Egngine Control for Battery Maintainence

[RogerAS]

Hey All,

I am about to embark on a home built system to control several factors in maintaining the battery bank, and I am seeking input. I will outline my ideas and plans and then wait for the AAA (Anti Aircraft Artillery).:-0  Actually I am very actively seeking input as I do not consider myself an electronics expert by any means!

I was thinking the center, or brain, of the system would be an LM3915 with a NE555 to allow for short term loads and avoid unrequired charge cycles. This would be setup in the dot mode. Nearly all actions and reactions will be based on results from this IC combo.

The genset:

Since I have a home assembled diesel/alternator genset the focus will be around that. When battery voltage drops to 11.5 I want the engine to start, warm up and then initiate the charging. When batteries get to 14.5V I want it so shut down. I need to allow for glowplug operation at lower ambient temperatures, but that can be full time summer or winter with a simple cutout switch during warmer months. To implement this I was thinking of a series of NE555's driving FETs and relays, in association with LM324 quad op amps acting as various sensor controls.

The low voltage signal from the master LM3915 could activate the relay H bridge to set throttle to a midpoint, determined by a cam/switchs on a geared dc motor disc that is like a small wench attached to the throttle lever with a spring to hold tension against the throttle drive.

At that point the system will activate the glow plug for 5-10 seconds, then activate starter until run condition is detected via a cooling fan pressure switch. If start fails after 10 seconds stop everything and wait 10 more seconds and then repeat the process for 3 tries. If all three tries fail an abort start signal is generated, and sound a buzzer, and or shut off the power to the entire monitoring system. That's the point that I'm having the most problem understanding.

How do I count those start attempts? Would latching relays, in series, one activated on each attempt and feeding the next in line, work? If so when the next master charging cycle is attempted can a signal from the throttle H bridge, at the midpoint, reset all of these? Could I tap into the running condition sensor to reset them? I've never worked with latching relays, so I really don't understand their operational needs. If these will work as I imagine can the master LM3915 send a reset signal to all of them at the next charge session before throttle reaches mid point?

To control the glow plug/start operation:

When the throttle midpoint is reached a master NE555(#1) set as a mutivibrator is powered up. This will drive the glow plug slave NE555(#2) set as a one shot or monostable mulivibrator whos output will drive a FET and a relay. Also driven by the master NE555(#1) will be the slave start NE555(#3), also a oneshot, with the same FET/relay setup, timed to delay until the glow plug NE555(#2) has been activated for 10 seconds. If the cooling fan reaches a pressure/speed high enough to activate a cutoff switch the master NE555(#1) is powered down, which should also cut the power to all the slave NE555's(#2 & #3).

To control warmup sensing/initiate charging:

My thinking is to tap into the water temperature sender for this sensor. If an LM324 quad op amp is used the variable resistor within the water temp sender could control one output. When the preset is reached a signal would be sent to the throttle H bridge to open to full power setting for the engine. If an LM324 quad op amp is used an over temp signal could also active the throttle H bridge to shut down the engine. There must be safegaurds! When thottle is opened to full power the charging system can then be powered up.  Since I'm using a simple rehostat on an automotive alternator to control alternator output the throttle H bridge position sensor, LM324 output can activate another FET/relay.

Safegaurds:

I've already addressed the over temp  issue, if that happens throttle is returned to full stop which will automatically disengage the charge. The oil pressure sender will have another leg of the LM324 quad op amp to monitor lubication. In the event oil pressure drops, or goes too high, a signal to the throttle H bridge will go to full stop. Any time the throttle H bridge is in any position besides full open the chargng system is deactivated. Also there will need to be a fuel level sensor. To run the engine dry isn't really a problem, but to reprime it after such an event is a hassel. So I was thinking a simple float level/rheostat built for this driving a leg of the LM324 quad op amp would work there.

So one leg of the LM324(#1a) quad op amp will detect warm running temperature,  one leg of the LM324(#1b) will detect over temperature, another leg of the LM324(#1c) to sense low oil pressure, another leg of the LM324(#1d) for high oil pressure (in case the bypass valve fails). Another LM324(#2a) quad op amp for low fuel level control, LM324(#2b) for throttle stop position, LM324(#2c) for throttle midpoint, LM324(#2d) for throttle full open.

Battery voltage:

The master LM3915 will iniate and stop all activity depending on the battery voltage. Homebrew optocouplers driving an LM324(#3a) quad op amp which will then drive FETs/relay(s) for master control. The remaining outputs of LM324(#3b,c & d) might be able to reset the latching relays on the start attempt count. A feedback from any of the safegaurds will de-energize the master M3915 and send the throttle position to full stop.

