Це відео не доступне.
Перепрошуємо.
445 Can we trust our mains' power meters?
- Додати в
- Мій плейлист
- Переглянути пізніше
- Поділитися
Поділитися
Вставка
Розмір відео:
- Опубліковано 15 сер 2024
КОМЕНТАРІ • 329
Наступне
Автоматичне відтворення
The search for the biggest shape in the universe.Stand-up Maths
Переглядів 33 тис.
New Tools That Can Make Your Life EasierAndreas Spiess
Переглядів 64 тис.
444 Lower the Energy Bill, Keep the Standard of Living (Consumption Optimization)Andreas Spiess
Переглядів 54 тис.
Я тоже так могуDouble Bubble
Переглядів 2,3 млн
Втрачене дитинство | GOVOR TikTok #govor #shotsGOVOR
Переглядів 402 тис.
Яшин - интервью после тюрьмы / вДудьвДудь
Переглядів 7 млн
Путин о боях в Курской области: Главное - выдавить противникаAKIpress news
Переглядів 1,5 млн
Smart Meters are worse than you think (UPDATED)ScottiesTech.Info
Переглядів 1,4 млн
460 Seven Sensors tested: Measuring Current with Microcontrollers (Arduino, ESP32, ESP8266)Andreas Spiess
Переглядів 97 тис.
467 Radar Sensors from $3 to over $100: Which one is Best?Andreas Spiess
Переглядів 170 тис.
(483) ESP32 precision GPS receiver (incl. RTK-GPS Tutorial). How to earn money with it (DePIN)Andreas Spiess
Переглядів 529 тис.
These Illusions Fool Almost EveryoneVeritasium
Переглядів 1,7 млн
The Most Important Circuit for our Electrical Future?! (PFC) EB#55GreatScott!
Переглядів 878 тис.
477 Use the very attractive new ATTINY chips for your projectsAndreas Spiess
Переглядів 204 тис.
451 Which Processor can kill the ESP32?Andreas Spiess
Переглядів 343 тис.
459 Radar Sensors and Summer BreakAndreas Spiess
Переглядів 119 тис.
«Це край, де я родилася й живу! І нікуди я звідси не поїду» #shortvideoСуспільне Запоріжжя
Переглядів 355 тис.
How I Did The SELF BENDING Spoon 😱🥄 #shortsWian
Переглядів 35 млн
SCHOOLBOY. Последняя часть🤓⚡️КАН АНДРЕЙ⚡️
Переглядів 10 млн
Jerry Heil & Volodymyr Dantes - ГУБИ У ГУБАХ (ПРЕМʼЄРА 2024)Jerry Heil
Переглядів 1,2 млн
«Приїхали Бандеру шукали. Який Бандера? Він помер взагалі-то. Ні, от сказали, у вас тут Бандера є»Суспільне Дніпро
Переглядів 103 тис.
Кадиров пішов НА ПЕРЕГОВОРИ із ЗСУ? АХМАТівці ТІКАЮТЬ! Київ захоплює НОВІ МІСТА РФ | НОВИНИ СЬОГОДНІУКРАЇНА СЬОГОДНІ
Переглядів 517 тис.
На Курщині НАСТУПАЮТЬ ПОЛЯКИ 🙈 Росіяни МАСОВО ТІКАЮТЬ З ПРИКОРДОННЯВікна-новини
Переглядів 152 тис.
It seems that all AliExpress links do not work if you chose Germany as a delivery country. If you chose Switzerland, for example, it should display (and you can choose the same product that delivers to Germany).
I usually use the links to the shops I purchased from.
Unfortunately it still does not work for me from Germany, even after changing shipment to Switzerland. The message from AliExpress is "Sorry, this item's currently unavailable in your location. Explore other Group Buy deals." Maybe they introduced geoblocking for these items. Tried in a browser in incognito mode, so no AdBlocker can interfere.
@Klaus Musch It seems to be a new thing. Not good. Did you find another link with the same product that works?
@@AndreasSpiess I was especially interested in the Mini PC. I think I found the same one as you've shown (from Topton), but it is sold for $247.37 (including shipping), and not for $187.75 (plus shipping) as in your video. Configuration is the same: 16 GB / 512 GB.
My experience is that Aliexpress links are a PITA when crossing markets, with things often not working, or jumping to undesirable language settings, sometimes not even with the same way of changing back to one I can understand. And them trying to follow rules may result in listed prices being very different due to them including VAT.
Might be worth to check for a Medion Laptop deal to replace such Mini PC cause their power consumption is even lower with the same chipset once the display is turned off .
Medion had sold them on their shop for less than 200€ this black friday incl. VAT and shipping.
I have gotten such a Lenovo i3 m73 which works but has a quite high consumption of over 14 W for its slow mode it is running.
I compared that with my 8 year old Medion 17 inch tablet cause that is connected to a powermeter socket / tuya reporting data to homeassistant.
Even the 17 inch screen on does not need 14W, less than 11W and once turned off it runs about 7W.
and yes, the sonoff are great just for the good readable display even though the V2 might come with a backlight or e-ink display and more opportunities. I have the 20 A version to measure and show everyone the amount of energy charged in the EV cause that is a one phase charger only running at 16A for a few hours to charge the car during sunlight and sometime during night. pretty powerfull but terrible at low energy scenrarios.
great video
For a DIY approach, there is the OpenEnergyMonitor project based on arduino. Quite good after manual calibration. And it has 4 channels.
