Dream GreenSaver Charger - The GreenBMS

sparc5's picture

I'm moving this from a post to a blog to keep people up to date with my progress.

For those who haven't been following the threads I've posted to here are the problems I'm trying to solve:

1) Battery charging- Make sure our batteries live long and prosper.
2) Equalizing- Make sure they discharge at an equal rate so you're range isn't limited by your weakest battery.
3) Cost- Keep the price lower than the alternatives.
4) Patents- Must do this all without violating patents.

After three weeks of heavy brainstorming here's the leading plan for what I call the GreenBMS:

More than anything, we need a good temperature compensated bank charger. So I'm building a percision, programable, multichemistry, bank charger that uses PWM to provide desulfication and active monitoring of each battery's voltage. Woah, try saying that with one breath (adjusts glasses). This is the heart of the GreenBMS. By using a chip that came to the market only a month ago, you can provide your own DC power source to the charger it can accept anything from 3-75 volts DC. (Anything above 14V means higher efficiency). It is small enough to fit inside the scooter. You can save money by using the charger that came with the bike to power it or buy something beefier. You can even connect it directly to solar panels if you wish.

Extra add-on 1 (An auxiliary battery):

The other benefit to having a power converter that takes such a large range of voltages is you can charge the batteries as you drive, like an alternator does, but it will get it's current from an auxiliary battery instead of a car's engine. You can run the 12V electronics off it and get rid of the inefficient DC-DC converter. The charger will switch to a mode that charges the weakest battery and brings it up to the average state of charge of all the batteries. If the batteries are evenly matched, the auxiliary battery will never discharge much. If all the batteries are very low, or if told to do so, it switches to an extend range mode, where the auxiliary battery charges all the batteries at a rate in which it will be depleted at the same all the other batteries are. When you plug it in, the auxiliary battery gets bank charged with all the other batteries. You can make the auxiliary battery any capacity and chemistry you want. A small cordless power tool battery might be all you need, I need to do testing once I build the prototype. The only cost of this add-on is the price of the battery and the wires to connect it.

Extra add-on 2 (a display, and user interface):

Since the microcontroller (uC) is already monitoring battery voltages, amps, and temperature, it's very simple to display this on an LCD. It will have a graphical user interface and some buttons to let you interact with the BMS' parameters.

Extra add-on 3 (wireless communications):

Everyone can change their programs if they wish to. I know everyone here loves to tinker and share their improvements. I'll offer a long range wireless (102.15.4) data communications add-on so you can program the charger, install updates, monitor the charging, and download data from your rides, from anywhere in your house. It's cheaper than bluetooth, and easier to interface.

This is all going to be opensource to keep costs for everyone to a minimum. I'm doing this project for myself, and making the knowledge on how to do it free for everyone. Early adopters of electric scooters have a soft place in my heart. I'll have extra circuit boards made up, buy enough parts to construct a dozen kits, and sell them in kit form or assembled for a bit extra.

This week I'll be drafting schematics and ordering components.

This project is much more complicated then initially planed. My fears are the cost of the parts, and the complexity of making them all work in harmony. In the past I've programed uC's. I went through a clock making phase. One clock flickered 8 leds fast when you attach it to the ceiling fan it displays the time. I've made a SMPS to power a nixie clock. It turned out well. This time I'll be programming a uC with the most sophisticated program ever. The SMPS this time is powerful enough if I put a wire in the wrong place it could start a fire. If you guys don't hear from me in a while, you can assume I forgot to ground something. :P

As a request: Everyone please take out your multimeters and tell me the resistance of your batteries and voltages before charging and after. Include the temperature as well. Thanks in advance.

-=Jeff=-

Comments

Sparc5,

I think this is an excellent idea. I have been thinking of building a charger for my Prius, using standard 12v pack as the source to charge the 230v pack. Using the 12v deep cycle as a source has several advantages, as you've already mentioned; but i would add that it is very commonly available and at decent price. It is also an easy way to add range, and I can tailor my battery load depending on my anticipated range. For example, if I am going to take a longer trip than normal, i can add 1 or 2 extra 12v as "luggage". And for shorter trip, i can off load them to save weight/space.

What is this new programmable chip you are referring to ?

I'll send you my batt spec the next time i charge them.

sparc5's picture

The LM5118 is my leading choice now for a buck boost converter. It's limited to 75V output so no help with your Prius as far as tying into the pack voltage with that chip. Has the Plug In Prius community made a boost converter?

