Cycle Analyst V3 preview and first beta release

[lostrack]

Great writeup teklektik. Lots of new options - looks great Justin!

Not sure if this will fall on deaf ears - yet another feature suggestion - but in terms of the throttle I have an idea

It would be really useful to have a calibrated 'sweet spot' which is around 25% throttle position, which is calibrated to sustain your speed, whatever amp draw required.

It would be really helpful as a built in cruise control 'sweet spot'. This way you could knock it down by putting your throttle lower, then continue at the lower speed. Or vice versa.

This may be an effective, simple mix of current / throttle, not too hard to program?

 

[Alan B]

Nice fast delivery, Justin!

One comment - I don't recommend any display alternation on the standard displays. My adult son crashed while looking at the CA waiting for the amp hours to come up. Yes it should not happen, but sometimes it does.

 

[Hyena]

I've got my V3 on the way too, hardcore fiddling ahead. Also playing with the CA :lol:

Justin, would it be hard to have one totally user customisable screen to scroll though ?
You can configure some of it now but it would be cool to have a totally custom screen for the parameters you wanted to look at at any given time (especially useful when beta testing new features you might specifically want to look at while riding rather than the usual stuff)
In the setup you could have position 1,2,3,4 and then a scroll through menu for each position with say 10 parameters to choose from. This might be useful too for people wanting to not have certain wattage or voltage parameters on display if certain people are looking at their bike, and it might be nice to have just current speed, distance travelled, Ah and somthing else fairly innocent :wink:

If you're looking to add other gimmicky features, somthing like car trip computers could be 'distance to empty' - calculated by subtracting the current ah from the programmed ah capacity divided by the wh/km(mi). This could either go off the average wh/km when stopped (or in general) or change in real time based on the current wattage consumption. So crusing along it could show 20km range but then if you punch the throttle it drops down to 5km or what ever. It could be useful for people who need to get a certain distance and can alter their riding style accordingly (you could argue this should be common sense but it might be good for noobs) Or the CA could even enforce it if you wanted to really be gimmicky - program your minimum range requirements and the CA would limit the current to maintain a desired wh/km (or just program a target wh/km)

 

[shorza]

If you're looking to add other gimmicky features, somthing like car trip computers could be 'distance to empty' - calculated by subtracting the current ah from the programmed ah capacity divided by the wh/km(mi). This could either go off the average wh/km when stopped (or in general) or change in real time based on the current wattage consumption. So crusing along it could show 20km range but then if you punch the throttle it drops down to 5km or what ever. It could be useful for people who need to get a certain distance and can alter their riding style accordingly (you could argue this should be common sense but it might be good for noobs) Or the CA could even enforce it if you wanted to really be gimmicky - program your minimum range requirements and the CA would limit the current to maintain a desired wh/km (or just program a target wh/km)

This is a great, idea. Of course there will always be variances such as battery temperatures, etc but i would love to see this feature.

 

[hillzofvalp]

It'd be cool also if the screen flashed the inverse pixels (everything but the original black characters) whenever an alarm or pack condition was met (such as high temp). Having the inverse colored display would be quite noticeable.

 

[justin_le]

If you're looking to add other gimmicky features, somthing like car trip computers could be 'distance to empty' - calculated by subtracting the current ah from the programmed ah capacity divided by the wh/km(mi). This could either go off the average wh/km when stopped (or in general) or change in real time based on the current wattage consumption. So crusing along it could show 20km range but then if you punch the throttle it drops down to 5km or what ever. It could be useful for people who need to get a certain distance and can alter their riding style accordingly (you could argue this should be common sense but it might be good for noobs) Or the CA could even enforce it if you wanted to really be gimmicky - program your minimum range requirements and the CA would limit the current to maintain a desired wh/km (or just program a target wh/km)

This is a great, idea. Of course there will always be variances such as battery temperatures, etc but i would love to see this feature.

I think this would less useful/valuable than it might seem when you consider the wide spread of Wh/km figures encountered over different sections of a trip. Something like a car can have a fairly predictable miles/gallon given that their huge inefficiency makes changes in grades and windspeed fairly insignificant. But with an ebike, a large hill at the end or a sudden strong headwind would make any prediction of range left fairly moot. And similarly, any attempt and dolling out power in a way that guarantees the rider makes a certain distance will also lead to a way less than ideal ride experience. A constant wh/km means having not nearly enough power going uphill and an unnecessary amount of power wasted on the downhills. With an inclinometer, GPS, and route plan tied in then in principle you could do an OK job of this, but the CA is nowhere there yet

I'm of the general opinion that it's best to show only stats that are clearly known, and leave the riders in charge of proceeding based on that info and all the other environmental stuff which a rider (and not the CA) is aware of. Is it just me coming from a bit of a tech/lab background or does that make sense? -Justin

 

[justin_le]

Got mine, thank you Justin!

