DIY_EFI Digest Friday, December 3 1999 Volume 04 : Number 672 In this issue: DAQ Re: QNX Floppy Tools? (Was DIACOM) Re: DIY_EFI Digest V4 #671 Re: DIY_EFI Digest V4 #671 Re: AC Klutch Re: Transients Re: KLUTCH!!! Re: KLUTCH!!! Re: transients ECU's and HEI RE: QNX Floppy Tools? ==> Linux Re:KLUTCH!!! Re:Transients See the end of the digest for information on subscribing to the DIY_EFI or DIY_EFI-Digest mailing lists. ---------------------------------------------------------------------- Date: Thu, 02 Dec 1999 15:01:11 -0500 From: Jay Wallace Subject: DAQ From: Andris Subject: Data Acquisition Folks, I'm looking to get a few more gadgets to help me design a better intake, and set up the fuel/timing maps on the Camaro. I would like to make a data acquisition system that can simultaneously record 5 channels at 10 Hz. I figure one K-Type thermocouple before each turbo, one in the intake manifold, one before the throttlebody and one just after a turbo (and before the intercooler). I'd like to know exhaust temps after the turbo for my own knowledge, and would just swap thermocouple inputs for that one... Snip... Andris, You might look around for a serial or parallel port A/D. They should have the bandwidth and resolution (10 - 12 bit) for your application if you don't mind building a couple of amps to boost up the thermocouple signal. One possibility is B&B electronics ($50 to $100; www.bb-elec.com - look for ADIO12 or 232SDA12) I'm sure that there are others, too. Jay ------------------------------ Date: Thu, 2 Dec 1999 15:53:55 -0500 (EST) From: Pat Ford Subject: Re: QNX Floppy Tools? (Was DIACOM) Previously, you (DIY_EFI Digest) wrote: { { RE: QNX Floppy Tools? (Was DIACOM) { Date: Thu, 2 Dec 1999 02:28:52 -0800 (PST) { From: Jon Watts { Subject: RE: QNX Floppy Tools? (Was DIACOM) { { > From: "Jim Velasquez" { > Subject: QNX Floppy Tools? (Was DIACOM) { > { > The QNX Demo Disk is one of my favorite examples of a small, { > focused embeddable product! (That, and the Linux Router Project's { > single-disk, run-from-ram concept). { - ------------- Yeap one of the best ways to gain public mindshare. { QNX is way too expensive for any kind of hobbyist to use (thousands { of dollars for a development license). What I had in mind was a graphical user interface, on a reliable os running from a single floppy, on most (read old 386 up) laptops. The expensive part may be changing soon. The serial engine reads a plain text aldl description file ( expandable w/o requiring a big dev. kit) and the ability to save logged data { The design is very simplistic I think the company line is elegant 8*) { (in a good way) but it mostly finds it way into video poker machines { and such. I am working on it at work ( I'm a programmer at QNX) { { What I'm surprised at is that there's not more talk of Linux on this { list. there has been alot of times linux has come up, I used it for years, it's good stable BUT I work at qnx and my play at home helps me at work { By itself it could handle an EFI system and host it's own { analysis software. With real-time Linux (www.rtlinux.org) this is now { a certainty. Sooner or later someone will build a PC-104 (or smaller) { EFI unit. Al Lipper designed a pc based efi, and the pc104 bus is just an isa bus with a funky connector( there is some minor electric differences but don't worry) and you can buy pc104 protocards (expensive). what I did was add a mini isa slot to a laptop using Al's pc-efi and ran it under QNX ( once again the main reason was size (ruled out M$), speed (same thing), and reliability. the choices I had was QNX (free from work) or Linux { { This would be a great project, but by the time you're done you could { easily have bought an Electomotive or MoTec setup. Unless Lego comes { out with EFI sensors for the Mind Storm....... :) { { { ===== { Jon { __________________________________________________ { Do You Yahoo!? { Thousands of Stores. Millions of Products. All in one place. { Yahoo! Shopping: http://shopping.yahoo.com { - -- Pat Ford email: pford@xxx.com QNX Software Systems, Ltd. WWW: http://www.qnx.com (613) 591-0931 (voice) mail: 175 Terence Matthews (613) 591-3579 (fax) Kanata, Ontario, Canada K2M 1W8 ------------------------------ Date: Thu, 2 Dec 1999 16:16:31 -0500 (EST) From: "Clive Apps Techno-Logicals 416 510 0020" Subject: Re: DIY_EFI Digest V4 #671 > > > The QNX Demo Disk is one of my favorite examples of a small, > > focused embeddable product! (That, and the Linux Router Project's > > single-disk, run-from-ram concept). > > QNX is way too expensive