I have used these to find faulty freezer, find out how much it costs to heat workshop and a few repairs as a sanity check. But, for example for the workshop if the heating bill is £400 for the winter, will more insulation make enough difference to justify the outlay and does knowign the running cost accurately make any difference to my decision?
You’re the only person who tested the meters against real lab kit. Excellent. Today, a half dozen companies copied the model and they all look the same. Do you have manufacture model or link to the one on the left at 1:35 with the blue buttons?
Thank you for this video it was very helpful with the teardown and showing the datasheets of the chips used and testing the accuracy. It would be interesting if you can make a video on the Kill-A-Watt P4460 that claims to be more accurate and is popular and see how it compares to the cheap power meters. It looks from other teardown videos the internals are different curious what chips they are using.
P4460 seems to only be available in 110V so not helpful for me to invest in, while MOR does get advertising revenue it is not much so I have to be selective! The teardowns seem to match the circuit on eevblog forum quite well (www.eevblog.com/forum/repair/confused-kill-a-watt/). They are using differential opamp for current amplification, another for voltage which is on the supply side of the shunt (so inacurate for calculating consumption but presumably placed there to detect phase difference vs consumprtion side of resistor. Transistors look like they are conditioning for frequency although they might provide handy data for phase difference calculations too. Then all the data processing done in a standard microcontroller (or DSP) of some type - unfortunately just an unidentified blob on the board, but anything with ADC sampling over about 2kHz and LCD driving would work. My initial thoughts are that it only uses a single current sense resistor, and since it runs to about 20A that means the resolution in the mA/uA areas are going to be both limited by the ADC resolution (probably 10-bit, 12 if lucky) and offset by noise in the opamps and high value resistors, there is also likely non-linearity in the amplifiers at this low level. The stated accuracy is 0.2% and if they do per-unit factory calibration I can see that would be possible in the mid-range. But some comments I've seen indicate real world is closer to 2%, also I've not seen any available ports on the board for that purpose so even with 0.1% current shunt and divider networks (unlikely) that would only be possible in mid-range say 2-15A. All of these devices are limited in the low current ranges by noise, non-linearity and ADC resolution, and in the high ranges by opamp compression. The only way I can see it could be improved is by introducing ranging using a second sense resistor in the same way that a multimeter does the job.
I mostly use these to measure the power consumption of my PC. Standby power doesn't bother me since it's extremely rare to have a device that consumes more than 5W on standby. I've heard that some new TVs use like 15W on standby though, and that is just ridiculous. But if it doesn't even measure less than 1W, it's not a problem for me.
it can be interesting to count up the amount of standby devices. with almost 8800 hous per year each 1w of standby can work out at £3 to £4 per year, but you don't need a meter to survey what you have and decide if you want to ore actively unplug things.
Very interesting. Had some very weird experiences with some. Mostly low power draws make them acting that weird that the value tended to be 50 or 60% off sometimes. What I would do then is to add a regular like 60 Watt light bulb and measure that way after things stabilized. My 'trusty' one that I used for a long time has just giving me some odd readings so I was wondering around TH-cam again. 211V while the multimeter (multiple) says around 238V. Seems I can probably throw this one away as well. I've seen people do funny things trying to save on their electricity bill. You should measure these wifi smart plugs or lights. Knowing that some people have like 20 or so of these devices in the house up and running for 24 hours a day while they will unplug a phone charger that consumes 0.5W when just plugged in. I actually wonder if they might use a wifi smart plug to shut these things down when not charging a phone.
I'm trying to get around to looking at the issues with low power devices. I have not found a cheep device that can measure them well. The issue with low power is that the digitisation loses precision as you get towards the bottom so even if you do get a reading it can be 40-100% inacurate. I've also seen a request regarding what power do they measure down to and I'm going to try and do a video about that if I can get to it. If it can't even measure the voltage properly then all bets are off! I did measure a smart plug at some point and it wasn't impressive!
@@MakeOrRepair Well they need to be able to measure an imense range of current/power. It's like measuring a Voltage between 1 and 2 Volt on a 10.000V range setting. Probably you could calibrate it by filing down the shunt or adding extra solder on it as a matter of speech. Of course that 211V instead of the real 238V was a not so good thing. I have now ordered the same type of smart switch (Tuya) in Zigbee. So when I will get it I will compare it to the BT model. From what I've heard is that Zigbee uses a lot less power. Lets wait and see.
@@Luke-san Yes exacly - dynamic range is the problem - no resolution left at the bottom end - I have a lab grade power meter that goes from about 20kW down to the micro watt - but it does it by switching between many current shunts depending what sort of range it is measuring. OF course it is massive and costs a fortune! It would be great if there was a low cost meter that had say two shunts so it could auto range down into the mW range without sacrificing too much rsolution and accuracy. The design is not much more on top of what is already there, but of course it would add about £5 to the cost and they lose competitiveness.
for first no review of timer accuracy useful to size solar/ battery capacity loads ? Timer on mine is grossly unreliable ? Other Mine stopped working: display : Lift display pcb clean the pressure contacts Tighten back up properly to squeeze the flexi contact strip . Back up and running
Strange about the timer, have not noticided an issue and certainly didn't expect to find a problem in that area certainly nothing worse than say 1 part per 100 which would be a 1% error and not really effect any decision making. The construction is certainly not wonderful, I was very woried having removed boards about getting enough pressure on the elastomeric strip without stripping the screw holes.
