PDA

View Full Version : Can i calculate the battery side current?



__Tango
16 June 2011, 1754
Hey folks

So, with my arduino and canbus setup, I can talk to my Kelly KBL brushless controller via CAN. from this, I can get the battery side voltage, motor side current, and motor side voltage (the motor side values come as triplets -- one for each of the 3 phase wires).

Is it possible to calculate the battery side current given this info?

I want to be able to calculate things like watt-hours and watt-hour/mile, but I feel like that info is better done from the battery side. I could always get a current sensor, but was wondering if I could do it without an external sensor.

Thanks!

Nuts & Volts
16 June 2011, 1815
Battery Current should be...

bat_I = (motor_V * motor_I / bat_V )*Cntlr_eff

I would say controller efficiency should be around 96-98% in general.

lugnut
16 June 2011, 1922
Battery Current should be...

bat_I = (motor_V * motor_I / bat_V )*Cntlr_eff

I would say controller efficiency should be around 96-98% in general.

I don't think this equation applies to three phase AC controllers. You have a factor something to do with the square root of 3 and also maybe the power factor (applies to sine wave but I don't know about other types like BLDC). The straight forward method would be to use a current shunt or hall sensor in the battery line.

__Tango
16 June 2011, 2215
I don't think this equation applies to three phase AC controllers. You have a factor something to do with the square root of 3 and also maybe the power factor (applies to sine wave but I don't know about other types like BLDC).

Yeah, this is what i was thinking about. I think i remember something about how AC or BLDC controllers send about 0.7 of the voltage down the phase wires. I'm really reaching here, so i don't know.

Anyone else know?

Thanks.

markcycle
17 June 2011, 0519
It's very difficult to calculate the RMS of a non sinusoidal or other well known AC waveform (such as square or triangle). The AC peaks can be as high as 5 times the DC current. On the controller I'm working on the AC peak is as much as 2.5 times or more the DC current. The average power is always a little less than the DC power or equal to when including the controller losses

frodus
17 June 2011, 0806
doesn't the kelly have battery current on the canbus? If not, seems a bit of a letdown.

DaveAK
17 June 2011, 0826
doesn't the kelly have battery current on the canbus? If not, seems a bit of a letdown.
My Sevcon only gives motor side current. It does spit out battery voltage, but not current. I'm going to have to do that separately as well.

frodus
17 June 2011, 0940
Measure it with Noah's circuit and show on the tach..... :)

DaveAK
17 June 2011, 0952
I have a hall effect sensor I'm going to use when I get around to hooking it all up. Then I'll be able to feed it in to a microcontroller along with the CAN stuff from the Sevcon and have it all synced up and dataloggable. I could throw in a cellular module and be able to give you a real time feed. :D

GUFF
17 June 2011, 1019
I have a hall effect sensor I'm going to use when I get around to hooking it all up. Then I'll be able to feed it in to a microcontroller along with the CAN stuff from the Sevcon and have it all synced up and dataloggable. I'll could throw in a cellular module and be able to give you a real time feed. :D

Haha! Now that would bad ass!

Facebook Status: 120MPH.....

I also vote measuring with the hall effect sensor. If you go that route I would be very interested to see what you come up with similar to Dave is describing above.

This thread has me rethinking if my Curtis spyglass is showing me motor amps or battery amps. Since the spyglass and the programmer share the same connection point I would wager that if I cannot see battery amps in the software program then no, its motor amps. I did not see a battery amps value in the monitoring function listed in the manual (Frodus, correct me if I am wrong).

DaveAK
17 June 2011, 1028
It's a shame they don't add that measurement, but since they don't actually need the current for their algorithms I guess it was deemed an unnecessary addition. I'll bet the Spyglass is showing motor amps.

frodus
17 June 2011, 1042
but they know the curent on each phase.... and can extrapolate....

