Cycle Analyst V3 preview and first beta release

hjns said:
Thanks for the reply. I understood from your CAv3 Manual that this additive power calculation is valid for the V+ pad on the CA. Now I understand that this is also valid for the dedicated power output on the CA PCB for the Thun. This makes me feel much more comfortable.

... I will meausure the current through pin 1 of the CA-DP connector without and with Thun.
Just to clarify - here's an annotated snapshot of p.44 from the Guide:

CAV3-maxCurrentCalc.png
So - the table values + 10ma should give the current readings seen at pin 1 of CA-DP.

Thanks for the comments - maybe the Guide needs a little revision to clarify this matter... :D
 
I broke my external shunt :oops: when moving it from the VW bus controller plate to the new one for my Dune buggy, so I am buying a new and bigger shunt. I always have problems calculating rShunt so I'd like to get a second opinion.

I am planning on buying a 1000a 50mv shunt. If I divide 50mv by 1000a, I get a resistance of .0500mOhm. IIRC the CA expects a 200mv shunt instead of 50mv, so I need to multiply by (4) to get the correct rshunt value, 4 x .0500 = .2000 mOhm.

Because the load exceeds 100a, I'll program the CA with Cal->Range=Hi. I think the minimum rShunt value under Cal->Range=Lo is .7630, but under Cal->Range=Hi the minimum is .0763.

Did I calculate the rShunt correctly, and am I correct that the CA will accept this rShunt value? :D

-JD
 
oatnet said:
I am planning on buying a 1000a 50mv shunt. If I divide 50mv by 1000a, I get a resistance of .0500mOhm.
That is correct

IIRC the CA expects a 200mv shunt instead of 50mv,

That is not quite correct. The CA will read up to +- 200mV across the shunt terminals but it doesn't care at all what the nominal mV rating is. A 50mV 100A shunt is identical to a 100mV 200A shunt. So in the case of a 50mV 1000A device (0.05 mOhm), the +- 200mV range would would mean it could use your shunt to measure up to 4000 amps of current. However, the lowest shunt resistance that the CA can accept is about 0.076 mOhm, so if you have a 0.05 mOhm shunt, you'll never be able to get it calibrated. What is the actual average current draw you expect in your vehicle? Most likely a 500A 50mV shunt (=0.1 mOhm, reads up to +- 2000A), will be fine, and you could possibly get away with a 200A 50mV shunt (0.25 mOhm) as well. The amperage rating on a shunt is not the maximum it can handle, it is just the current that will produce the associated voltage drop.

so I need to multiply by (4) to get the correct rshunt value, 4 x .0500 = .2000 mOhm.

No, a mOhm is a mOhm, there's no need to arbitrarily scale it.

-Justin
 
now that ive switched to summer tires there smaller and it changes the diametergoin from a 4.8 tire to a 3.8 shortins the tire .ive changed the numbers and the speedo is still wrong i test it with my gps and it says still .,25 mile short per 1 mile.in can change it back to 2259 or way back to 2136 and still no big change always short on miles .do i need to reflash my ca again.is there somthing im missing . thanks
 
justin_le said:
oatnet said:
I am planning on buying a 1000a 50mv shunt. If I divide 50mv by 1000a, I get a resistance of .0500mOhm.
That is correct

IIRC the CA expects a 200mv shunt instead of 50mv,

That is not quite correct. The CA will read up to +- 200mV across the shunt terminals but it doesn't care at all what the nominal mV rating is. A 50mV 100A shunt is identical to a 100mV 200A shunt. So in the case of a 50mV 1000A device (0.05 mOhm), the +- 200mV range would would mean it could use your shunt to measure up to 4000 amps of current. However, the lowest shunt resistance that the CA can accept is about 0.076 mOhm, so if you have a 0.05 mOhm shunt, you'll never be able to get it calibrated. What is the actual average current draw you expect in your vehicle? Most likely a 500A 50mV shunt (=0.1 mOhm, reads up to +- 2000A), will be fine, and you could possibly get away with a 200A 50mV shunt (0.25 mOhm) as well. The amperage rating on a shunt is not the maximum it can handle, it is just the current that will produce the associated voltage drop.

so I need to multiply by (4) to get the correct rshunt value, 4 x .0500 = .2000 mOhm.

