Dummy loads: when would you use constant power or constant resistance?

[Laogeodritt]

Hey guys! I've followed Dave for a couple of months now, but this is my first time joining and posting. I remember Dave mentioning in his video on his constant-current dummy load that his old battery tester (he used the PCB for the load) had firmware constant-power and constant-resistance mode. Constant-current mode's usefulness is fairly obvious to me; but in what cases specifically (for either batteries or DC supplies) would you want to test in constant-power or constant-resistance mode? What information could you get out of that? I've done a quick search and it doesn't look like anyone's asked that question yet.

Thanks!
Marc (Laogeodritt)

P.S. If this is in the wrong section, mods please feel free to move it over to Beginner. This seems like a fairly fundamental question, but at the same time I'm not looking for the overly basic/simplified/hand-wavey answer, so I'm not sure...

[JoannaK]

I'm in noway the best knowing, and I know my english is kinda broken. 

Normally when batteries are tested and the discharge curves and power values are calculated, the constant current mode is used. This has become standard method for battery manufacturers. This also approximates the system using linear power regulator.

Constant power mode is more realistic to use when making estimations for systems with switch mode power supplies. With those, voltage drop at the battery (due depletion) causes current to increase, thus keeping the total power consumption constant.

Constant resistance simply simulates the normal passive-Resistor -like power dispassiation curve, where voltage drop reduces the current.

[Laogeodritt]

JoannaK: Thanks for the reply. I understand how the modes work and how to implement them; what I'm asking about is how you'd use them in testing supplies or batteries and what information you can get out of them specifically (or as specific as possible for such a general question). For example, you said constant power is useful for "making estimations for systems with switch mode power supplies"—which estimations specifically?

By the way, your English is quite solid! You're just missing a preposition here and there.

EDIT: Wait, I figured out the constant power statement above. Regulated switch-mode supplies with a constant load draw constant power from a battery.

[BravoV]

Not an expert nor an EE, but here my take.

Constant power (CP) load usually more complicated than constant current dummy load, and also "again" usually came with certain protection mechanisms like cut off whenever reaching certain voltage or current, this is to protect the power source from over drained or over current/temperature and also a timer. Ideally its also equipped with data logging feature to measure the watt/hour.

Remember, constant power means whenever the voltage is lower, it will drain higher current in order to get the constant power and vice versa. (Watt = Volt x Amp)

Examples, using it on Li-Ion battery, the purpose is to see whether the battery capable of sourcing "constant wattage" for certain load level, say like a cellular phone which has a switching boost circuit in it to supply a constant voltage to the circuit. So the CP device will start to drain the battery to simulate the cell phone, and as we know the battery voltage will drop when draining started while the power switching circuit will eventually starting to draw higher and higher current as the battery voltage is getting lower to compensate the required constant voltage for the cell phone's circuit.

The extra protection mechanism will kick in when the battery voltage reached the minimum safe limit before damaging the battery or certain maximum current, and sometimes certain max temperature as well. Once done, the results from logged data + the timing will show the battery capability to give the required "constant" power for certain period. Thats how they came out the popular term called "talk time" for cell phone.

Hope this helps.

[JoannaK]

By the way, your English is quite solid! You're just missing a preposition here and there.

Thanks. We don't use prepositions (in Finnish language), so those tend to be a bit hard to understand and remember. Also the articles tend to be more hit/miss, at least for me. 

[penfold]

It is important to firstly distinguish what you are interested in finding out from the system.
If you are interested in battery testing then you are primarily concerned in the variation of conditions as affected by a constant current drain. Because the total charge extracted from the battery is dependant on the current drawn during the test the tests are standardised to use a tightly controlled current during test, really so that the test is easily repeatable. It is perfectly feasible to perform the test using varying current models (think vehicle battery testing and the loads that those batteries would be subjected to in practice). If these tests are performed under constant power or resistance modes then this current will change during the test because of the changing battery voltage, and because this voltage is not particularly well defined (dependant on cell ageing etc) these tests don't produce easily repeatable results, but the results are just as valid.

If you are testing the capability of a battery to supply power to a particular system then thats where constant power might be beneficial. If a switched mode power converter supplies power from a battery to a constant load at a constant voltage then the output power will be constant and thus so will the power at the input, but not necessarily the voltage and current. This kind of test is only useful to the system developer and thus standardisation isn't particularly important.

The fun part starts if you're testing a DC/DC converter. If you consider a regular DC power supply with current limit capability. Then if you connect the electronic load and start drawing a constant current, then because of the very low internal resistance of the supply a small change in current will produce in only a very small change in the supply's voltage but you can easily test the supply over its normal load range. If you reverse the situation and use a constant voltage load, then you have a very small output voltage change to test the supply over its constant voltage region (it may only change 100mV from zero to full load), so in order to test its full load range you'd be having to make very minute changes from its unloaded output voltage to its ever so slightly lower full load output voltage. That of course requires some prior knowledge of the power supply in order to determine what voltage range you're working over.

If you now have a device that you can set at any resistance you do desire, then it doesn't actually matter how much you know about the power supply and make a reasonable sweep of the output characteristics using a reasonable number of resistance values. You will get information about any constant current and constant voltage regions. It is a very useful tool for testing devices such as PV cells which have both a constant voltage and current region of operation where the most interesting point in the characteristic is where the CV mode transitions to the CI mode. In the PV case constant resistance modes are very useful because the output voltage and current are quite difficult to determine before any measurements are made.

The problem I have always had with constant power modes is the ability of the electronic load to maintain constant power for pulsating sources, such as a rectifier/capacitor with a lot of voltage ripple. Similarly for constant resistance modes. For those reasons when ever I have to make meaningful measurements of output voltage or current ripple, I just pick a nice power resistor to give the desired voltage and current, that way you can have reasonable confidence in the load condition.

Apologies for wordiness

[Laogeodritt]

penfold: Thanks! That pretty much answers what I wanted to know.

Examples, using it on Li-Ion battery, the purpose is to see whether the battery capable of sourcing "constant wattage" for certain load level, say like a cellular phone which has a switching boost circuit in it to supply a constant voltage to the circuit.

Thanks for the reply. This was pretty much the kind of thing I was looking for—an example of a test case and the information obtainable from it.

JoannaK: Heh, I understand. I live in Québec and went to high school here (native English speaker though), so I'm familiar with misplaced prepositions and articles thanks to the native francophones I've known.

[bfritz]

I perform testing of battery packs pretty often.  It is pretty common to test battery packs for the runtime you could expect from the device.  A flashlight bulb would be a good representation of a constant resistance load.  A constant power load is a good representation of a device with a switchmode buck converter, like a laptop.