__Tango
17 June 2011, 1055
doesn't the kelly have battery current on the canbus? If not, seems a bit of a letdown.

Nope. I got verification from Fany that they can't do battey side current monitoring. :(

__Tango
17 June 2011, 1101
I have a hall effect sensor I'm going to use when I get around to hooking it all up. Then I'll be able to feed it in to a microcontroller along with the CAN stuff from the Sevcon and have it all synced up and dataloggable. I could throw in a cellular module and be able to give you a real time feed. :D

What hall effect sensor are you using? I'm looking into getting one as well. Dimitri from MiniBMS uses a Tamura L01Z, but i don't think that will fit my 1/0 gauge.

Ooooohhh...cell connection. I hadn't thought of that. :)

Actually, that's going off in a slightly different direction that I'm thinking for mine, which may be more BMS like in the future...

__Tango
17 June 2011, 1102
Nope. I got verification from Fany that they can't do battey side current monitoring. :(

More specifically, Fany said:


the hardware can not support the sample of battery current

__Tango
17 June 2011, 1103
but they know the curent on each phase.... and can extrapolate....

That's where this whole thread started. Given mark's input, i don't feel hopeful that I can do this though. :(

frodus
17 June 2011, 1106
the sum of the instantanious current on each phase would add together, they should already be doing this inside, it should be a simple firmware update to get that data outside the controller.

For you, you get "average" current on each phase, so it doesn't really help get an exact number. You're estimating.

__Tango
17 June 2011, 1141
T
the sum of the instantanious current on each phase would add together, they should already be doing this inside, it should be a simple firmware update to get that data outside the controller.

For you, you get "average" current on each phase, so it doesn't really help get an exact number. You're estimating.

Travis, i'm not quite sure what you mean by this post, can you elaborate? Thanks.

lugnut
17 June 2011, 1219
the sum of the instantanious current on each phase would add together, they should already be doing this inside, it should be a simple firmware update to get that data outside the controller.

This sum should be zero. In fact, if it is not zero, you have a ground fault. At least that is the way sine wave 3 phase VFDs behave.

frodus
17 June 2011, 1232
Actually, not completely true Lugnut, I know how you love to hang onto every little detail......Let me rephrase .... the sum of AC currents going into the motor is Zero. But with an inverter, you take the value of each phase's instantanious current and add them together to get the total battery current. You're thinking with 3 phases, I'm thinking with 6 groups of FET's (or IGBT's) because that is really what is in an inverter. You're thinking outside, I'm thinking inside, where this stuff is calculated by the microcontroller.

For a 3-phase power stage, you have 3 positive, and 3 negative groups of FET's (Or IGBT's). Each one takes care of a "swing" of the sinewave (one positive, one negative for each phase). So if you're measuring the current for each phase, you may have 2 that are positive current, and 1 that is negative current, but what is actually happening, is that you have positive current flow through 2 FET's (IGBT's) doing the positive swing and positive current flow through the other which is creating the negative swing.

So yes, you add them together.

DaveAK
17 June 2011, 1253
What hall effect sensor are you using? I'm looking into getting one as well. Dimitri from MiniBMS uses a Tamura L01Z, but i don't think that will fit my 1/0 gauge.

Ooooohhh...cell connection. I hadn't thought of that. :)

Actually, that's going off in a slightly different direction that I'm thinking for mine, which may be more BMS like in the future...
Hi, I'm Dave. I'm a dataholic.

I ended up getting the Honeywell CSLA2EL from Mouser. There was another cheaper one that I was going to get originally but they were having supply issues. I probably should have just waited because I haven't got around to using it yet.

podolefsky
17 June 2011, 1320
This sum should be zero. In fact, if it is not zero, you have a ground fault. At least that is the way sine wave 3 phase VFDs behave.

Yes, if you sum the instantaneous voltage or current for three phase, you will get a sum of zero. But the thing you want for backing out the battery current is motor side power. P = I*V. For three phase it's a little complicated, but not that much.

The reason the power isn't zero is because the voltage and current have the same sign, so I*V will always be positive.

So with a BLDC controller, it depends on what the controller is spitting out for the voltage and current on each phase. I doubt it's instantaneous, the data stream for that would be way too dense. I'm guessing it's RMS, which *I think* means you should be able to do P = I*V for each phase and just add them up.