For a more industrial DIN mounted solution, I used the Orno WE-514 , it's calibrated in factory with some official certifications. So it should be very accurate. It also has an RS485 interface to plug to the computer. No cloud, no wifi.
Would be nice to get them compared. They are very interesting solution for those who dont want wifi.
If they have a MID class B certificate. So they should be 1% accurate. However, the data sheet says it has a 0.25A minimum current (60W)
Isn't DIY supposed to save money? This is expensive stuff with goofy smart apps
This is a fantastic topic. I see plenty of people think they have free energy when all they managed to do was fool their power meter.
Thank you. At least they have to "fool" the power meter of the utility company to save money ;-)
When flashing Tasmota on one of the compatible energy meters, you should likely also make sure to do the Tasmota calibration, using a purely resistive load (like that 60W incandescent bulb) to get accurate data. My experience with such a setup is that if all my consumption goes through such devices, I can get very close to the house meter used to calculate what I have to pay. Instead of trusting the resistive load to have a specific wattage, I used a "known good" medium-price meter. The point is that factory calibration may be lost when replacing firmware.
While I agree, the voltage can fluctuate quite fast where I live for more than one volt. So in the seconds from when you read your meter and configure the value in the device, it might have already changed.
@@AndreasDelleske If that was my situation, I'd do the calibration on the power from my UPS.
I did this today with the Sonoff. The problem of calibration is that you need a more accurate meter. A bulb is not a good "calibrator" in my eyes. The bulb I used had 60W@233.6V and 62.5W@240V. The 40W bulb had the 40W@231V. Both were Philips bulbs (I hope a decent quality).
A power resistor probably would be a better choice.
@@AndreasSpiess Yes, thanks - a great idea would be a fast electronic load with a resistor that tries to keep exactly 50 W, no phase shift (or 500) no matter what voltage is on... then calibration of the Shellys (where you have to input the power (W) directly). would then be easy. I think it is not that important if the voltage is exact down to 0.1 V for the user.
@@AndreasSpiess While I fully agree that using a bulb is does not give very accurate calibration, it may still be much better than losing factory calibration and then doing nothing to make a new calibration. As your numbers show, decent bulbs are not *way* off target. A major reason for suggesting an incandescent bulb was that the Tasmota calibration process needs to be done using a purely resistive load with power factor 1, which sometimes can be hard to recognize in a modern household. And accurate power resistors are a rare household item 😀
Anyway, like you, my choice was to do my own calibration using a "known good" meter. My load was an old 600W toaster, which was unsurprisingly not real close to what the sticker said. And getting very close to what I pay to use confirmed my process.
Great video (as usual).
For PSUs, I think it's extremely helpful that most of them have a active power factor correction. I don't fully understand how it works but the better it works, the more accurate would be the measurement of the device. For some devices it could be helpful to downsize a PSU for more efficiency.
I really liked the part about devices with constant power draw, no need to measure what's already known within a acceptable margin of error.
active power factor correction is a simple step up dc/dc converter that goes between the rectifier and the filter capacitor. Its trying to step up the rectified sine wave which goes between 0 and 170/340V (120/220V systems) to ~400Vdc. Because it can operate with a rather low input voltage it ends up drawing current for a bigger part of the sine wave, not just from the peak like a simple rectifier will do. The this type of circuit is also one way to get power supplies to operate with universal input voltage 110V/220V.
@@electrobob thank you.
Measurement devices that calculate Power Factor will display Watts correctly for any kind of load. Power Factor Correction makes no difference to consumer's power bills, since people generally get charged for kWh (Watts) rather than kVAh (Volt Amps).
So, the quality of the measuring device affects the accuracy, and not the Power Factor of the device under test.
Hello. thanks for the very professional video. I'm dealing with an inexpensive power meter 100% Chinese, PZEM-004T. I do not have a reliable meter to make a comparison but comparing energy measured over six months has showed less than 1% of deviation from my main provider's meter. Possibly both of them are affected by same error. It comes with a TTL serial output and I used a ESP32 to publish data over mqtt in my network. It worths all the 6€ i've spent!
The PZEM-004T uses a V9881 energy metering SoC chip which promises a 0.1% resp. 0.5% error. So I assume your observation is correct and your meter is ok. It just needs an MCU to transfer the values to the network.
I run a lot of Shelly devices using native MQTT support and nodered. They work really well.
Same here with my Shelly 3em. It works for the last 2 years or so.
I was testing an induction hotplate using a Killawatt meter. At the lower and higher power settings on the hotplate the Killawatt meter the watts consumed did not change. Later I had it hooked up to a SonoffPOW and that showed the power differences between all settings. The Killawatt meter just is not accurate enough , even though it seems to be the go to power testing device by so many people on youtube. Thanks for your video.
I do not like meters with big ranges because usually, they do not accurately display small values (which is what we need in our households).
@@AndreasSpiess The Killawatt is originally a very simple design using 1 transistor to multiply current and voltage.
Modern devices use a chip designed for commercial power meters, with 16 or more bits A/D and digital multiplication.
Nice video! Thank you Andreas!
You are welcome!
This is a superb episode. Learned enough to realize I need to go back to electrical 101 school 🙃.
Thank you! Sometimes it is good to get a short repetition...
A very important topic as we try to reduce our carbon footprint and save money. It certainly gives you pause as we, by our techie nature, consider automating everything in our homes. No automation draws zero power, so, is it the best thing to do? Certainly some automation projects increase efficiencies. An irrigation system with feedback can save precious clean water substantially and if it is connected to a weather service cut off watering completely when a storm is imminent. A thermostat that drops the temperature during sleep times or when a room is unoccupied can save quite a lot of energy. But it all rolls back to thoughtful application of tech and power factor is a key in understanding the merit of an effort. Thank you for this video.
