I think,
@JamesW that in our desire to give the OP some simple guidance we have made things rather confusing – especially across the last 15 posts or so

We could take our disagreement offline – but I’ll try to sort it out here. First, I do fully agree with your skepticism about the COP rating of 15.8 - as I too noted - and I appreciate your links to other crazy-high COP claims from other vendors – fun reading, thanks! Thanks also to
@DigitalGuru for your additional posts, nameplate info, etc. I also agree that the Aquastrong manufacturer is posting several other confusing specs that are, at best, misleading. But I think we’re probably nit-picking, for example, if we worry about some relatively minor things like what they meant by “minimum circuit ampacity of 2.5A” – as we already know it can draw 18 amps, or perhaps up to 30 amps in a locked rotor compressor condition. It may simply be that the foreign manufacturer intended to communicate that the unit never draws less than 2.5 amps when running. Or it could be they meant 25 amps instead of 2.5, as you suggested. True that most US manufacturers would never state something so awkwardly, but let’s move on…..
Of course it’s critical and true that in order to calculate COP - or anything really - one must pay attention to units. Let’s take the COP formula
denominator first (formula pasted below in two forms) as that is really simple. With all respect, you misunderstood me (I guess) when you wrote that one cannot get “total power” – at least in regards to that COP denominator. Run the unit for an hour or two with a reliable wattmeter that measures real power – aka true rms power (what we pay for in most cases) and you have your COP formula
denominator. In my example the wattmeter would have reported 4000 total
watts watthours over 2 hours (yes, about 2000watts
/hour or for 2 hours ie 4kwh
in popular jargon). One could argue about the accuracy of a $15 inductive pickup meter’s ability to report real power, but it’s simple enough to compare it to the service entrance meter and make any needed/minor adjustments to math.
The numerator in the COP formula is specified as the amount of energy (joules per second – the alternative definition of watts) imparted into the water over the same test period. I cannot claim to have found exactly how folks figured out that 1 BTU/hour “loses those units” to become 3.413 watts – but you can get that documentation in hundreds of places where they offer calculators and explain how they get from BTU/hour to joules per second in watts.
Starting here, but the converters are everywhere, divide by 3.413 (as you also stated) – or multiply by .293 if you prefer.
Back to summarizing my conclusion in the test example – this time for just a one hour test that increases 9000 gallons by 1F and used 2000 total watts of electricity per any favorite meter:
9000 gallons is 75,000 lbs of water. If it went up 1 degree F, by definition we put in 75,000 BTU (you’ll be lucky to get half that but it’s just an example).
Divide BTU by 3.413 and you put in 22,000 watts of energy to cause that temp rise.
Finally divide 22,000 by whatever the wattmeter said (that’s watts/watts, units fall out) and if the meter said 2000 total watts your COP was 10.98. If that were actually the case, the average current draw over that time would have been about 9 Amps – based on apparent power, so hard to say exactly because we don’t know the power factor. In any case our measurements are okay since the meter figures out power factor and gives us real power, usually, via some fancy chips that allegedly do waveform analysis
Of course if there is heat loss during my suggested test, COP will come out lower. That’s why I suggested tight controls on stable conditions 80/80/80, cover on, no wind, good circulation but no more than necessary, shortest test duration possible to get firm temp measurements, accurate thermometers. I don’t see the flaw in any of that, but I welcome correction.
The calculations you suggested,
@JamesW in post #29 are, in my opinion, flat out wrong, sorry to say – using juxtaposed inputs and incorrect use of the COP formula. You went awry, I believe, when you used “consumed electrical watts” to back into the BTU/hr for the COP
numerator, that presumes all the input electrical energy goes toward heating, which of course is not the case and corrupts the very meaning of COP – which is to tell us how much heat energy actually went into the water per unit of electrical watts consumed.
COP = heat output (in watts) / power input (in watts). To use BTU for the heat (numerator), it’s used this way: COP = Qh / (Pe * 3.413) where Qh is BTU/hr and Pe is electrical power input. To demonstrate or compare that in a resistive heater, the heat input is the same watts as the electrical energy input, so the COP is 1 watt of heat energy / 1 watt electrical input, therefore always “1” for resistive heaters like a typical domestic water heater. Please see some of the well documented web pages on COP.
That method as opposed to what
@JamesW was writing about that places thermistors, a new device to monitor them, then also flow measurement (usually pricey) that somehow includes accurate measurement of the amount of water heated. Hmmm ... In comparing the two methods, it’s my humble opinion that my method rocks! For more fun, and for what it's worth to perhaps save you time - some author(s) on
wikipedia here argue you'll never achieve more than a COP of 8.8 based on the first law of thermodynamics. Does the newer R32 rerfrigerant change the laws of thermodynamics? Doubtful
Returning to the earlier “circuit ampacity” topic, most manufacturers also specify minimum recommended circuit amperage and maximum breaker size, oftentimes specifying minimum wire size too. Oddly I don’t see that in the manual for this 33K model unit, nor do they tell you what type of connector is on the end of that power cord. Maybe ask them in a follow up? It seems there is a built-in GFCI circuit breaker on the cord. If the plug happens to be a NEMA 5-20, that would imply they accept a circuit protected by a 20 amp breaker, which could be #12 wire so long as the distance and temperature derating specs are considered met. I guess you’ll know what they intended when you see what kind of plug comes with the cord!
