Cutoff Voltage

[OneEye]

Is the low voltage cutoff on the controller adjustable?  (so you can change it based on which battery chemistry is being used)

What is the factory setting? (even if not adjustable)

-Mike

[myelectricbike]

I asked Philip Yao for a schematic diagram of the controller several months ago. Maybe if you ask him too he will provide us with one.

The cut-off problem involves not just the lowest voltage but how long the cut-off voltage is surpassed. The duration setting, when using lead acid batteries, may be even more important than the voltage level since even a half discharged lead acid battery pack can dip below full discharge of 10.5 volts and then come back up to 12.5 when the load is reduced or removed.

This morning I connected three good batteries and one bad one in series to my 48 volt controller. The bad one has a dead cell so I got a good feel for where the cut=off voltage is on this controller. I saw total voltage go as low as 40 volts in just a flash before the controller cut-off but it also cut=off at voltages as high as 44 volts. My meter only read in half second intervals so the actual momentary voltage low and long enough to trigger the cut-off seemed to average around 42.5 volts. Divide by 4 and guess what? You get 10.625 volts. Since full discharge is 10.5 volts my theory is that the controllers are intended to cut off below 10.5 for any duration more than half a second. That means 31.875 should be the expected cut-off for a 36 volt controller. Fortunately I took my 36 volt controller along and when there was not enough power left in the dead cell to keep the pack above 42.5 volts I switched to the 36 volt controller and maintained 40. volts the rest of the way home, hitting 38 volts only once on the next to the last hill climb.

 

[OneEye]

Too bad it is not a programmable or adjustable voltage cutoff.  Depending on the cell count, LiFePO4 or other lithium chemistries can have a lower cutoff voltage and still not kill the battery by over discharging them. 

I've been reading a lot of success stories from folks grabbing the DeWalt 36V tool batteries and using them as-is in their e-bikes (heinzmann, wilderness electric, and Crystalyte motors for the most part).  The packs have a much lower end-of-discharge voltage, about 22V.  If the controller cuts out at 31.5 volts you may be missing out on a significant chunk of the cells available energy.

Some remote control aircraft enthusiasts have been putting the lithium cells from the DeWalt packs through their paces and have run them at 30A through more than 1000 discharge cycles (to 100% depth of discharge) and they still had almost 80% of initial capacity.  That makes it sound like they will really stand up to a beating.

As with all Lithium ion batteries, the packs are a bit expensive (going rate on ebay is ~$105 shipped).  They deliver 2.3Ah each, so about 57.5Wh through 100% discharge.  A pair should give around 6 or so miles of motor-only range.  If someone actually continues to use their e-bike over the long haul they are probably 1/3 the cost of lead-acid batteries.  The lithium batteries are also much lighter, about 2.5lbs for each DeWalt pack.

On another forum, someone mentioned there is a way to replace resistors on a different controller (Crystalyte, I think) to change the cutoff voltage.  Maybe with the controller schematic we could figure out how to do a similar thing.  I'm guessing it would involve placing a lower R on a pull-up line.

The LiFePO4 packs GoldenMotor sells probably have enough cells to match the low-voltage cutoff of 31.5V at nearly 100% depth of discharge.

[myelectricbike]

The solution I discovered this morning for a single dead cell should be applicable for batteries where full discharge is a much greater percent of full charge. In other words I lowered my cut-off voltage by switching to a lower voltage controller instead of racing home I got home at normal speed with plenty of reserve due entirely to the good cells remaining in the bad battery. Another option I'm looking at in line with this idea is to use single cell (2.1 volts) SLA/AGM batteries intended for portable high current electronic devices. It might also be a lot cheaper in the end in regard to replacement than having to throw away a battery with 1 bad and 5 good cells.

However, I think the fuel cell may be the answer since its output is nearly the same as the line powered charger depending on what happens with cost for the unit and for the hydrogen and whether or not you can make your own since the storage system is hydride which I do not think is pressurized although I would prefer a fusion reactor. 

[OneEye]

They also have fuel cells that run on liquid methanol.  I believe they have some that run on about 500mL per hour for 500W continuous output.

[myelectricbike]

I've been to a web site of a company either in Denver or Provo that has a whole range of outputs and fuels including Methane and Methanol but the smallest output is 1,000 watts and it is too big and expensive for an ebike. Company did not respond to numerous emails requesting more info regarding fuel cells for ebikes.

[pdonahue]

Does anyone have a schematic of the controller yet???   I'd like to adjust the high voltage cutoff of my 36V controller for use with a 48V battery.

Pete

[Chris]

I am having the same problem using the Nimh packs, the cutoff is set too high so I am only getting about 30-40% from the packs before the controller starts to shut down the motor, looks like its back to SLA

[OneEye]

Anyone with a controller can start on a schematic by scanning their circuit board at high resolution and posting it.  We can do circuit traces and find out what most of the components on the controller are.  If we compare to other controller schematics (I believe there is a crystalyte schematic on the endless sphere forums) we may be able to figure out where the voltage reference / sense is, and perhaps someone who is good with electronics can recommend a modification to change how that is read.  I am told you can do a resistor change on a crystalyte controller to change the low voltage cutoff.  The same is likely to be true of this controller.

[myelectricbike]

Not necessarily. The new regen controller has a micro controller chip and the only stuff on the board is what's necessary to support the interface with the mosfets. Before you go tearing your controller apart you might want to just write down all the part numbers and then look them up. Some companies, however, use parts with assembly labels so in that case forget abut even looking up the parts. Hate to disappoint.   

[pdonahue]

I found the low voltage cutoff for the 36V controller...  I'll try to post a pic this evening when I get home...  Basically if you take a look at the edge of the board on the opposite side from the FETs there is a location that is marked (on the silkscreen) as a resistor, but is actually just a small piece of metal soldered in place.  To raise the voltage cutoff (i.e. for 48V operation) you can just add a resistor inplace of the wire.  (3.5K for approx 42V cutoff)  To lower the cutoff it should be possible to just replace the 11K resistor next to it with a slightly smaller one.

If anyone has found a way to disable the high voltage cutoff it would be most appreciated.  I tried e-mailing goldenmotor to find out how to order a 48V controller but there is still no response.

Pete

[OneEye]

Most controllers will have a micro chip, this drives the PWM on the fet's to control the motor.  You can sometimes guess about which pins serve which functions by comparing to similar models.  Admittedly it isn't a sure thing.

[pdonahue]

It's quite easy to trace back the Hall inputs and the FET outputs.   Also the LM56 temperature sensor.  I still can't figure out the over voltage cutout though. 

Pete

[OneEye]

You can find out which chip they are using for the microcontroller, and see if it has an internal voltage reference in its specification.  Occasionally the voltage comparison will be tied to a specific pin on the microcontroller.  Otherwise you would use another chip to serve as the voltage reference and it would provide a high-low signal to one of the microcontroller's I/O pins.

Of course EE is well outside my area of practice, so everything stated above is highly suspect.

[myelectricbike]

You guys are not hearing what I am saying... microcontroller means microcontroller as in software. Yep, that's right. Software to control both the low and the high (is there really such a thing?) cutoff. Of course there could be a dedicated onboard cutoff that's part of an overvoltage/undervoltage protection circuit but my guess is that for things like different type batteries or hookups its all controlled by the software.