What electronics do we need?

[Tiberius]

I don't mean what are the minimum electronics needed to make an e-bike. What I mean is "what electronics device/box/gadget do we need to make EVs work better?" What are we missing? What should the development engineers be working on?

Nick

[Miles]

Torque sensors, process and control, IMO.

[torker]

Maybe dc/dc converter built into the controller.

[amberwolf]

Process and control, yes. Motor controllers that don't just convert battery power into motor power, but also monitor the way that power is used and supplied, temperatures, voltage and current states, integrate this data over time and log it, and intelligently convert user's speed demands into system control.

Integral safety cutoffs (with options to defeat them) such as a tilt sensor that detects if the vehicle is no longer on-road (like a fallen bike) and brakes the wheel(s).

Not sure an integral DC-DC is a good idea because there are so many possible output voltages one might want. I suppose it could be done so that it has several outputs, but if the majority of users only use part of that it's a waste, and if they need more power than it provides they'd have to replace it anyway. I guess if one had to put one in, 12V automotive-level would be good, at 20-30A capability.

[mwkeefer]

Can we please have reliable and small high current and voltage controllers which work first?

-Mike

PS:
Then we need more tuneable or adaptable electornic control systems(variable throttle ramping, true current limiting, etc), more intelligent safety features and seriously better battery management and charging solutions for the uber high voltage systems (83v+)

[amberwolf]

Realistically, I think that some of the intelligence of the controller is necessary to make them more reliable.

True that using components *designed* for high currents would help, and improving board layouts and whatnot would help as much or more, but if the controller could detect and mitigate circumstances that would otherwise blow them up, they would be more reliable even if built out of TO92 transistors hand-wired on perfboard.

[Tiberius]

OK, a number of requests for better controllers, and that shouldn't really surprise anyone.
It seems to me there are two aspects of controllers that people want improving
- the power handling and robustness
- the functionality

Maybe we need a modular approach, where the power stage and the control board are separated and can be developed separately.

Nick

[Tiberius]

Maybe dc/dc converter built into the controller.

Dave,

Are you thinking of a low power, step down DC-DC to generate a supply rail for other stuff, like a Battery Eliminator Circuit in the RC world? Or are you thinking of a high power step up one, so the controller/motor can work at higher voltages than the battery?

Nick

[Lessss]

Controllers that match the efficiency of the motors, so your not wasting juice.

[Evoforce]

Better cooling for both the motor and the controller. And I know this is not electronics but a hub motor without spokes. You know one that bolts up to rim tabs.

[torker]

Maybe dc/dc converter built into the controller.

Dave,

Are you thinking of a low power, step down DC-DC to generate a supply rail for other stuff, like a Battery Eliminator Circuit in the RC world? Or are you thinking of a high power step up one, so the controller/motor can work at higher voltages than the battery?

Nick

I was thinking something to run 6 or 12 v loads such as lights but maybe it is better to have this separate.
A controller with a programming dongle built in might be nice though.

[Lock]

Bluetooth and WiMax interface for controller/bms/security/charger systems
tks
Lock

[johnrobholmes]

BMS chips that plug into balance taps with parallel-able outputs that can trip a 5v cutoff! Like a bm6, but with the cutoff outputs isolated.

Or maybe a chip that can take bm6 outputs and opto isolate them.

[Tiberius]

BMS chips that plug into balance taps with parallel-able outputs that can trip a 5v cutoff! Like a bm6, but with the cutoff outputs isolated.

Or maybe a chip that can take bm6 outputs and opto isolate them.

John,

Afraid I don't know what a bm6 is. Will the sort of thing in this thread http://www.endless-sphere.com/forums/viewtopic.php?f=14&t=18834 do what you ask?

Nick

[Tiberius]

Bluetooth and WiMax interface for controller/bms/security/charger systems
tks
Lock

Lock,

I can see possibilities for wireless links and control. But Bluetooth and WiMax are geared up for transferring large amounts of data over open standards. So for simple control and data they have huge software overheads, and they take time to set up the link. On the other hand, they are off the shelf solutions.

Bluetooth to your mobile phone? Bike is only in high power mode when its in range of your phone?

Nick

[Lock]

My "narrow" interest is leasing fleets of EVs as low speed urban transport. Think public bike share program on steroids. Wireless connectivity for "inventory control"/fleet management/customer convenience/service and maintenance.

