Beta CA build for R/C controllers

[mdd0127]

you could already do this by setting the throttle lower end limit just a little lower than the actual throttle output (cant remember the setting names) though I havnt tried to set the minimum speed, when I did this the drive was in an always on state.

you could change your riding style to not release the throttle completely unless you really have to :lol: . I like this because in theory your freewheels dont completely unload so there is no hammering effect when you apply the power again.

Rodger

With up to 14kw on tap, currently de-tuned to 5.5kw, the buck can be a little.....interesting at the least....

Changing the riding style is easy but I'm mainly concerned for people that are new to it or are less experienced riders.

I thought about the throttle lower limit thing and the only issue might be the HV160 needing to see zero throttle/throttle cycle to activate.......hmmmm...


It would be nice to have some way to keep the pulse running while in setup too. The power cycle issue seems easy to solve too and would make it much more user friendly when tuning.

 

[rodgah]

I thought about the throttle lower limit thing and the only issue might be the HV160 needing to see zero throttle/throttle cycle to activate.......hmmmm...

In my tests this didn't seem to be much of a problem. I assume when you set the minimum speed the throttle remains at zero till that speed is hit. If it however stays locked at the lower throttle limit till minimum speed is reached it would not work. I'm going to test mine again today so I'll try it.

 

[j3tch1u]

this is great news! i could use cadence/torque detection to activate just enough throttle to overcome motor cogging while pedalling. this would eliminate the need for a failure-prone sprag bearing in some cases.

 

[kfong]

I wasn't interested in the RC CA, but thought I would take a peek at this thread to see if it had any useful news. The torque sensor really has my interest. I would love to do away with the throttle and have a pedal type sensor that is truly linear. Justin, if you plan to offer a standard CA version with a Thun, I would buy one.

 

[Kepler]

you could already do this by setting the throttle lower end limit just a little lower than the actual throttle output (cant remember the setting names) though I havnt tried to set the minimum speed, when I did this the drive was in an always on state.

you could change your riding style to not release the throttle completely unless you really have to :lol: . I like this because in theory your freewheels dont completely unload so there is no hammering effect when you apply the power again.

Rodger

With up to 14kw on tap, currently de-tuned to 5.5kw, the buck can be a little.....interesting at the least....

Changing the riding style is easy but I'm mainly concerned for people that are new to it or are less experienced riders.

I thought about the throttle lower limit thing and the only issue might be the HV160 needing to see zero throttle/throttle cycle to activate.......hmmmm...


It would be nice to have some way to keep the pulse running while in setup too. The power cycle issue seems easy to solve too and would make it much more user friendly when tuning.

This intermittent surge on throttle up is a problem I am experiencing at the moment also and is something that needs to be solved. I am using K-force 100 ESC and thought the ESC may be partly at fault but it seems that the CC controllers are doing the same thing. This surge can be violent enough to loose motor sync and being unpredictable, makes every throttle up a lottery.

I have tried all sorts of tuning strategies to try and get around this and have managed to improve things. However, its still not perfect and at this point, due to its intermittent nature feel it may be an issue that needs to be fixed at code level.

Observations:
=> On first startup, the surge never occurs. I believe this is due to the ESC's inbuilt ramp is dampening it out.
=> If you coast at zero throttle for more then 5 seconds, then apply throttle, the surge does not occur. I believe this because the ESC ramp has reset within this timeframe and again is used to dampen the throttle up.
=> Most likely time to see the surge is if you come to a quick stop and then throttle back up again. This when I can see a surge voilent enough to lose motor sync.
=>Throttle up surges if you coast at zero throttle then throttle backup while the bike is still coasting. With my 1500W setup, its not too bad, but should be better.

Tuning improvements.
=> Set A-gain as low as you can for your setup. The lower the setting, the slower ramp. I have mine set to 10 on a 1500W chain driven standup scooter and still feels responsive enough.
=>Set i-term Min as close to the startup threshold as possible. Basically you are looking for the point that the ESC will stop at zero throttle but will also re arm on start up. Setting this as accurately as possible eliminates low throttle dead band. I run mine at 1.12mS.
=> Set Aux thresh to a higher figure. I have gone from a stock .99V to 1.4V. Setting this higher softens the throttle response at the lower part of the throttle movement. This has improved the surge greatly especially as you coast at zero throttle then throttle backup while the bike is still coasting. However, the big surge from quick stop to quickly back on the throttle is still very bad.

