Lithium Battery for Magic Mobility X4

[swalker]

I am not sure where to put this, so will put it here. Mods, please feel free to move this to a more appropriate forum.

As some of you probably read in the PALS checkin post, last Sunday we installed a home-built lithium battery pack in my Magic Mobility X4. I will provide some details in this thread.

As background, the X4 is a four wheel drive wheelchair that I primarily use for outdoor travel over rough terrain. We bought it off of craigslist for $500 about 3.5 years ago.

When we got the wheelchair, it was a derelict. It had no batteries, so we could not even see if it powered on. Each wheel was flat and had over 100 cactus thorns and the wiring harness was pretty much unuseable. Both leg canes were hopelessly bent, as was one of the armrests. We were hoping the motors and main control modules could be salvaged.

After about three months of work, my friend Greg and I were able to get it back into operation. I have used it extensively since then for some of my more hardcore off road adventures.

The X4 has several serious limitation, though. First, it has a very restricted range of about 4.5 miles. Second, it is very sluggish when starting up. Third, it is easily overloaded and the power will shutdown.

I suspected that all of the problems were directly caused by, or at least made worse by, the use of lead acid batteries. Even though it was fitted with what I think are the best gel batteries (MK group 24 gels), I suspected that installing lithium batteries could make a meaningful difference in all these areas.

About 2 years ago, I started doing research on lithium batteries to learn what might be practical for a wheelchair. I stumbled upon a wonderful site out of the UK which is run by a gentleman with substantial experience doing that. Following his lead, quite a few other technically-minded wheelchair users have successfully built lithium battery packs for their wheelchairs. There have been a few notable failures along the way as well.

Based on the knowledge gained there, I felt comfortable trying to built up a lithium battery pack for my X4.

To keep this more readable in small bursts, I will add posts to this thread with details on the build. For now, here are some pictures of the X4 in action (all with lead acid batteries).

Steve

 

[Nikki J]

I moved it here as I think it is of equal interest to CALS. Steve you are amazing

 

[swalker]

Thanks, Nikki

 

[swalker]

Here is another installment.

I want to start by saying such a project should not be done unless there is a thorough understanding of the dangers involved. There is enough energy in the lithium battery pack to do serious harm. I have a lot of experience working with electrical systems, so felt reasonably comfortable tackling this project.

In this thread, I will be providing a summary of what I have done. To see the technical details, you can follow a thread at wheelchairdriver dot com/board/viewtopic.php?f=2&t=8642

Our wheelchairs typically use two nominally 12 volt lead acid batteries connected together in series to make a nominal 24 volt battery pack. The voltage varies a bit with charge, but they are usually referred to as 12 and 24 volts.

In my case, I use MK group 24 gel batteries that have a capacity of 74 amp hours. The total energy in the battery is measured in watt hours, which is the product of the voltage and amp hours. A watt hour is simply one watt for one hour. Thus, a single 12 volt, 74 amp hour battery would have a theoretical energy capacity of 888 watt hours.

Two of those 12 volt batteries connected in series to make a 24 volt battery would have 24 volts * 74 amp hours = 1,776 watt hours.

A lithium battery pack is composed of separate pieces called cells. Different lithium chemistries have different voltage, so selecting the chemistry and number of cells to wire together in series to form a battery pack that operated >= to, but close to 24 volts needed to be carefully considered.

On top of this, I needed to pick a lithium chemistry that is super safe. I am sure many have read or heard about lithium batteries catching on fire, often with catastrophic consequences.

Since my ability to hop out of the wheelchair and run away from it if it were to catch on fire is, shall we say, limited, I really needed to focus on the safety of the chemistry selected.

I also needed to consider the form factor. It was important that the assembled battery pack would fit in the wheelchair's battery box.

I only wanted to undertake this project if I could build a lithium battery pack that would have substantially greater capacity that the lead acid batteries it would replace.

In addition to all that, I needed lithium cells that could handle the high energy draw that wheelchairs can place on the battery. This happens to some degree when going up steep slopes, but especially happens for brief moments when we turn the wheelchair from a stop or slowly go over a threshold.