Zip Cars illustrates some of this:


tks
Lock

[johnrobholmes]

Yep, that is what I want. The bm6 is just a battery monitor with settable low voltage alarms, display, and a buzzer http://www.hobbycity.com/hobbycity/stor ... oduct=8927

[liveforphysics]

A controller brain.

Something with an isolated and well stiffened switching supply to power itself (not more stupid voltage divider resistor circuits), and a separate isolated switching 5amp 8-10vdc supply to provide clean non-bouncy power to the gate drivers on the outboard external FET stage.

Something using a modern motor controller chip would be nice as well. No reason the same brain can't do brushed, brushless, sensored (encoder wheel capable as well), unsensored, and/or use sensors for starting then switching to unsensored for active timing control dependent on throttle position etc.

This would be the greatest boon that could befall DIY EV projects. I would also gladly help to fund creation of such a controller brain.

[spinningmagnets]

Integral temp probes in the motor that tell the controller that the motor nearing a dangerous heat, and the current is limited (not cut off suddenly) so that the motor can cool down some.

Temp probes in the controller that start a fan to aid cooling, two fans if heat rises more, and current limiting when it nears the danger zone of heat.

In spite of having color-coding and polarized connectors, still include diodes so that if its wired wrong, the controller or battery pack won't create plasma and magic-smoke.

Silicone high-temp wire insulation (its only a couple bucks more for each controller/motor). I'm a caveman when it comes to electronics, but these relate to many things I read on a regular basis.

[jag]

Can we please have reliable and small high current and voltage controllers which work first?

I'll second this. A 100A+ controller should weigh less than 500g, be simple and robust. I want a 100A in, 200 phase amp pedal first controller for my Astro RC build. Need 150,000+ electrical rpm to run Astro at 10-20,000 rpm. Can this be so hard to make??

And add:
Small lightweight on-board high power charging systems. I'd like to charge my A123's in < 1h, or as fast as possible on the road when I stop at a store or whatnot.

In terms of gadgets, I'd say thankfully eBikes are still minimalistic: Only an on-off switch, throttle and charger for the user to worry about. Audio equipment could have been as simple (basically need a volume control and a means to select the song/change disc), but in the past 30 years manufacturers have filled front panels and remote controls with ridiculous amounts unnecessary features.

[mwkeefer]

A controller brain.

Something with an isolated and well stiffened switching supply to power itself (not more stupid voltage divider resistor circuits), and a separate isolated switching 5amp 8-10vdc supply to provide clean non-bouncy power to the gate drivers on the outboard external FET stage.

Something using a modern motor controller chip would be nice as well. No reason the same brain can't do brushed, brushless, sensored (encoder wheel capable as well), unsensored, and/or use sensors for starting then switching to unsensored for active timing control dependent on throttle position etc.

This would be the greatest boon that could befall DIY EV projects. I would also gladly help to fund creation of such a controller brain.

I know I have been working on somthing similar (and many others too), personally I'm scavanging Infineons as the power portion and running them from various MCUs specific to the task which I have ordered and purchased over the past many months...

The issue with using the "pre-existing" or off the shelf motor controllers MCUs is they just aren't that great, even if EV was their intent... the waveform is distorted at best and they all have some quirk or other (seems like MS Windows XP) atleast with their provided reference designs used for our purposes (wide RPM range from 200 - 8 or 10K).

So far I've sampled about 5 different BLDC controller MCUs specifically designed for the "EV" and "LEV" markets and to be brutally honest, none of them are what we want as the heart of an "EV Brain" (imho).

I think if we want an intelligent and modular controller, we are kinda on our own... I've been playing with Atmega1280s (and the new one) which are nice for communicating with peripheral systems but just not fast enough to run some of the RC type and Geared high RPM motors in sensored or sensorless mode...

I came across a section of one of Matts earlier posts on his first recumbent project the other day, he pointed out that the controller he was using was a constant RPM type... so as he hit the hills it would ramp up current output for him to compensate and prevent loss of speed...

A good brain should implement this and also "traditional" style throttle via programmable curves or ramps with variable points...

In either case... we need good controllers, I am reasonable about size constraints and if we have optically isolated "Brains" - who cares about a dozens FETs on a PCB you can replace? (ohh... frame mounted FETs for heatsink:))

-Mike

[liveforphysics]

Mike- It's funny that we are both not-content with controllers, but doing opposite things to fix it. You're playing with controlling the infinion power stage from a different MCU. I find the infionion power stage to be a pitiful creation that gave up the ability to perform in favor of being as cheap as possible to mfg, but I use it's MCU as the brain to drive my own no compromise power stages. Just funny to see we are both working towards the same thing (a better controller), but flip-flopped in the way we are going towards it.