I did have one setup that just about nearly solved all the issues but then became unpredictable when first tuning the scooter on after charging.

These settings basically stopped the ESC from disarming completely so the ESC would never enter its startup phase during normal riding conditions.

Basically I set the the Aux thresh to .82V. This setting meant that the throttle was alway slightly on. I then would rely on the minimum start speed to turn the ESC off so once I was below 5kph, the ESC would shut down. This setup made the throttle very smooth in its operation and the slight amount of power on coast was acceptable.

However, the big problem was the first run after charging. What would happen is that as soon as you went above the 5kph threshold, the ESC would surge to a very high throttle setting and motor would loose sync. I imaging that on a high powered bike with a CC controller, this would be very scary.

Sorry about the long post, I figured I need to add as much detail as possible so that this issue can be solved.

 

[www.recumbents.com]

I have a RC CA with the beta revision 3 code. A few weeks ago I set it up with a CC HV160 ESC and a Astro 3215 motor. I used most of the settings that AJ noted a few pages back.

With no load it spun up easily up to max RPM. I must say that the "constant current" method of controlling the motor speed makes it a lot harder to set up because you cant use the throttle to vary to motor speed with no load.

I set it up on the bike and took a test ride. It seemed like the CA has some kind of feedback issue, because with a constant throttle the motor would be powered for a couple seconds, then loose power for a second, then powered again... ad infinitum. I tried a higher throttle and that appears to have broken my ESC. Apparently it did not like a full powered on off on off at low RPMs. The controller now shows a constant flashing red light. I am sending back to CC for repair.

I sent a couple e-mails back and forth with Justin but he didn't have any explanation for this behavior.

Has anyone seen a similar behavior with the new RC CA? Was there a way to switch to make it behave more like a server tester? I was hoping that the RC CA would give getter control at low RPM, but would rather go back to the old servo tester than blow my ESC up again. At least now I know what NOT to do

Thanks,

-Warren.

 

[Kepler]

The problem you're seeing is different to what I am experiencing but they still may related. Couple of questions.

What did you set your A-gain too?

What type of throttle were you using? Hall or Magura resistive?

What did you set your min start speed to?

What did you set your Thresh Aux too?

If using a Magura, the stock .99V setting would be wrong. Basically the throttle would be on all the time but held out by what ever you had the minium start speed set to. So as soon as you went above the min start speed, the esc would drive to what ever the throttle setting was. Now couple this with the surge issue I am seeing and you could really stress the ESC.

Any of that add up?

 

[www.recumbents.com]

Initially I had the start speed set to 0.
I got my throttle from ebikes.ca, with the RC CA. It looks like they use hall throttles since they say "It is also possible to use a potentiometer throttle, but we don't have any."

I started with the same settings AJ noted, except I used the potted shunt and used the value for that and 99 amps max.

AGain 300
Aux Thresh 0.75V

Then I changed the min speed to 5 mph, and I think I changed A-Gain to 60. After that when the bike hit 5MPH the motor started running at low speed with no throttle input.

Both settings had the same intermittent throttle issue as above.

-Warren.

 

[AussieJester]

Did i post my CC HV160 settings along with the CA settings i forget? i did alter these also..

KiM

 

[rodgah]

Your aux thresh is set too low most likely. This will cause the esc to fail to arm, because it is sensing that the throttle isn't 'zero' and cause the constant red flashes. Try setting it up around the 1.2v and see if it arms then. Then slowly reduce the value till you find the value that causes it to not work. Closing this gap makes the throttle feel like it has less dead spot at the bottom.

Be patient, Kim was going to throw it in the bin after 10minutes. Look at him now LOL :lol:

 

[Kepler]

Your aux thresh is set too low most likely. This will cause the esc to fail to arm, because it is sensing that the throttle isn't 'zero' and cause the constant red flashes. Try setting it up around the 1.2v and see if it arms then. Then slowly reduce the value till you find the value that causes it to not work. Closing this gap makes the throttle feel like it has less dead spot at the bottom.

Be patient, Kim was going to throw it in the bin after 10minutes. Look at him now LOL :lol:

Initially I had the start speed set to 0.
I got my throttle from ebikes.ca, with the RC CA. It looks like they use hall throttles since they say "It is also possible to use a potentiometer throttle, but we don't have any."