Please feel free to comment. I'd like to NOT buy piles of IC's and relays because I've overlooked something, which I'm almost sure I have. I think I've got a handle on most things, but again the start attempt count is something I'm very unsure about.

Thanks in advance,

[Nando]

THAT IS A MASSIVE CIRCUITRY I may say.

Many man hours putting it together plus the testing, it looks like several hundred hours.

Wiser would be to get one or two of those Stamps controllers to do the great majority of the work, the design will take less even if one needs to learn the programming of it.

You may have a clear idea but I see a lot of hills to overcome that do require time to solve and may it to operate properly.

Take your time and see if the stamps controller would be, long term wise, better to work with it to solve many additional problems and easy to modify with required parameter changes that will occur many times until You have a good handle of the problems.

Nando

[jimjjnn]

I feel that stamps need a full explanation on this Forum as many here do not understand pics and especially how to program them. Could somebody that is fully knowledgeable put a lesson here to show all how easy it supposedly is?

[RogerAS]

Jim & Nando,

I don't know much about stamps but I can say this. I run a Macintosh computer exclusively and I've yet to read about programing a stamp with a Mac. I used to play with a Radio Shack Color Computer III that I could program with basic, a little, but it's long gone. Yes I know this controller would be a bear to build, but I gotta automate it somehow. So let's have some info and direct suggestions!!

[redeyecow]

  Although your system will be electronically controlled it might be interesting for you to take a look at a fully automatic powerplant from the early days. There is a book by Mike Hackleman that has a company diagram of a machine that is fully automatic using old style mechanics and relays although it is a gasoline engine. I think its a cadillac version of the Delco Plant. Not sure if its in  Wind and Windspinners" or " A Handbook Of Home Generated Electricity". Sadly lent both these books with no return.

  Some old farmers around here used to put a power sensitive throttle on their welders by connecting the throttle to the spring loaded shaft of an old starter motor and running the output through the starter windings. This would open it up for welding as well as improving the arc. I like the load sensitive feature. Your project  sounds pretty interesting.  Keith  

[pepa]

Roger, this controller would be valuable to a lot of us, please keep us posted and make sure that you can duplicate the final product. put me on your list if you desided to sell some of these units as i would never be able to construct one on my on. pepa

[Nando]

Why don't You go to : http://www.parallax.com/ and investigate what You can really do in and easier way and much better, You may have to have decisions to be able to solve the programming of the unit, like thinking that there are more and different computers than the MAC.

JAMECO.COM Catalog 263 / AUGUST 2006 starting in Page 9 BASIC STAMP Controllers & Development Kits -- also the site has alternatives.

The most difficult solution starts to be easier when one does not try to be stubborn, obstinate or hard headed

Nando

[lohearth]

  Roger

  Maybe chech with a local chicken farmer.Alot of them use a similar set up to what you need. Seen a few that use John Deere deasel with a gen so that they don't loose whole sheds full of chicks in a power outage. Doesn't take long for them to die with out proper vent/heat. Might give you head start on design

  Lohearth

 

[TomW]

Nando;

The most difficult solution starts to be easier when one does not try to be stubborn, obstinate or hard headed

Oh, like you are about MPPT? Lets add in "arrogant" also, then.

T

[jimovonz]

I'm an advocate of the 'picaxe' micro and have used them for many projects over the last 4 years since I discovered them. The picaxe is a commonly available pic micro controller from the company 'Microchip' running an interpreter that allows the use of a form of BASIC along with a direct serial connection. The addition of the interpreter is what makes a pic into a picaxe and it is not done by Microchip. Most micro's require some sort of external hardware (such as a 'programming board') to allow you to connect it to a PC to upload software - usually written in a relatively low level language. One great  advantage of the picaxe is that you can connect it directly to a serial port (direct to TTL and via a couple of resistors to a standard RS232) and upload programs written in BASIC. Another great advantage is the cost. The most basic picaxe, the 08, costs me $NZ4.68 ~$US3 Another $NZ8.13 will get my a 'project board' which includes a socket for the chip, a jack to utilise a standard type serial cable for connection, the appropriate components for connection to a standard PC RS232 serial port and plenty of space for additional components for prototyping. So a mere $NZ13 ($US8) will get you started programming your own micro!

