DIY_EFI Digest Thursday, 10 October 1996 Volume 01 : Number 307 In this issue: Re: ignition reference Re: drive-by-wire Re: Ignition timing ref points RE: Using PC HW (& Ignition timing reference points) Re: Using pc parts Re: ignition reference Re: ignition reference Re: ignition reference Re: Ignition timing ref points Re: Volume of fuel injected Re: Ignition timing reference points Re: ignition reference Re: Using pc parts Re[2]: drive-by-wire Re: ignition reference Re: Using pc parts Re: ignition reference Software Re:FW: DIY Traction Control? OBDII info Disclosure See the end of the digest for information on subscribing to the DIY_EFI or DIY_EFI-Digest mailing lists. ---------------------------------------------------------------------- From: Kalle Pihlajasaari Date: Wed, 9 Oct 1996 12:34:32 +0200 (sat) Subject: Re: ignition reference Hi John, > My experience is that one is more likely to build a robust, high > quality timing disk with six or eight accurately spaced notches or > vanes than one with 360 or 256 notches and I do have access to > numerically controlled equipment. It is far better to improve the > quality of the inputs than to depend on many more, less accurate > inputs. Quantity certainly does not imply quality in this instance. > The most precise measurement you have available in a typical > ignition system is the record of time in a 16 bit counter. I would never have even tried to make such a disk (unless sending it out to be laser cut, quite cheap these days :-) A sealed encoder is what I had in mind for line counts like that. > If you divorce you mind from thinking in terms of degrees advance > and instead consider ignition TIME before TDC or some other datum, > ignition calculation is much simplified. The time advance bears a > near inverse linear relationship to manifold pressure. It is the > different rotational speeds that make calculation if the ignition > position touchy. No manifold but I get your point. Cheers - -- Kalle Pihlajasaari kalle@xxx.za Interface Products Box 15775, Doornfontein, 2028, South Africa +27 (11) 402-7750 Fax: +27 (11) 402-7751 ------------------------------ From: hoss karoly Date: Mon, 07 Oct 1996 08:55:15 +0200 Subject: Re: drive-by-wire dzorde@xxx.au wrote: > > I tried a mechanical linkage, cable and hydraulic for the clutch in a > 5 speed Supra manual bolted to a Chev, and trucs me, the only one that > you could use was the hydraulic, unless you have a left leg like an > elephant. Even with the hydraulic it took a while to find a > master/slave combination that would give a reasonably easy clutch to > operate. > > Pet hate.. cable clutches. > > Pet hate.. hydraulic clutches. What's wrong with a nice, self-adjusting > cable clutch? Which would you rather have fail on you while driving > through the Heart of Deepest Africa? > > let me add my two cents . the next project I'll make for my van is a power clutch operated by a deformation sensor on the clutch lever . so my Mcontroller can say if he's pushing the pedal with x force I move the pedal down a little if the the force is between the limits -keep it if under let it go up . I found a brand new servo-steering ele-pump in a salvage yard I'm not sure about the speed of movement :( but the nicest thing about it : if it fails it'll leave me with the orig. sys intact! bye charley ------------------------------ From: cloud@xxx.edu (tom cloud) Date: Wed, 9 Oct 1996 06:59:28 -0500 Subject: Re: Ignition timing ref points [ snip ] Is it just me, or does it seem >this way to the rest of you also, that GM has the simplest hardware >available. Smaller, simpler, ignition modules, ECMs with no "Power Modules" >as some companies use, changable EPROMS, etc. etc. I'd like to hear a discussion of using available OEM EFI units to modify for custom use -- and which are easiest // best for that purpose. Tom Cloud ------------------------------ From: cloud@xxx.edu (tom cloud) Date: Wed, 9 Oct 1996 07:32:31 -0500 Subject: RE: Using PC HW (& Ignition timing reference points) >>Time honored 'dumb' question: >>"Why is the sky blue?" >>answer: Clearly that's a two part question. ... "Why?" Well, we'd all >>like to know that. And "Is the sky blue". Yes. > >And just for those that do not know, The blue color of the sky is the result >of the light in the blue wave length is filtered and reflected by part of >the particles that exhist in the upper atmosphere. The light we actually >receive on earth is slightly deficient of blue. > >David J. Doddek really ? Tom Cloud ------------------------------ From: cloud@xxx.edu (tom cloud) Date: Wed, 9 Oct 1996 07:49:21 -0500 Subject: Re: Using pc parts talltom wrote: > I for one would be inteested in the pc parts approach, primarily on account >of entirely to much expierience with propietary junk. Yes, this means that >if it costs more thats to damn bad. When my chip burns up at 7:30am on Saturday >in Clovis New Mexico, I want to be back on the road within 30 min. after radio >shack opens! > Not only the much better parts supply situation, but the computing power >and flexibility should be much greater. > So you pc types, I'm all ears!