Automation increases energy consumption if you do not change anything (my last video was about that topic).
An active idle ESP will use around 0.5W while connected to wifi. If you had 10 of those, you're still only at 5W. 100 of them is 50W. Easily cancelled out with just 1 solar panel.
Of course many types of sensors don't need to be active. Water levels, temperatures, some buttons and switches, etc only need to wake up at intervals or on events. This can reduce the power consumption dramatically - sometimes low enough to run for 6 months or more off a pair of AA batteries.
Excellent as usual ! About the two PC, it seems difficult to compare them without understanding the needs, the Core i7 is twice powerful than those Celeron N5100 in single thread, but only +25% in multithread but those recent Nxxxx recognize more video codec so they could be more efficient than the Core i7 for encoding/decoding video especially if you use it with video surveillance application like the excellent Blue Iris (not free but so well made for only small part of the price of equivalent "pro" applications), 4 ETH is also cool as fanless.
In my video I used Proxmox with 2 VMS. One for HA and one for IOTstack.
@@AndreasSpiess HA and IOTStack ? so CPU power is not really needed, both CPU recognize all the VT-xyz for better virtualization, for someone with really low budget the I7 is still good (especially if it has a SSD) but if you have the money then N5100 could be better with more RAM support mean more VM at same time. I still use an old Core I7-2630QM 12GB for my Blue Iris video surveillance, Arduino IDE and Domotic things ;-)
Fujitsu power supply is made by Delta Electronics (written on right hand side of unit amongst all the certifications), which is the same manufacturer as the 'Chinese' one. Probably explains the similar results
Thank you for the info. This is what I thought. But still, there are many voices which advocate for "branded" products...
@@AndreasSpiess True. Delta is known to be generally good, just not really known. The real worry in the brick supply department is completely unbranded (apart from CE fake logos!) units. Few years ago had some of these anonymous bricks with old 3.5" and 5.25" external enclosures and the magic smoke came out after only a months!
Delta make almost every full size laptop charger you ever see... If it's a brand name laptop and has a soap-on-a-rope power supply, it's probably Delta.
Plugpack style are less likely to be Delta, but I think some still are.
The main purpose I use power meters isn’t actually to measure power, but for safety. For example, almost all the Extension Leads I have connected via. a Sonoff POW (various generations) because they have Over Current Protection and Over Power Protection. I often have both ON and a big safety margin because they might be connected to devices with a low power factor, but I’ve often wondered how accurate they are. The obvious reason is I want to make sure the extensions are not overloaded and I have a room on AirBnB.
I do not understand your point. We use fuses for over current protection because they are much faster and very reliable.
Would be interesting to see what the accuracy is on our home meters from Itron,L+G, Sagemcom...I have a spare French Enedis (Linky) SagemCom home meter and I am located in Switzerland. Tell me if you are interested to test it and compare it with these values.
Thank you for your offer. For the moment I have no plans for further power meter tests.
I got lucky with a AliExpress seller that got me an excel spread sheet showing the MODBUS protocol for the KWS power meter. I have a working Arduino sketch that can quarry it for all the parameters, and many that are not shown on the display panel. I did have to translate it. And figure out that it was MODBUS RTU protocol. These meters are based on a dedicated processor that was designed to be a power meter. There are many other commands in the spread sheet, that this meter does not support, so I assume the manufacture uses the processor with different firmware for other commercial versions of power meters.
It's to bad it has the poorest measurement accuracy. Thanks for doing these tests.
If you send me the Excel, I add it in the video description. Maybe it is also useful for others.
@@AndreasSpiess I attached as a comment to your github page, Documentation of sensors for "Energy Distribution" and "power-flow-card", and "Load"
@@f.hababorbitz Thank you! I already added the link to the description of this video.
@@AndreasSpiess I thought I had a breakthrough googling like mad when I just found the spreadsheet on github.. then I noticed the familiar face on the account lol. Good work guys!
I highly recommend when you compare a number of products to:
1. Put the name of the product on screen while showing it. Most have it on visible place, but at least one in this video was upside down. Using a cursive font.
2. When you list the products in a table, show a thumbnail of each product above its name. It makes it easier to recognize them, especially when they are visually distinct.
Thank you for the great video:)
Thank you for the tips!
+1 to the thumbnail! especially with such great great names that makes it hard to differentiate
Thanks for sharing Andreas, a very nice comparison. Schoene Weinachten und ein Frohes neuen Jahr.
Ebenfalls Frohe Weihnachten und einen guten Rutsch!
The Emporia Vue 2 is a very popular device for energy measurement. It is fairly cheap, has a lot of clamps and you can even flash it with ESP-Home or any other custom firmware to get rid of the cloud.
Thank you for the info!
As a human society, I think we should prioritize power efficiency, if we want to avoid (reduce if I should be realistic) more wars for it in the future. Every watt counts :)
I agree. If we do not find new energy sources close to us.
Thanks for this, I'll be adding power meters to nearly every device in my new apartment.
Maybe you have a look at the "quadrant theory" before you spend too much ;-)
@@AndreasSpiess I do like your quadrant strategy but I'm a sucker for data!
Thank you for this! Excellent topic because it is very helpful for those of us who are shopping!
You are welcome!