The ability of a battery to deliver power is expressed as the C rate. This is the ratio of the energy demand placed on the battery in one hour to the capacity of the battery, with both being measured in amp hours. A C rate of 1 would mean that the battery could safely and healthily withstand being completely drained in 1 hour. Most lithium cells have a C rate for continuous discharge and a different, higher C rate for momentary discharge.

Lithium cells typically are available as cylinders, but over the last several years, ones shaped like a box (called prismatic cells) have started to become available. Prismatic cells would allow me to pack more energy into the space available in my wheelchair's batter boxes.

Knowing that I needed all of the above really limited what I could pick for the lithium chemistry, form factor, and manufacturer.

For the chemistry, I selected Lithium Ferrous Phosphate (LiFePO4). I also soon realized I need to limit my search to prismatic cells. Finally, I had to find the highest capacity cells that could be assembled into something the size of a group 24 lead acid battery.

My lead acid batteries are 173mm wide and each battery box is about 174.5 mm. That became a real limiting factor.

I was finally able to locate a vendor with 176 Amp Hour LiFePO4 prismatic cells that are 173mm wide, 54mm wide, and 200mm tall. These cells could be assembled into something that would fit in the wheelchair's battery box, if the measurements and specs I was relying on were all accurate.

Those cells also had a C rate of 2 for momentary discharge and a C rate of 1 for continuous discharge.

Unfortunately, there is no US seller of those cells. I had to order the cells directly from the manufacturer in China.

That is enough for this installment. I will write more later.

 

[Firefighter58]

Hello, my scooter a buzz around EX come with two lithium batteries and I get about 20 kilograms per charge, if that helps.
Al

 

[swalker]

Al, there is a pretty good variety of scooters with lithium battery packs. Likewise, I can easily find compact, portable wheelchairs that use lithium. Unfortunately, those won't work for me.

The only full-up wheelchair that I know of available in the US using a lithium battery is the Bounder. Unfortunately, the Bounder is not covered by insurance in the US.

So, I am left to my own devices, which resulted in this project.

Here is the next installment of my lithium battery build saga.

In the last installment, I went over the factors that led to the selection of the lithium chemistry (LiFePO4), the cell packaging (prismatic), and the vendor. Since there was only one vendor for the cells I selected, the vendor choice was surprisingly easy.

Another facet of the project was to figure out how to charge and maintain the battery pack. LiFePO4 cells can be easily ruined with incorrect charging. The cells are fairly expensive, so ruining them quickly is pretty undesirable.

After a lot of research and learning, I decided that a very high quality charger with characteristics particularly well suited to charging and maintaining LiFePO4 cells was necessary. I followed the recommendation of others who had gone before me and selected a Revolectrix PowerLab 8 Version 2 (PL8) charger.

That charger has the ability to charge all the cells in the pack and to carefully ensure that the cells are balanced against one another. This is very important so that one or more cells do not accidentally drop below the lower voltage threshold for the cell, which would ruin the cell.

That charger does not plug into the wall. Instead, it is fed from a Direct Current (DC) power source, such as a power supply. I need a pretty high capacity power supply to drive the PL8 and charge the battery pack relatively quickly. The lithium battery pack takes much more energy than the lead acid batteries. The directions for my lead acid batteries say to charge them for 15 hours. If I were to transfer energy to the LiFePO4 battery pack at the same rate, it would take days to charge.

For the power supply, I selected the Revolectrix 1200 watt, 5V to 24V, 50 amp power supply.

The next part of the project was to figure out how to connect the charger to the cells, especially in a way that would not damage any of the wheelchair's electrical components and would be absolutely safe.

I could not use the XLR 3 prong plug found on many wheelchairs. I need 11 total wires going from the charger to the battery pack and I need to charge the wheelchair at a much higher current than would be appropriate for lead acid batteries.

In principle, it is simple. The 11 wires must run from the charger to the 8 cells making up the battery pack. Two of these are wires that must be able to carry 40 amps at a nominal 24 volts (really more like 28 or 29 volts) and 9 wires used for balancing that will typically carry just a small amount of current, again, at nominal 24 volts).