I started with the same settings AJ noted, except I used the potted shunt and used the value for that and 99 amps max.

AGain 300
Aux Thresh 0.75V

Then I changed the min speed to 5 mph, and I think I changed A-Gain to 60. After that when the bike hit 5MPH the motor started running at low speed with no throttle input.

Both settings had the same intermittent throttle issue as above.

-Warren.

That Aux Thresh setting suits a Magura but not a Hall throttle. That would definitely cause the motor to start as soon as the bike was over 5 mph. You need to set this .9V or above. A higher setting will make the throttle softer over the first half.

AGain of 60 is probably OK but I would make this as low as possible. Lower means a softer ramp on acceleration

 

[Kepler]

Your aux thresh is set too low most likely. This will cause the esc to fail to arm, because it is sensing that the throttle isn't 'zero' and cause the constant red flashes. Try setting it up around the 1.2v and see if it arms then. Then slowly reduce the value till you find the value that causes it to not work. Closing this gap makes the throttle feel like it has less dead spot at the bottom.

Be patient, Kim was going to throw it in the bin after 10minutes. Look at him now LOL :lol:

Actually I think the ESC will still arm with a low Aux Thresh setting. Its the iTerm Min that determines if the ESC arms or not. I think plugging in with this low Aux Thresh will all sorts of havoc.

 

[mdd0127]

The throttle surge IS an issue. Something weird's going on. I tightened up my slipper clutch to eliminate it's behavior for trouble shooting and turned the amp limit up to 160. Every few throttle attempts, the motor loses sync, or a surprise power wheelie happens. I really wish I had a dyno to load the bike up on so I could watch things with my meter! I've messed with all of the settings and it just randomly freaks out no matter what. The HV160 isn't flashing or beeping but I need to get it off the bike and check out the logs. Also, it's started cutting out occasionally when cruising at around 45 amps and when it resumes, it resumes at whatever the throttle was at before it quit which is really scary, causing wheelies again. Again, there are no error indications on the HV160 and I know I'm not hitting LVC.

Something's wrong.

I'll post up datalogs and exact settings as soon as I catch my tail.

Those with crazy power setups need to BE CAREFUL until this gets sorted out!

 

[Kepler]

We are definitely seeing the same problem here. Worth noting the same thing is happening on 2 different types of ESC's so that rules out that its a compatibility issue with the HV160.

I have tried to tie it down to a certain set of events but its quite unpredictable. What I am confident of is that if the ESC has not gone back to dead zero, then the throttle input stays smooth and predictable. Its after the throttle has been zero'ed and the ESC re-sets its internal ramp when things get unpredictable again.

Just thinking, The HV160 controllers have a a few extra setting that may help. This setup might be worth tying.
In the throttle tab, select Helicopter and fixed Endpoints. Set the initial spool rate and head speed change rate to max. (20 and 20)
Set Auto Rotation to Enable.

By enabling Auto rotation, you get a small window from 1.0mS to 1.1mS that will stop the motor but will give you a faster rate of change then the normal spool up. You would need to set your iterm min to 1.1mS also.

I dont have a CC controller in a bike at the moment but I have one in a large Heli and use the Auto rotation feature there so I know it does work.

 

[www.recumbents.com]

Ok, it looks like I have some stuff to try then. I haven't sent the HV160 off yet so I'll go out to the cold garage and test it with some of these suggested parameters.

Thanks,

-Warren.

 

[www.recumbents.com]

I put the bike on a trainer in the basement so my fingers won't freeze when I'm testing out the RC CA.

I hooked the HV160 back up to the servo tester, motor, and batteries. It does work but I get the initialization tones, 11 beeps, then the arming tones. This is different than I an used to hearing, but am I using a different motor than I had before. It used to have the initialization tones, then the two similar arming tones.

Next I will reconnect the RC CA. Fingers crossed.

Hey AJ, do you remember what setting changes you made to the HV160?

-Warren.

 

[Kepler]

11 beeps is the cell count. It thinks you have an 11S pack. Probably your voltage is around 44V at the moment. Sounds like the ESC is working OK. Make sure you set the Aux thresh back to stock. (.99V)

Justin is aware of the issues we are having. I am doing a few tests for him at the moment. I've got a feeling its going to take a code modification to sort this out.

 

[AussieJester]

Hey AJ, do you remember what setting changes you made to the HV160?