As another low load I would suggest building a "dummy switch mode power supply". Bridge rectifier, 100 ohm resistor, 10 uF smoothing cap. Bridge the smoothing cap with the desired "stand by power" leakage resistor. 300k would be around 1 mA, so 300 mW. But modern power supplies consume less in standby, perhaps 1M or 2M is more suitable.
Swap bridge rectifier for single diode for an additional test.
I do not think that the tested power meters are good to measure idle power. The Metrahit has several ranges. Because of that fact is can measure mW precisely.
I am monitoring my ShellyEM power meter, which I have installed in the past 6 months, against the main power meter installed for my apartment. Each month the kWh values of the two devices are exact matches
That is a very good sign!
Great video as always.I was kind of disappointed due to the lack of Shelly EM from your tests. And by the way, because you are an HA user now, check Powercalc add-on from HACS.
Maybe I will cover Powercalc in a next video.
In a future video, please can you include the shelly EM? It Is a specialized device, hopefully more accurate. I would be intersted in knowing the behaviour of the ready to use Shelly plug s, or the shelly plus pm based on esp32 instead of the esp8266. Indeed there Is room for a video completely focused on shelly devices! And yes, i am an entusiast shelly owner :)
For the moment I have no plans for a part two because there are so many devices out there...
@@AndreasSpiess ok, thanks for the update. Hope the plan will change due to the increasing attention (at least in italy) on arguments related to energy transition like photovoltaic systems and energy monitoring, due to climate change and energy costs raising
In Switzerland you can get a PM231E that was tested by "ABE" from RTS and gives good accuracy even with small loads (when you check the standby power of devices) and big loads.
Thank you for the info!
Interesting topic and great reference data.
Thank you!
Have you tried the PZEM-004T V3? I am using some with esp8266 on each consumer unit. Unfortunately it doesn't discriminate by where to where power goes, though, so one needs to take that into account with storage and production.
I had a PZEM-004T V1 in my last video and it did not work :-(
Indeed interesting and usefull. This will inspire me for choosing the right device/meter.
Thank you! And have a great Sunday.
Nice video, thanks!
Would be interesting to do two test more:
-Flash one smart plug (sonoff, Shelly) with ESPhome and Tasmota, perform the calibration without the Metrahit, and then repeat the test.
-Do the same with an Arduino/esp8266/esp32 attached to a CT clamp. Those use the ADC converter, and I am wondering how bad (or good) will measure current (not power, offcourse) on switching power supplies
For the second one a circuit like the board produced by Motramlabs...
The good thing is you can build an inexpensive current logger (5$) with ESPHOME/Emonlib... of course, it is needed also a Voltage logger to have the power... but I am wondering how good or bad those work.
How do you calibrate a device without one that is more precise?
@@AndreasSpiess I am not asking this… I am a step before… when I need to calibrate an instrument and I do not have a Metrahit…
As Tasmota and ESPHome suggest, with a device with a known wattage… but this method is highly inaccurate… of course, this the worst case… so will your regular amperimeter/voltimeter (did you already forgot about video #99???) and then compare results with Metrahit… the challenge here is to reply the question: How good (or bad) is the method proposed in the video 99?
Then, the second question that would be interesting to reply is how good or bad an ADC of a microcontroller can calculate the power (area below the curve shown at 4:32 ) of a highly noise signal of a switching power supply… if the microcontroller calculate the area of the upper envelope or the lower envelope of the current shown at 4:32 (good old true color bulb”).. the wattage will be quite different…. (the voltage is quite sinusoidal… but the current…)…
I mean, arduinos/esp ADC’s technique may work well with sinusoidal or quasi-sinusoidal waves… but how do they perform with the dirty waves of the good old color bulb?
@@andresdans You are right. You can use a true RMS volt and an ampere meter for a resistive load. That should be good enough.
I would not try to use a microcontroller. The prices of these specialized meters are quite low. You also need an amplifier for the current because a shunt should not have a high voltage across. And keep in mind that the GND of the Arduino might be at 240 volts (as it is the case with the Sonoffs)
Next time also test one of those new fangled gallium nitride USB power supplies, also test them with different output voltages and power loads, to get a better understanding of their efficiency curves.
Maybe I will look into this new technology because it is very interesting.
@@AndreasSpiess GAN Chargers seem to be a huge media hype right now. Would be interesting to see if they actually come over 80% efficency or if Chinese sellers just put GaN on the advertisement to sell it.
I have used such devices to gather info about power use of different gadgets (not permanently monitoring, but plugging them in and reading the values), and my major gripe is that many of them do not go down to the power levels that todays equipment uses in standby mode.
As you already noticed, one of them cannot measure 1.85W. But mandatory standby power is 0.5W and you found that small devices idle at 0.05W.
So the powermeter should be able to measure that...
OTOH, I am less interested in measuring the power of a vacuum cleaner.
So maybe a specific device for low-power usage could be recommended.
I do not agree that measuring low power accurate is very important if it comes to energy optimization. Brewing one less coffee can easily offset months of standby energy of a modern device. Not to talk about not to drive a short distance with the car.
It is interesting to know, we should not loose perspective.
@@AndreasSpiess Indeed it is not important for energy optimization, but of course we always want to measure and know everything 🙂
When I don't want to measure it I don't need a power meter at all.
(I borrowed a HOPI from someone and checked all my equipment)
OPTIMISING ELECTRICITY CONSUMPTION
New remote controlled electricity meters appearing in many countries and several of them use WiFi transmission to collect the data. As l advised in another comment, these are easy to disrupt.
