Powersoccer chair that really kicks

[Burgerman]

Once balanced, and you may need to restart the charger, to complete it, the next charge will be fast.

You did fully charge each cell before assembly?

[Burgerman]

I just looked at your file...
You didn't assemble the pack with fully charged cells. Quite the opposite. SOME are charged.
So...

You now have 1 set of cells fully charged at 3.600V.
Most are at 90 percent full at 3.4xx.
1 set is at 95 percent discharged... At 3.2v

Your max charge rate is about 45mA so it will never charge or will take several months... You have not charged it at all.

So either you have charge rate set to a silly low figure (use my preset)
Or more likely it is seeing an error due to the massive unbalance.
Or you chose MONITOR instead of CHARGE.

You need to fully charge cells in group 8 separately as a single cell. Then charge normally.

The charger will error, and revert to a really low safety charge current if the difference between highest and lowest cells exceeds 250mV. It thinks there's a battery fault. Which is true since you have full, and empty cells in the same pack.

[sepos]

can I use the same presets, just connect charging cables into single group of 8 and naturally disconnect bridges between groups? and no balancing there.

[Burgerman]

With 1 cell(s), best way is to connect no balance at all, and use the attached pre-set. Delete the .zip from the file name.

ONLY connect from the charger main charge wires to the group of cells required. Set 8. So its charging only one set. Use jumper cables that are 10swg or 12swg approx.. It will charge at 18 amps.

Then go back to charging as before with normal connections and the other 8S pre-set.

[sepos]

Thanks again.

in graph cell 7 was lowest, should I charge that or all?

[Burgerman]

I thought it said cell 8 low?

Just checked. Cell 7 and 5 are the lowest.

Feel free to do every one. They should have been done before assembly really. THEN connect and charge all 8 together. And wait till it ends. Then all will be charged, matched, ready to go.

[sepos]

i have done all separately and number 7 second time alone.

now after 30min charging of all together cell7 have dropped back to 3.132V, cell 8 has risen to 3.849v. others remain neatly around 3.455v.

I don't know if this will work as current is now 5-7mA. Or should I still wait?

or should I separate each unit in cell7 and charge them one by one?

[woodygb]

cell 8 has risen to 3.849v.

Why did you let that happen?..

B.M. gave you these charging specs.

CHARGE
Charge Amps: 30.00 (whatever amps you want based on your battery capacity, or power supply. The charger can do 40, most power supplies cannot.)
Increment: 5mA: NO
Charge Voltage: 3.600V (3.50 to 3.55 will take longer to charge, longer to balance but healthier for the battery. 3.65 volts is faster to balance, less good for the battery. All end up full! if time allows)
Conv Frequency: 125khz
Cold Weather T: NONE (reduces charge voltage in winter for safety, mostly for lipo batteries we don't need it)
F. Average past 1min: NO

[sepos]

I have been using those presets. so next thing I try is to take cell7 and charge each unit in it separately.

[sepos]

weird. I separted all units in cell7. measured voltage separately and each has 3.44V

[sepos]

any ideas why graphics show differently?

[Burgerman]

One cell, or 50 cells, all behave and charge / discharge as one great big cell. No need to take it apart. Just charge each parallel GROUP fully. To 3.600v.

Then charge as normal.

[Burgerman]

God knows what you doing. Read this FULLY until you understand it! Over and over.

http://www.revolectrix.com/support_docs/item_1377.pdf

This will tell you everything in detail.

If there's anything you don't understand in there ask me here!

weird. I separted all units in cell7. measured voltage separately and each has 3.44V

Then that wasn't 7. And all were connected together so have to be the same voltage.

[Burgerman]

i have done all separately and number 7 second time alone.

now after 30min charging of all together cell7 have dropped back to 3.132V, cell 8 has risen to 3.849v. others remain neatly around 3.455v.

I don't know if this will work as current is now 5-7mA. Or should I still wait?

or should I separate each unit in cell7 and charge them one by one?

Then you have a wiring/connection issue. Somewhere on cell 6, 7 or 8.