No, a mOhm is a mOhm, there's no need to arbitrarily scale it.

-Justin

Thanks Justin,

Great explanation - I think that, for the first time, it makes sense to me, you just cleared out a lot of disinformation I got along the way. Thanks for saving me from ordering the wrong part, plus I realize I had the wrong rshunt value when I was running this hardware on the VW bus, so all my data from it was wrong so now I don't know what my average current draw was. :oops:

The controller for the existing 9" DC motor should pull <400a, but in the next phase I will upgrade to HPEVS ac motor, which will be 1 (ac-51) or 2 (AC-35 * 2) 650a controllers - I was thinking of the 1300a peak from dual controllers, but it occurs to me I should have a separate CA and shunt for each controller so 650a should be my target. I also went for a big shunt because saw a few sites recommending not exceeding 2/3 of rated current in continious discharge, but you are right a 500a-600a shunt is probably more appropriate.

I'm looking forward to leveraging the CA V3's advanced features on a super-lightweight automobile. :mrgreen:

-JD
 
botz244 said:
now that ive switched to summer tires ... and it says still .,25 mile short per 1 mile.
Hmmm - 25% error?

The numbers you are quoting are quite reasonable for the size tires you indicate. However, with a nominal circumference in the 2200mm neighborhood, you would need to increase it by 33% or about 700mm to correct the error (i.e. (1.0-0.75)/0.75 * 2200 ~= 700) - this is why your comparatively small adjustments had little effect. Clearly there is something else wrong and minor tweaking of the circumference setting cannot fix it.

Unless your CA is demonstrating some other very bad behavior, a re-flash is not indicated - you most likely have something else awry. I would suspect either:
  • incorrect pole count setting or
  • badly misaligned or faulty wheel pickup that is failing to close consistently with each spoke magnet swipe.
 
I ,ME ,tOM HAVE A CROMOTOR AND ZOMBIESS CONTROLLLER AND CA3. THE TIRES ,WHEN I FIRST PUT IT TOGETHER WAS A 4.8X26 SALSA LOU.it read fine,or close.now that i put on summer tires the tire is a 3.8, an inch and a half shorter.the measurement went from 2286 down to 2155cm circumference. the only calculation is to change one measurement .i did that,and still out of wack. i change the number all the way down to 2000and it is still under milage. i have a measured mile i check it with or 2 miles round trip . i use the pole count at 32.cause this has halls.and no sensor on the wheel.plus i double check it with my gps. i cant ignor it cuse that messes up thw watt/miles . thanks Tom from Alaska
 
So previously I had an LM35 temp sensor on a MAC motor so I performed hte LM35 sensor mod to my CAV3 as seen here: http://www.endless-sphere.com/forums/viewtopic.php?f=4&t=37964&start=270#p573801

Now I am running a crystalyte 4065 which has a 10K NTC thermistor. I started to convert my CAV3 temp sensor input back to stock but OF COURSE I've lost that tiny stock 4991 surface mount resistor. I was able to source another surface mount resistor in the 4.7K range so I installed it. RIght now the bike is sitting at 23C but the CAV3 is reporting it as 65C! I'm assuming this is due to the resistor I used.

So I was wondering what my options are? Can I do a custom sensor calibration in the CAV3? How do I know the other values? OR should I just simply source hte correct 4991 resistor? :shock:
 
teklektik said:
cal3thousand said:
What are the signs of corrupted EEPROM?
Typically an inability to set the RShunt value (or other values), or a loss or scrambling of settings.

cal3thousand said:
Would lack of throttle response even after doing the green wire mod on a 'Large screen compatible' controller fall in this boat?
Nope - something else at work there - probably settings...

Exactly what does "lack of throttle response" mean?
Inoperative, poor power, laggy performance?