By jamming the WiFi during high-cost metering periods and stopping jamming about 15-30 minutes when the low rate kicks in, the accumulated high-rate consumption is uploaded during the economy period.
As the "jammer" l use is a standard piece of WiFi test equipment, it can be purchased from many sources. WiFi is, as many know, a shared, licence free service.
Well done electricity authorities. Going cheap doesn't pay.
Great review. Thank you for making this comparison video.
You are welcome!
Hi Andreas!
Would be really great if you could compare cheap chinese DC-UPS vs "non-chinese" ones to see if they are at all efficient. They are used as simple backup PSUs for Routers and such.
Also maybe conventional UPS's?
Best regards!
I ordered one, and maybe it will appear once on the channel ;-)
This analysis is great. Many thanks!!
You are welcome!
I have heaps of the thin clients here Andreas, pitty it cost so much to send stuff. Has a cool display on it. Was made in England.
But for whole house meter? Like the connected BTcino F20T60A on DIN rail?
Capture the consumption of the whole house via WiFi+MQTT or ZigBee can be useful
I agree that it is useful. Still to optimize you probably need more informations. I used the Shelly 3em for that purpose before I got my PV,
Thank you for your work ! Nice to have good info 👍🏻
My pleasure!
I think you was comparing the power consumption relative to the power meter from the energy supplier of your home. Is it taxing correctly?
I did not do that. _Two weeks ago, I got a new power (smart) meter. Because they have to be calibrated from factory, it should be accurate.
No gold this weekend, but special vlogs coming on my upgraded power supply.
Cool! I will have a look at it.
Thanks! Very helpful.
You're welcome!
I was expecting to see being tested the EmonCMS along with EmonPi and EmonTX hats and since you have lots of Pi's doing various things in your home/lab it's a good idea to add on one of them the EmonPi hat and make it a power energy meter. It also supports a AC/AC adapter for RMS and power factor correction plus it's open source and you can edit the code according to your needs.
A few years ago, this was an interesting product. With the new specialized chips available it makes no sense to build a sensor using an Atmel chip. Better to concentrate on the "what to do with the data".
I think a PZEM-004t 100A v3 is worth considering, to use as an IOT Energy Meter. I have 4 of them connected to my 3-phase home connection and on-grid solar power plant (UART Daisy Chain; assigned address) with ESPhome and HomeAssistant. It's almost line with my Energy bill I get from my utility company [Difference was about 2%]. Considering its price, I believe it's a decent trade off. Also found this video testing its accuracy: ua-cam.com/video/j0_y8dPfpKc/v-deo.html
Thank you for your information!
@@AndreasSpiess Your videos are just amazing, thank you for making such great content.
Would love to see your take on the pzem sensor! Sharing a link that delivers to Germany:
m.aliexpress.com/item/4001102546990.html?spm=a2g0n.productlist.0.0.60a3tLvgtLvg0G&browser_id=1a568f791ce04a7690f6cf09a6e727bd&aff_trace_key=dabaf000658d44659b585b0d8c8c5191-1665895968831-07583-_A99xzP&aff_platform=msite&m_page_id=rkishhcxcyecaazh1853f4b40a111feac391d51c44&gclid=&pdp_npi=2%40dis%21INR%211756.99%21941.59%21%21%21%21%21%402101e9d416718041746225780e91ed%2110000014401715359%21sea&curPageLogUid=1qDs6H6noMuP&isseo=y&algo_pvid=6c2efaff-68b8-4505-9ce1-4581791faeb8&_universallink=1&m_page_id=rkishhcxcyecaazh1853f4b40a111feac391d51c44&gatewayAdapt=Pc2Msite
www.amazon.de/s?crid=2C0IL4T35GA7H&k=pzem%20004t&ref=glow_cls&refresh=1&sprefix=pzem%20004t%2Caps%2C231
Update: fixed typo
The question I am always concerned about is, how accurate is the power company meter that monitor mains power usage,
They should be designed to under-count the consumption because anything else would get them into trouble.
Here they have to be calibrated. So they are quite accurate
flashing tuya devices: it depends on the used wifi module. i got some wifi sockets of digitec, they dont use an esp but some tuya developed wifi interface, cant be flashed with esp-home or tasmota.
Thanks for the info!
Nice! Wonder if calculating the real power using an oscilloscope will not yield more precise results than measured by Metrahit Energy (according to the datasheet the intrinsic uncertainty for active power at 10-65Hz AC is 0,4% or 1% +20d)?
PS. Wonder if Metrahit Energy measures to it’s specifiaction? Mains probably not clean/stable enought to get more precise measurements..
I do not know. Keep in mind that oscilloscopes do not have good ADCs e. g. 12 bit)
Very interresting video !
Glad you think so!
The good question is how can we trust national energy provider power meters ?????
At least here they have to fulfill hard regulations...
Probably not, but you don't get a choice either way.
On my powermeter the optical LED got broken after 5 months (Holly DTZ541). This also happened to the powermeters of my flat neighbors. I wonder how the overall quality and the accuracy for measurements is of this chinese product is when such simple things break.
The meters are all calibrated at the factory with traceable standards and sealed to protect against tampering.
Here in my country Hungary, I experienced too high readings on my utility meter provided by the electricity company. This was a dual meter with consumption and feedback counter because of my PV system installed on my roof.
To track down the suspicious meter behavior, I separately measured all my devices for one day with a commercial power meter.
When I finished the individual measurements sequence, I summarized the individual daily consumptions, and take this number as an accepted average daily consumption.