[sepos]

reassembled wiring & connections. start of charging looked nice. highest cell8 3.48v, lowest cell5 3.383v others 3.424v.

after 45s charging cell8 rises to 3.912v and cell7 drops to 3.203v.

still more work with wiring, I guess...

[woodygb]

The software should prevent the cell rising to that excess voltage.

[Burgerman]

The software should prevent the cell rising to that excess voltage.

Yes it does. If its set correctly. But if a balance wire has a bad or intermittent connection then one cell reads high, and one low even if ACTUAL cell voltages are correct. And these are right next to each other... So this is the likely issue.

[sepos]

once more rebuilted wires 8 and 9. I don't know what was wrong earlier, but now it seems working.

All cells at level of 3.605v after 30min charging.

Although current level is still low only 90mA.

[sepos]

charge complete in 43minutes.

[Burgerman]

90mA termination?
What was battery state of charge? How much returned?
90mA "safety charge" due to greater than 250mV cell differences?
90mA charge during CC or reduces to 90mA during CV?
Need a bit more info...

Or the graph data!

[sepos]

battery is fine now.

finally managed to get wheels rolling! amazing feeling after these few months of waiting parts and replacing them. like Chinese converter giving only 0.8v instead of 12v for breaks... now got 30x more expensive norwegian one, but seem to work also few times better

starting to work with roboteq configuration. I just connected new joystick and calibrated it. it has a bit wider range than in your original configuration.

I suppose some maximum value is exceeded when pushing it to extreme positions as chair is stoping.
Should I replace max & min I got from calibration with few % bigger numbers?

Other funnier thing is that turing joystick left, left wheel starts move first and right wheel follows long time later. Also I will get no reponse, if I push joystick backwards or forwards. connections should be ok this time as I see values changing in roboteq "run"-page for each of these analog inputs.
Need to study more of these configurations.

[Burgerman]

Are you using:

a. the SCRIPT? Without that it will not be usable.
b. the SSR to control brake, via a 12V or 24V supply?
c. the current sensors? These are essential or you can forget any idea of control...
d. speed pot?
e. 24V, 36V, other?
f. what is the Roboteq model, and firmware version?

And.

Did you read and understand the complete manual?
Did you calibrate the current sensors manually, joystick by auto method, etc accurately? And then allow some "extra" at each end before it hits the error (which makes it stop or whatever you have choses in "action".) (approx. 5%)

Did you set up all the other hundred Roboteq input and control parameters correctly?

For e.g, your joystick is set to do what if over calibrated limit ? Same with speed pot, current sensors, RC inputs etc? And not set to "absolute" for eg?
How about the Roboteq power settings, and is script set to auto run?

How about the MOSFET fail output pin > SSR > power off? Status lamp?

Do you have all of this configured correctly?

[sepos]

- comments in text

Are you using:

a. the SCRIPT? Without that it will not be usable.
- should be in
b. the SSR to control brake, via a 12V or 24V supply?
12v for brakes
c. the current sensors? These are essential or you can forget any idea of control...
- not yet, building next. currently script should use internal roboteq estimates
d. speed pot?
- no
e. 24V, 36V, other?
- 24v
f. what is the Roboteq model, and firmware version?
- hdc2450, firmware version1.4a
And.

Did you read and understand the complete manual?
- hope so:)
Did you calibrate the current sensors manually, joystick by auto method, etc accurately? And then allow some "extra" at each end before it hits the error (which makes it stop or whatever you have choses in "action".) (approx. 5%)
- current sensors no, using roboteq estimates, but will do that later
- joystick by auto method, but have not added any extra. I was assuming that might be needed (action choices from your script so it will do "safety stop"
Did you set up all the other hundred Roboteq input and control parameters correctly?
- changed so far only voltage limits in general section: under 14v and over 31v
- did not figure out any other that I should change. any hints here?