For BLDC, it's going to be 2x the power for any one phase (at any given time, two phases are on and one phase is off). So P_total = 2*I*V for one phase. Then do I_battery = P_total / V_battery

This image shows how V and I are lined up (in phase) for BLDC. (here, solid lines are voltage (e) and dashed are current (i))

http://www.appliancedesign.com/AM/Home/Images/ACM%20Fig-4_large.jpg

frodus
17 June 2011, 1337
Depends on the BLDC....

Mark, how does kelly do the BLDC? Is it a trapezoidal wave? sawtooth? or pure sinusoidal?

podolefsky
17 June 2011, 1358
Kelly BLDC is trapezoidal. Sinusoidal is generally referred to as AC (or PMAC), as in the Sevcon Gen4 controllers. (I shouldn't say generally, people use BLDC to refer to everything...but if we're being picky)

I only know this because I asked John Fiorenza about running the ME0913 with the Kelly, and he said the trapezoidal would be noisier than the sinewave from the Sevcon.

podolefsky
17 June 2011, 1419
It's very difficult to calculate the RMS of a non sinusoidal or other well known AC waveform (such as square or triangle). The AC peaks can be as high as 5 times the DC current. On the controller I'm working on the AC peak is as much as 2.5 times or more the DC current. The RMS or average power is always a little less than the DC power or equal to when including the controller losses

RMS values for sine, square, and sawtooth are simple:

http://en.wikipedia.org/wiki/Root_mean_square

And sorry for being picky...but, RMS refers to voltage or current. There's no such thing as RMS power, just average power (usually averaged over time). For 3-phase, both instantaneous and average power are constant and equal.

You are right that average power out has to be equal to DC power in (minus efficiency losses).

Just clarifying - not trying to be a d!ck.

__Tango
17 June 2011, 1439
Noah. Thanks. That was a great explanation that my pea brain can kind of understand....er...maybe.

I have data from a ride today. it's in excel spreadsheets at the moment. The first one is a summary of the whole ride at one sample per second.

http://electriceptor.files.wordpress.com/2011/06/00019-summary.xls

It's got all of the data from GPS and Canbus (most everything is labeled including units up till column O. At that point, look at the kelly docs (http://kellycontroller.com/mot/downloads/KellyKBLUserManual.pdf) for the unit informatino (for CCP_A2D_BATCH_READ1, CCP_A2D_BATCH_READ2, CCP_MONITOR1, and CCP_MONITOR2)

The second and third are approx 1 second dumps of data as fast as i can get for the motor side current and voltage (which currently is about 150 samples/sec).


approx 2.18mph, accelerating: http://electriceptor.files.wordpress.com/2011/06/00019-slow-raw.xls
approx 55.65mph, getting close to the top end speed: http://electriceptor.files.wordpress.com/2011/06/00019-fast-raw.xls


I was trying to calculate power based on this information and noah's explanation, can i take the max of the I and V for a timeslice, multiply it by 2 to get power?

__Tango
17 June 2011, 1445
Hi, I'm Dave. I'm a dataholic.

Ditto. I'm in a 12 step program ("Hi Dave!") and am working on it.



I ended up getting the Honeywell CSLA2EL from Mouser. There was another cheaper one that I was going to get originally but they were having supply issues. I probably should have just waited because I haven't got around to using it yet.

What was the cheaper one you were considering?

frodus
17 June 2011, 1448
But I wonder why Kelly can't add that to the firmware? I mean, isn't battery current REALLY important for us EV'rs? I guess they think the BMS is gonna do that.... or some other sensor.

They're already calculating current on the output (they have to in order to current limit), so why not just store those currents and add them and spit that out on canbus? It's already IN the software (partially), just not implemented.

podolefsky
17 June 2011, 1502
But I wonder why Kelly can't add that to the firmware? I mean, isn't battery current REALLY important for us EV'rs? I guess they think the BMS is gonna do that.... or some other sensor.