Interestingly the utility meter always showed 20-30% more consumption compared to my summed up average.
I started an official complain with my electricity company, to ask a calibration for my meter.
They came and replaced the meter with a new different type, and take the old meter for calibration check.
After 2-3 months, the official result arrived, saying the old meter was good, and I was due to pay a small amount of money because of the "unverified" measurement complain.
However the new meter is showing more realistic consumptions, the 20-30% extra readings are gone.
It can be stated as a lost interaction with the utility company from my point of view, since my complain was rejected, but finally my original problem was solved, right?
I invested a small money for the "meter replacement", and the result is 20-30% less consumption....
I just shared this story, because even the utility meters can be a scope of questioning.
As they are produced in vast amount, where every cents counts, and this pressure of cost savings can result in unreliable metering.
Modern devices are supposed to have pfc now, and if they want actual certification to be on the grid, they must be within certain limits and the current curve must be corrected. As not only PF but the form of the current curve has a severe effect on transformers. What happens is that companies submit a diagram, and if it simulates ok they get the certification. But as we all know, simulating and actual testing are different.
I assume there is a minimum power consumption for this law. The750W power supply I showed had PFC and the small bricks did not have one.
In the end, the large consumers are interested in good power factors, because it saves them money.
Main voltage is just a little spicy reminder of what electricity can do
I agree!
For the AC power adapters, you should check which DOE level they have. Since level 6, they need much lower power consumption in no/low power applications.
Good info. Thank you!
How to check that DOE level?
@@TishSergThe level should have either a circle with roman numbers inside, or just DOE with the number next to it.
@@tverspeelt the circle made of two arrows? I have "10" in that on notebook PSU and router PSU.
But on 1+ charger I found "VI" inside plain circle. Looks like it is what you refer to.
Thanks
If one of your viewers comes up with the RS-485 API for that one meter, it would be cool to learn about. I understand a lot of household inverters, chargers, and even batteries use RS-485 for communication. I wonder if there's some kind of standard?
I do not think they use standard messages. A viewer sent me the commands and an Arduino program. You find a link in the description.
Could you test one of the new GaN USB power devices and show if the claims that they are vastly more efficient is actually true?
I might have a look into this new technology...
My KWS-AC300 seems to read high by about 5%. But mostly, it doesn't display anything under 1 watt. I would say it is a failure from my perspective,
Same here. That's quite annoying - I ordered it after viewing Andreas' test and hoped for an accurate reading of small currents. Obviously AC300 is not suitable for this purpose.
I did not test them below one watt because it is asked for too much. My Metrahit has several ranges to get below 1W. I made some videos about "burden voltage" which is important to understand in this respect.
Can you comment on whether you trust the power consumption as measured by the meter installed by the electricity supplier?
I trust them because they have to fulfill strict standards. But I only can speak for Switzerland.
Steve M. Potter, In my state it's codified. The utility will test for free once per annum if requested. The board will do it for 5 yankee bucks. If it's beyond 2% (1.5% for water) error you get a rebate. If it's under billing there is no back charge. If the meter was tampered with, the underage is due immediately. If it reads 0 or you should reasonably have known it was under billing the difference is paid back over time.
@@AndreasSpiess Same here. Residential customers also get a pass on demand and reactive power.
@st1300r Wow, which state? Here in Ireland, many meters' clocks are off, a problem if you have day-night rates. And if you call them out to complain about your meter and they don't find a fault, they charge you €200 for the visit. My clock is only off by 11 minutes, but some are over 2 hrs off.
@@StevePotter New Jersey. Here, at least in my town, I don't have any time dependent rate. I think the other side of the country (where everything causes cancer) does have time dependent rates for residential power. I have a dumb meter it doesn't have a clue about time. 60 cycle here on avg is fairly accurate but a meter using mains frequency like an old school electric clock will drift over time without a method to realign it. Off hand I don't know how places with "smart" meters derive time.
leaving the usb bricks in the sockets is a very big concern, so i'm not sure why you said that... these things produce dirty electricity like there is no tomorrow!
Mine were ok.
@@AndreasSpiess You are probably one of the exceptions with a big emi filter right at the entry of the house... otherwise: no
Regarding the quality of "cheap Chinese products" I can mention than one of the most common meters that power grid companies in Norway use for their private home deliveries is a Kaifa automatic energy meter. It is made in - you might have guessed - Shenzhen!
Thank you for the info. I think, most of us know the facts. Still I get sometimes such comments...
Your Videos would benefit from chapters. I've jumped back to the device capabilities after the efficiency comparison and needed a while to find the right device again. This could probaply have an positive effect on Affliliate-Link-Clickrate.
So far I decided not to add chapters because I do not like if people ask me questions because they skipped chapters :-(
It’s a shame that the Swiss type J plug wasn’t adopted more widely. It’s more sensible than the type F.
:-)
Thanks
You are welcome!
Does anyone know if we can expect similar results for the Shelly plus 1pm compared to the 2pm in the video?
I do not know. Maybe you check if they use the same chip?
How come the KWS-AC300 and KWS-AC301 produced such very different results at 10:00 ? They look like they should be nearly the same hardware under the hood.
Is it just variability in calibration??
I do not know. I also did not find a calibration function for users. The AC-301 has double the current range. Maybe this is part of the reason.