For e.g, your joystick is set to do what if over calibrated limit ?
- "safety stop"
Same with speed pot, current sensors, RC inputs etc? And not set to "absolute" for eg?
- not used atm, set to "no action"
How about the Roboteq power settings, and is script set to auto run?
- script autostart enabled
- power settings, I don't know what you refer to with this?
How about the MOSFET fail output pin > SSR > power off? Status lamp?
- this is connected to brakes, mosfail "light" is clean in "run"-window
Do you have all of this configured correctly?

[Burgerman]

c. the current sensors? These are essential or you can forget any idea of control...
- not yet, building next. currently script should use internal roboteq estimates

Well mine is un-drivable like that... (In comparison to sensors). And does some freaky stuff if you turn it up higher to try and make it work.

- changed so far only voltage limits in general section: under 14v and over 31v
- did not figure out any other that I should change. any hints here?

Set it below 14V if you can, because with 24V lead and a pretty used battery I often see voltage drop below 18V under load and then the mobility controller reduces amps (power) to keep above 18. See how easy it is to get it to drop to 18... http://www.wheelchairdriver.com/gopro/voltdrop.mp4

And again, 32 to 35V (as on a mobility controller) is safer. Because the last thing you need is to exceed 31V going down a steep ramp or hill at speed via regeneration, and the Roboteq cut power, short out motor, and apply brake. Because it stops extremely rapidly!

For e.g, your joystick is set to do what if over calibrated limit ?
- "safety stop"

Mine is set to auto calibrated, plus 5 percent extra at each end. In all 4 directions. Because again the Roboteq will see full stick as a fault and stop the chair dead with locked wheels if not. Or turn it off. (do nothing). But that's not safe. You want it to go over the limit in case of fault/wire off etc.

MOSFET FAIL pin, and Roboteq 5V and a protection diode, should connect to (in my case) a 2nd SSR. This in turn can switch OFF main power via the latched contactor. SSR switches the always on 12V inverter for contactor control power to do this. Then in event of failed output MOSFET the pin becomes active and turns off contractor.
You must also have a on/off momentary button to enable this power on/off relay on the control pod too. So 12V to the pod and a momentary button. And another for on. In case of runaway for other reasons.

http://www.wheelchairdriver.com/gopro/roboteq2.mp4

Script http://www.wheelchairdriver.com/BM3-con ... Script.txt

My Roboteq settings for you to look at. Delete .zip

[sepos]

now I have built the current sensors. although cannot get any values in calibration.

[Burgerman]

You can get the CENTRE point via calibration. That is with zero motor current.
Then set the other ends to:
0V
Centre (Calibrated) will be about 2350
5V
Action: NON
Deadband 0%
Action on min, max NON.

Make sure you have left and right set correctly.
Make sure you have polarity set correctly...

Get this wrong above and you will have some very odd behaviour, or OPPOSITE behaviour on one or both motors.

Be sure to set motor compensation LOW in script, then increase in small steps while testing. Do this OUTDOORS in open space. Small steps... Go too far, it WILL accelerate away, or spin uncontrollably. And you then need the emergency stop contactor... And if you are too slow probably some bandages!

[sepos]

calibration gives only 27.

left and right should be ok. giving power to sensor from pin14 and ground from pin 9.

[Burgerman]

27 is 27mV. (Effectively zero)

It should be HALF WAY between 0V and 5V approx. You did get the right sensors?

BI-Directional? 100A?

http://cdn3.volusion.com/pgsmd.pvnjz/v/ ... 1347323401

[Burgerman]

giving power to sensor from pin14 and ground from pin 9.

One on each motor hopefully!

Acceleration and deceleration is set low on the script. You will want to increase these a little at a time once you get the motor compensation set correctly. And then try turning up speed with the speed pot.

Do initial testing with chair on blocks, and motor compensation while sat in it on the ground... Then fine tune all the settings. They will all be wrong as you are using 24V and 1/4 the power.

[sepos]

27 is 27mV. (Effectively zero)

It should be HALF WAY between 0V and 5V approx. You did get the right sensors?

BI-Directional? 100A?

http://cdn3.volusion.com/pgsmd.pvnjz/v/ ... 1347323401

yes, bi-directional 100a hall effect.