Building the Best Controller

Get all your technical information about electric bikes here.
Post Reply
User avatar
Kingfish
10 GW
10 GW
Posts: 4275
Joined: Feb 03, 2010 11:23 am
Location: Redmond, WA-USA, Earth, Sol, Orion–Cygnus Arm, Milky Way. Age: > yesterday < tomorrow
Contact:

Building the Best Controller

Post by Kingfish » Aug 16, 2010 5:18 pm

Preamble:

Greetings –
I am on a primal quest to go farther faster better, as I’m certain others are swimming in the same direction. The component choices we have to get there are off-the-shelf, with modification, and by innovation. Like most, I have acquired, tampered, and twiddled my way forward through pain and suffering reaching for that endorphin-laced taste of freedom. The pleasure and joy that surrounds personal achievement on each stretch of road, each hill mastered is second only to the knowledge that we are truly a rare breed, a cadre of few that can do what we do with our custom-tailored machines. Even so, I think we can make them better, smarter, more reliable, and extensible.

Image

Straight-up: I am a Design Engineer though not Electrical Engineer, therefore I ask that you forgive my naivety and accept my open hand instead towards resolving the elementary puzzle, and build the best controller from scratch.

This challenge is not for people who eat Wonder Bread, Dinty Moore Stew, and drink Budweiser; no. It’s for people that have no fear in making something from a mysteriously dark and unknown void, to “roll their own”, and smoke-test it! It’s like gardening or cooking: I like to think of it as crafting a tasty beer from a full-mash with my own small-scale hand-made brewery... Or to go out and buy something that is interesting, take it apart, examine, and figure out how to make it better; a habit I’ve been doing since I was a kid, and well… I’m still a kid at heart 8)

Not long ago I examined the Infineon controllers, reviewed the software, and attempted to reverse – or at least understand the conditions upon how the firmware was set through the programming interface. This is an ass-backward approach to engineering a solution: Replace the foreign application with an open-source to affect changes on foreign-built controllers. No – that is the wrong direction. Ambitious edge-case driving people need more control, period. Therefore I’m on the hunt to build the best controller, and I’m looking for other hunters to join with!

Defined:
Every project should have a Plan. The function of the Controller can be defined as being the sum of three parts:
  • Power Supply – Modulates the raw battery power into serviceable channels that can be used by our System.
  • Driver Circuit – Manages the power output to the Motor and rectifies the regeneration.
  • Brain – The configurable controlling module that responds to User Input demand, regulating performance the System, and providing time-critical feedback
Any other feature that a controller might provide is an ancillary application that dovetails into one of these buckets. Examples:
  • Headlights – Source of power is branched from the Main Power Supply. Control is localized or augmented through the main.
  • Alarm System - Source of power is branched from the Main Power Supply. Control is managed through the Brain which tells the Driver to lock up, or produce some other audible deterrent.
  • External Device, such as CA or GPS - Source of power is branched from the Main Power Supply. Data is provided through the Brain.
Image

“Form Follows Function”
The brevity of this statement was extracted from a phrase first expressed by architects Louis Sullivan, and extended by his apprentice Frank Lloyd Wright; it is a principle that advocates a way of defining the shape of an object based upon the intended use or purpose.

The controllers we use today are prepackaged in tight little boxes affording little room for expansion. Though fine for the lion’s share of the consumer market, for the hobbyist it presents a mild-to-difficult challenge to augment. It doesn’t have to be this way: The goal of this thread to break this device up into three parts that are accessible, extensible, and reliable.

The needs for my next ebike, and the ones after that, demand that I abandon off-the-shelf controllers as being too inflexible and limiting to my creativity. To be fair, they do their job nicely. However, I have learned enough, and now I’m ready to move on. I can’t be the only one with these strong aspirations.

I ask you: Let us agree on the features, design the circuits, map the layouts, cut the boards, and stuff them. Let us build a product that is reliable and rocks! As part of that creative process we require the development of a specification, to lay the foundation as it were.

more...
Last edited by Kingfish on Mar 04, 2011 3:31 pm, edited 1 time in total.
* My 2WD Garden Wall
* Kinaye MotorSports
* Primary ride: 2WD Disc 9C 2806-equiv / Dual Lyen 12FET / 20S7P LiPo.
* Epics: Going to California: 2011 8)
* 50-mph, 101, 10k-Club. 12,527 miles-to-date, 7037 as 2WD.

It is by caffeine alone I set my mind in motion.
It is by the beans of Java that thoughts acquire speed.
The hands acquire shakes, the shakes become a warning.
It is by caffeine alone I set my mind in motion.

User avatar
Kingfish
10 GW
10 GW
Posts: 4275
Joined: Feb 03, 2010 11:23 am
Location: Redmond, WA-USA, Earth, Sol, Orion–Cygnus Arm, Milky Way. Age: > yesterday < tomorrow
Contact:

Feature Specification

Post by Kingfish » Aug 16, 2010 5:35 pm

Each module should be stout enough to stand on its’ own, with characteristics personalized to the User need, and flexible in utility. There very well could be different versions of a particular model, though all are derived from a master plan. Imagine now that this is your birthday, you can have whatever you want, and we present the all-inclusive list.