A micro allows you to take a number of inputs and produce the desired outputs based on what ever criteria you specify. This sometimes complex mapping between inputs and outputs is infinitely flexable and can be altered by simply uploading new code. Your typical picaxe has a number of digital inputs (can detect if an input pin is at 0 or +5v), a number of analog inputs (can read the voltage on the pin between 0-5V in 256 or 1024 steps), can accurately measure the length of a pulse on a pin, can read serial data comming in on a pin (from another micro or PC) and can read data from various 1-wire devices on a pin (such as a digital temp sensor, date/time from a RTC, data from an EEPROM etc). Outputs include digital (pin is set either 0 or +5V), pwm (the pin is rapidly alternated between +5 and 0V with the ratio of 'on' time to 'off' time being some specified value - with the appropriate filtering this can become a 'pseudo' analog type output) and serial data (can include a cheap serial LCD display), 1-wire devices (such as sending data to an external EEPROM).

A device such as the picaxe would be ideally suited to the task described in this post. Responding to battery voltage, controlling sequences of actions, counting attempts, monitoring results, timing events - all infinitely flexable without so much as changing a resistor!

[Nando]

Yes Tom like you.

Nando

[RogerAS]

Nando,

Well I really don't know what brought that on about being stubborn & hardheaded. I don't have the coinage to go buy a PC just to play with something I'm not sure I could program to do what I want anyway. Heck, I've only operated a PC a very few times and I'm not sure I could even run one good enough to post here! Hell, If I could afford another computer why buy the damned stamp modual? A simple I-O board would do way more than the stamp anyway. My solar tracker works without a stamp modual and it ain't that much different than this device.

I use a Mac because for what I do to pay the bills it's the best platform to use. Graphic Design. If you don't believe that then ask any major newspaper or magazine here in the US. That's just the way it is. I'm not asking you to like it Nando, nor am I all that wild about being judged by the likes of you for trying to make a living.

Yes, you pissed me off big time Nando ole son! Why don't YOU go to http://www.kissmybackside.org!

Since you obviously do understand these stamp modual devices and know how to program them why don't you just build these and sell them? I'll buy the first prototype. NOT!!!

Oh, and BTW Nando ole pard, what ever became of the HUGE files I sent you? Surely you remember the Adobe Illustrator version 9.0, for Mac, that added up to 15+ megs and had to be segmented into hunks that would go through my email server. You remember, the one YOU asked for. You know, the one where I had to go find software and download so I could divide it up and make a self extracting archive. Yes the one that took several hours to get done and used up a good portion of my satelitte FAP (Fair Acces Policy). The one I use to draw schematics and nearly everything else. Yes, the $800+ retail software I sent you as a favor and you never responded to any of my followup emails or even bothered to say thanks ONE TIME! I may be an ass but by God I've just met my better in that regard.

Give me a freakin break you condescending piece of dung! I don't know what your F___k*#) problem is but get over yourself! You MIGHT know a good deal about a lot of things but you sure as hell don't know much about being a worthwhile human, or how to treat one.

Sorry everyone else but I've been degraded for the last time without speaking my mind. I may have screwed up a month or so ago, but I'm not a target for abuse. I was trying to post my idea for a non-digital-logic operated system to keep my batteries charged without having to babysit the entire system. Now this??

Now I think I'll build the damned thing and spend the rest of my life making it work! I still welcome input from those that might actually know something based here in the real world. Heck, just simple ideas about where I might be screwing up are welcome. However being called names by somebody that obviously isn't my better ain't what I was lookin' for, and it stops here!

TomW, I think you and I need to get together and hoist a few to know-it-all fools, or maybe in spite thereof. Man, how did you take it for so long?

[Volvo farmer]

Tom,

You got some way of posting that doesn't show this message at the top of the screen?

Please be nice. Flame wars and rude comments are NOT tolerated here, and the offending comments will be deleted immediately. Keep in mind that sarcastic and cynical comments that you might make in person (with a smile on your face!) often appear to be very rude on a discussion board.

Where's that moderator when you need one? ;-)

[dinges]

Roger,

Did you have a look at Commanda's dumploadcontroller?

http://www.fieldlines.com/story/2006/7/21/93147/7405

http://www.fieldlines.com/story/2006/6/22/04750/5496

my feeble attempt at it:

http://www.fieldlines.com/story/2006/7/1/185910/1212

I'm building it right now, so far it seems to work. Overdesigned, well-thought out. I understand how it works and can modify it myself. Praise the lord, no micro-controllers. Serviceable till doomsday, well, almost. All parts I have in my parts-bins. And I even understand how it works, which is more than I can say about many a u-controller.

