GEVCU – Gesundheit and Geilheid

This week we have several balls in the air. On Monday after the show was posted, our crew spent an exhausting day squirreling away 100 Siemens motors we received by truck. All in spanking new looking factory boxes. They came in 8 large crates of about 2700 lbs each with a dozen motors per crate. It was a grueling day just opening crates and forklifting motors – each about 200 lbs into position inside.

In the show this week, I introduce a little side project to develop a battery module for the 2014 Club Car Precedent. Recall that I actually put together 17 cells to try to match the 62 volts we observed in charging the six 8-volt Crown lead-acid wet cells. As it turns out, the 2014 Club Car model was probably the worst golf car/cart selection we could have made.

[jwplayer streamer=”rtmp://s3einxnpkaij93.cloudfront.net/cfx/st/” provider=”rtmp” file=”news092813-iPhone.m4v” hd.file=”news092813-1280.mov” image=”http://media3.ev-tv.me/news092813.jpg”  width=”950″ height=”584″ html5_file=”http://media3.ev-tv.me/news092813-1280.mov”]

The reason goes to why they use lead acid at all. The reason is the expense. For you lead heads who clung to the lead acid cells for EVs because they were inexpensive, I don’t think you’ll ever know HOW inexpensive they can be. The $200 per battery you were paying was a horrendous screw job.

We found a golf course lease online largely because it was a public municipal golf course in Florida. It was for 160 golf carts from the admittedly number three in the industry, Yamaha. For a four year lease, which included maintenance of the carts AND the batteries, which were guaranteed for 1000 rounds, this city was paying Yamaha precisely $63 PER MONTH lease rate per cart. That’s cart, batteries included. AGMS, gel cells, VRLA have made essentially NO inroads on golf courses because of the extremely low cost of the wet lead acid cells. They typically go for 6% above the spot market cost of LEAD. In all cases it costs more to ship them than to buy them.

In our case, the 2014 model, for the very first year, uses a brand new DeltaQ IC650 battery charger. Of course, Club Car simply branded it as their own battery charger. Had I known it was a DeltaQ product, I would never have bought this golf cart at all. My distaste for DeltaQ goes back years and knows no bounds. They are the most whiney, CAN”T DO company in North America (Canadian) and are apparently in a state of constant terror over possible lawsuits.

My experience with them goes back to an exchange on the EverlDull (EVDL) discussion list. One of the project engineers for DeltaQ was whining verbously about the CHinese ripping off their charger design and he was in full North American nationalism noting all the thousands of “American” jobs this cost as these guys were underpricing him with his own design.

I felt obligated to point out that if they had indeed ripped off DeltaQ, there was some good news and bad news. First, the TCCH 3kW charger he was referring to did 144 volts at the time, while his product line topped out at 96 volts and the TCCH charger did higher currents as well. Second, DeltaQ needed to brush up their user manual because the TCCH version sure was a comical version in Chinglish. And finally, WHAT American jobs, DeltaQ was having their chargers manufactured in CHINA – which is probably how they got ripped off in the first place. I suspect the only thing copied was the yellow and black color scheme.

We had of course wanted to make the lithium battery module work with the stock manufacturers charger that comes with. I knew there was a wee bit of a potential problem, but I thought it would work itself out. It revolves around the concept of “equalization charge” for lead acid cells and dV/dT.

Lead acid cells have an internal shuttle mechanism that allows a kind of constant self discharge. Brand new lead acid cells will simply bleed down in charge with no connection at all at a rate of about 4% per week – more on older worn cells. But they do this at varying rates. After just a couple of charge discharge cycles, they are all “out of balance” at different states of charge.

The cure is to “equalize” them by overcharging. This is actually the historical stuff of which lithium top balancing mythology derives. The open circuit voltage of a Pb cell is 2.12volts. You can quite ably charge these cells at 2.25v. But if they reach 2.4 volts, they cease to charge, and the charge energy is rather used to hydrolize hydrogen and oxygen out of the water in the electrolyte. And so by overcharging them to about 2.5v per cell, you can cause ALL the cells to achieve hydrolysis. At that point, they are all fully charged and all at the same state of charge (SOC).