Image

Main Power Supply:
  • Main Battery Power, Unregulated, 36V to >100V, with voltage test pads
    • Option to have high-charge and discharge buffer (SuperCap or High-C batteries)
  • 12V Branching Regulated PS
    • Driver PS, Regulated, Isolated, 12V for FETS
    • Controller/Brain PS, regulated (5V or less)
    • On/Off circuitry (cut power to Brain or Driver supply), with slow-ramp/spark suppression
    • Auxiliary PS, regulated for lights and accessories (may be a mix of 3.3, 4.1, 9V), with optional capacitance when main power is lost. Example: Headlights still work for another hour if the main batteries loose power and you have to pedal home.
    • Scalable power output to meet or exceed need.
  • Separate Charger & BMS connections
Image

Driver Circuit/FETs:
  • Modular driving circuit to provide fast-switching/high-resolution and high-current requirements
  • Manage Regeneration
  • Location of this circuit needs to be close to the Battery for best performance
  • Audibly/RF-quiet; supports either Trapezoidal or Sinusoidal waveforms
  • Supports low-to-high kV
  • Designed to handle 40A minimum out of the box, with built-in options to dramatically scale-up.
  • Flexible thermal cooling options
Controller/Brain Module:
  • Low Voltage, low power demand
  • Manage low-to-high kV motors
  • Audibly quiet
  • Basic Features:
    • Throttle; ramp, resolution, delay, dual-throttle support
    • ebrake-Regen; minimum, maximum, resolution, dual-input support
    • Multi-speed switching on the fly (ala 3-way)
    • Smart-Cruise; on/off, set, coast, accel/decel, one-to-many presets stored in non-volatile memory.
    • Alarm/Wheel-locking
    • Reverse
  • Advanced Features:
    • Waveform: Trapezoidal, Sinusoidal, or possibly custom profile
    • Power Management profiles for hills and flats
    • Advanced Throttling; by current-limiting, by voltage-limiting, by phase-shifting/clocking
    • Multi-motor support, non-collision
    • Sensorless support
    • Set poles and slots, limits for V/I
  • Event Logging
  • Connectivity via Bluetooth, USB
  • User-Programmable through .NET
    • Support for multi-language/localization
    • Multi-platform Support for Windows, Apple, Linux, Android
    • User-defined configuration parameters
    • Set and Get configuration
    • Help Guide online
Image

Give the People What They Want:
The principle behind this controller is a modular product packaged just for you, for experimenters, for racers – for those of us that simply want to go farther, faster, better!

If you can think of a missing feature, please post. There are lots details to review and many ways to contribute. The very next task on my agenda is crafting the System Diagram: To map out the components and features from a high level.

With humility I realize that there have been pursuits like this before. Well, no – that isn’t exactly true: No one has ever tried to make a controller as an assembly of three distinct modules. I’m not the first or the only one with this concept either. It is however a hill that needs climbing, a novel approach to flexibility, and one designed for ultimate performance in mind.

My wager:
<shoving all chips in> With over 30 years of engineering experience, I can do printed circuit board design, packaging, and software development. I am really great at documentation and online support too. My weakness though is EE and it is here in particular where we need input and validation.
I ask the graybeards of this forum with their years of experience, and I reach out to the fearless young bloods with their ambitious moxy: With your support, we can make this kick-ass open source project; let’s hit a home run first time out!

Image

The Cost:
We place orders, burn boards, split cost. Parts and assembly are on you. Can you think of a less-expensive way? :)

Wouldn’t it be awesome to begin with an ambitious goal and aggressive schedule:
  • Before the year is out, ten of us will have functioning prototypes.
Who’s with me?
~KF
Last edited by Kingfish on Mar 04, 2011 3:33 pm, edited 1 time in total.
* My 2WD Garden Wall
* Kinaye MotorSports
* Primary ride: 2WD Disc 9C 2806-equiv / Dual Lyen 12FET / 20S7P LiPo.
* Epics: Going to California: 2011 8)
* 50-mph, 101, 10k-Club. 12,527 miles-to-date, 7037 as 2WD.

It is by caffeine alone I set my mind in motion.
It is by the beans of Java that thoughts acquire speed.
The hands acquire shakes, the shakes become a warning.
It is by caffeine alone I set my mind in motion.

User avatar
Whiplash
1 GW
1 GW
Posts: 3079
Joined: May 10, 2010 2:55 pm

Re: Building the Best Controller

Post by Whiplash » Aug 16, 2010 6:14 pm

IF YOU BUILD IT, THEY WILL COME....



Oh an make sure it can handle R/C motors as well! It would be a shame not to since most will eventually go there anyway!
Power is a fascinating thing, the more you have, the more you want, but the real power is having the restraint not to use it all at once...............Um...Yeah..

The harder you work.....The luckier you get!!

