Jump to content

Total Wh drained from the battery


Recommended Posts

2 hours ago, meepmeepmayer said:

All these numbers, while the tendency may be clear, are just unreliable.

This is really disappointing to read.  I would I have hoped that my $900 unicycle would at least take as accurate of measurements as that of say a couple hundred $ battery charger that measures the current going back in, but maybe not.  Going to do what I can to evaluate the accuracy of my glide 3 data at least. I don't have a fancy charger like that for my unicycle though, only one for hobby RC. <sigh>

Link to comment
Share on other sites

22 hours ago, Blueblade said:

This is really disappointing to read.  I would I have hoped that my $900 unicycle would at least take as accurate of measurements as that of say a couple hundred $ battery charger that measures the current going back in, but maybe not.  Going to do what I can to evaluate the accuracy of my glide 3 data at least. I don't have a fancy charger like that for my unicycle though, only one for hobby RC. <sigh>

I see your point, but to be fair, a very large portion of that "$900" is used on the motor and batteries (quoted the price because we should actually compare manufacturing costs, but...). Comparing the price to a couple of hundred battery charger doesn't really work.

Another problem is that it's relatively easy to measure the current / calculate watthours when charging a battery with steady constant current, or at least it's changing very slowly. In the wheel, the current is changing very rapidly all the time as the motor turns (coils turning on and off very fast, transient current spikes etc), so for accurate measurements, they'd need to add a lot there (high speed ADC's/current measurement ICs, even a separate MCU?). I don't think many people would be willing to pay much more just so that the wheel could better measure the current, if it only just gives you better estimation of energy consumption, but nothing else.

Link to comment
Share on other sites

3 hours ago, esaj said:

Another problem is that it's relatively easy to measure the current / calculate watthours when charging a battery with steady constant current, or at least it's changing very slowly. In the wheel, the current is changing very rapidly all the time as the motor turns (coils turning on and off very fast, transient current spikes etc), so for accurate measurements, they'd need to add a lot there (high speed ADC's/current measurement ICs, even a separate MCU?). I don't think many people would be willing to pay much more just so that the wheel could better measure the current, if it only just gives you better estimation of energy consumption, but nothing else.

@Blueblade

Internally there are sufficient good measurements for self balancing.

But for this only something like the afair in the past stated 100 measurements per second are enough. Great for balancing - could work a bit for energy consumption measurement, but not really accurate.

Also for self balancing many/most values imo don't have to be accurate (calibrated) - they are just used for a control loop to "straighten" the wheel. So absolute numbers are not that important.

The reported values are just a subset of whats available internally and just happened to be sent, so the app can show something. Imo this values were never ment to do domething really "usefull".

The only real shortcoming of internal measurements were with some ?early charges of the KS14? were voltage measurements were off and the wheels blocked too early due to overvoltage conditions.

So - the wheels are sold as self balancing devices - this task is nicely fullfilled. They are no accurate measurement equipments and don't need to be. Some addutional features in this directions would be great but not really necessary for riding and having fun!

Link to comment
Share on other sites

On 3/8/2019 at 9:39 AM, esaj said:

Another problem is that it's relatively easy to measure the current / calculate watthours when charging a battery with steady constant current, or at least it's changing very slowly. In the wheel, the current is changing very rapidly all the time as the motor turns (coils turning on and off very fast, transient current spikes etc), so for accurate measurements, they'd need to add a lot there (high speed ADC's/current measurement ICs, even a separate MCU?).

Yeah, i guess I didn't think about that, probably a situation much like a 4-ohm speaker coil, they measure static resistance for rating it, but actuals during operation vary widely depending on input frequency, resonance effects of the air/enclosure etc.

My first thought was wheellog doing several samples per second would be pretty granular for understanding trending. I noticed looking at actual log data though, the sample intervals aren't super consistent. I see anywhere from 4-7 samples per second looking at various periods.  Might be able to do an average current grouped by second (truncate the ms portion before grouping) to convert into a new series of "amp seconds" samples, and just add those up and convert to amphours to estimate power consumption.  Could try to cross-reference that against starting/ending battery % for same time range, and or measured recharging data.

I'm not sure if it would be more accurate to take charger amperage * time, or try to use something like a kill-a-watt on the charger for a poor-mans version of a measuring battery charger. (Would need to estimate the amount of inefficiency in the charger's ac-dc coversion first of course).

It may be all for naught, but it would be interesting to know the range of error if it's not completely accurate might still be useful if not 100% accurate.

Of course it probably varies significantly from Wheel to wheel on how accurate the sampled data is.

Maybe do all that all just to end up confirming @Chriull when he says "it was never really meant to do something useful" 😀

Link to comment
Share on other sites

34 minutes ago, Blueblade said:

I'm not sure if it would be more accurate to take charger amperage * time, or try to use something like a kill-a-watt on the charger for a poor-mans version of a measuring battery charger. (Would need to estimate the amount of inefficiency in the charger's ac-dc coversion first of course)

Would probapbly be the easiest way.

36 minutes ago, Blueblade said:

Maybe do all that all just to end up confirming @Chriull when he says "it was never really meant to do something useful" 😀

The problems are:

The wheels measure something around 100 values a second and just report 10 (KS) or 4-7 for your wheel.

One needs the battery current for power calculations. But most wheels report motor current. And/or report some "preprocessed/modified/from time to time extremely seldomly" reported battery voltages like ninebots. For more on battery vs motor current take a look at 

Additionaly the reported current has just some proportional value to the motor current, as one can not directly get acceleration values from it...

(The "problem" i discoverted and did not solve in 

)

For electric motors as the used BLDCs one has the relation U (?emk?-?induced?)=kv x rpm and M (torque) = km x I Motor. With kv x km = 60/pi. Which was not true for reported values from my KS16B and S. So the currents are somehow not the effective motor current...

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

×
×
  • Create New...