(this comes from a guy that drives a >cornbinder!) [ snip ] Don't work that way. Rat-Shack ain't gonna have *NONE* of the stuff you need. PC mudderboards use proprietary (read if it breaks, forget it) ASIC chips. Very few of the chips (like the cpu, memory, bus drivers) are "standard" chips -- and rat-shack don' carry none of that stuff. Rat-shack does carry the leetul battries fer yur ceemos RAM. Now, the good side: I've had very few mudder boards fail! The problems I have experienced have generally been with poor hardware / software design rather than component failure. No doubt, the modified PC approach is probably the most appealing for someone with experience in other areas (notice how diplomatically I handled that ?? -- and they said I was insensitive!). Only problem I see, is that by the time you've got it working, will you be happy? You will definitely have learned some very useful *stuff* (and I'm not being facetious). But, one of the things you'll learn is that you could have spent less, and your result will be B-I-G . That's why, in another post on a diff. thread (I think), I asked for discussion about modifying OEM EFI controllers. Seems like that may be the most efficient, quickest, satisfying way to go. Tom Cloud ------------------------------ From: cloud@xxx.edu (tom cloud) Date: Wed, 9 Oct 1996 08:01:41 -0500 Subject: Re: ignition reference >> My experience is that one is more likely to build a robust, high >> quality timing disk with six or eight accurately spaced notches or >> vanes than one with 360 or 256 notches and I do have access to >> numerically controlled equipment. >I would never have even tried to make such a disk (unless sending >it out to be laser cut, quite cheap these days :-) Has any thought been given to making the disk from a PCB?? The artwork would be relatively simple using AutoCAD or such. I can't see how to do it right now -- maybe plate the copper with iron so a hall sensor would work with it. Long ago I used a similar approach to make optical encoders: encapsulated a film negative artwork in a lexan disk. Tom Cloud ------------------------------ From: Kalle Pihlajasaari Date: Wed, 9 Oct 1996 16:22:29 +0200 (sat) Subject: Re: ignition reference Hi Tom &c. > >> My experience is that one is more likely to build a robust, high > >> quality timing disk with six or eight accurately spaced notches or > >> vanes than one with 360 or 256 notches and I do have access to > >> numerically controlled equipment. > > >I would never have even tried to make such a disk (unless sending > >it out to be laser cut, quite cheap these days :-) > > Has any thought been given to making the disk from a PCB?? The artwork > would be relatively simple using AutoCAD or such. I can't see how > to do it right now -- maybe plate the copper with iron so a hall > sensor would work with it. Long ago I used a similar approach > to make optical encoders: encapsulated a film negative artwork in > a lexan disk. You could photo etch a copper or iron disk. It would probably be quite easy to do as well. We were going to make soft copper head gaskets for high compression conversions for older off-beat engines that there are no kits for by using sign writing vinyl which can be cut to 0.25 mm accuracy without any trouble. We have a set of vinyl masks which would be about right for a chainsaw but have not had time to try out the setup. With copper it is easier, you take some ferric chloride and sulphuric acid and then REVERSE electroplate the holes into the sheet with the vinyl mask protecting the gasket. Making a iron rotor with 36 - 1 holes or blades for a hall interrupter should be almost as easy. I would tin or zinc plate the iron afterwards to prevent corrosion. The laser people would not cut copper for us as the IR LASER beam is reflected and