DIY_EFI Digest Monday, 26 August 1996 Volume 01 : Number 246 In this issue: Digital EFI How? Re: EFI musings Re: CPI vs. TBI Re: EFI Conv. Re: Camshaft/crankshaft sense Honda Prelude problems F.Y.I GM mass airflow sensors for sale Re: Digital EFI How? Re: Data Logging Re: Data Logging Number of rotation sensors. See the end of the digest for information on subscribing to the DIY_EFI or DIY_EFI-Digest mailing lists. ---------------------------------------------------------------------- From: Jonathan Lloyd Date: Thu, 22 Aug 1996 20:38:33 +0100 Subject: Digital EFI How? I will carry on "messing" with my EECIV but I would also like to work in parrallel on my own ECU system. I have been wondering for a while about what principle to adopt and have been a little peturbed about the algorithms and tables neccessary to do the job right. I have been reading some detail recently about neural networks implemented on some fairly complex tasks on some pretty ordinary hardware. I like the idea of a system which is programmed to do no more than "learn how to make the car run best and then do it". Presumably with the correct parameters for definition of "best" one could rely on the system to be self correcting to the extent that a "limp home" was inbuilt and ageing taken care of. Anyone know whether this is being considered or if current micros can process enough to do it? - -- Jonathan Lloyd john@xxx.uk ------------------------------ From: "John Faubion" Date: Sun, 25 Aug 1996 10:43:03 -0500 Subject: Re: EFI musings > -> Well Dave with some prudent shopping you could have done it cheaper > -> too. I picked up a GM #730 ECM, harness and most of the sensors from > ... > -> especially if you can adapt factory parts. Ford and GM already spent > -> the money to make them reliable, why not use them? > > What do you intend to do about reprogramming the ECM? > Actually very little. The programming for the '90-'92 F-bodies is a speed density system. By sizing the injectors correctly you can adjust for the size difference between the engines. For example, the factory 5.7L used a Lucas #5207011 injector that flows 20.73 lb./hr at 44 psi of fuel pressure. My 427 is 427/350 = 1.22 or 22% larger. If I maintain the same RPM range of idle-5000 RPM that the factory used, then I would need to use an injector that flows 22% more fuel to maintain the same injector pulse width. So my injectors need to be 20.73 x 1.22 = 25.29 lb./hr. in this case I could use the injectors from a Nissan 300ZX (Lucas #5206004) and adjust the fuel pressure. These flow 24.73 lb./hr at 37psi. To get 25.29 lb./hr I need to use 25.29/24.73 = 1.02 x 37 = 37.84 psi of fuel pressure. By sizing the injectors this way and maintaining the same injector pulse width, the ECM fuel maps should be close. With the learning ability built into this ECM, it should be able to compensate for the small variances. If you are using a V8 that had the same type injection you want to use, (i.e. your going to TPI and running a 350 small block) then your programming is easy. Even if you have modified the engine with aftermarket pieces such as a cam, headers, better heads etc., just contact someone that already programs PROMs for these. Get information from them about their chips on a version that would be close to what you have already built. If like me you are building something the factory never even considered like port EFI on a 427 big block, this could still be an option. You only need a chip that will support the gear ratios, rear axle ratio, approximent weight, type of injection (TPI, MPI, TBI) and most importantly RPM range. Since I will probably not maintain the same RPM range (my 427 is built for 1500-6500 RPM) I would need a stage 3 or 4 chip from Hypertech. This will also require an increase in the size of my injectors to support the horsepower level of this engine. Even if I buy an aftermarket chip, I'll still be under $400 for my total project. The factory systems are much easier to adapt than to build your own controller. I like to idea of building my own but I just don't have the money to sink into the project. I'm also trying to decode the factory ROM as well. I have access to the hardware to do this but as I'm sure everyone knows, this is not a simple undertaking. Until then though, there are several options open. John Faubion jfaubion@xxx.net ------------------------------ From: "John Faubion" Date: Sun, 25 Aug 1996 12:08:15 -0500 Subject: Re: CPI vs. TBI > valve. CPI is essentially a poor-mans SFI system, with fewer injectors. > You should be able to find better descriptions and pictures in some of > the car magazines from last summer or fall. Of course finding injectors that flow 56 lb./hr to support 500 horses is much more difficult as well. John Faubion jfaubion@xxx.net ------------------------------ From: "John Faubion" Date: Sun, 25 Aug 1996 12:22:32 -0500 Subject: Re: EFI Conv. > The only thing I was under the impression a throttle body requires peak and > hold and the 165 uses a mondo CMOS fet Xistor to drive the injectors > usually saturated. The fact that they work indicates