It could serve as the basis for the system you have in mind. One of my plans is to add capability for a 24V grid connected battery charger; very easy to implement with this design; thus, I expect it to be easy to be used with a genset too.

It's based on the LM3914/15. There's a temperature module in there too (for battery temp. measuring & controlling), which could perhaps be adapted for your use. Only you can tell.

You want to do much more than this dumploadcontroller can, but I think it would be a way of starting the project.

As far as the rant. I know how you feel, mate. I don't like discussing things by private mail, esp. unrequested mail that slam other people's designs. Designs, that I happen to like for the exact same reasons that they don't like it. The guy reading this will know what I mean.

[TomW]

volvo;

Well, excuse me, but he is a total A$$ and if you can't see it, I feel sorry for you.

I tried to be nice and if you read closely, I did not say anything untrue or viscious. Just stating facts as I see them. Besides, he initiated the attacking and I dislike him anyway so decided he needs to be chewed out.

Not going to get into a tirade here but I am sick and tired of all these know it alls with not one RE device in use giving overeducated advice that nobody here can use or wants. It has turned into an engineers urine spraying contest here and that type has pretty much run off those who do.

Replies to comments always appear under the comment it is in response to.

T

Just my opinion.

[commanda]

Peter (Dinges) beat me to it. Look at my dump load controller.

Use an LM3914, the 3915's are logarithmic. Great for a VU meter on your hi-fi, not great for a battery voltage monitor.

If there's anything in the circuit you don't understand, just ask.

As for the counter for the restarts, have a look at a 4017 johnson counter. This is a 10 step sequential counter, advances with each clock pulse.

Picaxe is a great chip, used them many times. Not sure if the basic compiler/programmer will run on a mac or not.

One thing all microcontrollers are not great at; getting up & running again quickly after being taken out by lightning. That's why I like the discrete approach.

Amanda

[Volvo farmer]

You have some good points there, no need to feel sorry for me. I just wondered if you couldn't make those points without calling people names. It truly does come across as rude (to me anyway). It's quite a challenge to debate people's ideas without resorting to ad hominem attacks, and I think calling someone a name (even if it's true) lowers the converation to a fourth-grade level.

I personally try to be extra nice on this board because it appears our hosts want it that way. If you want to see a master at being nice while disagreeing, look at some of the posts by DanB, he was even nice to that kckclass guy a few weeks back.

Just my opinion too :-)

[oztules]

Hi Roger,

Well it looks like you have a good understanding of the required parameters, and are comfortable with your chip selection. For your own homebrew this I think is a good thing.

The sticking point at this stage appears to be the start count. As you say, relays could do this, or a 4 bit shift register may solve the problem of counting (each attempt, clocks/shifts the attempt 1 step until 4th step is shutdown/alarm etc. Maybe  a 40194 bidirectional shift register will do the trick. It has asynchronous reset as well to restart the whole show.

In our diesel water pump "set and forget" system, it uses a electrovalve right before the fuel injector pump to shut down in emergency (oil/temp) and after the preset time (few hours) has elapsed. The engine can be controlled easily, and doesn't require re-priming using this technique.

Perhaps the fuel tank level, oil pressure, overtemp,charge reached parameters could all be put inseries with this valve. Their outputs could all be in the normally closed condition in normal operation, and any one opening will shut off the fuel.( A simple parallel across it for starting should get around the low oil pressure at startup.)

It would simplify things if you could fashion a govenor on the engine (briggs do it with the cooling fan on their little motors, depends how stiff your injector lever is i guess), then only on/off via the fuel shut-off will be needed to control the beast.

You may find that simple R/C networks in conjunction with your 324's will do most of your timing for you, there by eliminating the 555's in most cases (restart timing etc.)

A handful of dpdt relays will complete the picture, and allow for simple interlocking of functions..... With leds across the inputs, I like a visual display of whats on and whats not, makes fault finding easier.

Looks like an interesting project.

.............oztules

[RP]

Roger,

I agree with the others that a Basic Stamp would be suitable for this.  You can get one at RadioShack in the states.  I too am a Mac user and a quick search revealed that there is a Mac interpreter available:

http://www.parallax.com/html_pages/downloads/software/software_basic_stamp.asp

If you'd really like to avoid programming or a whole lot of electronics then consider a cam mechanism.  They've been sequencing washing machines and other industrial equipment for 100 years.  Although it's a bit old school, I believe there is a certain elegance in a well designed cam sequencer.  Depending how you do it, it can be sort of programmable by adjusting cams and worst case with a grinding wheel!