I should have been more clear, but I meant that there was no throttle output to the controller. With the throttle directly connected to the controller, it would work. Plugging the throttle into the CA, the input was readable on the screen when actuating the throttle, but no joy on the output.

What are the details of "the green wire mod"?
If from the earlier Guide version I'm guessing reconnecting the CA-DP cable ThrottleOut wire to controller ThrottleIn?

I tried several ways of hooking it up including taking the CA-DP ThrottleOut to the controller ThrottleIn (swapped pins)

I ended up installing another V2 back in there for now.
 
cal3thousand said:
teklektik said:
Exactly what does "lack of throttle response" mean?
I should have been more clear, but I meant that there was no throttle output to the controller. With the throttle directly connected to the controller, it would work. Plugging the throttle into the CA, the input was readable on the screen when actuating the throttle, but no joy on the output.

teklektik said:
What are the details of "the green wire mod"?
I tried several ways of hooking it up including taking the CA-DP ThrottleOut to the controller ThrottleIn (swapped pins)

I ended up installing another V2 back in there for now.
Sorry that didn't work out, but since your V3 was detecting the throttle properly but not developing an output voltage, I'm thinking a simple Setup issue. An important indicator is the OUT field on the Diagnostic Screen.

Next time you want to give this a go, follow the steps in the Guide and implement just one throttle mod :D . If the throttle is inoperative, post which mod you used from the Guide and either post up all your ThrI and ThrO settings or (better) dump your settings to file using the CA3 Setup Utility (link here) and post the file here.

I'm certain this can be remedied straightaway....
 
botz244 said:
...now that i put on summer tires the tire is a 3.8, an inch and a half shorter.the measurement went from 2286 down to 2155cm circumference. the only calculation is to change one measurement .i did that,and still out of wack.
You need to relate some actual measurements - so far you've basically said: "it used to work, I tried a lot of new values and it doesn't work." The only measured value you provided was a 0.25mi error for traveling 1 mile (25%) which does not make sense for a minor circumference change of ~10%. You also seem to be adjusting the circumference in the wrong direction to correct a distance shortfall error - reducing instead of increasing it.

Since things seemed to work before or "were close", presumably the pole count is correct.

So- please do the following:
  1. Use the tire measurement technique described in the Guide on p.53.
    Be sure to actually ride the bike (not just roll it) so the tires will deform appropriately.
    Note this value.
  2. Forget your GPS - use the measured mile.
  3. Go to the start of the measured mile and do a Trip Reset.
  4. Ride the mile gently without tire-distorting maneuvers/acceleration.
    Note the mileage.
  5. If the distance error is more than a few percent you have some non-Setup problem - contact Grin Tech Support and tell them the noted circumference and distance measurements.
    If the error is minor but you want to tweak it in closer, follow these steps:
    • Code:
      Calculate the actual tire circumference setting based on the present CA setting and the fractional distance error:
      
        Cact = ( Dact / Dca ) x Cca
      
        where:
          Cact = Circumference - actual 
           Cca = Circumference - as set in CA Setup
          Dact = Distance - actual (measured mile = 1.00mi)
           Dca = Distance - measured by CA
    • Adjust the CA to the calculated actual tire circumference (Cact).
    • Retest the measured mile using steps (3) and (4).
If you don't want to do the measurement in step 1, then guess or use a published value. Start with step 2 but do the calculation and recalibration in step 5 before evaluating your distance measurement. The calculation should 'fix' your initial estimate. If the results using the re-calibrated circumference are still really far off, then you will need to contact Grin...
 
Bump

Here are pictures of the reading I am getting. Bike is cold at room temp which is 25C. Why is it so far off?
 

Attachments

  • IMG_20140425_125741.jpg
    IMG_20140425_125741.jpg
    117.7 KB · Views: 5,085
  • IMG_20140425_125746.jpg
    IMG_20140425_125746.jpg
    132.3 KB · Views: 5,085
m52 power! said:
So previously I had an LM35 temp sensor on a MAC motor so I performed hte LM35 sensor mod to my CAV3 ...