I don't off hand no, i just got in need me some sleep will hook the Data logging cable up and note the hv160 settings i am using later this morning/afternoon. I haven't had time to keep up with much so i don't know if you guys are using hall throttles or Maguras? Mines a Magura and it was also adjusted, Rodgah may recall the adjustments?

Best of luck getting it sorted mine works flawlessly zero complaints here, would use this over the 'servo tester' setups hands down....

KiM

 

[www.recumbents.com]

I am feeling better now. Apparently I don't have to send my HV160 ESC back to CastleCreations.

I made the suggested changes and it started working properly. At least in the basement on the trainer, but I was able to stress it to 2500 watts with no cut-outs or other funkyness.

Using ebikes.ca Hall Throttle
HV160 settings: http://recumbents.com/wisil/e-bent/rc_drive/castle_esc_settings.htm
Astro 3215

RC CA settings
Amps Limit 99amp
Low Volts limit 40v
Main Display Watts
Set Range Low (0.1A)
Averaging 5
Set RShunt .1.420 m0hm (Using potted shunt)
Volt Sense 30.95 V/V (adjust this value until CA voltage agrees with your high quality VOM.)
PSGain 0.08 V/kph
ISGain 200
AGain 60
VGain 800
ITerm Max 1.99mS
ITerm Min 1.21mS
Aux Thresh 1.20V
Vshutdown 10V

I found that if I used the CastleLink software to dump the logs, the 11 beeps went away (yes batteries were down to about 44V, thank you for explaining that).
The ESC boot up now has the expected beep sequences.
The motor does seem to run at low RPM with less jitter than when using the servo tester.

I'll solder the connectors back on and do an outside test within the next few days to make sure it works properly under "real world" conditions.

Thanks all!

-Warren.

 

[lostrack]

Happy anniversary thread, you're now just over 1 year old.

It would be really helpful to put the findings of this into a wiki. I have read it but it's just hard to retain the 13 pages of data here, and the conclusions of it all.

I have a large screen cycle analyst, version 2.27 and would love to hook this up. I'm not quite sure if I have to have a new version - or it's just a software change. Will pop the back off and see if I have the throttle/pwm pin.

The aussies seem to be running it fine, which is a good sign - what are your best settings?

 

[amberwolf]

Happy anniversary thread, you're now just over 1 year old.

more like 6 months, as the op was "Posted: Tue Jul 12, 2011 6:36 pm".

It would be really helpful to put the findings of this into a wiki. I have read it but it's just hard to retain the 13 pages of data here, and the conclusions of it all.

that's true of many very useful threads around here. :lol: you're certainly welcome to start the article, in the es wiki at http://endless-sphere.com/w/index.php/Main_Page

 

[rodgah]

I'll solder the connectors back on and do an outside test within the next few days to make sure it works properly under "real world" conditions.

How did you go?.......its always unsettling seeing someone post they are going out to test something and then not check back in :lol:

I have now started to notice a bit of jitter in my servo tester setup, kind of like a ICE engine missing is what it feels like. Makes me more motivated to change over to the RC-CA.

 

[www.recumbents.com]

Sorry, "Real World" is now snow and cold. I decided it was a good time to send my ESC back to Castle for the recommended upgrade to V2 before I soldered the connectors back on. I did test it inside on a trainer which put a pretty good load on it, so I think it's working right now but I won't know for sure until I get it all back together...

-Warren.

 

[Alan B]

Excellent development. Looking forward to the release model.

 

[www.recumbents.com]

So, real world finally is getting warmer and I was able to take it for a ride around the block (actually several laps around the block). The controller works well, though it seems like minimum throttle cuts in a little harder than with the old servo tester / BEC system. Performance at very low speeds seems to be comparable to the old servo tester / BEC system, so I still have to pedal up to 5 MPH or so before throttling up. I had also replaced my Astro 3210 with a 3215. The new motor has a bit more grunt and longer legs. I had the CA set to cut off at 99A and 35 MPH, and the power definitely shut off at 35MPH and did not allow more throttle until I got below 35. With the 3210 it maxed out at 30 MPH. When I looked a the IMax (amps) later it showed a max of 99 amps. 80 amps was max with the old motor..

Next I need to remove the speed limitation to see how fast it goes, and see if the watts / mile is significantly worse than the 26 wh/mi average with 30 MPH cruising that I was seeing last year.

-Warren.