Lithium cells can’t be overcharged in this way, but the thinking persists and so the many idiotic efforts at devising a scheme to “top balance” lithium cells without overcharging them.

ALl that said, the actual voltage at which lead acid cells go into hydrolysis actual varies with both temperature and the age of the battery. But what does not change is that once they hit hydrolysis, the voltage ceases to climb. And so the change (delta) in voltage (v) over the change (delta) in time (T) goes to zero. dV/dT.

SOOOOOooooo… if you could simply detect that the battery pack voltage is not going up any more, you could use that to terminate the charge and avoid continously overcharging them and excessive hydrolysis, which lowers the water level and just wears the cell out. So when the voltage of the pack ceases to increase in the face of current applied, stop charging.

My fear of course, is that lithium batteries at the end of charge do not CEASE changing voltage, in fact as it climbs the steep part of the charge curve the dV/dT accelerates heroically. But my bet was that they had a terminate anyway at about 61 volts. SO by adding a 17th cell, we would be in like flynn.

Enter the DeltaQ IC650. What I didn’t think of is that as lithium cells charge from low levels, the change in voltage decreases and once we are up on the flat part of the curve charging, the voltage remains quite stable for most of the charge – increasing from 3.1 to 3.4 over the course of several HOURS. the DeltaQ IC650 charge algorithm reads that as zero dV/dT and shuts down, terminating the charge. So I would plug in the charger, it would charge about 9Ah and terminate. If I inplugged the cord from the cart and plugged it back in, I get ANOTHER 9 Ah and terminates.

Of course, your average golfer is going to plug in the charger and come back the next morning ASSUMING the pack is fully charged and off he goes to his round with about 9Ah onboard instead of the 60Ah we told him. Bad news. Bottom balanced, we’ll survive it. But I don’t think the golfer will really be impressed with the five holes he gets on a charge.

DeltaQ makes HUGE play out of how they have a USB port on the device and 25 charge curves INCLUDING LITHIUM available. Better, they have done a kind of cunning thing. You don’t need a laptop at all do change charge curves. Just plug in a thumbdrive with the new algorithm on it and the charger will automatically load it and overwrite the existing one. So with NO special equipment, you can change the algorithm in the field just by plugging in a cheap thumb drive.

Great idea. Why they went to all that trouble I have no idea because they don’t let anyone change them “in the field”. I suppose if they HAD to promulgate a firmware update, this would work. But these control freak whackos are terrified of anyone actually using this feature. A rep at DeltaQ claims they discontinued support for lithium because this charger has no temperature sensor and without temperature there is no way they could possibly terminate at teh proper voltage. He might watch an episode of EVTV now and again. There is no federal LAW requiring you to overcharge lithium cells. In fact, you can just undercharge them. In any event, it’s my cart, my charger, and my cells. They don’t care. Can’t take the liability.

We also spoke with a couple of their authorized dealer techs. They are terrified of DeltaQ who constantly berate them about what they can do and can’t do and if caught how they’ll lose their status as distributor. These guys are the first example of CHARGER NAZIS I’ve ever heard of. Absolutely ridiculous. COmplete with some totally BIZARRE views on lithium cells and battery management systems. This company has subscribed to every battery superstition available on the planet for ALL the chemistries and combined it into one gigantic algorithm set pretty much guaranteed to screw up batteries. We now understand the ridiculously high 62 volts.

Unfortunatelyh, that kind of terminates my concept of making the module work with the stock charger. I really didn’t like the idea of having to change the algorithm anyway. The Club Car has a standard Club Car charge port. The charger isnt’ actually onboard. You hang it on a wall. And what if they plugged into somebody ELSE’s charger.

I’ll probably continue to play with this on the side. I’ve ordered the earlier P3 charger that Club Car previously used, and we’ll test that, but we’ll more than likely drop back to 16 cells and we’ll just have to supply a little fixed voltage TCCH 1500 watt charger with the battery upgrade. Not at all my first pick. Fortunately, they can use it as an onboard charger.