"People who say it can't be done are often interrupted by those that have already figured out how to do it!"

Email me @ currentcycles@gmail.com
OR check out http://www.CurrentCycles.net

User avatar
Thud
100 MW
100 MW
Posts: 2809
Joined: Aug 20, 2009 6:20 am
Location: West Michigan,USA

Re: Building the Best Controller

Post by Thud » Aug 16, 2010 6:36 pm

Kingfish,
I know there are a few controller oriented side projects on going atm. I am anxious to see LFP & Jeremy Harris chim in on this endevour.

I wish I had more than a throtle twisters ration & the scares that go with that to offer on a project like this.....as a guy who has streached his imagination with electric bicycles (& beyond) I am rooting for success. (purly selfish reasons 8) ) I will be lurking in the back ground waiting for an oportunity to be able to contribue.
Good Luck!
get some......

All information & advice provided by Thud are "Open Source" & free for personal use & distribution under the following agreement linked below.
http://creativecommons.org/licenses/by-nc-nd/3.0/

User avatar
fechter
100 GW
100 GW
Posts: 11822
Joined: Dec 31, 2006 3:23 pm
Location: California Bay Area, USA

Re: Building the Best Controller

Post by fechter » Aug 16, 2010 7:59 pm

There are certainly challenges in the project, but most of the modules can be 'borrowed' from proven designs. A decent sensorless startup algorithm combined with a trapezoidal output would rank high on my wish list. Regen is good too, if properly implemented.

The dang 12v power supply turns out to be pretty challenging if you want a wide input range. Linear is no problem but wastes quite a bit of heat. Some kind of switcher is really the ticket, but I haven't seen a good setup that's simple and handles over 100v. I was trying to apply some off-line integrated switching chips, and never got a good solution, but I belive it's possible since I have switching power supplies that will fire up at under 20V.
"One test is worth a thousand opinions"

John in CR
100 GW
100 GW
Posts: 13764
Joined: May 20, 2008 12:58 am
Location: Paradise

Re: Building the Best Controller

Post by John in CR » Aug 16, 2010 8:53 pm

Kingfish,

Good luck with this. There's so much talent on ES, and you sound like just the guy to pull that talent together and keep it pointed in a common direction that is for the benefit for all. I love the idea of a modular system that is the "last controller you will ever need to buy."

Some food for thought-

It seems like the direction you're headed is for it to be the electronic control system of the bike. As such, I think it should also manage the battery too.

Since this is to be a smart control system, I think it should handle both sensored and sensorless operation. Then in the event of either a hall sensor failure or hall sensors aren't available that it doesn't miss a beat.

Also since it is smart, we should be able to simply connect the hall and phase wires, click on new motor in the program, and spin the wheel or motor to learn the physical firing order sequence and save. No more guesswork or changing wires around. Instead the system recognizes which wires are which. I have a nothing little tester that does this, so it would be a snap for the controller.

Control over timing advance or retardation could be very beneficial to optimize performance in different conditions, as well as multiple settings for current limits in a addition to the usual speed limits.

Diagnostics in the event of a failure. Tracking down a loose wire or minor or major failure can be a nightmare for the non-electrically oriented. The perfect control system should report what the error is, so it will need an onboard means of both input, output, and display.

Performance recording, reporting and management. Bikes don't have much room, and any duplication of processing or memory or display capability is a waste. Instead of reinventing the wheel in some areas, maybe we can adopt a standard device like the Cycle Analyst to be the on board display and I/O and memory device for this small EV central control system. In exchange for that exclusivity, it only makes sense for additional user functionality. eg A simple and programmable fuel gauge, as in the user enters his battery capacity and what he wants to be his typical capacity available for use (80% of total or whatever), and then a simple bar gauge always displaying remaining capacity is shown prominently in the display. To make it perfect add in the ability to preset alarm notifications, and we might as well get real time consumption and remaining range too, since it's a matter of simplicity.

You mentioned a security alarm, so go ahead and include an electronic lock of the motor and a password protected vehicle shutdown in the event of any tampering, along with a programmable message for the display for thieves to see should they physically score the bike that they can't make function. Whether it's a big F U or instructions for reward upon return or whatever is up to the owner to set.


Like Thud I can twist a throttle and plug in a few wires. When you get to the stage where financial input is needed you can count me in. It's a great idea, so I'm counting on you to manage the project and not let it become vaporware.

John

jsplifer
1 W
1 W
Posts: 55
Joined: May 21, 2009 9:16 am

Re: Building the Best Controller

Post by jsplifer » Aug 16, 2010 10:06 pm

Neat!

Great feature list right off the bat.

John in CR makes some good points.

Couple of mine: ability to parallelize the "driver circuit" per phase for monster motors... Temperature monitoring and cut off! Maybe allow for some motor/fet cooling switch. Might be liquid cooled motors on the horizon. Umm delta/wye switching, variable regen (perhaps this is what "resolution" means). Also allow for multi motor drives, such as dual RC, dual hub or more for velos. (OOPS see you mention that)...