the cutting is not fast and the reflected beam probably causes the cutters some concern. Cheers. - -- Kalle Pihlajasaari kalle@xxx.za Interface Products Box 15775, Doornfontein, 2028, South Africa +27 (11) 402-7750 Fax: +27 (11) 402-7751 ------------------------------ From: cloud@xxx.edu (tom cloud) Date: Wed, 9 Oct 1996 11:19:47 -0500 Subject: Re: ignition reference >> Has any thought been given to making the disk from a PCB?? The artwork >> would be relatively simple using AutoCAD or such. I can't see how >> to do it right now -- maybe plate the copper with iron so a hall >> sensor would work with it. Long ago I used a similar approach >> to make optical encoders: encapsulated a film negative artwork in >> a lexan disk. > >You could photo etch a copper or iron disk. It would probably be >quite easy to do as well. > >We were going to make soft copper head gaskets for high compression >conversions for older off-beat engines that there are no kits for >by using sign writing vinyl which can be cut to 0.25 mm accuracy >without any trouble. We have a set of vinyl masks which would >be about right for a chainsaw but have not had time to try out the >setup. With copper it is easier, you take some ferric chloride >and sulphuric acid and then REVERSE electroplate the holes into the >sheet with the vinyl mask protecting the gasket. Whoa, dude. My experience etching thick stuff is not happy. Also, what's this with the sulfuric and ferric chloride? Ferric chloride or ammonium persulfate are normal for etching PCB's. Can also use alkaline ammonia and chromic-sulfuric acid, but they are dangerous. Just plain nitric acid can also be used. But, as I said, the most commonly used for PCB's are ferric chloride and ammonium persulfate. If anyone cares, I can elaborate on pros and cons of each. Don't have much success with anything very thick (that requires spraying the etching solution to lessen 'undercut'). Tom Cloud ------------------------------ From: Todd King Date: Wed, 09 Oct 96 09:54:00 PDT Subject: Re: Ignition timing ref points <<< From: Kalle Pihlajasaari Date: Wed, 9 Oct 1996 02:56:28 +0200 (sat) Subject: Re: Ignition timing ref points ... I want to do the timing with no interpolation so that the counter can be very dumb and just count pulses ant any RPM and then go blip. >>> Good point; but I think you may find that a ref mark of say 50 deg BTDC for each cylinder combined with a *simple* interpolation will do the job within a few degrees as well, plus drastically reducing the interrupt overhead. But let's see, I guess after every firing event you could determine the current rpm and load; consult a lookup table for the correct "tooth count" of the next fire event and initialize the count; when the count is up again it's "Fire in the hold! <*ka-boom*>" time, plus set up the count for the next event. FYI the spark map in one of the GM PROM's only has a granularity of 400 rpm increments (Y axis) (if memory serves) vs a load (X axis) granularity of something like 12 or 16 or so columns. But what about dwell time for your coil(s)? Kalle, can you discuss any of your plans for this with the list? Todd Todd_King@xxx.com ------------------------------ From: Stephen Dubovsky Date: Wed, 9 Oct 1996 15:50:03 -0400 Subject: Re: Volume of fuel injected At 07:03 PM 10/8/96 +0100, you wrote: >>Here's my problem. I'm trying to calculate the volume per burst for a >>typical fuel injector. I've downloaded some specifications, but I'm at >>a loss to break them down. Ok, here's my understanding of injectors - the flow rate (cc/min or lb/min) is rated at 100% duty cycle. As for other duty cycles you need to do a little testing, but it helps to understand how the injector works. FI are little solenoids. When you first apply the voltage, it takes some time to built up enough current (and magnetic feild -> force) to start the valve opening. This is dead time and can be corrected for easily (add it to the required 'on' time). The valve then starts to open and assume the flow increases linearly until it is wide open (and has reached its 'rated' flow). At this point you can reduce the voltage on the injector to save power and off time - more on this later. You leave it on for some desired time at rated flow - this is about the only linear part of the curve. When you go to turn off the injector, you do something to try and stop the magnetic field holding the valve open. Disconnect it. The problen is that the injector is a coil of