These 11 wires then connect to the battery via ring terminals that are bolted onto the battery, much like on a flat top group 24 battery found in many wheelchairs.

The charger uses two different types of connection: two 4mm banana plugs for the high current wires, and a JST-PA connector for the 9 balance wires.

I did not want the entire wiring harness to always be resident on the chair, so I needed to divide the wiring harness into two pieces with a single connector. When charging, the wires connected to the charger could be simply plugged into a connector on the wheelchair.

For this, I followed the recommendations of others to use a 13W3 dsub connector (the same kind used to connect monitors on some Sun and Silicon Graphics computers, though I needed a higher amperage version).

One end of the 13W3 connector would be semi-permanently mounted on the seat of the wheelchair and the other end would be on the cable coming from the charger.

I also wanted to have a connector that would allow me to disconnect the wiring harness so the seat could be removed (necessary for any major maintenance on the X4 wheelchair). I selected JST-PA connectors for the balance wire and Anderson Powerpole SB50 connectors for the two high current wires.

It was a lot of work to put together all those components of the wiring harness, along with a lot of research into connectors and the tools needed to make the connections correctly.

The result is shown in the pictures below:

1. The cable that goes from the charger to the wheelchair seat.
2. The cable that goes from the wheelchair seat to the chassis.
3. The cable that goes from the chassis to the battery cells.

That is it for this installment. Next time, I will talk about how I would monitor the condition of the cells and overall battery pack during a ride.

Steve

 

[swalker]

Here is the next installment of the lithium battery build.

It is critical that the LiFePO4 cells I selected are not drained below a certain voltage. For safety, I plan not to go below 2.9 volts, though they could probably survive 2.7 volts OK.

To monitor the cells, I found an 8S cell meter that is sold to the hobby industry (typically used by RC model airplane enthusiasts who can have 8s cells). That cell meter will handle the LiFePO4 chemistry just fine.

To connect the cell meter, I needed to run nine 22 AWG wires from the battery pack to the cell meter. I was able to purchase two premade cables that were 36 inches long for this. These cables came with JST-PA connectors. I needed ring terminals on the battery end and a JST-XH connector on the cell meter end.

So, we cut off the end of one cable and crimped on 1/4 inch ring terminals with heat shrink. We cut of the other end of the other cable and crimped on female JST-XH pins, which were then inserted into the JST-XH nine-pin housing.

This worked out great, because the two cables are joined together by the remaining male and female JST-PA connectors. This allows me to disconnect at that point when the seat needs to be removed for maintenance. Just about all serious maintenance on the X4 requires the seat to be removed.

This arrangement has worked out great for the limited test I have done on the wheelchair.

One of the challenges I faced was coming to grips with all the connector types involved.

The ring terminals on the battery end were pretty straight forward. We used 1/4 inch, heat shrink, tinned, copper terminals for all the 22 AWG wires (the balance wires and the cell meter wires) and 10 AWG main power wires.

We used 1/4 inch, 8 AWG tinned, copper ring terminals with heat shrink for the 6 AWG wire for the seven cables used to interconnect the 8 cells into an 8S configuration.

Those were easy to figure out, except we learned the hard way that the 6 AWG ring terminals were too large for the 6 AWG wire. After we figured this out, we found the fine print on the manufacturer's site telling us to use the 8 AWG ring terminals, and they worked perfectly. That delayed us for a few days while we waited for the correct terminals to arrive.

The battery charger has a male JST-PA connector for the ballance wires. I had never heard of that kind of connector before. I researched those and found that Japanese Solderless Technology (JST) makes a wide range of connectors, including the PA flavor. As mentioned above, I was able to order premade 22 AWG cables with JST-PA connectors, which saved us a lot of work.

And, as mentioned above, we had to replace one of those JST-PA connectors with a JST-XH connector for the cable to the cell meter.