I have used these to find faulty freezer, find out how much it costs to heat workshop and a few repairs as a sanity check. But, for example for the workshop if the heating bill is £400 for the winter, will more insulation make enough difference to justify the outlay and does knowign the running cost accurately make any difference to my decision?
You’re the only person who tested the meters against real lab kit. Excellent. Today, a half dozen companies copied the model and they all look the same. Do you have manufacture model or link to the one on the left at 1:35 with the blue buttons?
Sorry the answer was in the description. The model is Nevsetpo B07H1ZFVFM.
no problem, glad you liked it
Thank you for this video it was very helpful with the teardown and showing the datasheets of the chips used and testing the accuracy. It would be interesting if you can make a video on the Kill-A-Watt P4460 that claims to be more accurate and is popular and see how it compares to the cheap power meters. It looks from other teardown videos the internals are different curious what chips they are using.
P4460 seems to only be available in 110V so not helpful for me to invest in, while MOR does get advertising revenue it is not much so I have to be selective! The teardowns seem to match the circuit on eevblog forum quite well (www.eevblog.com/forum/repair/confused-kill-a-watt/). They are using differential opamp for current amplification, another for voltage which is on the supply side of the shunt (so inacurate for calculating consumption but presumably placed there to detect phase difference vs consumprtion side of resistor. Transistors look like they are conditioning for frequency although they might provide handy data for phase difference calculations too. Then all the data processing done in a standard microcontroller (or DSP) of some type - unfortunately just an unidentified blob on the board, but anything with ADC sampling over about 2kHz and LCD driving would work. My initial thoughts are that it only uses a single current sense resistor, and since it runs to about 20A that means the resolution in the mA/uA areas are going to be both limited by the ADC resolution (probably 10-bit, 12 if lucky) and offset by noise in the opamps and high value resistors, there is also likely non-linearity in the amplifiers at this low level. The stated accuracy is 0.2% and if they do per-unit factory calibration I can see that would be possible in the mid-range. But some comments I've seen indicate real world is closer to 2%, also I've not seen any available ports on the board for that purpose so even with 0.1% current shunt and divider networks (unlikely) that would only be possible in mid-range say 2-15A. All of these devices are limited in the low current ranges by noise, non-linearity and ADC resolution, and in the high ranges by opamp compression. The only way I can see it could be improved is by introducing ranging using a second sense resistor in the same way that a multimeter does the job.
I mostly use these to measure the power consumption of my PC.
Standby power doesn't bother me since it's extremely rare to have a device that consumes more than 5W on standby.
I've heard that some new TVs use like 15W on standby though, and that is just ridiculous.
But if it doesn't even measure less than 1W, it's not a problem for me.
it can be interesting to count up the amount of standby devices. with almost 8800 hous per year each 1w of standby can work out at £3 to £4 per year, but you don't need a meter to survey what you have and decide if you want to ore actively unplug things.
Very interesting. Had some very weird experiences with some. Mostly low power draws make them acting that weird that the value tended to be 50 or 60% off sometimes. What I would do then is to add a regular like 60 Watt light bulb and measure that way after things stabilized.
My 'trusty' one that I used for a long time has just giving me some odd readings so I was wondering around TH-cam again. 211V while the multimeter (multiple) says around 238V. Seems I can probably throw this one away as well.
I've seen people do funny things trying to save on their electricity bill. You should measure these wifi smart plugs or lights. Knowing that some people have like 20 or so of these devices in the house up and running for 24 hours a day while they will unplug a phone charger that consumes 0.5W when just plugged in. I actually wonder if they might use a wifi smart plug to shut these things down when not charging a phone.
I'm trying to get around to looking at the issues with low power devices. I have not found a cheep device that can measure them well. The issue with low power is that the digitisation loses precision as you get towards the bottom so even if you do get a reading it can be 40-100% inacurate. I've also seen a request regarding what power do they measure down to and I'm going to try and do a video about that if I can get to it. If it can't even measure the voltage properly then all bets are off! I did measure a smart plug at some point and it wasn't impressive!
@@MakeOrRepair Well they need to be able to measure an imense range of current/power. It's like measuring a Voltage between 1 and 2 Volt on a 10.000V range setting. Probably you could calibrate it by filing down the shunt or adding extra solder on it as a matter of speech. Of course that 211V instead of the real 238V was a not so good thing.
I have now ordered the same type of smart switch (Tuya) in Zigbee. So when I will get it I will compare it to the BT model. From what I've heard is that Zigbee uses a lot less power. Lets wait and see.
@@Luke-san Yes exacly - dynamic range is the problem - no resolution left at the bottom end - I have a lab grade power meter that goes from about 20kW down to the micro watt - but it does it by switching between many current shunts depending what sort of range it is measuring. OF course it is massive and costs a fortune! It would be great if there was a low cost meter that had say two shunts so it could auto range down into the mW range without sacrificing too much rsolution and accuracy. The design is not much more on top of what is already there, but of course it would add about £5 to the cost and they lose competitiveness.
for first no review of timer accuracy useful to size solar/ battery capacity loads ? Timer on mine is grossly unreliable ? Other Mine stopped working: display : Lift display pcb clean the pressure contacts Tighten back up properly to squeeze the flexi contact strip . Back up and running
Strange about the timer, have not noticided an issue and certainly didn't expect to find a problem in that area certainly nothing worse than say 1 part per 100 which would be a 1% error and not really effect any decision making. The construction is certainly not wonderful, I was very woried having removed boards about getting enough pressure on the elastomeric strip without stripping the screw holes.