Used Prius packs go for around $800 on ebay, http://hybridinterfaces.ca/assemblies.html looks interesting.

XM-3000...
-DC-DC converter replaced with a Dell D220P-01 power supply.
-72V mod
-Expensive bank charger until I come up with something better... Still trying.
-

sparc5's picture

If all the batteries are very low, or if told to do so, it switches to an extend range mode, where the auxiliary battery charges all the batteries at a rate in which it will be depleted at the same all the other batteries are.

Same "time as" rather. I really need to start proof reading. :D

XM-3000...
-DC-DC converter replaced with a Dell D220P-01 power supply.
-72V mod
-Expensive bank charger until I come up with something better... Still trying.
-

I'm not exactly technically challenged, but neither can I discuss the latest IC developments with you :) I am also planning on getting a 3500Li so some of this work might not be valid for me (very interesting though!).

Here is my question, Given your "rig" or some part there of, Would I be able to mount solar collectors on a board ( at home) that collects friendly little electrons and stores them in the 12 v batery all day long; Then when I return, plug your charger in to the battery ?

Is that the most efficient (cost and power) or is there some obvious hole?

Mik's picture

4) Patents- Must do this all without violating patents.

Hi, good ideas there!

It might be equally important to stop others from patenting the ideas, lest they get locked up again!

Mr. Mik

This information may be used entirely at your own risk.

There is always a way if there is no other way!

sparc5's picture

Hi Dave,

You can plug any DC power source up to 75V into the GreenBMS. Your method of charging would work given the battery you're using to store the solar energy stores enough energy to fully charge the scooter. It's also possible plug the solar panels directly into the GreenBMS, and you'll get higher energy efficiency from the solar panels.

Since the scooter comes with a BMS, we'll just have to see how the charger works then and write a customized charging script. It's likely to be beneficial especially for solar charging of the batteries since you won't need to convert DC to AC back to DC to power the charger it came with. If you're just charging from AC, it's unlikely to benefit you much except for the individual cell monitoring.

XM-3000...
-DC-DC converter replaced with a Dell D220P-01 power supply.
-72V mod
-Expensive bank charger until I come up with something better... Still trying.
-

Thanks for the reply. For me the issue is that my scooter will be at my office during the day and I can not put up solar panels there.

I don't really expect to fully "top off" the scooter batteries but I may be able to get enough to get me to and back from my office.

SO this GeenBMS can actually be programmed ? I am a software geek so that sounds exciting!

sparc5's picture

Yes it's completely programmable. You'll get full control over the hardware.

As for display, what kind do you guys think I should go with? Listed from cheapest to most expensive:

A)Two line monochrome LCD, like the PakTrakr uses.
B)A two color OLED 0.85"x0.44"
C)Palm III.
D)Small color Nokia LCD like on a cell phone 1.2"x1.2",

XM-3000...
-DC-DC converter replaced with a Dell D220P-01 power supply.
-72V mod
-Expensive bank charger until I come up with something better... Still trying.
-

reikiman's picture

Yes it's completely programmable. You'll get full control over the hardware.

As for display, what kind do you guys think I should go with? Listed from cheapest to most expensive:

A)Two line monochrome LCD, like the PakTrakr uses.
B)A two color OLED 0.85"x0.44"
C)Palm III.
D)Small color Nokia LCD like on a cell phone 1.2"x1.2",

The most recent Linux Journal has an article on connecting a gumstix unit to an e-ink display. Surely that's an option ;-)

Two colors are always nice t when you MAY need to alert to an error condition.

sparc5's picture

I went with the color OLED. They are amazing. The epaper is damn cool, and low power, but no backlight and slow refresh. Are you into uC's David?

Guys, I have some crappy news to report, the charger I wanted to build won't be much cheaper than the vector chargers everyone is buying. Big 500W-1000W DC power supplies, the big mosfets, and the rest of the parts, as well as assembly, makes the vector chargers a price I won't be able to beat. By the time we change our batteries, I suspect there will be a lot of other batteries with new chemistries that will be more affordable.

I'm still trying to optimize the circuit. Looking at the way modern RC battery chargers function has inspired me.