wire around magnetic stuff and has inducance so the current (and the magnetic field) wont go to zero instantaneously (would take infinite negative voltage). Dont know if your an EE, but there is this rule that goes V=di/dt... Anyway, thats why you usually need some sort of clamp to keep from killing your drive circuit. So, after some time delay the magnetic field drops below that required to hold the valve open and the valve starts to close (due to that little spring and maybe fuel pressure). Again assume some linear curve (which I again doubt - but hey, these are supposed to be negligible times right;) And finally the valve is closed. So the things to acct for are: 1) delay until beginning of turn on 2) time required to turn on the injector (some fixed amt of fuel will be squirted during this period. If you assume linear flow increase, this amt is quadratically (or worse) related to this time.) 3) actual time the FI delivers 'rated' flow 4) delay from when you remove power until valve starts closing (again some twisted relationship of flow vs time) 5) the valve actually closes. At high duty cycles, period 3 should swamp all others and you can assume 'linear'. Time 1 is easy to figure out. See how small of an injector pulse delivers NO flow (it never gets to open). The other times might be taken into account (although I dont know of anyone who does this, but its possible) if you plot flow vs. on time. This is probably only interesting at very low flow rates. Hope this helps. If anything here is too complicated, just ask. We build things that drive inductors and actuators all the time and sometimes I confuse myself. If you want more detail, ask and Ill scare you w/ math;) (albeit simple calculus) SMD ------------------------------ From: Stephen Dubovsky Date: Wed, 9 Oct 1996 16:04:50 -0400 Subject: Re: Ignition timing reference points >> >>>I was thinking of adding a 256 or 360 slot optical interrupter disk into >>>my distributor to give double the that number of edges every rev >>... ... > >I have often wondered about using a resolver on the distributor drive to >measure the position. Although the new price is frightening, there is a >lot of old equipment around with them in at give-away prices. At least >near me in Nottingham with its big military surplus dealer. (Anchor >Surplus). > >My worry is whether they would tolerate going round at half crank speed >for hours and hours. I know they have top quality bearings but is the >life adequate? Its not too much problem digitising the sine and cosine >output signals. >-- I have read several responses that say this much accuracy is required, but Im going to comment on the technical aspects and not the reason to use it... Even the 'sealed' optical encoders can have problems. One is on the end of an induction motor (for electric vehicles) and microscopic drops of oil can get in cause false edges. A REAL pain. We just have not had good experiences w/ them. Several people wrote me asking about resolvers so... I believe that are reliable enough to use on a motor. There are no wiping contacts (i.e. no brushes) so only the bearings can fail. Many resolvers are designed to be put on the end of permanent magnet three phase machines (at least this is where we find them most common) so you can determine the field position at any time w/ good accuracy (so we can bulid electronics and turn them into 'DC brushless' machines). Since they were designed to mount to an electric motor, rotational life is very long and they are designed to turn very fast (have seen them rated well past 12-14k rpm which is our top speed). Anyway, the little things should be able to be used to get crank position w/ good accuracy and still be really reliable. Some one said they are analog in nature (agreed) so he couldnt use them w/ a micro (here's where I disagree). We use an Analog devices R2D converter (resolver to digital) that gives you a nice 10-14 bit word (depending on dynamic perf. required) giving absolute poistion. Not too bad -> 0.35 to 0.02 degrees resolution. (The R2D we use is about $60) They are pretty much little rotating transformers. The is a primary on the stator (outside part that doesnt turn) that you excite at, say 10kHz. There are 2 secondary windings 90deg apart on the rotor. If you look at the signals induced on these windings, you get the carrier (at 10k) sin and cos modulated. The signals on the rotor are then coupled back to the stator useing more little