There is an open source brushless dc motor controller site already, perhaps a fork to their project for e-bike needs. http://open-bldc.org/wiki/Open-BLDC

All that comes to mind at the moment. There seem to be some really neat motors in development, and it be great to have a controller that can keep them fed. As it is I see the speed demons blowing controllers left and right and these big suckers aren't even out yet. Like the Colossus, and Recumpance's. I hope a reliable all around solution can be had. Cheers.

John in CR
100 GW
100 GW
Posts: 13764
Joined: May 20, 2008 12:58 am
Location: Paradise

Re: Building the Best Controller

Post by John in CR » Aug 16, 2010 10:34 pm

jsplifer wrote:Neat!

Great feature list right off the bat.

John in CR makes some good points.

Couple of mine: ability to parallelize the "driver circuit" for monster motors... also allow for multi motor drives, such as dual RC, dual hub or more for velos. (OOPS see you mention that)... Temperature monitoring and cut off! Maybe allow for some motor/fet cooling switch. Might be liquid cooled motors on the horizon. Umm delta/wye switching, variable regen.

All that comes to mind at the moment. There seem to be some some really neat motors in development, and it be great to have a controller that can keep them fed. As it is I see the speed demons blowing controllers left and right and these big suckers aren't even out yet. Like the Colossus, and Recumpance's. I hope a reliable all around solution can be had. Cheers.
Good points. The beautiful part is that I believe Kingfish fully understands the magnitude of the project, and his carrot for achieving the goal is a nice slice of pie that just makes this world a better place. As he starts to pull it off, I'm going to try to think of a way to make a similar contribution on the mechanical side where I can do something useful. I can see the ES membership open sourcing a whole new class of different EVs, that will kick the shit out of the 3-4 thousand pound behemoths that the big manufacturers would have everyone buy, and are just a further waste of the world's resources just for profits sake. Can you imagine DIY transport vehicles becoming commonplace? I can, and ES is the best place I've seen to turn that viral due to the concentrated pool of talent in such a wide range of disciplines. It's like a few thousand who "got it" after just a few minutes on an electric bike (or just thinking about their possibilities) all migrated to one spot, and concentrated talent pool across such a wide range of disciplines is truly staggering.

Knightmb and Fechter, there's a t-shirt idea for you
ES
I "got it"

John

User avatar
Whiplash
1 GW
1 GW
Posts: 3079
Joined: May 10, 2010 2:55 pm

Re: Building the Best Controller

Post by Whiplash » Aug 16, 2010 10:39 pm

I think we should watch the list of options here before we get into the 1000 Dollar controller region! Now if its modular, like say an MSD Ignition system, then those of us that need certain things can add them and others that do not don't have to spend the money...
Power is a fascinating thing, the more you have, the more you want, but the real power is having the restraint not to use it all at once...............Um...Yeah..

The harder you work.....The luckier you get!!

"People who say it can't be done are often interrupted by those that have already figured out how to do it!"

Email me @ currentcycles@gmail.com
OR check out http://www.CurrentCycles.net

User avatar
liveforphysics
100 GW
100 GW
Posts: 14455
Joined: Oct 29, 2008 1:48 am
Location: Santa Cruz, CA, USA

Re: Building the Best Controller

Post by liveforphysics » Aug 16, 2010 11:12 pm

I've got the FET stage and gate drivers handled (for people looking to do multi-hundred amp phase currents). I've been working on this side of things for a few grand in parts and the last 6 months or so of freetime. Somebody else should put together a ~50 phase amp option for folks who are making commuters. This can but build to be extremely small and economical (like 1/4 the size of a normal infinion).

Things I need folks to volunteer on are the brain and the control voltage power supplies.

I need fully isolated high-side and low-side switching power supplies for the FET gate drive power. I do not want to boot strap for the high side. Boot strapping is a band-aid to save $5 in parts to do it right, and boot strapping can fall all apart when you starting getting some inductive bounce on the high side power bus to the FETs. When the FET stages have real parts cost them (rather than a handful of TO220's), it doesn't make sense to trust it to boot-strapping. I know the penny-pinchers hate the idea of an extra $5 in parts... Deal with it. :) The MCU chips power supply should also be isolated and stiff enough the power leads can jump, spike, bounce, ripple to their hearts content, and the MCU thinks everything is smooth-sailing. The idea of having a built-in 12-14vdc supply in here as well for folks axillary loads (lights, phone charger etc) sounds great to me. :)

We need a huge focus on the brain side of things. I don't have the skills to do that side of it, and my current method of piggy-backing off infinions for the brain is crap.
Each carcinogen vapor exposure includes a dice roll for cancer.

Each mutagen vapor exposure includes a dice roll for reproductive genetic defects in your children.

Each engine start sprays them into a shared atmosphere which includes beings not offered an opportunity to consent accepting these cancer experiences and defective genetics life experiences.

Every post is a free gift to the collective of minds composing the living bleeding edge of LEV development on our spaceship.