More to come... I built the auxiliary battery for testing. It's made from 8 used lithium laptop battery packs I bought off eBay. I would need a minimum of 40 packs to replace all the greensavers in the scooter, and that's assuming all the cells in them are good, which they are not. It's the cheapest solution, but require discharge current limiting, they are the kind of batteries that explode if discharged too fast. A metal box is required also just in case :-)

XM-3000...
-DC-DC converter replaced with a Dell D220P-01 power supply.
-72V mod
-Expensive bank charger until I come up with something better... Still trying.
-

reikiman's picture

Are you into uC's David?

No.. I've got nearly 20 yrs experience as a software engineer, on desktop or server systems.. but have had some thoughts along the lines of building a microcontroller gizmo to do what the paktrakr or cycleanalyst does but integrate the functionality more cleanly. I like the paktraker from a gear-head perspective because it lets me explore what the battery pack is doing during a ride. But the UI is next to horrible and I'd also like a spedometer on the same display just as in the cycleanalyst. And the CA UI is also next to horrible. It seems possible to make an embedded linux gizmo (linux since I despize windoze) that has real time data acquisition and displays this data in a user friendly way. It maybe could also in real time send commands to the controller to modify controller behavior. And heck if it had an itunes equivalent it could also be the center of an entertainment system.

Whatever... actually doing that kinda gizmo is so far afield from my work experience but reading through the gumstix and/or buglabs web sites makes it seem possible to do it. Whatever... I clearly don't have time to actually do it. But it's fun to think about.

sparc5's picture

Some uC's allow you to program them in Java. The Propeller I believe is one of them. If I was to make something easily hackable I'd write it in Java or Basic, but for various reasons (mostly time and money) I'm writing it in C++. I'll heavily comment it so you'll be able to easily reverse engineer it.

I drew out my first user interface on paper. I want it to take advantage of the color screen and use the color for some functional way. I'm also planning on incorporating 3D graphics in it, just to give the GreenBMS a high quality feel. Writing your own charging script is the hardest part to keep simple, too many variables and just five buttons.

It's my first time making a user interface. If it blows, people like you can change it. Don't worry it will have the PakTrakr monitoring and Cycle Analyst speedo by tapping into a hall sensor.

XM-3000...
-DC-DC converter replaced with a Dell D220P-01 power supply.
-72V mod
-Expensive bank charger until I come up with something better... Still trying.
-

Mik's picture

Could you use an existing hand-held device as the basis for programming your interface and the charging and other algorithms?

Something with trans-reflective touch-screen, weatherproof, GPS/navigation, MP3 player, bluetooth, maybe even a phone...

The sort of thing many people drag around anyway! It could maybe serve as the electronic key to the vehicle as well.

Mr. Mik

This information may be used entirely at your own risk.

There is always a way if there is no other way!

sparc5's picture

I was just working on the software for the GreenBMS speedometer. I can have it show the usual boring MPH, KPH, RPMs, but just for fun I think I'll do meters/sec, and football fields per hour. Am I being too silly?

XM-3000...
-DC-DC converter replaced with a Dell D220P-01 power supply.
-72V mod
-Expensive bank charger until I come up with something better... Still trying.
-

reikiman's picture

If you do footlongs per fortnight you'll satisfy a dream one of my high school buddies kept wanting.

But.. uh.. for practical use, don't you think it should show measurements in units typically used in normal use?

My high school buddy was turning his homework in calibrated to footlongs per fortnight I think because the teacher kept walking us through units conversions and he had some screwy wild hair up his backside about tweaking the teacher.

sparc5's picture

I mean it will have the standard plus some non standard units of measure. Just for you I'll put in footlongs per fortnight. I'm also going to put in cubits per hour, I'm gonna target that Vatican City market.

XM-3000...
-DC-DC converter replaced with a Dell D220P-01 power supply.
-72V mod
-Expensive bank charger until I come up with something better... Still trying.
-

jdh2550_1's picture

So what's your guesstimate on cost and on availability?

You say the long range wireless you've chosen is easier to interface - but easier for whom? Bluetooth would be nicer IMNSHO.

BTW, having this as a blog makes sense - unfortunately I don't always notice when blogs get created and or updated.

Are you shopping at Sparkfun? They have a lot of good displays and various interface technologies.

Is the design modular? i.e. can I just add extra modules to go to different cell counts?