Now I am running a crystalyte 4065 which has a 10K NTC thermistor. I started to convert my CAV3 temp sensor input back to stock but OF COURSE I've lost that tiny stock 4991 surface mount resistor. I was able to source another surface mount resistor in the 4.7K range so I installed it.
...
RIght now the bike is sitting at 23C but the CAV3 is reporting it as 65C! I'm assuming this is due to the resistor I used.
m52 power! said:
Why is it so far off?
Well - there are only two parts in play here: the thermistor and the pullup resistor:

What is the resistance of the thermistor at room temperature?
(We know it should be 10K at 25degC.)
  • If it measures 10K then the pullup resistor may be damaged.
  • If it does not measure ~10K then the thermistor is either damaged or is not the proper part
m52 power! said:
So I was wondering what my options are? Can I do a custom sensor calibration in the CAV3? How do I know the other values? OR should I just simply source hte correct 4991 resistor? :shock:
As far as the 5K resistor is concerned, you can just parallel two thru-hole 10K resistors across the NTC and +5v pads. There's really no reason to tinker with SMD on the crowded PCB (unless you're very comfortable doing that). 10K parts are readily available. Or stack a couple of 10K SMD parts...

You cannot switch the V3 to linear mode and calibrate it to work properly using a 10K NTC device because the linear mode expects a PTC device and entry of negative coefficients is not supported. That is, with an NTC (Negative Temp Coef) device the resistance decreases with increasing temperature. For example, if you go to the SETUP TEMP SENSR screen you should see a voltage of about 3.33v for 25degC and about 1.48v for 65degC.

10k_ntc_b=4300-2.png
In any case, you can do the math, but the small difference between your substituted 4.7K part and the specified 4.99K resistor cannot account for the large error you are seeing....
 
teklektik said:
m52 power! said:
So previously I had an LM35 temp sensor on a MAC motor so I performed hte LM35 sensor mod to my CAV3 ...

Now I am running a crystalyte 4065 which has a 10K NTC thermistor. I started to convert my CAV3 temp sensor input back to stock but OF COURSE I've lost that tiny stock 4991 surface mount resistor. I was able to source another surface mount resistor in the 4.7K range so I installed it.
...
RIght now the bike is sitting at 23C but the CAV3 is reporting it as 65C! I'm assuming this is due to the resistor I used.
m52 power! said:
Why is it so far off?
Well - there are only two parts in play here: the thermistor and the pullup resistor:

What is the resistance of the thermistor at room temperature?
(We know it should be 10K at 25degC.)
  • If it measures 10K then the pullup resistor may be damaged.
  • If it does not measure ~10K then the thermistor is either damaged or is not the proper part
m52 power! said:
So I was wondering what my options are? Can I do a custom sensor calibration in the CAV3? How do I know the other values? OR should I just simply source hte correct 4991 resistor? :shock:
As far as the 5K resistor is concerned, you can just parallel two thru-hole 10K resistors across the NTC and +5v pads. There's really no reason to tinker with SMD on the crowded PCB (unless you're very comfortable doing that). 10K parts are readily available. Or stack a couple of 10K SMD parts...

You cannot switch the V3 to linear mode and calibrate it to work properly using a 10K NTC device because the linear mode expects a PTC device and entry of negative coefficients is not supported. That is, with an NTC (Negative Temp Coef) device the resistance decreases with increasing temperature. For example, if you go to the SETUP TEMP SENSR screen you should see a voltage of about 3.33v for 25degC and about 1.48v for 65degC.


In any case, you can do the math, but the small difference between your substituted 4.7K part and the specified 4.99K resistor cannot account for the large error you are seeing....