I also wound up learning about Anderson Powerpole SB50 connectors. It turns out the X4 came with a pair of those for the cables going to the motors. We added an SB50 connector for the 10 AWG main charge cables from the batteries to the new charge connector I mounted on the wheelchair.

The new charge connector mounted on the wheelchair is a female 13W3 dsub connector, which is the same configuration used by monitors on some old unix workstations (I have used them on SUN and Silicon Graphics monitors for years). I needed ones that could carry much more current than would be used for a monitor, and was able to find some rated at 40 amps. 40 amps is plenty for my application.

The 13W3 connectors I purchased have solder cups. We soldered the 10 AWS main charge cables to the 2 of the 3 larger connections and the nine 22 AWG balance wires to 10 of the smaller connections. That was a bit tedious, but we got it done.

We used heat shrink at the ends of all the cables we made and encased the cables in looming material for a protection and a clean look.

That is it for this installment. I should be finished after a few more of these.

Here is a picture of the cell meter cable.

Steve

 

[Nuts]

Steve, I am in awe. Seriously.

 

[KarenNWendyn]

I’m also in awe. Glad you and your friend are so resourceful.

 

[swalker]

In this installment, I will describe the process of buying the batteries.

If you have followed my saga so far, you will know that the characteristics I deemed necessary for the battery cells limited me to a single vendor, which is in China.

I could not find any resellers of that particular product in the US. I corresponded with the vendor via email for several months and finally decided I needed to just dive in and buy direct from China.

This is not the first time I have ordered directly from a vendor in China, and my experiences have left me jaded. Prior to this, I have had one successful transaction and several failed transactions. I always used paypal and always received a refund for the failed transactions. But, I had never purchased anything directly from a vendor in China for more than $100.

LiFePO4 lithium cells are expensive and expensive to ship. The vendor charged $110 per cell and $290 to ship all 8 cells. The total price was $1,170. I was wary of spending this much with a Chinese vendor, but, as I said, I had no real choice (except to abandon the project, which I was not about to do).

The vendor accepted Paypal, so I placed the order and waited, a bit nervously.

After about a week, the vendor set an email saying that the shipment of the cells would be delayed by 2 weeks due to the Chinese new year. They were supposed to ship before the Chinese new year.

I waited anxiously for those two weeks and then most of another week before the vendor finally sent an email saying the cells had shipped. About two weeks later, the cells arrives.

I was pleased to see that the cells were well packed and that there appeared to be no damage during shipping. Unfortunately, I also observed that the plastic coating on each cell showed significant wear and marring, as if they were used cells. I also noticed that each cell was bulged.

The marring and bulging were clear indications that the cells had been abused, either by extended prior use (as in used cells) or by improper charging and discharging.

I communicated the issues to the Chinese vendor via email. They made a few excuses, but, without any particular prodding from me, offered to correct the situation. They asked that I find out how much it would cost to ship the cells back to them in China.

After quite a bit of investigating, I did not find any way to ship damaged cells to China. Most carriers will not ship damaged LiFePO4 cells at all. I found one carrier that would, but they would only ship within the 48 contiguous US states.

I communicated this to the Chinese vendor. To their credit, they asked me what I would like to do to resolve the situation.

To me, these looked like used (potentially well-used cells). In the US, comparable LiFePO4 cells sell for at least a 50% discount, so I asked for a 50% refund. They agreed and refunded me the money.

I give credit to the Chinese vendor for making this right. Obviously, I would have preferred to have unused, unmarred, unbulged cells. But, I really had few options. There were no other cells that fit my criteria.

Most LiFePO4 cells advertise a life span of 2,000 charge cycles. The ones I bought advertise a life span of 6,000 cycles. I suspect 2,000 is the right number. The cells I received probably have a significantly degraded life, either from prior use, from abuse, or from both.

I calculated how often I charge the X4 wheelchair, and decided that even if I only get 500 charge cycles out of my cells, they will likely last longer than I will. So, I am satisfied with the set that I have.

I was able to fully charge each cell, fully discharge then recharge one of the cells, and then partially discharge each cell. This effort confirmed that the cells could at least be charged and would hold a charge adequately and would likely have close to the advertised capacity.