John H. Founder of Current Motor Company - opinions on this site belong to me; not to my employer
Remember: " 'lectric for local. diesel for distance" - JTH, Amp Bros || "No Gas.

sparc5's picture

Hi John,

My target is $200 assembled or $150 as a kit with display. It will be modular so if you want just the minimum "keep my batteries safe price" will be even lower. I think I'll offer 500w and a 1000w versions. I want people to be able to use their string charger as a supply to keep price even lower. I've not finalized the specs yet, the biggest cost with these is the step down transformer. The more powerful the more expensive. Wireless will be longer range and cheaper then bluetooth, also zigbee is longer range and easier to interface on my end. Sparkfun has great ideas, but awful markup. If you want to buy their serial bluetooth adapter for $60+shipping it's doable, but you just raised the price of your GreenBMS an extra $55. My display will be a GORGEOUS large color oLED. I placed a large bulk order from these to get them at more then half retail price. If my BMS doesn't sell, I can sell them on online with my custom interface and recover my loss.

I want this unit to be professional and full featured while balancing price. I won't sacrifice quality. Period. That's why it's modular, you can get what you need and pay for nothing you don't want. It will outshine the competition by combining all the features people want in just one system, while offering redundancy, open source hardware and firmware and design with an eye for detail. I'll prioritize the development so that you guys with the XM-3500 can protect your batteries ASAP, then I'll ship the accessories when they are made.

Since my first post a month ago, I've been spending all my free time catching up on all the latest charging methods, state of charge measuring methods, various battery chemistry's, as well as looking at what the other chargers are capable of. I want it expandable so it can be used on any size pack with minimal upgrade cost, or even on solar and EV markets.

I hope to have a prototype ready this week, I'll test for 3 more weeks and if it goes well I'll have the circuit boards made and shipped. One month from now is the earliest for the main system. Two months is the latest.

XM-3000...
-DC-DC converter replaced with a Dell D220P-01 power supply.
-72V mod
-Expensive bank charger until I come up with something better... Still trying.
-

jdh2550_1's picture

Jeff,

It sounds great - too good to be true in fact ;-)

When I said modular I was more thinking about number of batteries managed. Not features added. So, what's the difference in price between a 5 battery PbA version and a 20 cell LiFe version?

Rather than wireless, how about an SD Card. You can log to the card and you can download updates to the card from the PC and then loaded from the SD card by the uC. You'll need some form of "mass storage" for data logging, right? So, using removable storage might negate the need for a wireless interface entirely. A lot of your audience will already have SD readers - those that don't can get them for around $10. Just a thought.

Good luck - and let me know if I can help in some way.

John H. Founder of Current Motor Company - opinions on this site belong to me; not to my employer
Remember: " 'lectric for local. diesel for distance" - JTH, Amp Bros || "No Gas.

sparc5's picture

Hi John,

Yes, it's modular in the number of cells also, that's a major design requirement so that it doesn't become obsolete. I'm expecting it to cost about $2-3 dollars a cell. It has two levels of redundancy to protect your investment, without costing more than your investment. The other inexpensive shunt based BMS don't have that redundancy.

Tell me how you'd like it configured for the xm-3500. Monitor each cell individually, or each group of four?

XM-3000...
-DC-DC converter replaced with a Dell D220P-01 power supply.
-72V mod
-Expensive bank charger until I come up with something better... Still trying.
-

jdh2550_1's picture

Each cell individually please and damn the expense! :-)

Are you using a PIC on each cell communicating back? Or one central controller? Can I see the design somewhere?

Can it also measure amperage draw?

Thanks!

John H. Founder of Current Motor Company - opinions on this site belong to me; not to my employer
Remember: " 'lectric for local. diesel for distance" - JTH, Amp Bros || "No Gas.

sparc5's picture

Hi John. Your wishes are my command. Each cell will get tender care. I will also use an SD card instead of wireless, Anyone else want to give me a wishlist?

There will be a PIC on each cell, they'll communicate with each other over a single communications bus to keep the wire nest to a minimum. Stay tuned everyone, a prototype working on lithium and lead acid cells should be on display by the week's end.

-=Jeff=-

XM-3000...
-DC-DC converter replaced with a Dell D220P-01 power supply.
-72V mod
-Expensive bank charger until I come up with something better... Still trying.
-

Mik's picture

Could it be adjusted for use with the 102 cell NiMH battery in my Vectux?

Mr. Mik

This information may be used entirely at your own risk.

There is always a way if there is no other way!