transformers that dont care about position. You have then put a signal on the rotor, modulated it w/ the rotor angle and gotten the info back to the stator w/o any electrical contacts. Cool huh. Then you can use some PLLs and counters to get the rotor angle from the modulated info (or use a R2D w/ everything in one chip). Thats my 2 cents worth. Hope this helps. SMD ------------------------------ From: Sandy Date: Wed, 09 Oct 1996 14:33:49 -0700 Subject: Re: ignition reference All this talk of making/using an encoder, just use a steel wheel with a VR pickup, like most everybody else is doing. If you like you could also use a hall effect pickup. One thing again, is that if you are getting timing from the distributer, it will be subjected to a bunch of error, many from chain (belt) stretch, and other cam related jitter. Ask someone how their electromotive distributer pickup works! (have has several bad stories) But, back to the problem, you can easily get a laser cut wheel made of steel (or plastic) that can be made to your fancy. The cost will be < $50.00 QTY 1. All you need is a DXF file from your CAD program, and you have it, dosent matter what shape or size. Sandy ------------------------------ From: "George M. Dailey" Date: Wed, 9 Oct 1996 18:56:00 -0500 (CDT) Subject: Re: Using pc parts At 10:21 PM 10/8/96 -0700, you wrote: > > > I for one would be inteested in the pc parts approach, primarily on account >of entirely to much expierience with propietary junk. Yes, this means that >if it costs more thats to damn bad. When my chip burns up at 7:30am on Saturday >in Clovis New Mexico, I want to be back on the road within 30 min. after radio >shack opens! > Not only the much better parts supply situation, but the computing power >and flexibility should be much greater. > So you pc types, I'm all ears!(this comes from a guy that drives a >cornbinder!) ... Talltom > I've gotten lots of info on the PC approach to fuel injection (or control) lately. I'm going to stick with my OEM brain box though. Several months ago I posted about using a PC to control some auto "body" functions. Did not get nearly as much of a positive response as the most recent posting. Most responders believed it to be the wrong way to go. Now I think I'm dead on the money. Here is a repeat for the newbies. ...I neglected to tell you folks out in DIY FI land, what my true reason was for wanting to read ALDL data. Here's my confession. I have constructed a mid engine, rear wheel driven sports car from scratch. The engine is a '85 PFI 3.8L Buick V6. The plan is to transmit the ALDl data to a dedicated PC, peramanently mounted in the car and connected to a small VGA/LCD monitor. Software would convert the data to usefull information that would be displayed on the monitor. Boom, no additional instruments needed and a wealth of data on demand. The car (Bianca) was designed to house a PC. Think of her as a streached Lamborghini Contach with all the room and comfort of a full size '70's Olds. Data logging would also be utilized. The PC would also control some BODY functions. It's not that I'm to cheap to by Diacom or XP240 (I just might by the XP240) but, if I'm going to make this work I need to get some type of simple PC to ALDL communication established..... So, what's available for those fancy graphical interfaces in a PC controlled i/o system? GMD ------------------------------ From: dzorde@xxx.au Date: Thu, 10 Oct 96 08:15:43 Subject: Re[2]: drive-by-wire Could you elaborate on this, it sounds like a brilliant idea, but how does it interface to the original system ? and what is the original system ? Does the pump operate a slave cylinder that operates the release arm in the gearbox ? I can see the idea, but I can't picture how the pump connects to the system and operates. Dan dzorde@xxx.au ______________________________ Reply Separator _________________________________ Subject: Re: drive-by-wire Author: diy_efi@xxx.edu at INTERNET Date: 10/9/96 9:24 PM let me add my two cents . the next project I'll make for my van is a power clutch operated by a deformation sensor on the clutch lever . so my Mcontroller can say if he's pushing the pedal with x force I move the pedal down a little if the the force is between the limits -keep it if under let it go up . I found a brand new servo-steering ele-pump in a salvage yard I'm not sure about the speed of movement :( but the nicest thing about it : if it fails it'll leave me with the orig. sys