no problem in > this area. >From what I understand, the ECM designed for the peak and hold type can use the saturated injectors with out problems but not vice versa. However from the looks of it should be a simple process to design a board to use a peak and hold injector driver to take the signal from the ECM and fire the injectors. This way we won't hurt the ECM and you could use either injector type. > Oh and tnx for VATS info. Changed a bit in code to defeat Vats in my > 730 project. Have confirmed data on timing, idle parms, and fan control. And you didn't send me any info on it?? If you don't mind I'd like to get the info. John Faubion jfaubion@xxx.net ------------------------------ From: "John Faubion" Date: Sun, 25 Aug 1996 12:27:46 -0500 Subject: Re: Camshaft/crankshaft sense > electronic ignition controller, does the controller take (up to) one > rotation to make sure that its assumptions about TDC are correct before > starting to fire the plugs? Yes but this occurs quite rapidly since it rarely ends up just past the sensor. > A second question is, does anyone sense crankshaft and camshaft positioning > separately? This sounds silly, but it would let one sense timing > belt/chain failure almost instantly, so one could inhibit plug firing which Actually if you only used the camshaft sensor you would accomplish the same thing. Once the cam stops turning you won't fire the plugs or the injectors again. John Faubion jfaubion@xxx.net ------------------------------ From: "Woodd, Michael" Date: Mon, 26 Aug 1996 08:03:00 +1200 Subject: Honda Prelude problems F.Y.I Hi all, Thanks to all (bar two, you know who you are) for the astounding number of replies to my last post, not too many Honda drivers out there eh?!! ;-) Anyhoo, found a major cause of some of my problems - in trying to locate the window washer hose which has fallen off (DAMN those honda engineers) I had my wife in the hot seat - I tell her to pull the washer lever, and she turns the key and starts the thing.... I yell turn it off!! and she does so, and I hear a sucking noise which lingers for a second after the engine stops turning. Frustration forgotten, I get her to repeat, and follow the noise to........ A largish brass screw with an O-Ring on it, sitting completely wound out of its hole in the throttle body. Ah ha!! I say. Whats that, my wife says, and I flick the screw up and over the intake clamp, which is *just* stopping the screw from falling out and away into oblivion, and into my hand, and show it to her. I've had this problem for about 6 weeks!!! Tried to screw it back in, but it wouldn't go, I inspect closer and see that it has been clogged up with CARBON of all things, not something I'd expect to find on this side of the works.... Any ideas anyone? I have discovered that it must be the idle air bypass screw, when it is turned in, the idle drops (god i'm good) It was sucking tons of air in, upsetting god knows what. Also, I think it must be a MAP F.I. system, no evidence of what I imagine a FLAP AFM looks like, but heaps of important looking little hoses. Where does a FLAP AFM usually live? (i.e. manifold side or before T.B.?) Took it for an Italian tuneup this w/end, 190 km/h with much more ease. Still that speed limiter, come through for me guys.... Any ideas why virtually all power and torque is found during the first 1/3 accelerator travel? Will a new or reprogammed chip improve things? Enough for now, stay tuned for the next episode, where I delve further into the labyrinthine depths of the Honda Prelude PGM-FI engine bay insearch of the holy grail (the computer)... Mike Woodd (wooddm@xxx.nz) ------------------------------ From: "Bill Lewis" Date: Sun, 25 Aug 1996 19:31:25 +0000 Subject: GM mass airflow sensors for sale I have two used GM mass airflow sensors for sale. They are logically from an '86 or so Chevy 305 or 350 Tuned Port motor. They are manufactured by Bosch. They have a 70mm inside diameter and accept an 80mm diameter boot. I have the connector pigtail. One of the protective end screens has a small tear. I also have an air cleaner I machined up to use with these during my experiments. It is basically two aluminum end plates, separated by steel rod strechers and uses the filer element from an early Porsche 911. I have a rubber boot from a BMW about 6 inch free length that connects the air cleaner to the MAF sensor. It's not untralight, so shipping that may cost a wee bit, but I'll throw it in free with the sensors. I also have a set of 8 injectors from the same engine. They are the high impedance style intended to be run with 12 volts. I don't know if they are from a 305 or 350 (or even if there's a difference). They came from a running engine and have been stored in the house (for about 8 years), but beyond that I can't guarantee them. I'm not trying to get rich off these parts. I'm cleaning house and want to see them go to a good home. Make me an offer. Bill (bill@xxx.com) ------------------------------ From: Darrell Norquay Date: Sun, 25 Aug 1996 19:44:24 -0700 (PDT) Subject: Re: Digital EFI How? At 08:38 PM 8/22/96 +0100, Jonathan