John in CR
100 GW
100 GW
Posts: 13764
Joined: May 20, 2008 12:58 am
Location: Paradise

Re: Building the Best Controller

Post by John in CR » Aug 16, 2010 11:40 pm

Whiplash wrote:I think we should watch the list of options here before we get into the 1000 Dollar controller region! Now if its modular, like say an MSD Ignition system, then those of us that need certain things can add them and others that do not don't have to spend the money...
Whiplash,

We want what would be a $2-3k controller if a company came up with it for sale, all for the shared cost of parts. Options are just programming and maybe the cost of some connectors and little electronic bits. The significant costs will be components in the Driver Circuit which the user can upgrade to suit their specific needs. The topic included "best", so it should have everything appropriate and reasonable that everyone can think of, no more no less. The more important consideration is recruiting enough of the vast talent on ES to break this thing up into manageable pieces, so the burden on any one person isn't excessive. I can guarantee there's plenty of talent, and the burden is on Kingfish to pull it all together and manage the project, and I suspect that's what Kingfish excels at, so let's hold his feet to the fire. The hard part is managing the project, because so much of the talent is accustomed to doing it all themselves as one man start to finish projects. We have a volunteer for that hard part, since it was his idea. The price of the physical pieces is likely to be insignificant for something that could forever change the industry. Welcome to open source for electronics baby! :mrgreen:

Maybe I'm being naive, but why not push for the ideal?

John

User avatar
amberwolf
100 GW
100 GW
Posts: 23954
Joined: Aug 17, 2009 6:43 am
Location: Phoenix, AZ, USA, Earth, Sol, Local Bubble, Orion Arm, Milky Way, Local Group
Contact:

Re: Building the Best Controller

Post by amberwolf » Aug 17, 2010 2:17 am

This might not be relevant or useful, but this is some of the "specs" I'd started to write for my control system idea some time ago.
http://www.adafruit.com/forums/viewtopi ... 6&start=15
Unfortunately no one that offered to help ever got back to me, so nothing ever happened and I gave up on writing any more of the specs. That's page two of the thread; there is more discussion of the idea on the previous page.

But it would have been a modular design, so that bikes without motors woudln't have the motor controller/power section, those without lighting (hopefully none, but...) wouldn't have the lighting board, etc. All could be upgraded or added to anytime, with a simple firmware update or control panel setting change to enable those new modules.

User avatar
Arlo1
100 GW
100 GW
Posts: 8494
Joined: Apr 26, 2009 10:36 pm
Location: Nanaimo BC Canada
Contact:

Re: Building the Best Controller

Post by Arlo1 » Aug 17, 2010 2:30 am

Cant wait to read this!
My Leaf motor controller build. https://endless-sphere.com/forums/viewto ... 27#p963227
My YSR build http://www.youtube.com/watch?v=BRo8r5g4NBg
RC and most types of Lithium batteries you MUST know your individual cell voltages charging and discharging.
Don't keep them were you cant afford smoke or fire!
Never above 4.2v never below 2.7v EVER!!!
HI power controller design. Game Changer

curious
1 kW
1 kW
Posts: 460
Joined: Dec 29, 2007 1:46 pm
Location: NJ, USA

Re: Building the Best Controller

Post by curious » Aug 17, 2010 7:38 am

Are you guys thinking of a commercial project or an open source DIY kit ?

I am not radically against the later but usually this path ends up in a group buy or two and then enthusiasm and support dies out.

If you are thinking about a joint commercial project we need to talk the business side of it first. Most people here would want a fairly high-end controller but this market segment is not large, and it would be extremely difficult to compete with Chinese controllers on their own turf of low-end devices. Additionally if the design is successful it will be copied in no time out there. That is why I am fighting strong temptation to get into the EV engineering business myself :).
bike1: hardtail frame, 5304 rear, 850wh a123 66v/33v switchable, Kelly KEB72601, proportional regen;
bike2: bikeE FX recumbent, sensored bafang front on a 16" rim, 16s a123;
bike3: work in progress (slow), long range SWB touring recumbent, >1.5kwh pack TBD, Kelly KBS72101, likely HT35 hub, proportional regen.

User avatar
fechter
100 GW
100 GW
Posts: 11822
Joined: Dec 31, 2006 3:23 pm
Location: California Bay Area, USA

Re: Building the Best Controller

Post by fechter » Aug 17, 2010 8:58 am

Having a modular power stage has been suggested before. I think this makes some sense. The 'brain' part would be common, and you could have a selection of power stages that meet your needs. Optically coupled might be good.

What about synchronous rectification? Why not. This gives you regen anyway.
"One test is worth a thousand opinions"

User avatar
Whiplash
1 GW
1 GW
Posts: 3079
Joined: May 10, 2010 2:55 pm

Re: Building the Best Controller

Post by Whiplash » Aug 17, 2010 11:10 am

OK I understand, get it right first, then try to get the cost down if possible...Makes sense!
Power is a fascinating thing, the more you have, the more you want, but the real power is having the restraint not to use it all at once...............Um...Yeah..