OK wow, thanks to Adam here for getting the shipments out fast, I wasn't expecting anyone to recieve theirs till this coming week!

Anyways, I just want to make sure people who did receive understand that this is beta and that many firmware upgrades will be happening in the coming 3-5 weeks as we get everything flushed out in preparation for the production release. The firmware that was shipped is V3Beta12 which is functional but has a lot of things still incomplete, and I'll try to list most of it here so people aren't surprised:

1) Battery SOC Gauge: At the moment, the battery fuel gauge has been reverted to being based on the open circuit pack voltage (calculated from V-IR) rather than doing the combination of this plus the accumulated Ah as described in one of my first posts. I'm still working out the best way to implement the algorithm efficiently with the new chip and wasn't ready with this at time of shipping. With LiPoly, LiMn, and SLA this compensated voltage based approach works quite well. Just be sure to input your battery internal resistance by looking at your pack voltage before and after applying full throttle, then use RBatt = (V1 - V2) / A, where V1 is the voltage with no throttle, V2 is the voltage at full throttle, and A is the current limit of the controller. For those with LiFePO4 batteries, you'll have to wait until the full algorithm is implemented before the fuel gauge display works in a meaningful way.


2) Display Screens: Some of the display screens are anything but 'polished' and are there more for diagnostic purposes during development. This will be obvious as you scroll through them. An ability to select or deselect which screens show up (both while moving and while stationary) is on the agenda.

3) Units: It was pointed out by Telektik that if you select units of miles instead of km, the units text are not necessarily carried through to various setup options. I just haven't got around to that yet, but in your head you can read V/kph as V/mph etc. It doesn't change anything, just the display

4) Initialization: Some things like the accumulated torque for pedal assist mode aren't properly initialized yet on startup, so if you have a PAS torque sensing mode you can see for a few seconds some random human watts on the screen shortly after turning on.

5) PAS Quadrature: I just realized too that the PAS direction input is set only for quadrature PAS encoders. If you have a PAS sensor that only sends one set of pulses, then the CA actually won't pick anything up for your pedal cadence because it will think that every falling edge you are pedaling backwards. In the next firmware there will be a new item in the PAS Direction setup for enabling or disabling quadrature mode

6) Temperature Probe: You have the option of adding a 10K thermistor, a linear temp probe, or no temperature probe. Once a temperature probe is enabled, then the temperature will show up on the first display screen toggling along with the Ah and Dist. It currently shows units as oC, but I will have a oF option listed sooner or later.

7) RShunt: In the calibration menu in high range mode, you can set RShunt up to 9.99 mOhm, rather than just 0.99 mOhm as before. If you do set it above 1mOhm in the high range mode, and then you switch to low range mode, the first digit will be lost. So say 2.36mOhm in high range becomes 23.6 mOhm in low range, which will show up as just 3.6 mOhm, and if you change this to say 1.8 mOhm it will actually change to 21.8, because the 2 was still there just hidden.

I think that's it for the obvious shortcomings. As well, there were some things that were successfully added since I started this post

Ebrake Graphic: If you engage the ebrake levers by shorting the Bk pad to ground, then the throttle slider is replaced with an animated brake handle, to make it super clear why there is no power. As well, when the brake lever is engaged the CA's throttle output goes all the way to 0V, rather than to the min output voltage (which would typically be set to ~0.9V). This is done so that in principle you could use the throttle signal going to 0V in order to activate a regen input on the controller with just a simple comparator circuit, and eliminate the need for a separate wire going to the controller for that.


V/kph: I added a 'KV' parameter as one of the more advanced throttle control options based on earlier discussions here as a way for the throttle to jump to a value correlating to the current wheel speed before the throttle rate limiting kicks in. However this is not yet implemented, so changing the value has no effect yet.

Ramp Down Rate: There is now a throttle ramp down rate limit, similar to the ramp up rate limit, as requested by Adrian and hopefully useful to a few others too
View attachment 1

 

[adrian_sm]

Ramp Down Rate: There is now a throttle ramp down rate limit, similar to the ramp up rate limit, as requested by Adrian and hopefully useful to a few others too

Awesome. Thank you, this will be very useful for helping to control the disengagement of my friction drive. But best to have it default to 0 or a very low number, as most people will want instant off. I wouldn't have said anything, but it looked like from teklektik's post above, that it may be non-zero by default.

- Adrian

 

[justin_le]

It'd be cool also if the screen flashed the inverse pixels (everything but the original black characters) whenever an alarm or pack condition was met (such as high temp). Having the inverse colored display would be quite noticeable.

The current plan is to flash just the parameters affected. So for instance, if LVC is reached then the voltage display would flash. If the speed limit is reached then the speed readings would flash. Overtemp would cause the oC display to flash, etc. To do the entire screen is probably overkill, as some of these limits are hit on a fairly routine basis throughout the trip.