That is it for this installment. I am almost to the good part where we start putting things together.

Steve

 

[swalker]

Here is the next installment.

With "sort of good" LiFePO4 cells in hand along with all the necessary bits and pieces, we got to work building up the overall battery pack.

We wrapped each of the eight cells in fish paper, which is electrical insulation paper. We then encased each cell in PVC heat shrink material.

After that, we created the seven cables that would connect the cells together in series.

Finally, we wired everything together on a desk and tested charging the battery pack, discharging the battery pack, and then charging it to full again. The wiring consists of two main charge cables (plus and minus), 9 balance wires, 9 cell meter wires, along with the 7 wires that connect the cells together in series.

Connecting them in series allows the nominally 3.2 volt cells to create a battery that is close enough to the nominal 24 volts needed by the wheelchair's electronics and motors.

The battery pack consists of two parts, each with 4 cells. Each of those parts fits roughly into the space that was consumed by the group 24 lead acid gel batteries that we were replacing.

Here are some pictures of the process.

Steve

 

[swalker]

After bench testing the battery pack, we were ready to install it in the wheelchair.

After all the expense and work invested so far, we still did not know for sure that the LiFePO4 cells would fit into the Magic Mobility X4's battery boxes. The only way to find out was to go through the laborious effort of removing the entire seating system and then the fiberglass shroud.

My friend Greg did all the lifting and twisting while I supervised. We had a lot of fun working on it together, but it took longer than anticipated. We spent about 5 hours on the install.

The first thing we did after getting access to the battery boxes was to drop in a cell and see if the battery boxes were wide enough. They were, just barely. We had measured them as carefully as we could before embarking on the project, but without disassembling the wheelchair, we could not be absolutely sure the cells we selected would fit. I was very relieved when that first cell nestled in there with about 1/4 of a millimeter to spare.

Though we had bench tested the batteries in a configuration I thought would work, we decided on a slightly different configuration in the installed cells. Greg placed the cells in position and then we started wiring them up. That was a bit exciting, because there were quite a few wires to wrangle.

We had to reroute the wheelchair's circuit breaker and make sure all the terminal bolts on the cells were secured. Making a mistake here could have led to destroying cells or worse.

All went well. We plugged in the cable to the joystick (with the seat still off the chair) and were relieved when we successfully powered on the chair.

We spent an hour or so cleaning things up and reassembling the chair. There were a lot of wires to fit in a pretty tight space. In the end, we got it all assembled and tested it in my garage before Greg took off for the night.

I felt great that night, knowing that the project was wrapped up and looked forward to start testing how far I could go on a charge.

I slept for a very long time that night, not getting up until well into the afternoon. Once I was up, I made my way to the garage and turned the wheelchair on. I was greeted with flashing LEDs and no other response from the wheelchair!

After a moment of panic, I reviewed in my mind what we had done the previous night. It turned out that the last thing we did was lower the seat elevate actuator so that the seat was in its lowest position.

After regaining my composure, I raised the seat a bit and then the chair started working normally (on that wheelchair, you can raise and lower the seat even with the wheelchair powered off).

After investigating the wiring a bit, I found that when the seat was completely lowered it put stress on an electrical connector. I was able clean that up and reroute a few wires so that it should not happen again.

After that, I was off for a test ride.

That is enough for this installment. Next time, I will will reveal the results of the first test ride.

In the meantime, here are some pictures of installing the LiFePO4 cells. My friend Greg is the guy in the pictures.

Steve

 

[swalker]

In my previous post, I described installing the lithium battery pack into the X4 wheelchair. This time, I will tell you about my first ride in it.

After recovering from the panic caused by the X4 not powering up, I was able to diagnose and resolve the wiring issue that was responsible. I then set out on a short test ride just to see how things handled.

I was fully prepared to need rescuing, as we had done quite a bit of disassembly and reassembly to fit the lithium cells into the wheelchair. So, I was planning on going on a ride of a few hundred yards in our neighborhood.