The harder you work.....The luckier you get!!

"People who say it can't be done are often interrupted by those that have already figured out how to do it!"

Email me @ currentcycles@gmail.com
OR check out http://www.CurrentCycles.net

User avatar
Kingfish
10 GW
10 GW
Posts: 4275
Joined: Feb 03, 2010 11:23 am
Location: Redmond, WA-USA, Earth, Sol, Orion–Cygnus Arm, Milky Way. Age: > yesterday < tomorrow
Contact:

Answers to Questions

Post by Kingfish » Aug 17, 2010 12:00 pm

I am humbled... Thank you for the encouragement! :D
Really great input people – there is certainly lots to think about. This is your time to vote on what you want; let’s hear it! :)

As Richard has suggested, I believe the large part of this effort has already been documented in one form or another and it is just a matter of corralling the resources to provide the best viable solution. Invention will still be required.

Since everything on this forum is open-source, open to public scrutiny as we scrutinize publically the products we purchase – and then have the audacity to publish the modifications, it would be tempting fate to commercialize any product developed. That said we could follow down the path of how the BMS is developed… :idea:

Forgive me; I’m a passionate brewer so I like to use the brewing analogies:
  • I can go out and by a Budweiser.
  • I can learn how to make a Budweiser clone with a kit and brew it on my stove.
  • I can learn how to build equipment to do a full-mash, with cold-conditioning to produce a Budweiser clone.
  • I can build a brewpub and hire a brewmaster to make a Budweiser clone and manage the business.
  • I can go to school, like UC Davis, Seibel Institute, Weihenstephan, etc. and learn from the masters how to properly manufacture Budweiser, and then go work for an industrial complex specializing in chemical engineering as we mass-produce Budweiser world-wide.
  • The problem is – I don’t like Budweiser. I like thick chewy brews that put hair on my chest and muscle on my bones – you know, what the ATF doesn’t want you to have: That which is fortifying, chock-full of healthy goodness, pacifying, and able to stave off the majority of what ales ye. I can make beer that is better than what you can buy in a store, better than what is made by most breweries, and second only to the guy that has the brewpub that is just a tad more twisted in his enthusiasm to push out a fine product to his local patrons. A commercial brewery can’t compete at that level by virtue of the inherited bureaucracy, the cost of boutique marketing, and the thin profitability.
DYI flattens the playing field, takes the wind out of the competitive sails, and sets us distinctly forward on the bleeding-edge of affordability because we inherit all the features of an upstart rebel start-up: Ambitious, light, fast, self-invested (as in no VC), plumbing the best with liquidity.

We have a twisted enthusiasm to create something that either is not on the market already, or so expensive that it’s out of reach. We all become employees, working for stock, the value of which is translated into freedom of choice and expression: To go farther, faster, better.

I believe that the costs should be clamped to reasonable affordability; the guy on Main Street could buy one like he might buy a nice tool that is going to last him for years. And in the spirit of say… a classic old car, can be maintained, upgraded, augmented with horns and whistles – literally, and hopefully even traded. Wouldn't that be cool 8)

There are two forks on the road of Development: Theoretical and Practical.
  • Theory says we can do X, Y, and Z but it might cost an arm, leg, and first-born child.
  • Practical says “Hey I can grab this widget off the shelf, twist it for this amount of effort and get you brother of X, sister of Y, and cousin of Z for less than $250”.
I think if we break down the costs by module, the affordability will become apparent and relative to the features. I expect there will be a dichotomy of affluence in versions offered; this isn’t a hobby for moth-bearing wallets. Personally, I want the best that money can buy… within reason.

Rolling up the sleeves:
I need a place to update the working wish list and place tracking docs, therefore I can either edit the original or maybe we can create/leverage a space for working specs. The Technical Reference is not designed for discussion, and the Online Market is not where we want to be either. What about an ES Projects sub-forum? Thoughts?

~KF
* My 2WD Garden Wall
* Kinaye MotorSports
* Primary ride: 2WD Disc 9C 2806-equiv / Dual Lyen 12FET / 20S7P LiPo.
* Epics: Going to California: 2011 8)
* 50-mph, 101, 10k-Club. 12,527 miles-to-date, 7037 as 2WD.

It is by caffeine alone I set my mind in motion.
It is by the beans of Java that thoughts acquire speed.
The hands acquire shakes, the shakes become a warning.
It is by caffeine alone I set my mind in motion.

jsplifer
1 W
1 W
Posts: 55
Joined: May 21, 2009 9:16 am

Re: Building the Best Controller

Post by jsplifer » Aug 17, 2010 4:09 pm

Ok, hmm I suggest asking the people with the other brushless controller site to allow us some space. This could have the benefit of attracting some of their talent as well. On phone so please see link in my above post, will edit in later.

Perhaps it is possible to protect this from copy with some kind of open source software license? GPL etc.

+1 on nailing down the wish list first.


Please see the link I posted earlier.