intact! bye charley ------------------------------ From: "George M. Dailey" Date: Wed, 9 Oct 1996 19:18:09 -0500 (CDT) Subject: Re: ignition reference At 02:33 PM 10/9/96 -0700, you wrote: ...snip...you can easily get a laser cut wheel made of steel >(or plastic) that can be made to your fancy. The cost will be < $50.00 QTY >1. All you need is a DXF file from your CAD program, and you have it, dosent >matter what shape or size. > >Sandy Where do I send my check and DXF file to?? GMD ------------------------------ From: Thor Johnson Date: Wed, 9 Oct 1996 21:09:00 -0400 (EDT) Subject: Re: Using pc parts On Wed, 9 Oct 1996, George M. Dailey wrote: [mucho snippage] > So, what's available for those fancy graphical interfaces in a PC controlled > i/o system? I was planning to use LabView (cost mucho dollars... but check out the alliance discount). Has one of the easiest to use interfaces out there... Not only that, but is really nifty as a programmming method. Drawbacks: 1. For big things, you need BIG computer. (Runse under Windows) 2. Cost.. $2K. But for a $100 alliance membership, you get 40% off (I think). Thor Johnson johnsont@xxx.edu http://falcon.mercer.peachnet.edu/~johnsont Have you seen the WarpMap lately? http://falcon.mercer.peachnet.edu/~johnsont/warpmap ------------------------------ From: Sandy Date: Wed, 09 Oct 1996 19:40:36 -0700 Subject: Re: ignition reference For laser cutting, I have just found these two places - AZ Manufacturing 1-800-959-0992 Dick Bogue Desert Metal 1-520-888-4500 Paul The quote I have so far is from AZ, and it was for a 8" 58 tooth wheel, at $50.58 each. The size matters, and they can do most anything from DXF files. If you need missing tooth (or added) I have a program written that will generate a DXF with just a few parameters. So If you know the # teeth, depth of tooth, wheel diameter, and #missing/Added teeth, I can send you a DXF of what you need. Sandy ------------------------------ From: dzorde@xxx.au Date: Thu, 10 Oct 96 10:36:21 Subject: Software Fellow DIY's, With all this talk about changing the code in OEM systems. Does anyone have a copy of Microsoft Quick Basic 4.5, or know where I could download a copy from ? I need to make a small change to an old program (not fi code), but without the actual Quick Basic software I can't recompile it. If somebody has a copy they wouldn't mind sharing, please mail a zipped version. Thanks Dan dzorde@xxx.au ------------------------------ From: Clint Sharp Date: Tue, 8 Oct 1996 18:29:09 GMT Subject: Re:FW: DIY Traction Control? In your message dated Tuesday 8, October 1996 you wrote : > > > I know dropping random cylinders is the way used to limit RPM and > > regain traction, etc. But I dread everytime I hit the rev limiter in As I read it, a few aftermarket traction control systems retard the ignition to a point where the power drops off until the wheels are (roughly) at the same speed, the more expensive ones use three wheel speed sensors and can control a car with an LSD. Would you use traction control in anything other than 1st and 2nd gear ? - -- If you have a problem with excess cash, mail all those unwanted notes in plain packing to; clint@xxx.uk ------------------------------ From: gwalton@xxx.com Date: Wed, 09 Oct 1996 23:16:47 -0700 Subject: OBDII info Disclosure Hello, Just joined the list, read over the archives and found a post about the EPA requiring car manufacturers to make ecm info publicly available. Anyway, I didn't see anymore posts on it so I went and found it http://thorplus.lib.purdue.edu/vlibrary/reference/gpo/GPOAccess.cgi?1995_register/TEXT/176923/3=150900%20176923%20/diskb/wais/data/1995_register/fr09au95.dat.wais;7=%00; It's a pretty lengthy document. If you have any trouble finding it let me know I can email it to you. AND if this is old news for everybody, sorry! One last item from the archives....Someone posted about a possible FAQ for the "ALDL to PC" and "ALDL Data Stream" topics. Was this ever accomplished, if so I would be real interested in getting a copy(Or any other ALDL datastream info). Later Greg Walton gwlton@xxx.com ------------------------------ End of DIY_EFI Digest V1 #307 ***************************** To subscribe to DIY_EFI-Digest, send the command: subscribe diy_efi-digest in the body of a message to "Majordomo@xxx. A non-digest (direct mail) version of this list is also available; to subscribe to that instead, replace "diy_efi-digest" in the command above with "diy_efi".