Thank you for the most excellent reply! I have measured the resistance between ground and the thermistors output, at 25C I am getting a reading of 1.968k Ohm. Does this mean it is really a 2K NTC thermistor? I am so upset over this with crystalyte, the info they've told me is its the KTY81-210 thermistor:
http://www.nxp.com/documents/data_sheet/KTY81_SER.pdf
 
Through some crazy calculations I don't even know how I came up with, I entered: 1.20V for temp unit, and 100 degrees/volt and the CA is displaying the correct voltage. Does that sound ok?
 
m52 power! said:
teklektik said:
You cannot switch the V3 to linear mode and calibrate it to work properly using a 10K NTC device because the linear mode expects a PTC device and entry of negative coefficients is not supported. That is, with an NTC (Negative Temp Coef) device the resistance decreases with increasing temperature.
I have measured the resistance between ground and the thermistors output, at 25C I am getting a reading of 1.968k Ohm.
Does this mean it is really a 2K NTC thermistor?
...the info they've told me is its the KTY81-210 thermistor:
http://www.nxp.com/documents/data_sheet/KTY81_SER.pdf
Sort of - It looks from the spec like this part (KTY81-210) is a PTC 2K thermistor - which is good because the V3 linear mode can only handle PTC not NTC devices. This is Good! :D

KTY81-210-spec3.png
m52 power! said:
I entered: 1.20V for temp unit, and 100 degrees/volt and the CA is displaying the correct voltage. Does that sound ok?
...working out some numbers:

  • From the spec the resistance at 0degC is 1.63K
  • Considering the 5K pullup resistor, at 0degC the SET TEMP SENSR screen should show this voltage:
    1.63K/(5K+1.63K) x 4.96v = 1.22v
  • From the spec the resistance at 100degC is 3.392K
  • The voltage at 100degC should be:
    3.392K/(5K+3.392K) x 4.96v = 2.004v
  • The scaling factor is then:
    (100degC - 0degC) / (2.004v-1.22v) = 127.55 degC/V
This gives us these V3 settings:

  • Setup for KTY81-210 Thermistor:
    • Temp->Sensor = LinearType
      Temp->Units = 1.22 Volts (This is actually an error in 3.0p6 - it's the voltage at 0degC not 'Units')
      Temp->Scale = 127.6 Deg/V
A room temperature test voltage reading should be:
  • From the spec the resistance at 0degC should be 2K
  • Considering the 5K pullup resistor, at 25degC the SET TEMP SENSR screen should show this voltage:
    2K/(5K+2K) x 4.96v = 1.42v
Give these values a try and see how the voltage matches the indicated temperature at room temp.

Although the V3 'Linear' setting uses simple linear scaling, this part does not actually have a linear R/T relationship. In this calculation I used 0degC and 100degC - the temp should be accurate at those points and somewhat off elsewhere. If you are going to run your motor hot, you might do better to repeat the scaling calculation using a higher temp as the top end to get the best accuracy in the region where it really counts (e.g. 120degC or higher - see thermistor datasheet) ....

For example, here's the same scaling calc but at 125degC:
  • From the spec the resistance at 125degC is 3.915K
  • The voltage at 125degC should be:
    3.915K/(5K+3.915K) x 4.96v = 2.178v
  • The scaling factor is then:
    (125degC - 0degC) / (2.178v-1.22v) = 130.48 degC/V
This gives us these V3 settings:

  • Setup for KTY81-210 Thermistor:
    • Temp->Sensor = LinearType
      Temp->Units = 1.22 Volts
      Temp->Scale = 130.5 Deg/V
Really just a few percent difference, but use the value that gives you the most usable accuracy... :D

EDIT - Ah Jeez - I forgot you have a 4.7K not 5K part. I'm going to leave the above post as is for foks with a stock V3 - you can run the calcs again using your part value - sorry 'bout that....
 
Telektik,

Thank you so much for showing me the light (so to speak ha)!

I studied your calculations and how you came up with them and understand fully what's going on here. I've never dealt with different types of temperature sensors (PTC, NTC, LM35 etc) until this so it's been a bit of a learning experience.