User avatar
amberwolf
100 GW
100 GW
Posts: 23954
Joined: Aug 17, 2009 6:43 am
Location: Phoenix, AZ, USA, Earth, Sol, Local Bubble, Orion Arm, Milky Way, Local Group
Contact:

Re: Building the Best Controller

Post by amberwolf » Aug 17, 2010 8:45 pm

I think it's already in the right forum, E-Bike Technical, since there isn't a controller-specific subforum.

If you do update the original post, I'd really recommend keeping all of the original content intact, and just appending anything new to it, so the context of the thread is left intact.

Alternately, first "build" the controller concept within this thread, and only when the concept is "complete" in at least it's first essence would the first post then be appended to with the "full specifications". At that point a new thread could be started for each module or portion of the controller, to separately develop the ideas necessary for each, with their more detailed specifications. Links to those threads would be appended to the first post of this thread.

User avatar
fechter
100 GW
100 GW
Posts: 11822
Joined: Dec 31, 2006 3:23 pm
Location: California Bay Area, USA

Re: Answers to Questions

Post by fechter » Aug 17, 2010 9:58 pm

Kingfish wrote:
Rolling up the sleeves:
I need a place to update the working wish list and place tracking docs, therefore I can either edit the original or maybe we can create/leverage a space for working specs. The Technical Reference is not designed for discussion, and the Online Market is not where we want to be either. What about an ES Projects sub-forum? Thoughts?
One way is to keep editing one of your posts, usually the first one in the thread so it can serve as a reference.

Otherwise a separate thread for that stuff, but it tends to get trashed up with comments. I could create something like the Technical Reference area with limited posting permissions.
"One test is worth a thousand opinions"

User avatar
Kingfish
10 GW
10 GW
Posts: 4275
Joined: Feb 03, 2010 11:23 am
Location: Redmond, WA-USA, Earth, Sol, Orion–Cygnus Arm, Milky Way. Age: > yesterday < tomorrow
Contact:

Re: Answers to Questions

Post by Kingfish » Aug 17, 2010 10:59 pm

fechter wrote:
Kingfish wrote:
Rolling up the sleeves:
I need a place to update the working wish list and place tracking docs, therefore I can either edit the original or maybe we can create/leverage a space for working specs. The Technical Reference is not designed for discussion, and the Online Market is not where we want to be either. What about an ES Projects sub-forum? Thoughts?
One way is to keep editing one of your posts, usually the first one in the thread so it can serve as a reference.

Otherwise a separate thread for that stuff, but it tends to get trashed up with comments. I could create something like the Technical Reference area with limited posting permissions.
Thanks AW/Richard. Allow me to parrot back:

If I understand correctly,
  • Go to the Technical Reference Area and create the specs and docs for the “Project”.
  • These docs would be kept current, with revision history, and have limited-access.
  • Continue on then with this thread, and other threads as required to complete the design and implementation.
Does this work for you? :?: :)

I can still foresee a need for “Source Depot” as we say in Micro-speak; a place to put source code. We’re not there yet so perhaps in time it will sort itself out.

So much to do, so little time: More docs, diagrams, and decision trees coming up. 8)
~KF
* My 2WD Garden Wall
* Kinaye MotorSports
* Primary ride: 2WD Disc 9C 2806-equiv / Dual Lyen 12FET / 20S7P LiPo.
* Epics: Going to California: 2011 8)
* 50-mph, 101, 10k-Club. 12,527 miles-to-date, 7037 as 2WD.

It is by caffeine alone I set my mind in motion.
It is by the beans of Java that thoughts acquire speed.
The hands acquire shakes, the shakes become a warning.
It is by caffeine alone I set my mind in motion.

whatever
100 kW
100 kW
Posts: 1134
Joined: Jun 03, 2010 2:16 am

Re: Building the Best Controller

Post by whatever » Aug 18, 2010 5:39 am

you could go through the list and find references in es forum to 95% of the features required or how to do the functions required. Its possible to achieve what you want, but the time and expense involved will be way beyond what you imagine it will be. Modifications of xie cheng controllers ( not much point calling them infineon controllers any more as they haven't used infineon chip for a long time) is the way to go in my opinion.
I find most valuable very detailed explanations of modifications of the xie controllers as way of example knuckles work.
You might add to the list : capacitor placement/size/type on the high amp section for protecting the fets.

User avatar
Thud
100 MW
100 MW
Posts: 2809
Joined: Aug 20, 2009 6:20 am
Location: West Michigan,USA

Re: Building the Best Controller

Post by Thud » Aug 18, 2010 5:44 am

In the name of progress & to inspire some enthusiasm for the project, I thinks it is important to get a "win" under your belt very quickly. Just a small cornerstone set to build on. Its a huge phsycological boost for the comunity as a whole.

