect the pzem to VFD it's make a fire. I don't know why, is the pzem affected by the acceleration of voltage and frequency ? because when the VFD is running, it accelerates each of them from 0 to 230V and 0 to 50Hz If there is a solution to measure voltage between two phases under 250V i want also modify my pzem to read Frequency 200Hz

[@TheHWcave]

@@abdelkarimkhelifi863 VSD are quite hard to measure . If you use the original non-modified PZEM-004 it wont work because it gets its supply from the voltage difference between the L and N or in your case phase 1 and phase 2 and is furthermore frequency dependent. It will get too much power at 200Hz. I am not sure what your setup is varying with respect to the phases. Check what a good (!) true RMS (!) voltmeter show if used instead? That would give an indication if there is any chance to use the PZEM-004 powered from USB. There are just too many unknowns to give any further advice in comments. That could end up to be really dangerous. Please get someone professional to look at your setup..

[@apollorobb]

As a note on this i found that the power management on all 5 of the units i have will activate from 8.5-12v

[@marcus3d]

I don't understand, why does the DC-DC converter need to be able to withstand continuous high voltage isolation? Isn't it connected to neutral? Where would the high voltage come from, and in what circumstances?

[@TheHWcave]

The converter is all that separates and protects you (and your PC/laptop) from mains voltage. 240VAC (RMS) is 340V peak or 680V peak-to-peak, so the voltages are substantial. Add to that momentary spikes and transients that you get when somewhere large inductive or capacitive loads are turned on or off. Its true that specifying a higher break-down voltage is maybe a little belt and braces but the costs are not that much higher and I'd rather have a good safety margin between me and mains voltage. Ok that's me, other people think maybe differently. The main thing is to get whatever isolating DC-DC converter you chose from a reputable distributor and not some questionable source on eBay, aliexpress... A lot of components these days are fakes that do not meet even minimum standards.

[@marcus3d]

@@TheHWcave Again, where would that high voltage be coming from? I mean, what's its path from L to one of the DC-DC output pins? Wouldn't that require at least one of the other components breaking down first? And if so, could we make a fuse there to detect that (or actually the mains fuse should get tripped anyway, because it gets shorted to N, right?), so that if there is suddenly high voltage then the whole thing would disconnect, so the DC-DC would never need to isolate for more than a millisecond or something?

[@TheHWcave]

@@marcus3d Not sure how to put it and please don't get me wrong but after reading your comments I can only I strongly recommend that you should not try this mod or work on anything that is directly connected to mains until you learn more about such circuits and the dangers involved. I assume you will ignore that advice but its all I can do.

[@marcus3d]

@@TheHWcave Fair enough. Still, which path is the constant high voltage AC reaching that DC-DC?

[@TheHWcave]

The whole chip and everything that connects to it has to be regarded as being on mains voltage. The chip is directly connected to neutral and only a very dodgy and not mains voltage rated resistor is between it and live. If that resistor fails or if someone accidentally swapped neutral and live anywhere in the wall or power cord wiring (I have seen many examples of that) ,,,

[@zaynabds]

The schematics is slightly different with brochure which the Line placed on 3rd pin not 4th pin

[@TheHWcave]

in principle it does not matter and many countries don't have polarized mains plugs where L and N can interchange depending how the mains plug is inserted. The UK has polarized mains plugs where L and N are always the same. I Ilke connecting N to the common and L on the pin that is subdivided by resistors. Call it personal preference but I find it neater and makes the rest of the circuit (marginally) safer ... but applies only if your plugs and sockets work that way (and are wired correctly)

[@marcus3d]

@@TheHWcave It also applies if you're not using plugs, but wire it directly, in which case it's particularly easy to see which is which.

[@youv4656]

why R12 is connected between the ground ?