They're already calculating current on the output (they have to in order to current limit), so why not just store those currents and add them and spit that out on canbus? It's already IN the software (partially), just not implemented.

They probably can, they just didn't. My Alltrax spits out battery current, motor current, and battery voltage...but not motor voltage (but you can get that from the throttle %).

Speaking of things they should do .... that HSR controller I was looking at (that does series motor regen) has separate motor and battery current limits. I think most of the other Kelly's just have motor. I don't know why, it seems like a really useful feature. it will let you protect your batteries without limiting motor current at low RPM (when motor current > battery).

markcycle
17 June 2011, 1509
great data Alex
I plotted phase current VS battery voltage just to get a feeling for voltage drop under load
https://www.sugarsync.com/pf/D336079_7080457_82587

markcycle
17 June 2011, 1524
The Kelly KBL controller has both motor and battery current limit

Just clarifying - not trying to be a d!ck.

podolefsky
17 June 2011, 1531
The Kelly KBL controller has both motor and battery current limit

Just clarifying - not trying to be a d!ck.

Ah, cool. Good to know.

podolefsky
17 June 2011, 1534
Noah. Thanks. That was a great explanation that my pea brain can kind of understand....er...maybe.

I have data from a ride today. it's in excel spreadsheets at the moment. The first one is a summary of the whole ride at one sample per second.

http://electriceptor.files.wordpress.com/2011/06/00019-summary.xls

It's got all of the data from GPS and Canbus (most everything is labeled including units up till column O. At that point, look at the kelly docs (http://kellycontroller.com/mot/downloads/KellyKBLUserManual.pdf) for the unit informatino (for CCP_A2D_BATCH_READ1, CCP_A2D_BATCH_READ2, CCP_MONITOR1, and CCP_MONITOR2)

The second and third are approx 1 second dumps of data as fast as i can get for the motor side current and voltage (which currently is about 150 samples/sec).


approx 2.18mph, accelerating: http://electriceptor.files.wordpress.com/2011/06/00019-slow-raw.xls
approx 55.65mph, getting close to the top end speed: http://electriceptor.files.wordpress.com/2011/06/00019-fast-raw.xls


I was trying to calculate power based on this information and noah's explanation, can i take the max of the I and V for a timeslice, multiply it by 2 to get power?


Wow - lots of data. Hard to tell exactly what's going on, since I think it's y-connected the a,b,c voltages and currents aren't straightforward to read. But I see 3 columns labeled B+(V), iAvg and vAvg. Could those be battery voltage, motor current and motor voltage?

DaveAK
17 June 2011, 1649
But I wonder why Kelly can't add that to the firmware? I mean, isn't battery current REALLY important for us EV'rs? I guess they think the BMS is gonna do that.... or some other sensor.
My guess would be that if they don't need it in their algorithm they're not going to bother to measure it. Sure it would be nice, but it's also going to cost a couple more dollars to implement. My Sevcon is exactly the same way. Everything the controller needs to know to operate is available to the end user, anything else and you're **** out of luck.

DaveAK
17 June 2011, 1650
What was the cheaper one you were considering?
I'll see if I can find it when I get home.

__Tango
17 June 2011, 1701
Wow - lots of data. Hard to tell exactly what's going on, since I think it's y-connected the a,b,c voltages and currents aren't straightforward to read. But I see 3 columns labeled B+(V), iAvg and vAvg. Could those be battery voltage, motor current and motor voltage?

yeah, lots of data...definitely. Let me elaborate some.

That set of data was from a ride that is 3.5 miles long. It took about 5.5 minutes.

On the "Summary" sheet, the data taken once per second (the EM406 GPS only produces data at once per second, so that seemed a good sample interval).

I realize that there is a lot of current and voltage data that is missing for motor spinning at high rpms, so that's where the second set of data comes in. i also get the motor current and voltage readings as fast as i can (which is, as i mentioned before) about 150 times per second. The "Raw" (what was the "fast" and "slow" sheets in the previous post) are slices of that set of data. if you're interested, the entire raw data from the entire trip is at: http://electriceptor.files.wordpress.com/2011/06/00019-rw.xls (beware, it's 1.5MB).