Mik's picture

As a request: Everyone please take out your multimeters and tell me the resistance of your batteries and voltages before charging and after. Include the temperature as well. Thanks in advance.

Is there a way to measure the resistance of batteries with a standard DMM?

Mr. Mik

This information may be used entirely at your own risk.

There is always a way if there is no other way!

sparc5's picture

Is there a way to measure the resistance of batteries with a standard DMM?

There is, it's called the DC load test, where you measure the voltage sag which correlates to a known load. or AC Method (No, not for finding roots of a polynomial :D) where you watch a phase shift between voltage and current.. I realize this is too much work for everyone. I've only had one person give me voltage readings even. So the GreenBMS will have to be able to take these measurements and learn while in operation.

XM-3000...
-DC-DC converter replaced with a Dell D220P-01 power supply.
-72V mod
-Expensive bank charger until I come up with something better... Still trying.
-

sparc5's picture

Could it be adjusted for use with the 102 cell NiMH battery in my Vectux?

Yes, it will work on NiMH, it has all the necessary hardware. The only negative is I am writing the script to this one last. From what I've read it's more complicated than lithium and lead acid because it gives false voltage peaks. Each BMS has an optional temp sensor input, you can use that with good results but there are even more complicated ways that pulse the charger and apply some algorithms have even better results. There may be even better ideas out there, I'll have to experiment if no one else will.

This is open source, for this very reason everyone can expand on the functionality. If you're able/willing to write in C++, this can be your nearly immediate solution.

Which battery does it use? I'll design the BMS so it can screw onto the top of your battery if possible.

XM-3000...
-DC-DC converter replaced with a Dell D220P-01 power supply.
-72V mod
-Expensive bank charger until I come up with something better... Still trying.
-

HCT's picture

Delighted to see interest in battery/cell impedance or resistance , it made my day ! !

The only way impedance or resistance can be measured is by :
Sending a varying know DC current through a battery (cell).The Delta Voltage change from the two DC currents ( can be charging or discharging ) delta V divided by the Delta Current Yield the impedance or resistance.
In the DC method,the current change must be significantly different ( 10/1 ) to get any reliable resolution .
Also the charge state of the battery will be effected which may introduce error.

The AC method , the concept is the same ,expect a 1 KHZ Sine wave is used at very low current so that the battery ( cell ) is not affected .This very small delta V AC measurement is difficult to measure among all the other noise present.

AC method is very fast and is the one used by most if not all manufacturer in testing their cell on the production line for impedance and a good indicator of its capacity.The lower impedance usually conrespondant to high capacity regardless of chemistries .

Exception is , if a cell has a internal short or even partial short , its impedance will appear lower and its value useless .

My humble opinion,believe that the DC method need better connection,will require a significantly longer test and need power to charge or power dissipation both a waste of energy.The AC method does not consume any power so to speak.

HEAT & LOSS & INEFFICIENCY RELATED DIRECTLY to IMPEDANCE

sparc5's picture

Hi HCT,

Glad to have made your day and to have your contribution to the design. My plan was to use the DC method because it's more simple to monitor current draw. When you're driving or charging the scooter the current and/or voltage will in the process very and this will give points to use in calculating resistance. I put a scope on a battery and operated the scooter, lots of noise. It will wreak havoc on the ADC. Filters are inevitable for voltage and current measurements.

What do you think?

XM-3000...
-DC-DC converter replaced with a Dell D220P-01 power supply.
-72V mod
-Expensive bank charger until I come up with something better... Still trying.
-

HCT's picture

Happy that I able to contribute to your design.

The noise that you see on the scope is a good indicator of the each cell impedance .,
To convert it into a actual impedance number may not be to easy , but thinking about what you said running the motor creating noise , this come to mind .

The lower the noise level,the lower the impedance,so if you can design a noise level detector you will have a bad cell quality comparator and detector based on less noise mean low impedance , the cells with the highest noise are either discharged or have permanently high impedance , to be changed .

If the cell is shorted ( extremely low impedance ) you will not see any noise or ripple across a short .

In the charge mode , would think that ripple level will give the same results.

Regards

Andre

HEAT & LOSS & INEFFICIENCY RELATED DIRECTLY to IMPEDANCE

Pages


Who's online

There are currently 0 users online.

Who's new

  • eric01
  • Norberto
  • sarim
  • Edd
  • OlaOst

Support V is for Voltage