for any kind of hobbyist to use (thousands > > What I'm surprised at is that there's not more talk of Linux on this > list. By itself it could handle an EFI system and host it's own > analysis software. With real-time Linux (www.rtlinux.org) this is now > a certainty. Sooner or later someone will build a PC-104 (or smaller) > EFI unit. check out uclinux.com they use a coldfire chip and a baord about the size of a simm it has verious seral and 10bT in/outputs cost about $ 200 US has ~ 8 mb ram and os on prom cost ~$200 maye it is time ofor a linuxEFI list? and guys WATCH THE TRIMMING no one wants to see the whole digest posted again and it makes the reply difficult to find This would be a great project, but by the time you're done you could > easily have bought an Electomotive or MoTec setup. Unless Lego comes > out with EFI sensors for the Mind Storm....... :) that would be just too way cool espcially if you printed up a giant vinyl sign that said "Powered by LEGO" and stuck it on your vehicles' rear fascia Clive > ------------------------------ Date: Thu, 02 Dec 1999 17:39:36 -0400 From: Bill Shaw Subject: Re: DIY_EFI Digest V4 #671 Jon Watts wrote: > > What I'm surprised at is that there's not more talk of Linux on this > list. Guess you missed the EFI332 mailing list, that's where the Linux stuff is. Bill ------------------------------ Date: Thu, 2 Dec 1999 14:54:45 -0800 (PST) From: Carter Shore Subject: Re: AC Klutch The horsepower is not so important as the torque that the clutch is rated for. The typical A/C units do not suffer clamping force degradation like typical the driveline diaphragm clutches. There is also the matter of the inital start torque. A/C units must be able to endure a large startup torque, because there can be a substantial residual gas pressure in the compressor. A centrifugal supercharger on the otherhand, probably has very low initial torque requirements. You might need to use some kind of 'check valve' or flapper system in the intake tract. > > ------------------------------ > > Date: Thu, 02 Dec 1999 10:09:18 -0500 (EST) > From: Andris > Subject: Re:KLUTCH!!! > > > Date: Thu, 2 Dec 1999 01:50:53 EST > > From: A70Duster@xxx.com > > Subject: KLUTCH!!! > > Could one use an AC clutch to engage and disengage > the impeller. At idle and > > cruise, the impeller doesn't turn, but for gettin' > giggy, fire up the > > "blower" and blow past the NIGHT RIDER!!!!! > > > > Can an AC clutch "hold together" when attached to > a blower??? > > Granted, it would be a cool gizmo, and all, but I > really don't think > you'll gain anything by it. At the lower rpms and > cruise, the blower isn't > really working too hard, so your parasitic losses > would be quite small. If > you still want to do it, I doubt the AC clutch would > hold. Making a WAG, > I'd say the AC compressor might take max 20hp to > turn (guessing since my > timeslips are .2sec and 2mph slower w/ the AC). At > full boost, I've heard > === message truncated === __________________________________________________ Do You Yahoo!? Thousands of Stores. Millions of Products. All in one place. Yahoo! Shopping: http://shopping.yahoo.com ------------------------------ Date: Fri, 3 Dec 1999 00:22:18 +0100 From: "Espen Hilde" Subject: Re: Transients Hi All! I think transients is the weak link in EFI systems to day. I dont know what the Indy or F1 guys are doing but the regular EFI system is more conserned about steady state emisions . In a realy fast car you can allmost trow away every sensors exept TPS witch gives input for acceleration mode. The acceleration enrichments lasts longer than going from one gear to another, in the lower gears offcurse. For the efi controller to try to estimate how much air is trapped inside the cyl with cam with lots of overlap is like trying to estimate the lottery numbers. Look at the 2 strokers they have great problems using efi, especialy with expansion cambers. One cyl that is not ignited pours the unburnt mix into the expansion camber ,the expansion camber cools of and gives no socktion wave.The result is to rich mixture next time around. The new 2 stroke direct injected outboards is running very nice at idle much like a 4 stroke...... The air trapped inside the cyl on a 4 stroke with much overlap is much dependent on the pressure waves.If a syl is not ignited ,maybe the mix is ignited into the exaust system, it will change the pulses and the enviroment compleatly. All OEM manufacturers stay away from this "field" Relaying on TPS on this self EGR state is stabilacing tings .Maybe using