So, for the "Summary" page:

The B+ voltage (column G) is reported by the controller directly (as a number from 0-255, and the docs say:

Vreal = Vcontroller * 1.84

"Vreal" is the real voltage, and "Vcontroller" is the voltage output from the controller.

the "Ia", "Ib", "Ic" (columns T, U, V) are reported by the controller and I think the units are Amps.

The raw Voltage numbers (Va, Vb, Vc -- columns W, X, Y) are reported by the controller as an integer between 0-255 (described above).

On the summary page, all of "Ia", "Ib", "Ic", "Va", "Vb", "Vc" are snapshot numbers for exactly that instant where the reading was taken.

The "iAvg" (column J), "vAvg" (column K) are calculated by me (on the arduino) as follows:

For each summary timeslice (approx 1 second), I take the max of the Ia, Ib, Ic for the 150 or so CAN readings and average that number. The same applies for the voltage.

If you notice in columns M and N are lableed "wh/mi" and "m/kWh" (which i realize are effectively the same thing). These numbers are based on IAvg and Vavg, so given that i'm not sure if how i calculated them is correct, i'm not sure if my other caluclations are correct as well.

podolefsky
17 June 2011, 1931
Tango, that raw data helps a lot. The summary data didn't seem quite right. Here's what's happening (check me if I'm wrong, as usual).

You have a wye connected motor. See p 10 of the Kelly manual, or image below and ignore the Neutral.

http://www.teal.com/images/ILL_AN15c.gif

The voltages recorded are at the points A, B, and C. At any given time, two windings are on and one is off. This is done by putting, say, 50V at A, 25V at B, and 0V at C. This makes current flow from A to C. No current flows through B - this is because windings A and C form a voltage divider where the center is at 50 / 2 = 25V. Since this is the same voltage as B, no current flows through winding B.

So the total power will be Va*Ia. Actually, in your data it records the current going into the point at 0V, so it would be Va*Ic (it's just their convention I guess). The voltage at B doesn't contribute to the total power...it just holds winding B neutral so no current flows.

When you found Vmax and Imax, those were the values you need. They are equivalent to Va and Ic in the example above. Pmax = Vmax*Imax is the total power output from the controller at that point in time. That should be the same power as the battery puts out.

All said and done, I think the formula you want is:

I_battery = P_max / V_battery = V_max * I_max / V_battery

podolefsky
17 June 2011, 1943
By the way, if you want to see how this works, try this simulation.

http://phet.colorado.edu/en/simulation/circuit-construction-kit-dc

And build a circuit like this one:

1337

With these voltages, current is running through the loop on the left. There's no current through the upper-right resistor.

DaveAK
17 June 2011, 2015
What was the cheaper one you were considering?
Turns out it was the Tamura L01Z. They're still out of stock though. Maybe Dimitri has a source for them, but like you said, I don't think it will take a 1/0 cable.

__Tango
18 June 2011, 1818
Turns out it was the Tamura L01Z. They're still out of stock though. Maybe Dimitri has a source for them, but like you said, I don't think it will take a 1/0 cable.

Great. Thanks. Other than the one you got, any ideas on getting anything bigger that will fit my 1/0 Gauge?

lugnut
18 June 2011, 1833
Great. Thanks. Other than the one you got, any ideas on getting anything bigger that will fit my 1/0 Gauge?