The wheelchair handled wonderfully. Because the lithium cells deliver higher voltage than lead acid batteries, the wheelchair was noticeably faster (always good for me!). Also, because the lithium batteries can deliver power faster than a lead acid battery, the wheelchair was much more responsive.

Finally, the wheelchair had much more power available, again because lithiums can deliver power faster than lead acid batteries.

I was having so much fun testing the wheelchair that I decided to go for a longer ride.

Unfortunately, I was not dressed properly for the weather, which was a few degrees above freezing. I am very sensitive to the cold and normally wear two down jackets along with overpants and heavy gloves even when the temperature is way above freezing. For this test ride, I was not wearing nearly enough.

I wound up going over 5 miles and had a wonderful time. The wheelchair worked perfectly. I could not have asked for anything better.

when I got back home, I was thoroughly cold soaked. It took me a very long time to warm up, but I eventually did.

After the ride, I charged the wheelchair and was surprised at how little energy I had pulled out of it during the ride. I did some quick arithmetic and computed that the wheelchair's potential range was going to be over 25 miles. That was way better than I had though would be possible when I started the project (with lead acid, the range is about 4 miles).

It was a very successful and satisfying test ride.

I have since ridden it close to 10 times. Today I finished an endurance test to see just how far the wheelchair would go on one charge. It took 3 days of riding to drain the battery. I was able to go 33 miles. Absolutely amazing!

That is it for this installment. I will write more later.

Steve

 

[Nikki J]

Steve I will say it again. You are amazing! Thank you for sharing

 

[swalker]

In this installment, I will share a story that is funny in hindsight, but was far from funny as it was unfolding.

After my fourth ride with the lithium battery pack, I went to charge it up and the charger made a series of chimes, displayed two error codes, and then did nothing.

This is a very special charger. It was developed for the radio controlled (RC) model community and is used extensively by RC plane and helicopter enthusiasts. I am not aware of any other charger on the market that will do what is needed.

I had read many reviews about the charger and the company making it before committing to the project and was very impressed with what I learned.

I notified the company of the issue. They had me try a few things and then had me ship it back to them for a warranty replacement. I have since received a replacement, which I tested and confirmed works well.

In the meantime, I consulted with others who had done a lithium conversion on their wheelchairs and the general consensus was that you absolutely need a backup charger. The charger requires a 24 volt power supply, so you need a backup one of those as well.

The company (Revolectrix) was having a March Madness sale, so I went ahead and ordered up a new set. They were scheduled to arrive on a Monday.

Unfortunately, the package was delivered early, on Saturday evening. It was delivered by the US Postal Service. To save about 10 steps, The US Postal Service left the package behind the rear wheel of our car instead of placing it by our front door.

Carnage ensued. The box was pretty well destroyed when the car backed over it. Two of the items in the box seemed undamaged (though their boxes were squished). The third item, the power supply, bore the brunt of the impact and was "distorted".

I was more than mildly upset with the Postal Service. I am convinced this is their failing and they should have the liability of making things right. But, the shipper has to file the claim.

I called Revolectrix and explained what happened. They said in their experience, filing a claim with the post office would be fruitless and instead they would just ship me a replacement. I was clear with them that we were the ones who actually ran the box over.

In the end, we agreed that I would try to "undistort" the power supply. They offered to send me new pieces for the case as necessary. They also said that if the power supply failed any time during the warranty period they would replace it at no cost.

Their customer service was wonderful! I chatted with the owner a bit and found out that there is a good chance we used to hang out together as kids. My father was an avid RC airplane and helicopter enthusiast, as was I. His father and my father were in the same RC club at the same time (we both lived in the Washington DC area). It is highly likely that we ran into each other and hung out at the flying field on a regular basis, though I have no specific memory of him.

All this came out after he had made the generous offer regarding the power supply.

We did surgery on the power supply and were able to straighten out the case and restore the innards fairly well. I have been using it for several charge cycles over the last week or so and it has performed flawlessly.

So, I now have a backup charging system so that if any component fails I will still be able to charge my wheelchair.

That is it for this installment. I will finish off with a picture of the squished box.

Steve