Lloyd wrote: >reading some detail recently about neural networks implemented on some >fairly complex tasks on some pretty ordinary hardware. I like the idea >of a system which is programmed to do no more than "learn how to make >the car run best and then do it". Presumably with the correct >parameters for definition of "best" one could rely on the system to be >self correcting to the extent that a "limp home" was inbuilt and ageing >taken care of. Seems to me engine controllers would be a prime candidate for fuzzy logic controllers. Especially the EGO system. There are fuzzy logic tools available for most microcontrollers these days, along with some controllers specifically designed for it. Might be something to look into. Neural networks would be a bit overkill, though, IMHO. regards dn dnorquay@xxx.com ------------------------------ From: David Crocombe Date: Mon, 26 Aug 1996 13:54:37 +1000 Subject: Re: Data Logging I have been reading up on the Analog Devices ADXL-05 accelerometer and it looks promising. The trouble now is supply. I have been told that major car manufacturers have bought up all the stock, and that supply will take 6 to 8 months. Apparently there is also to be a new version called ADXL-76 which is to combine the 5g and 50g versions in a DIP package (as opposed to the metal can package for the ADXL-05). The ADXL-05 is quoted as $35 Australian. This sounds good to me if only I can get my hands on one. I think I will try to get one of their evaluation modules. Hopefully these will be available. Those interested should check out the Analog Devices pages for data sheets and good application notes. http://www.analog.com/publications/magazines/Dialogue/29-2/brief11a.html http://www.analog.com/techsupt/application_notes/application_notes.html I was interested to read of the Amiga program. It is similar to some ideas I had. My big problem (and why I want an accelerometer) in predicting vehicle performance is that with the Drag car the wheel diameter is not constant as it increases with speed and also there is no speedo for measuring speed against a stopwatch. Also no passenger seat for the second person to take the readings! My brother also has a hot transit van in which we can try some of the suggestions that have been made on this list. Current performance is better than 13.608/102 mph over the 1/4 mile. With 351 engine and nitrous is getting just under 13 seconds I believe. I am currently developing an analogue control system for the nitrous for both vehicles. When I get up to speed on microcontrollers I will implement a digital version with lots of features including the accelerometer. Well that's about it from me for the moment. David Crocombe Sydney Australia ------------------------------ From: Chris Howard Date: Mon, 26 Aug 1996 16:48:48 +1000 Subject: Re: Data Logging At 13:54 26/08/96 +1000, you wrote: >I have been reading up on the Analog Devices ADXL-05 accelerometer and it >looks promising. The trouble now is supply. I have been told that major >car manufacturers have bought up all the stock, and that supply will take >6 to 8 months. >Apparently there is also to be a new version called ADXL-76 which is to >combine the 5g and 50g versions in a DIP package (as opposed to the metal >can package for the ADXL-05). >The ADXL-05 is quoted as $35 Australian. This sounds good to me if only I >can get my hands on one. I think I will try to get one of their evaluation >modules. Hopefully these will be available. I tried to get samples of the ADXL-05 in Australia and had no luck. I instead used some IC Sensors accelerometers which are very nice. The manufacturer is: IC Sensors 1701 McCarthy Blvd Milpitas California Phone: (408) 432-1800 Fax: (408) 432-7322 The Australian distributors are: Duff & Macintosh Pty Ltd 13/45 Leighton Place Hornsby NSW 2077 Phone (02) 482 1411 Fax (02) 482 1489 Sorry I've lost the data sheet, but they have a large range, some 8mm square surface mount models other types have all the conditioning and amplification circuitry built in. The type I used was a PCB mount model No. 3028-002. Ring Duff and Macintosh, they were really helpful. It only took about a week for them to get three free samples from the usa in for me as compared to months for the AD device. Regards, Chris Howard Computer Science University of New South Wales email: s2184002@xxx.au Web: http://www.usyd.edu.au/~choward ------------------------------ From: einarp@xxx.no (einarp) Date: Mon, 26 Aug 96 09:46:57 CET Subject: Number of rotation sensors. The Weber Marelli unit uses crankshaft and camwheel sensors. It does need to phase in before firing. I can't imagine any system that would fire on the first compression stroke unless using a coded disk or a distributor. It will stop firing if cambelt breaks, because it looses phasing information. That doesn't help your valves a bit, as many can attest. The damage is already done. - -- einarp@xxx.no ( Maserati Biturbo Spyder ) ------------------------------ End of DIY_EFI Digest V1 #246 ***************************** 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".