I solved this problem once by fashioning a short copper bus bar which would fit through the current sensor and connect to the 1/0 lug and to the battery terminal. It could go on a contactor terminal or fuse block post also. In another case, I stripped the insulation off of the cable back far enough from the terminal lug such that the current sensor could fit over the bare strands. Obviously I crimped the lug after slipping on the sensor :-)

podolefsky
20 June 2011, 1317
That's what I did with the Hall effect sensor for my BMS (blue box on the right). Aluminum bar attached to B+ cable and other end to fuse. The sensor has a rectangular hole, the bar is machined just to that size.

https://lh6.googleusercontent.com/-gvqUjCLeKMw/TXR7AWBkghI/AAAAAAAAEFw/X6cwko10qBI/s800/DSC_72371.jpg

__Tango
20 June 2011, 1743
While we're on the topic....any reason one would use a hall effect sensor vs. a shunt resistor (e.g. what the CycleAnalyst uses or http://www.powerwerx.com/tools-meters/current-shunt-resistors-500-amp-max.html)?

lugnut
20 June 2011, 1818
While we're on the topic....any reason one would use a hall effect sensor vs. a shunt resistor (e.g. what the CycleAnalyst uses or http://www.powerwerx.com/tools-meters/current-shunt-resistors-500-amp-max.html)?

It depends on what you want to do with the signal. Some systems, like the Pak Trakr, need to use the supplied hall sensor. Others like the Cycle Analyst need to use the shunt. Each has an advantage. The hall sensor is isolated from the battery pack, introduces zero resistance and is light weight. However it requires an external power supply. The shunt does not require an external supply, but is electrically connected and heavy.

podolefsky
21 June 2011, 1141
By the way, did any of the stuff I said make sense? I realized it was a big dump of information...

Basically, you can use I_max*V_max to get the motor power at each time step (in the raw data). You don't need to multiply by 2.

lugnut
21 June 2011, 1716
By the way, did any of the stuff I said make sense? I realized it was a big dump of information...

Basically, you can use I_max*V_max to get the motor power at each time step (in the raw data). You don't need to multiply by 2.

I don't know. It sounds too simple. It sure doesn't work that way for sinusoidal 3 phase machines. I haven't dealt with BLDC that much. Or taken a good look at his data. I'll put it on my to-do list :-)

__Tango
21 June 2011, 1749
Thanks Noah. Yes, it did make a bit of sense to me, but based on Mark's comments, it too seems a bit simple...almost too much so. But that is showing my obvious ignorance about this stuff. From what you're saying, the motor Power (Imax and Vmax from the three phase wires) taken at any given point is basically the same as the battery Power (battery voltage * battery current). Well, except for losses in the controller.

I'm currently working through a GPS issue in my arduino/canbus/gps logger, but once i get that sorted out, i should have better data to post. I'm still considering getting a hall sensor to record the battery current directly though.

podolefsky
21 June 2011, 1810
It does sound simple, but I'm 90% sure it is correct. It's not anything like 3-phase AC. In 3-phase AC, you have all three phases non-zero at the same time. With BLDC, you only have two phases on at once...but really you just have current flowing in one direction through two of the windings.

I think what Mark said is that it's hard to calculare RMS current from a non-sinusoidal signal. That's true if it is some really weird signal - but for the square wave coming out of a BLDC controller, I'm pretty sure the RMS current is just equal to the peak current.

Anyway, one way to find out is to try it. I have a BLDC hub motor on my scooter, and a clamp current meter. I'll see if I can get data to see if what I was saying actually works. (or if you manage to get battery side current measured, that would also tell us).

markcycle
22 June 2011, 0425
Here is a video of the phase current at full RPM its a lot messier at low RPM


http://www.youtube.com/watch?v=brCg0krnLBc

podolefsky
22 June 2011, 0813
Nice, thanks Mark. That helps me get my head around what's going on better. Just for my education, mind answering some questions?

I take it the two traces are the two motors? Why is the lower one so much messier?

Is the data that Tango has the actual instantaneous voltage and current? If so, then I see the problem you were talking about, the signal isn't a nice square wave.

But, it still seems like you should be able to calculate RMS over a few cycles and get a pretty good estimate. It should wash out the sawtooth variations.

I dunno - just trying to figure this out because I think it should be possible...somehow.

__Tango
22 June 2011, 1000
Is the data that Tango has the actual instantaneous voltage and current? If so, then I see the problem you were talking about, the signal isn't a nice square wave.