I should mail you a stack of the power stages from infionion boards, and you can mail me a stack of the MCU side of the boards.

[jag]

Mike- It's funny that we are both not-content with controllers, but doing opposite things to fix it. You're playing with controlling the infinion power stage from a different MCU. I find the infionion power stage to be a pitiful creation that gave up the ability to perform in favor of being as cheap as possible to mfg, but I use it's MCU as the brain to drive my own no compromise power stages. Just funny to see we are both working towards the same thing (a better controller), but flip-flopped in the way we are going towards it.

I should mail you a stack of the power stages from infionion boards, and you can mail me a stack of the MCU side of the boards.

Separating the signal and power stages in a motor controller is a good idea. Here is an example. Power stage is to the left and signal card on the right side. It normally plugs into the header seen at the front of the power board. Some interesting features: Control is two stage. Instead of a simple PWM it uses a pseudo-sine coded into an EPROM. First the DC is regulated down to the desired RMS voltage (the leftmost FET stage on the board and the big on-board inductor to the left), then the psudo sine 3-phase is generated by the FET stages to the right. This suppresses higher order harmonics which generates less RFI and can also increase system efficiency (The PWM modulation happens internally in the first stage and no HF harmonics to go into the motor and heat up the core laminations).

DSCF6815.JPG (112.08 KiB) Viewed 512 times

Another unusual feature is that the FET and driver circuitry is isolated by a small transformer. The transformer (unlike and opto-coupler) passes both the pseudo-sine modulated signal and the power to drive the FET. It also gives very good electrical isolation. Isolation is a good thing with the high voltage motors used in process industry.

DSCF6812.JPG

(109.15 KiB) Downloaded 96 times

I originally built the controller while a high school summer intern at Kamyr. It was was intended to drive a mixing paddle in a Kamyr continuous cooker, an invention by one of my parents' neighbours. The cooker is the 120m tall tower to the left in the picture.

The boss of the automatic control department at Kamyr took us on a field trip and all the summer interns got to climb to the top of the tall cooker:

DSCF6817.JPG

(111.51 KiB) Downloaded 96 times

I took the controller out of the basement because I was curious to try if it would be more efficient than the standard single stage ones we use. Unfortunately I didn't find any power FET's. I may have lost them or just blown them all. The schematic also seems lost. The FET driver circuit is just ripped from a power electronics journal article by some Chalmers dudes though, so it should be findable. The rest one would probably do differently with a modern MCU these days.

[mwkeefer]

Can't wait till they start spamming us with Come to Jesus booklets

Luke,

I use the infineon power board to develop the brain, normally starting with blow controllers and replacing / upping the FETs to somthing useable but yea... I'm going at it from the other side, not for the idea of production alone but simply as the most rapid manor I can acheive a better controller..

Besides ... if your working out a power section and I'm working out a brain, we will likely meet in the middle... agree on a standard of interface for the gate drive and whalla...

BUT - It must have a 5v 500mah USB power port, that is one damn feature I am so sick of eBikes not including... what am I supposed to do install an RC power system and then one of those gen front hubs to power / charge my iPhone? - although this seems silly and trivial, I'm actually not kidding and with packs of the average eBike size and a need to do a solid 5v rail in the controller brain... it is one feature I am not going to leave out

More seriously... one goal of my "tinkering" (the best way to describe how I work) is to mimic those Quadracopter or Mirocopter controllers - 2x BLDC controllers in sync.... so basically I want to include either a sync bus for multiple controller brains or design from scratch to handle 2 seperate (and physically different) motors - I owe Jeremy Harris for this thought which was actually inspired by his latest project a diy self balancing scooter (requiring perfect sync and better control than PWM could deliver I believe)..

I am no where near that level yet, I am barely able to spin up a BLDC motor in sensorless mode (doing ok in sensor mode though) via raw code on an Atmega but I am still sampling EV controller chip offerings from many vendors looking for somthing with a better base than the old infineon - not just more modern but better commutation and design architecture...

Enough from me tonight.
-Mike

[Microbatman]

Smaller more reliable controllers with an easy to program Low voltage cuttoff.