[@TheHWcave]

The AD converters inside the chip need a common potential for measuring both the voltage and the current

[@tobienmarc]

Hello Thanks for the good video and the explanations. I have a question which is not yet properly understood by me. The removal of C15, R15, R7,D1,Z2 and E4 prepares the new power supply 12 V. I am not concerned about the power supply but idle current and heat generation. Can anyone give me information on this? What is the difference between original and modification here? Point 2, I may have missed the first time. Why is R17 replaced by 17a and 17b, the part with the approximation of the original values because of the calibration I have understood. Is it because of the height of the voltage drop? Why then not also R16, the voltage drop is L>R17>16>N. Does anyone know the operating voltages of R17. I intend to connect 240V. THX

[@TheHWcave]

In the original the power supply for the electronics is created through a capacitor dropper (C15 in series with R7 and the Z-diode). This is done before rectification by D1 which means at one half-wave the Z-diode is actually in forward direction and that energy is simply converted into heat. Its not terribly large but unnecessary and wasteful. Most of the voltage drop happens in C15 which is physically large and can handle the generated loss easily. R17 and R16 form a divider for mains voltage with a ratio 1M/750 = 1333:1. At 240V, R17 drops about 239.8V and R16 0.2V. A tiny SMD resistor for R17 is not rated to drop that kind of voltage safely. It may simply creep across. You need a physically larger resistor or a series connection of multiple resistors where the drop across each is less. For safety, I recommend both, larger (and properly rated ) and splitting it into 2 series resistors. To maintain calibration the replacement R17a + R17b must match the original value of R17 (which wasn't 1M in my version)

[@sparsh_editx]

​@@TheHWcavehey i saw you said your resistance was 1 M my chip has 100 ohm resistance only brown black brown golden colour code

[@TheHWcave]

@@sparsh_editx That would be very strange because that high-ohm resistance is needed. Have you measured it (with nothing connected to the module)?

[@vvg2010]

Hello. Pay attention to the top mounting diagram of 10A in the minute 3:55 to 4:40. There is an error connecting N to PCB. The manufacturer's erroneous diagram connects L and N to the central contacts that are together in the PCB (see minute 7:24 to 7:29) causing a short circuit. To solve this, connect N to the lower connection where the load is connected. Regards.

[@Mr_ToR]

The diagram you've mentioned is for the 10A model. However, the device in the video is the 100A model. The circuit differences between these different models were mentioned in the video.

[@kemalaryapanduwicaksana8402]

Great videos 🤩. Can you help me out by explaining the use of each component (eg resistors, capacitors, diodes, etc) that make up this pzem module?🙏

[@TheHWcave]

This circuit is a bit too complex to explain what all the components do. Also most of the good stuff happens inside the V9881D chip and the 7133 and all you see are a few external components. From those that do something more interesting, the big yellow capacitor C15 acts as a resistor (a few Kilo-Ohm at AC mains frequency) to limit the current through the large 100 Ohm resistor and the Z-diode. A capacitor is used here instead of a resistor because it can do this without getting very hot (like a resistor would). This method is known as a capacitor dropper circuit and very common to drop (=reduce) a large part of the high mains voltage. Capacitors do this by shifting the phase (timing) of the AC current relative to the AC voltage, but that's another story. The circuit of the Z-diode and D1 is unusual because the Z-diode is driven with AC current so D1 sees an AC voltage capped to +12V while the negative half-wave is much higher, but D1 will only let the +12V pulses through to charge the big electrolytic caps which form the input to the regulator. The 470K resistor in parallel to the big yellow capacitor is only there to discharge it quickly when AC power is removed. Otherwise you could get a nasty shock even after you disconnected everything. There is more, but comments are not a good way to explain it. If you look at other channels, EEVblog or Bigclivedotcom, you find good beginners videos.

[@pepethefrog7193]

Why did u use a 5V to 12V DC-DC and not a 5V-5V? The 12V will be regulated down to 3.3V anyway and vdcin already switches at 1.0V.

[@TheHWcave]

Because the V9981D takes the 12V (through a divider) to its VDCIN pin to do power management. It will go to sleep if it does not see the right voltage.

[@pepethefrog7193]

@@TheHWcave Would the B0505S ($1 on ali) be suitable adapting VDCIN divider?