On the Siemens front, I do love it when a plan comes together and it appears it just might. In this episode, Michael Neuweiler has reached a signficant milestone in the development of the Generalized Electric Vehicle Control Unit or GEVCU. And subsequent to this episode, he’s reached another.

The concept from the beginning to put the basic “controls” brake and throttle, along with features, such as precharge, cooling, BMS interlocks, forward/reverse, etc in OUR open source software instead o using what you are given in the inverter software. Almost all inverter developers hanker after the OEM sale. Few actually achieve it. Even relatively SMALL OEM’s tend to do their own power switch electronics because it is not hard for them to develop from the power module providers reference designs. Case in point is Azure Dynamics, hardly an OEM by most standards but they had their own DMOC (Digital Motor Controller) which they tuned to the Siemens motor.

But in seeking the OEM model, almost all inverter developers have adopted the CAN control model where a Vehicle Control Unit provides the interface to the vehicle controls and instrumentation, and drives the inverter by CAN bus. And so the main function of GEVCU is to wrest control of controls and features away and put it in open source software where anyone can change it.

Part two is to develop a kind of complicated but rich configuration environment where YOU can change it within certain parameters without even having to code in C++.

But the third leg of this thing is to modularize the code so that different child objects can be developed in entirely sealed software units to interface to DIFFERENT inverters. Each inverter has its own peculiarities and CAN bus message digest. True, most will require SOME level of customization of the firmware in the inverter – in DMOC645 this is done by changing variables in ccShell over a serial port. But the actual driving is controlled by torque and speed commands TO the inverter and certain status messages are received FROM the inverter, usually providing pack voltage, pack current, motor voltage and current, inverter temperature and occasionally motor temperature, RPM, actual torque, etc. These are really pretty similar in nature from inverter to inverter, but the addresses and data formats can vary a lot.

We have GEVCU actually driving the VW THING test bed pretty well. I have braking regen and throttle regen and I can map the throttle and brake and tune it up where it feels pretty good. This week, Michael Neuweiler has developed a similar “motorcontroller” child object for the BRUSA inverter/motor he is using in his VOlvo 80 build. At least well enough to spin the motor on the bench which is the main hurdle. After that, it’s a matter of fine tuning and featuritus.

That kind of proves the model. We have no limit on the number or variety of CANbus controlled inverters we can add to the tree. We’ll have two examples that come with, and so almost anyone that wants to add a third or a fourth etc will be able to do so without a lot of wheel reinvention. Most of the inverter developers openly publish a document on their CANbus messages and data formats.

My idea so naturally I would think it brilliant. But I really do think it’s kind of huge. DIY community now has its OWN VCU and we cannot be held hostage for one. You can have any feature you want. And the switching electronics kind of becomes a rather boring commodity item.

I guess I should speak of motor/inverter power drive trains. The inverters have to have the code to optimize the driving of the motor. So the inverters typically have “configurations” for several motors. You can’t use a UQM inverter to drive a Siemens motor or a DMOC645 to drive a Rhemy motor. But we’ve broken off a large chunk of the utility and brought it into the GEVCU.

Mr. Neuweiler has since Friday announced a further development. Our GEVCU hardware uses a MiniSocket iWiFi module from Connect One. This is actualy a pretty mature module that has been out there since about 2003. It’s a tad expensive as these things go. But they’ve worked out a lot. It does WiFe 802.11 b/g and offeres a lot of TCPIP functions, telnet, FTP, SMTP etc. But for our purposes, it also has kind of a prepackaged web server. The interesting thing about that is you can include “tags” in the code that if the end users changes a variable in a form, it spawns what’s basically a hardware interupt and then allows you to retrieve just what changes. This is kind of a neat feature in that we don’t have to check up on teh web server a lot in code, slowing us down, and when something does change, we don’t have to scan an entire post XML form, but rather just retrieve the variables that changed.
With WiFi, you can connect of course via wireless which is now pretty much built into everything, including your phone.

We still have to design some pages. We intend three:

The first is a status screen. This will show things like motor torque, current, voltage, temperature, etc. You can access this with any browser as you see it. Of course, it is quite doable to make an Android or iPhone program that simply makes an ad-hoc connection and reads this XML page and uses the data to manipulate onscreen meters and dials, integrate Ampere hours and kilowatt hours, etc.