I believe so. i don't have any proof of this, but the way the sequence goes, it seems to be numbers at that instant.

markcycle
22 June 2011, 1053
Nice, thanks Mark. That helps me get my head around what's going on better. Just for my education, mind answering some questions?

I take it the two traces are the two motors? Why is the lower one so much messier?

Is the data that Tango has the actual instantaneous voltage and current? If so, then I see the problem you were talking about, the signal isn't a nice square wave.

But, it still seems like you should be able to calculate RMS over a few cycles and get a pretty good estimate. It should wash out the sawtooth variations.

I dunno - just trying to figure this out because I think it should be possible...somehow.

Yea the lower one is messier I switched sensors and it follows the motor. But as the load increases there is load sharing. I don't want to go off on this as it isn't relevant to the thread. I'll repost the video in the EnerTrac thread for that discussion.

The thing is this is a full RPM when the controller is not doing PWM or very little PWM. At low throttle it is a more complex waveform. I'll post pictures of that on the next dual motor test - a few days from now.

The peak voltage the controller is reporting must be the expected peak of that lob or cycle as a function of the PWM duty cycle.

podolefsky
22 June 2011, 1216
I was just reading back over this. These two statements don't make sense together.


Nope. I got verification from Fany that they can't do battey side current monitoring. :(

But...


The Kelly KBL controller has both motor and battery current limit

So how can they do battery side current limiting if they can't do battery side current monitoring? Something is weird - they know the battery side current, either directly, or they are calculating it from motor data. If they're calculating it, you should be able to do the same thing (in principle at least).

frodus
23 June 2011, 0952
yeah, if they are limiting battery current, then they are monitoring/calculating it.... and it shouldn't be hard to ask them to throw that on the canbus..... ask kelly to update the firmware.

__Tango
23 June 2011, 0955
yeah, if they are limiting battery current, then they are monitoring/calculating it.... and it shouldn't be hard to ask them to throw that on the canbus..... ask kelly to update the firmware.

Yes, they certainly are limiting the battery side current. I'll put in a request to see what they say.

DaveAK
23 June 2011, 0955
yeah, if they are limiting battery current, then they are monitoring/calculating it.... and it shouldn't be hard to ask them to throw that on the canbus..... ask kelly to update the firmware.
It seems to me that Kelly are quite happy to talk to customers. Do they actually respond to requests like this? That'd be pretty cool.

__Tango
23 June 2011, 0957
Yea the lower one is messier I switched sensors and it follows the motor. But as the load increases there is load sharing. I don't want to go off on this as it isn't relevant to the thread. I'll repost the video in the EnerTrac thread for that discussion.

The thing is this is a full RPM when the controller is not doing PWM or very little PWM. At low throttle it is a more complex waveform. I'll post pictures of that on the next dual motor test - a few days from now.

The peak voltage the controller is reporting must be the expected peak of that lob or cycle as a function of the PWM duty cycle.

Super interesting mark. I was wondering though, given that I'm able to sample at around 130-150Hz, if i do average out the "highest value" numbers, do you think it would even out the peaks? And thus allow for a calculation?

podolefsky
23 June 2011, 1005
Super interesting mark. I was wondering though, given that I'm able to sample at around 130-150Hz, if i do average out the "highest value" numbers, do you think it would even out the peaks? And thus allow for a calculation?

Seems like it. Might work better to use RMS than straight average. RMS is what you need to calculate average power.

http://en.wikipedia.org/wiki/Root_mean_square

markcycle
23 June 2011, 1136
Super interesting mark. I was wondering though, given that I'm able to sample at around 130-150Hz, if i do average out the "highest value" numbers, do you think it would even out the peaks? And thus allow for a calculation?

I'm not sure. I think if you sampling that fast then maybe average every 10 samples. I didn't look at the definitions when they say phase current or voltage is it measured or a calculated value. They don't have three current sensors in that controller. My guess they read the voltage drop across the fets and assume a know value for "RDS" on and get a crude current number. This is a old trick
This is why everything is in percentage when making current settings they really don't have hard numbers. Using RDS on of the FET's means they can do approximate current limiting but they couldn't output a accurate number