 

[justin_le]

I've got my V3 on the way too, hardcore fiddling ahead. Also playing with the CA :lol:

Justin, would it be hard to have one totally user customisable screen to scroll though ?

Hard? Yes. Doable? Yes. Likely to happen? Yes as well! With the many additional parameters being sensed and measured it's getting tough to figure out "one size fits all" display screens, and customization will be hugely appreciated. To some extent I'll have that automatic based on what features are enabled (no point showing temperature if no temp probe is connected, no point showing a human power bar if no torque sensor is configured etc).

Your suggestion is spot on for how to do a fully custom screen. 4 parameters really is the most you can show on a 16x2 LCD screen without clever graphics, and the way you described setting it up is exactly as I was envisioning. You'll notice that in the "Setup Display" menu there isn't a lot there yet, but expect this to get filled with many custom display options in upcoming firmwares.

You can configure some of it now but it would be cool to have a totally custom screen for the parameters you wanted to look at at any given time (especially useful when beta testing new features you might specifically want to look at while riding rather than the usual stuff)
In the setup you could have position 1,2,3,4 and then a scroll through menu for each position with say 10 parameters to choose from. This might be useful too for people wanting to not have certain wattage or voltage parameters on display if certain people are looking at their bike, and it might be nice to have just current speed, distance travelled, Ah and somthing else fairly innocent :wink:

 

[Hyena]

I'm of the general opinion that it's best to show only stats that are clearly known, and leave the riders in charge of proceeding based on that info and all the other environmental stuff which a rider (and not the CA) is aware of. Is it just me coming from a bit of a tech/lab background or does that make sense? -Justin

Yeah I guess. I suppose it's something you quickly get the hang of but I've had a few noobs say to me "man, I was out having a blast then suddenly ran out of battery and had to pedal the rest of the way to work/home" (oh, the horror! :lol
I guess once that happens a few times they learn to keep an eye on the ah!

If you engage the ebrake levers by shorting the Bk pad to ground, then the throttle slider is replaced with an animated brake handle ...This is done so that in principle you could use the throttle signal going to 0V in order to activate a regen input on the controller with just a simple comparator circuit, and eliminate the need for a separate wire going to the controller for that

Cool, but shorting bk to ground automatically cuts out the throttle and activates regen (if enabled) anyway - atleast on my controllers it does. Or are you referring to a bk pad on the CA itself ?
Next up, an ebikes.ca throttle with only a 6" long cable on it that plugs straight into the side of the CA

the way you described setting it up is exactly as I was envisioning.

Excellent!
Now it'd really make my day if you could program the over temp alert caused the whole screen to flash
"WARNING: Danger to manifold" :lol:

On a serious note, if you're going to have temperature rotating with 3 other parameters having it just flash might not be sufficient if it only displays temp for 2 seconds anyway. Maybe that parameter could stay on and flash if it's causing the CA to cut power. Though I guess if you're really worried about temperature (primarily in hod rodding applications I assume) you'd set up one of the above mentioned custom screens to keep an eye on it, and you'd notice power ramping down before it cut right out.
Or if that parameter isn't on the current screen, maybe you could have a flashing exclamation mark icon or something (like the brake lever display) that instantly tells you the CA has has cut your power rather than a BMS, wiring or hardware fault.

 

[AussieJester]

the way you described setting it up is exactly as I was envisioning.

Excellent!
Now it'd really make my day if you could program the over temp alert caused the whole screen to flash
"WARNING: Danger to manifold" :lol:

Well...at the very least..."WARNING: Frocked motor imminent"

 

[shorza]

I am unfamiliar with thermistors.
10k NTC thermistors on Ebay vary in size and price. Can you recommend one?

 

[lostrack]

Here's the throttle map idea.

More than a curve, it's got an adaptively set mid-point, based on the energy needed to sustain the same speed.

Outside this zone, it's current throttle.



Thoughts?

 

[justin_le]

Yeah I guess. I suppose it's something you quickly get the hang of but I've had a few noobs say to me "man, I was out having a blast then suddenly ran out of battery and had to pedal the rest of the way to work/home" (oh, the horror! :lol
I guess once that happens a few times they learn to keep an eye on the ah!

I'm hoping that the battery gauge graphic will go a long way to eliminating that surprise for riders who don't yet get the gist of Ah. But trying to say "you have 15km of charge left" is potentially asking for trouble when a change in grade or wind could quickly reduce that to 7-8 km.

If you engage the ebrake levers by shorting the Bk pad to ground, then the throttle slider is replaced with an animated brake handle ...This is done so that in principle you could use the throttle signal going to 0V in order to activate a regen input on the controller with just a simple comparator circuit, and eliminate the need for a separate wire going to the controller for that

Cool, but shorting bk to ground automatically cuts out the throttle and activates regen (if enabled) anyway - atleast on my controllers it does. Or are you referring to a bk pad on the CA itself ?