The seconds screen will be a configuration screen and it will be a fight to keep this both brief and comprehensible. But our intent is again to allow most users to make a lot of configuration changes from any wireless browser, instead of the USB serial port we’re currently using. So you won’t be restricted to a particular operating system, a particular computer, or a computer at all. Again, most smartphones and tablets can do ad-hoc WiFi now from an onboard browser. We don’t have to write and update a configuration “program” that you then have to install and of course have the proper operating system on your laptop for. Any browser device that can do wifi.

As we have an onboard EEPROM much larger than we need, I can see adding logging to this EEPROM that can be downloaded by browser eventually as well. A data format you can import into Excel and graph your drives to see what kind of energy you were using and when.

The third screen will simply be an about screen with links where you can get software updates and so forth. Something like the following:

ABOUT GEVCU

Version 1.00001

The Generalized Electric Vehicle Control Unit – GEVCU is an open source software project to develop a Vehicle Control Unit (VCU) specifically for electric vehicles. Following the OEM automotive model, standardized components are used in multiple cars, with a single component (VCU) containing software to integrate operation of those components specifically for that make and model vehicle.

The purpose of GEVCU is to handle throttle control, regenerative braking, and such peculiarities as precharge, cooling system control, instrumentation etc. – essentially the driver interface of the car. GEVCU then manages torque and speed commands to the power electronics via CAN bus to actually operate the vehicle in response to driver input. But it also provides outputs to instrumentation, cooling systems, and other controls specific to that vehicle.

GEVCU is both open-source and object-oriented allowing easy addition of modules to incorporate multiple standard power components. It is easily modified to incorporate additional features peculiar to any electric car conversion or build. For most operations, it is easily configurable by non-programmers to adapt to a wide variety of power components and vehicle applications.

GEVCU is designed to operate on any Arduino Due hardware with two CANbus channels and appropriate analog and digital inputs and outputs.

GEVCU was originally conceived of and proposed by Jack Rickard of Electric Vehicle Television (http://www.EVtv.me) who wrote the original design specification. The main source of the program was developed and is maintained by Collin Kidder and the latest version is always available at http://github.com/collin80/GEVCU

Major contributors to the original project include:

Michael Neuweiler
Charles Galpin
Ed Clausen
Mark Weisheimer
Paulo Almeida

Collectively, this group operates loosely under the name EVACI – Electric Vehicle Ass Clowns International – a global team of software and hardware developers dedicated primarily to not electrocuting themselves or burning any buildings to the ground.

GEVCU is open source and forever classified EXPERIMENTAL – USE AT YOUR OWN RISK. It is offered strictly for experimental and educational purposes and is NOT intended for actual use in any commercial product for any specific useful purpose.

 

You can add your name to this list pretty easily. Check out the software from the github address above and jump in. The obvious need is for object modules for UQM, Rinehart, Siemens, and other inverter manufacturers. But the number of other useful features that can be added is nearly endless. There are really only 10 kinds of people in this world: those who understand binary and those who don’t. If you do, and have a little C++ ecperience, join us.

And it is a truly International team of Ass Clowns. Paulo Almeida of the University of Lisbon in Portugal has done something marvelous to our hardware in just a scant few weeks. He’s shrunk it. There is no Arduino Due separate board. He put the input protection, the output MOSFETs, the isolation, and even a port for the MiniSocket iWiFi on a single board that also features its own 3.3v converter and the AMPSEAL 35-pin connector. The ATMEL SAM chip is onboard. He did lose the programming port but I’m really enthused that he replaced that crappy little micro-B connector with a USB – B printer port connector – the strongest of the lot I think. I break off the little micro-Bs very easily and do not think they are sufficiently robust for an automotive environment.

We should be receiving our first boards this week hoepfully in time for Friday’s show.

I have to say we are still seeing a lot of change in both hardware and software. That makes it very difficult to go have a hundred of them run off so we can make them available to our existing DMOC645 and Siemens motor users. But I’m confident we’ll get there – maybe in a month. These hardware/software design projects always go faster than you think they will….don’t they?

Jack Rickard