[bob g]

i just thought i would pop in,,, and i see the full moon must be out here as well.

group hug?

anyway,, on the subject at hand,,

my business partner just picked up an old kohler diesel power plant,

it has all the controls you are looking for, but in the older mechanical version

auto sensing, auto start, engine management, safty features, the works

he got the manual with it, and i think the schematics as well.

if you are interested i can check with him and see if he does indeed have the schematics

as an aside

not sure i like waiting for 11.6 volts to start the charging regime, and i know i don't like the 14.5 shutdown.

if you are going to the trouble of a pic/stamp system, then go with a marine controller to take over the charging regime.

such as the offering from balmar

but i am sure you have your reasons

* also if anyone can get a tutorial going on the use of a stamp or whatever, i too would like to be on the learning end of that discussion.

sign me up!

bob g

[WXYZCIENCE]

Roger, my first suggestion is to abort after one try with your start sequence. If every thing is working properly it should start first time else something should be checked. Thirty seconds is a lot of cranking. Joe

[wooferhound]

TomW

Chill out man

I don't normally call anybody names but,

you are becoming a Cranky Old Fart

There was a time when you made comments that helped people.

[maker of toys]

there are quad comparitors out there that will fill the niche you've put the 324 in. . . the one thing about using an op-amp as a comparitor is that you either need a negative bias supply or the output does not swing to ground, which can complicate the output interfacing a bit.

 many comparitors are specified to run from a single supply.

I think Parallax makes a USB interface for the STAMP; and last I checked, AllElectronics (.com) makes a USB programmer for the PIC series.

the STAMP is more user freindly (basic instead of assembler) but I'm sure a GUI interpreter is out there for the PIC.  I haven't bothered to find it, though.

so it's a cost vs. convienience thing.

that said, however, there is a LOT to be said for understanding the equipment you depend on.

PS: there are ways to harden uControllers and PCs against big RF spikes and the like that result from nearby lighting and other big sparks. I'll compile some of the easier/cheaper ones I use at work and post them as I get time. the basic idea is to make the expensive parts look more difficult for a fast transient to flow through than some other path.

To get you started: ferrite toroids are your friends; Allied electronics (and others) sell a large selection.  place them as close to the uController as you can to protect your inputs from fast transients; you want to wind the cable  leading to the i/o pin as many times as you can around the ferrite without overlaping any turns.

-Dan

[solarengineer]

well, after 20 minutes of reading the flamewar... heres a quick basic pic tutorial.

microcontrollers like microchip's PIC series..i.e 16f819 ect can be programmed with basic or c compilers. i use proton+ and microengineering labs compilers.

i wont get into circuitry here just a quick and dirty basic view of how you can use these devices easily. I manufacture many device that sell worldwide using these chips..

So lets say we want to read battery voltage and then when the voltage reaches 14.8V turn on a relay that has some form of dump load attached.. im going to omit certain things to make the flow easy to understand, but basically its as easy as i write it.

the circuit:

The program to run it, In the schematic i did not show the voltage divider, this isnt a full tutorial. simply to show those that have no experiance how easy it is to use controllers. you can substitute a basic stamp in place if you dont own a compler.

Start:

adcin 0, battvolts         this reads the voltage divider(divide by 10) so 12v=1.2V

if battvolts >=1.48 then shunt - if batt volts before divider=14.8 then turn on relay

if battvolts <=1.43 then shuntoff - if battvolts <=14.3 turn off the shunt load

goto start - do it all over again

shunt:

high portb.0 - this turns on the mosfet and the relay energising the shunt load

goto start

shuntoff:

low portb.0 - turns off the mosfet and relay

goto start

again this is a simplification, but NOT BY MUCH! the programming is how you read it minus a few small details. but easy to learn. roll your own tacho's controllers.. u name it...

Jamie

[WXYZCIENCE]

Roger, I was thinking about your plan again and I was looking at those remote car starters, they are cheap and would contain all of the glo-plug, tach, start and alarm functions. They should be easy to modify for your purpose. Joe

[RogerAS]

Joe,

Hey, thanks for the heads up. I've seen those advertised and never thought anymore about them. I could maybe leave the remote functional and eliminate a bundle of wires and switches inside the house. Thanks again.