That said, the focus should be on the primary objective. Its a motor controller before anything else, so Lets get a bomb proof controller working.Its fine to plan for all the diagnostics & periphials, but a working, scalable, open source controller alone is a pretty big deal (and worthy) all by itself.
(said the guy who blows up everything :lol: ) Any thoughts on setting a goal for getting the parameters list set for v1.0

In closing,
I think Methods example of thread managment worked very well. OP, then reserve the next 4 or 5 posts in line for adding tech information.links to downloads, revisions & fixes as they become aparent.
if & when it all comes together the mods can sticky it for eternity.
Thanks for listening, T
get some......

All information & advice provided by Thud are "Open Source" & free for personal use & distribution under the following agreement linked below.
http://creativecommons.org/licenses/by-nc-nd/3.0/

User avatar
drewjet
10 kW
10 kW
Posts: 920
Joined: Jan 23, 2008 4:34 am
Location: Orlando, FL USA

Re: Answers to Questions

Post by drewjet » Aug 18, 2010 6:02 am

Kingfish wrote: [*]The problem is – I don’t like Budweiser. I like thick chewy brews that put hair on my chest and muscle on my bones – you know, what the ATF doesn’t want you to have: That which is fortifying, chock-full of healthy goodness, pacifying, and able to stave off the majority of what ales ye. I can make beer that is better than what you can buy in a store, better than what is made by most breweries, and second only to the guy that has the brewpub that is just a tad more twisted in his enthusiasm to push out a fine product to his local patrons. A commercial brewery can’t compete at that level by virtue of the inherited bureaucracy, the cost of boutique marketing, and the thin profitability. [/list]
DYI flattens the playing field, takes the wind out of the competitive sails, and sets us distinctly forward on the bleeding-edge of affordability because we inherit all the features of an upstart rebel start-up: Ambitious, light, fast, self-invested (as in no VC), plumbing the best with liquidity.
Forget the controller, I want some of your beer!!!

Seriously, this is a great project and absolutely needs oversight by one individual who can take the many talents from here and divvy up the work and pull off a success. Kingfish you are the man, thank you for stepping up to the plate. I am with Thud. Make a great bullet-proof controller first, keeping in mind all the extras you want to add. Make the Zilla of the brushless world!

User avatar
rhitee05
10 kW
10 kW
Posts: 842
Joined: Apr 07, 2009 9:08 pm
Location: Upstate SC

Re: Building the Best Controller

Post by rhitee05 » Aug 18, 2010 12:20 pm

The good thing is that most of the extra features are just a matter of software. As long as they hardware has the necessary bells and whistles, they can be added later. You can test the hardware first with a stripped-down bare-minimum feature set and then start adding the nifty bits.

Success in a project like this has as much to do with a good design process as it does with the technical aspect. Even more so if it's a collaborative effort. If you plan things out well and work your way through a logical process, it becomes easy to divide up the problem and put the pieces back together later. Let me suggest a process flow for the project based on my experience. It's very useful to take a top down approach which starts at the very high level and gradually works down to the detailed level. If you start thinking about resistors and ICs too soon, it's easy to get confused.

The first step is brainstorming, which is basically where we're at now. Everybody tosses out their crazy ideas and we see what sticks to the wall. The end product is a wish list of functionality, possibly sorted out into "must have", "like to have", "nice is possible", etc. Note that we're not talking about the hardware necessary to implement anything yet, that's comes next.

The second step is to figure out at a high level what hardware is necessary to implement the desired features. We don't care about the details yet, just blocks of functionality (current sensor, gate driver, throttle input, etc) and how they're connected. You also want to decide how the hardware will be physically grouped at this stage. I'd suggest making 2 or even 3 separate boards. One board would be the power electronics (gate drivers, FETs, capacitors,etc) and the second would be the digital brains. You could make it even more modular by separating out the supporting electronics (voltage regulators, analog signal conditioning, interface) into a 3rd board. Keeping those components on a separate board would make it very easy to mix-and-match different power and logic boards.

Once you have a block diagram and know which components will live on which board, you need to define interfaces between the boards. This is VERY important, especially if you want the ability to swap different board boards for different needs, etc. However, it's very easy to do once you have the block diagram set up, because that will tell you what signals need to cross. You do have to take the time to define the electrical characteristics of the signal, though.

At this point, it should be fairly easy to divide up the design into sections that individual people can work on. It doesn't really work to design a circuit collaboratively, but you can divide up discrete chunks for each person. One obvious division would be by each circuit board, but you could split it more finely too. Having the functional blocks and interfaces defined makes it easy to break it apart.

Now everybody does their designs, collaborates, etc. You probably want to have a "lead" person for each board. Eventually the design for each board is complete and you can do the PCB layout and get some prototypes. Each board can be tested separately, then put the whole stack together for testing.

Also, the software side can be developed in parallel with the hardware. Once the block diagram is complete and the interfaces are designed, you know everything necessary to write the software. I'm not a software developer, but ideally you'd want to follow a similar development path. Eventually the software and hardware come together and tada, you have a functioning prototype. Ideally each portion of the design is documented during the process, so you end up with a product that people can easily mod for their particular needs.
Eric

Trek FX 7.3 hybrid - Cyclone 650W setup in process! Still...

Post Reply