1:40 so that's were those nasty plugs and sockets we're using here in Brazil came from 😂😂😂
Why do you call them "nasty". I like tham. Just not the fact that none of the other European countries use it ;-)
@@AndreasSpiess that's exactly the reason. I remember your video about RISC V when you said something like "standardisation on the lower level leaves you free on the higher level". People should decide on one type of connector and be done. If you haven't already, you should check out Adam Something's video on why European international rail sucks. Same exact premise. Cheers
Thanks for this excellent video!
You are welcome!
I have a plug-in meter and some wifi (meross) ones... turns out the plug-in measures quite high for low (
If you connect then in series, one measures also the power of the second. This is relevant for low power consumptions. I would not do that.
Cat hairs on the oscilloscope are a sign of a properly setup lab.
:-))
Thats a good video.... Andreas, is there any for 3 phase / 4 phase power measurement?
I am trying to make one with pzem-004t and use an Arduino / ESP, so that i can use the parameters to control my motor (3 phase..)? Really appreciate if any better alternative that you have come across...?
I use the shelly 3em for 3 phases.
Fantastic the way you organized the results.
Question: what's a good way to (once) measure the power usage of a high power item, like a burner on a 20A circuit?
A clamp meter is what you want for that, assuming you mean for manual readings like you'd do with a multimeter.
As shown in the video, you get transformers for both of the "manual" meters which go up to 50A and more.
@@AndreasSpiess Oh I see! There are 50 and 100 amp versions of the KWS-AC30x. Thanks!
hi, i'm searching for one of your videos, you explained how to prevent any micro controller from crashing on startup, (simple circuit) can you help me pinout this video ?
i have the same problem once i switch the power on, somtimes it doesn't work.
thank you sir.
it was video #436
I bought a 'smart' plug with near real-time monitoring as one feature to see roughly how much power my new PC uses and it dawned on me that I didn't know if it was accurate. I found how much power my fan uses (50W) and saw it only using about 37.5W and my heart sank. Then I turned it up to max speed and it hit 50W.
As you saw, all meters were quite accurate. So 37.5W definitively had to be wrong ;-)
@@AndreasSpiess I realised afterwards that it was 50W at the maximum speed. 37.5W was the lowest which I didn't know beforehand.
When you say wolfram bulb are you talking about your average tungsten filament incandescent bulb or some other kind of bulb that I don't yet know about?
Yes. Tungsten and Wolfram is the same
Is the KWS-AC300 calibrated? Does each one give the same readings, or did you get lucky?
None of these cheap sensors are calibrated, and it is well possible that I just was lucky.
Since the voltage may change over time during the day, did you measure each along with the Metrahit meter at the same time?
Otherwise you may have significant errors in your measurements as the mains voltage may easily fluctuate 10 - 20V, sometimes within a short interval.
You are right. I did the measurements at the same time.
I ordered Shelly EM (the one with current clamp) couple of days ago. Hope to get it soon. I plan on using it for 24/7 measuring apparment's power usage. The utility company I purchase power from doesn't provide this info.
I used this for last two years before I got my PV. It worked fine. However, for optimization, often, more granular info is needed.
For me it does not matter what cheap meter shows - it is important what is the quality of the meter installed in my home by the Power Utility Company. If they meter cannot handle modern devices with the dc/dc power supply we will pay more...
I Hope (and assume) that they have to use certified meters.
@@AndreasSpiess Yes, they are certified fore pure resistive loads. But unfortunately LED, and other devices use DC switching.
@@AndreasSpiess See the report:"Electronic energy meters’ false readings almost six times higher than actual energy consumption" of the Amsterdam University of Applied Sciences
Thx
:-)
I am curious about the accuracy of Atorch devices. Also availble on Aliexpress. Maybe in a next video?
So far I have no plans for a second part :-(
Excellent comparison! Not sure why anyone would care about monitoring low power consumption??!! Ex Tuya @ 2.4kwhr/yr x $0.15/kwhr = $0.36/yr Who cares?!
In Germany, they pay $0.8/kWh. So they started to care...
Andreas muy bueno el video lastima que no estas traducidos al español, como loas demás
ya que el traductor de YT traduce como quiere algunas cosas......
Luciana, an Argentinean lady does the Spanish translation for the subtitles. Was her translation bad?
@@AndreasSpiess cómo estás en primer lugar gracias por responder, recién veo tu msj, pero no es que ésta chica hizo mala la traducción sino que no está en el vídeo como en otros, ya que la traducción la genera YT, y no una traducción que haya sido subida por vos.
FYI the N5100 is 66% of the performance of the i7-4765T
You are right. I showed this fact in one of the last videos. However, for running Home Assistant, this is not very important.
👍 - _auch_ für den YT-Algo! 😊
Wieso meinst du?
Hi!
What is the cheapest device (method) to measure active and reactive 3phase power used to test capacitor bank of pfc panel?
If you want to know reactive power you need a meter which displays at least active power and the power factor. Only my Metrahit had a display for VA and VAr
Very interesting!
One thing I'm curious about is, how did you produce the load of the power supplies?
I found, crappy cables (thin aluminium core) can have quite an impact on their efficiency.
I used the IT8512A electronic load I showed in the video. Crappy cables can be offset by using the 4-wire method.
Tasmota on the SonOff Elite should be much better, calibrate it with your meter and the 60W bulb. Also increase MQTT frequency, I suspect will be better than stock.
Of course, I can calibrate all meters. But what about all viewers who do not want to spend 1000 Euros?
@@AndreasSpiess
The Aneng multimeters are pretty accurate for budget viewers... Mine was around $20US from Banggood if I recall. And it's so close to my Fluke 179 that I don't even bother taking my Fluke into work anymore, even for things where I need a finer measurement.