[@TheHWcave]

I take this you would make your own divider that uses only 5V but delivers whatever the VDCIN pin sees at the moment? Yes I suppose that would work, I believe the VDCIN voltage is less than 2V but I can't remember exactly. I found a number of different data sheets for the B0505S and they are all a bit different in terms of isolation voltage which is a bit worrying as you don't really know what your are getting. www.kosmodrom.com.ua/pdf/B0505S-1W.pdf www.elipse.eu/wp-content/uploads/2020/01/b_s-1wr3_datasheet.pdf As a start, you don't need 200 or 300 mA, 10mA should be good enough. The MORNSUN datasheet lists a minimum current of 20mA. Some switch mode supplies do need a load to work. If so, the load presented by the module is too low to kickstart the converter. Of course you could put a 250 Ohm resistor in parallel of the output to draw more current... The other datasheet doesn't mention it. The isolation voltage of this thing isn't great in either datasheet but the MORNSUN device is at least testes for 1 minute at 1500V (the other one tested only for 1 second at 1000V), so the MORNSUN B0505S-1WR3 is a clear winner over a MICRODC B0505S-1W I personally would be very weary of dodgy sellers trying to sell you a fake MORNSUN 1WR3 when its really only a MICRODC 1W . I would get such a safety critical item from a proper distributor (RS/Mouser/Digikey ... ) and not from Ebay or Ali .. but that is me. From a pure technical point of view, it should work if you do the VDCIN adjustment.

[@pepethefrog7193]

@@TheHWcave Bought the " DDS238-2 ZN/S single phase Din rail KWH Watt hour energy meter with RS485 MODBUS-RUT" which should be able to measure reverse amps. $20 with split-core. Times 3 phases :(

[@TheHWcave]

@@pepethefrog7193 Nice.. Can you post a comment how you like these once you receive them?

[@electroboyIND]

If Phase/Neutral is Interchanged , You can see some fireworks on BUCK converter/Secondary MCU

[@TheHWcave]

Please explain why you think that.

[@electroboyIND]

Are you saying ,the incoming L and N , if interchanged doesn't affect the circuit ? Assume a bad electrician connects phase to nuetral and neutral to phase or some wiring interchange happens by some mistake. Am not saying your correction/upgrade is wrong. Am saying that the product as it is gets damaged when L and N gets interchanged.

[@TheHWcave]

@@electroboyIND I understand and appreciate your concern. Yes, electricians and more likely non-professionals make wiring mistakes (in fact I once found a shop-bought extension cable with L-N swapped!) but the mains part of this circuit is completely floating and not connected to anything else (opto-couplers and isolated DC-DC converter) so a L-N swap has no effect

[@raflyakhsani]

@@TheHWcave oh, I found the answer.. thank you 😄

[@jensschroder8214]

If you want to improve the circuit then all you have to do is replace the capacitor C15 with an X-type with higher withstand voltage.

[@stevenm45]

Firstly, a very helpful video, thank you. I have a couple of PZEM-004T v3 on the way to me. I shall modify at least one of them as described in your video. I plan to use one for up to 10A measurement with either 10 secondary turns (ref. kzbin.info/www/bejne/hpOcd6Nrf5yfiqs) or, and the reason for writing and my preference, increasing the value of the 0.05 ohm burden resistor R18 to 0.5 ohm. Had you considered doing this? I also have a desire to be able to monitor Earth Leakage Current in real time by coupling the current transformer around the live & neutral conductors of the target device or distribution board. For example I would like to be able to monitor earth leakage currents up to say 500mA (full scale) as I have a pump that the manufacturer requires a 300mA RCD. In the 500mA case this would require a 10 ohm burden resistor. I appreciate that there are dedicated Earth Leakage Current monitoring and SoC chip devices but the PZEM router might be quicker and cheaper - and I can bring the readings in to Home Assistant.

[@TheHWcave]

R18 is already 0.5 Ohm, I made a mistake when doing the schematics. On my board it is labelled R500. To make the system more sensitive, it is probably easier to use the trick I used in mine, which is doing 10 turns through the current transformer instead of of one turn.