Should have been more clear. Yes, I am referring to the EBK pad on the CA itself. I need to check what the logic thresholds are like for the ebrake signal in most of the controllers, because there's a chance you could do this even more simply than a comparator circuit by just linking the throttle to the controller's ebrake input with a diode. So suppose the controller turns on regen when "BK" falls below 1.2V. If you link it to the CA's throttle output via a signal diode, then at a low throttle of 0.9V, the ebrake would be ~1.5V and hence not activated, but the moment the CA's throttle goes to 0V, then the controller's ebrake input falls to about 0.6V, which would enable regen.

Having the CA enable regen has some other benefits too, since it could then take care of regen over-charge protection, and enable regen via other means without needing ebrake cutoffs, (like negative pedal torque or backwards pedalling). It's also in that case possible for the CA to do a low frequency PWM on the ebrake line and potentially modulate the regen force a little bit.

On a serious note, if you're going to have temperature rotating with 3 other parameters having it just flash might not be sufficient if it only displays temp for 2 seconds anyway. Maybe that parameter could stay on and flash if it's causing the CA to cut power. Though I guess if you're really worried about temperature (primarily in hod rodding applications I assume) you'd set up one of the above mentioned custom screens to keep an eye on it, and you'd notice power ramping down before it cut right out.

In my head the idea of overtemp protection limits was to eliminate the need of the rider having back off from the motor overheating, as the CA itself will do that automatically. If all it had was a temperature display, then yes a fairly proactive alerting system would be on order. But since it's also taking care of ramping power down way before temps get critical, then it seems less urgent no? It wouldn't be too hard to have whatever display screen is currently periodically flash to show the temperature whenever it is between the threshold and max value.

-Justin

 

[justin_le]

I am unfamiliar with thermistors.
10k NTC thermistors on Ebay vary in size and price. Can you recommend one?

Good question that I should have addressed. Anything with a beta constant between 3800 and 4000 should be pretty accurate. At some point we'll make the thermistor lookup table programmable via a PC interface so that you can calibrate it to different models, but for now it is fixed and hard coded assuming B~3900.

-Justin

 

[electricwheels.de]

Finally I received mine too.
Sometimes Germany seems to be a loooong way away... :wink:

 

[Tench]

Got mine today too Big thanks to Justin and Adam for getting these out so quick!
I will get the bike rewired later this week and start playing

Justin you had better get ready for the myriad of questions! :?

Simon.

 

[hjns]

Mine arrived today as well! Yippiaeee!
Oh, and it came with a cycle analogger as well (cause I ordered it...)

 

[oatnet]

Oh, and it came with a cycle analogger as well (cause I ordered it...)

Another great tool I am happy to have, but it is in serious need of a new name - the term analogger suggests other images to me. On the other hand, odds are good that the trademark hasn't already been taken like drainbrain was. :lol:

-JD

 

[electricwheels.de]

Hi Justin and others,

I would like to use this high precision digital 1-wire thermometer:

http://datasheets.maxim-ic.com/en/ds/DS18S20.pdf

would this work with the V 3.0

 

[Alan B]

Two problems with the DS18S20 type digital temperature sensors:

1: they require a 1wire protocol to read them out which would require special software in the CA

2: they can't handle high temperatures, a hot motor can damage the sensor before the motor is damaged

There are high temperature capable thermistors that will survive in a hot motor and work with the existing programming in the CA.

 

[miro13car]

About speed sensor on CA
I never installed sensor and I don't know how accurate it is but very unlikely
More accurate than Cateye bicycle computer
So how accurate it is after you enter correct wheel tire curcumstance

 

[hjns]

About speed sensor on CA
I never installed sensor and I don't know how accurate it is but very unlikely
More accurate than Cateye bicycle computer
So how accurate it is after you enter correct wheel tire curcumstance

Why don't you turn it around? Measure a distance of several hundred meters. Then adjust the wheel circumference so that the CA provides you with the correct information.

 

[hjns]

Hi all,

I am trying to hook up my CA v3. I am runnng 30S 126V HOC.
Anybody can give me a hand with the following?

External VPack: Finally, the last input is for a separate battery voltage signal. For people running >100V, this gives the option of wiring up a resistive divider at your battery side for sensing the pack voltage and feeding just this signal to the CA, with the CA itself being powered from a lower 12V bus or similar. This is a much safer approach for higher voltage systems than powering the CA directly at those levels, and it means up to 650V packs could be supported.

I do not completely understand what a "resistive divider" is, and how to set this up. Justin or someone else, please help me create this?
Thanks very much in advance!