I was also expecting you'd test the Shelly EM or 3EM. Also both Tuya devices are din rail mountable
There is a lot of choice...
@@AndreasSpiess That's true, but as those are specifically made for power metering it could have been a good candidate. However I don't think Shelly devices will change that much, however the gen1 vs gen2 might.
Thanks Andreas for all your great content,
I know this is hard to do right, but is comparing only one of each device a valid method to compare accuracy?
I believe this is not how to measure/compare accuracy, there should be at least 3 of each device
I agree.
What kind of zigbee2mqtt device would I need in order to use the Tuya meter with home assistant? Will Sonoff zigbee 3.0 dongle work?
I use the Sonoff Zigbee dongle and it works.
@@AndreasSpiess thank you very much for the response. I'll get the sonoff dongle then. Smart home adventure begins!
@@Zamsky39 Order a USB extension cable to mount the dongle away from the PC. If you mount it into the PC it will only have a short range.
Great video as usual
Thank you!
Does the Tuya include any switching function? The Tuya energy monitoring plugs that include switching have/had a great feauture that as "protection" they turn off the output if the line voltage is exceeding 250v (and don't turn it back on, even if the voltage eventually drops below 250), not great if you are monitoring a fridge/freezer/server, basically anything you actually need to be turned on.
The one with Zigbee has a switch.
I do not know if they have such a function and if it is fast enough to protect your devices. Overvoltage is a fast killer
You can solder a thick copper wire across the switching relay in the Tuya, effectively disabling the switching function.
@@Zamsky39 Exactly what I did, the “protection” takes about 6 seconds so is in no way protecting the device.
Nice video. I liked it 👍
Thank you!
Hi Andreas, thank you for the fantastic job you are doing. We have an FPGA-based development board that we would like to offer you for tests. Please get in touch with me via UA-cam if you are interested. Happy Holidays!
FPGAs unfortunately are not very interested for a general public because their applications are narrow and its usage too complicated. This is why I decided not to use it on this channel. At least not till I find a compelling reason.
Is there a reason why you only tested the tuya zigbee version and not the wifi version (DIN rail mounted)? Unfortunatly there are no results shown in the spreadsheet.
The ATMS1603 had an average error of 4.6% and a power consumption of 0.4W
Would be nice to test them against the meters that have been installed in the house by the energy providers. I have a Contigea meter, probably from the 1950's. I have been measuring a lot of devices over the last couple of months and I am still missing a couple of KWh a day somewhere it seems.
In the fifties, the loads were very different. So maybe these meters cannot properly deal with the fast- changing current curves.
@@AndreasSpiess Would be nice to compare since of course our electricity bills are being calculated on the readout of that meter. Worth looking into in my case.
Yes. I was thinking the same. I wonder just how the domestic meters measure the V / I values. Since this is the meter that creates our bills, it would be informative to see what contributes and by how much. I'd be surprised if a domestic smart meter had the same measuring circuit as a €1000 test meter.
@@petelynch9468 The $1000 come also because it has several ranges (which enhances low power measurements).
Now imagine you discovered you're using a lot more kWh than the domestic meter shows. Would you notify your energy provider?
Legrand 412015 its my choice, please test it i want part 2 :)
For the moment I have no plans for a part 2. Too much choice out there...
Thanks for this comparison. One question: Where did you find the Mini PC with 16GB RAM for this low price? The Link doesn‘t work unfortunately.
On AliExpress. It seems that AliExpress applies some sort of geoblocking. Search for G30b and N5100
@@AndreasSpiessPerfect, found it. Thanks Andreas
Go shelly! Would you mind test also Shelly Pro 4PM? :)
So far I have no plan for a part2.
Almost all links to Aliexpress are broken (Sonoff POW Elite, Tuya Zigbee, DIN Rail mounted, CP2100a, DP20003, Mini PC). Maybe this is something which was not intended ...
I just re-checked with "Germany" as country and it did not display. So if you change to "Switzerland" as delivery country, you should see the product. I usually use the links to the shop I ordered.
Can someone tell me the exact name of the tuya device? The link in the description isn't working for me. It says the product ist not available in my country.
Sometimes it helps when you change the delivery country. Which Tuya device do you mean?
@@AndreasSpiess did not work even when selecting Switzerland or US. I mean the tuya zigbee Power meter visible at 8:54.
Still looking for a precise power meter to measure how much my PV inverter actually harvests.
After recognizing that the Shelly 1PM is not measuring the voltage for watt calculation it's out of the game for me as the voltage here ist 230V +-10%.
I found one with these words: „Tuya Smart Life ZigBee Switch“. Mine has the button on the front.
@@AndreasSpiess Thanks...honestly I didn't find something that looks exactly like yours. I hope I can improve the precision of the Shelly 1PM I already ordered before the following way:
- Get the power value from the shelly that was calculated with the voltage set in the webinterface (e.g. 220V) via MQTT
- Divide the power through the hardcoded voltage to get the current
- Multiply the current again by the actual voltage read out from my smart meter from the corresponding phase (SML/D0 protocol field 1-0:32.7.0*255, 1-0:52.7.0*255, 1-0:57.7.0*255)
Does that sound reasonable to you? Do you have any experience regarding precision of the voltage readouts from Smartmeters?
By the was i guess the reason why many cheaper shellys are not measuring the voltage ist because the esp8266 used internally has only one analog digital converter which ist probably already use to measure the current with a shunt.
Maybe the more expensive ones have an esp32 with more ADCs.