Anyways, with my 4.7K resistor I came up with:
1.28V @ 0degC
2.25V @ 125degC
128.8 degC/V

Entered those into my CAV3 and it worked beautifully! Thanks again :mrgreen:
 
Hi

I have a bit strange condition on my cav3 prelim. I have just changed to Hi shunt settings and I have set it up with three modes (low, PAS 250w 25kph- medium, PAS 700w 70kph- high, throttle only 7000w. I have a 3speed switch hooked up to the ca so I can use on all three settings.
When I turn the key and power up the bike, everything is fine until I use the throttle and pedal at the same time, then it kicks out in error : PC02117 U100000
T_Err U200000

If i turn the key to on, and toggle the CA buttons through all the modes (low medium high) before I start to pedal and using the throttle, everything is normal and no error occurs.
I'm thinking if the error will go away if I flash the unit with the newest software?
I can't see anything else than this to be a software bug. Everytime I start up the bike, I have to do the same to get around the error, but then it's gone for the rest of the trip..
Does anyone know why this happens? or is it just the way electronics works?:p
In advance thnx for the feedback
 
fsmuggen said:
I'm thinking if the error will go away if I flash the unit with the newest software?
It looks like you are running some really old stuff that had debug messages to report main processing loop timing overruns that were suspected in rare device freezes.

Suffice it to say that no version prior to 3.0p6 is supported- issues on obsolete pre-release firmware are just not worth investigating...

Re-flash ASAP.
 
Ok so thanks guys for all the help in the past but I'm afraid I still need some.
The main reason I bought a CA which happens to be a V3 was for power or current throttle. I'm running a MAC 7T on 12S lipo with a 40 amp infineon controller on a 26 inch wheel. This setup peaks at 45amps and about 2500 watts.
I wanted a smoother, more gas engine like, throttle response.

The only way I can get my throttle to behave normally is to set it to PASS THRU

My IN min and max thrtl V is .8 and 4.79
My OUT min and max are similar and I played with the min out to minimize the initial dead band.

1. When in PASS THRU mode, the diagnostic screen shows logical numbers for in and out throttle voltages and the controller behaves normally.
However, if I try either POWER or CURRENT throttle, the diagnostic screen shows:
Right after I leave setup about 2.2 volts and it creeps down gradually until around 1.7 volts which still makes my wheel spin.
What am I doing wrong?

2. I calibrated the zero amps but is it normal that the CA sees negative watts at stand still. Not many below 5 but it tells me I had some regen when I have a MAC???

Thanks guys
 
Are you seeing this behavior when riding or just when the bike is on the stand? Does it ride okay? Current and Power throttle can sometimes give odd behavior under low load situation on the stand.

First, I would try revising the throttle adjustments so you have a little dead zone at ZERO as recommended. This will ensure that the PI controller for Current Throttle is actually being told to set the current to 0%. If this is not the case, then the V3 will continue to increase Throttle Out to whatever voltage is required to get whatever small percentage of total power is required to meet an erroneous non-zero Throttle In setting (e.g. 1% throttle). On a stand where the motor draws little current, the CA can run the voltage up quite high trying to get a very small current - but on the ground, the current develops quickly and the seemingly odd behavior is not apparent.

If this doesn't remedy the issue, you will need to post your settings when testing Current Throttle mode. These days the best means is for you to use the CA Setup Utility and download the settings to a file and post it up here (I haven't tried this yet - you may need to zip it up or add a ".pdf" extension to trick the site. Please advise if you cannot run the utility - there's always the manual method if necessary.... :D
 
Previously, I was testing on my bike stand...

OK so I put my throttle out V at 1.3 and the motor starts spinning at around 1.45V so I have a small dead zone.
So still in Pass thru > Works fine.

With the bike on the ground and immobilized, I set the CA to Current throttle and Just after leaving the setup.
The motor wants to spin and I see about .5 amps on the CA and my Wattmeter.
If I lift the wheel, the wheel accelerates a little???
 
Ya - okay, looks like you may have some wonky config in there someplace.

Please extract and post the settings as mentioned above. You can either zip up the .hex file or add '.txt' the end to bypass the site file-type exclusion logic. See sample below.

CaSetupUtilitySaveSettings.png
 
Back
Top