sequential EFI you are stopping some fuel going out into exhaust and ruining the steady pulses. I think the solution is to measure each cylinders air consume and maybe how much that is dumped out again , and inject the apropriate amount of fuel . Another way is to have variable cam system or control the air with additional electronical valves that is capable of controlling the air flow and not the high compression pressure.Maybe read valves in the inlet is the DIY aproch.Then you have widend the dynamic range of the valve system so you dont get self EGR. I think the enrichments should have different output from where in the map you are acclerating from , acclerating from leanburn aria is different than from a rich aria.Maybe a accleration map..... This is whats called rambeling isnt it?........ Espen Hilde (used to wearing the CSH , its not that ugly......) Flame on! ------------------------------ Date: Thu, 2 Dec 1999 18:25:58 -0500 From: "Clare Snyder" Subject: Re: KLUTCH!!! Exactly what Toyota did on the Supercharged MR2. The clutch is engaged when power requirements excede a given level - sensed I believe by throttle position vs RPM. Worked fine for Toyota, so no reason why it can't be done. > Date: Thu, 02 Dec 1999 10:09:18 -0500 (EST) > From: Andris > Subject: Re:KLUTCH!!! > > > Date: Thu, 2 Dec 1999 01:50:53 EST > > From: A70Duster@xxx.com > > Subject: KLUTCH!!! > > Could one use an AC clutch to engage and disengage the impeller. At idle and > > cruise, the impeller doesn't turn, but for gettin' giggy, fire up the > > "blower" and blow past the NIGHT RIDER!!!!! > > > > Can an AC clutch "hold together" when attached to a blower??? > > Granted, it would be a cool gizmo, and all, but I really don't think > you'll gain anything by it. At the lower rpms and cruise, the blower isn't > really working too hard, so your parasitic losses would be quite small. If > you still want to do it, I doubt the AC clutch would hold. Making a WAG, > I'd say the AC compressor might take max 20hp to turn (guessing since my > timeslips are .2sec and 2mph slower w/ the AC). At full boost, I've heard > that superchargers take 75+hp to spin. You would need something a bit > stronger. A small puck/disc-style clutch (like in the drivetrain) should > hold. Maybe something from a racing bike? Anyways, I've rambled enough, > and don't think you'll see much gain. > > Andris Skulte > Z28tt - 89 IROC T56 DFI Twin Turbo ------------------------------ Date: Thu, 2 Dec 1999 19:02:36 -0500 From: "Jonathan Davis" Subject: Re: KLUTCH!!! > > Could one use an AC clutch to engage and disengage the impeller. At idle and > > cruise, the impeller doesn't turn, but for gettin' giggy, fire up the > > "blower" and blow past the NIGHT RIDER!!!!! Another way to do this, perhaps would be a large bypass valve that recirculates any boost until WOT. Something like that would be intoxicating to drive - nice fat kick in the pants at WOT... Jon Davis ------------------------------ Date: Thu, 2 Dec 1999 19:19:52 -0500 From: "Clare Snyder" Subject: Re: transients > Date: Thu, 2 Dec 1999 10:59:46 -0600 > From: "G. Scott Ponton" > Subject: Re:Transients > A few things to consider. Stoich is the point where all of the fuel inhaled by the engine is combined with all of the available oxygen in the cyl. Any leaner means less than optimum power. Any richer means less than optimum power per lb of fuel used, as some is not burned, as well as less power overall. There is a fine line between too lean for maximum power and too rich. The transition from Idle to full throttle throws more than one monkey-wrench into the works. Many have been discussed here already. The ones I see are, in no particular order: 1] Reversion -at low rpm when the throttle is cracked open, cam design can cause/alow some compression back into the intake. This dilutes the intake charge, as well as weakening the venturi signal and pushing the fuel back the runner, delaying it's ingress to the cyl. All the way around this means less A/F mixture in the engine, and at a reduced strength. 2] Inertia - OK, I've got my Nomex on - I got flamed for this once before. The density of gasoline as compared to the density of air? well let's say there is no comparison. The gasoline is MUCH heavier than the air, so has much more inertia. Crack the throttle open and you "instantly" get more air. The fuel follows - quite a bit behind. This causes the mixture to weaken again. 3]Gravity - OK, not gravity in the strictest sense of the word, but hey, I've got my Nomex on. The fuel tends to "settle out" of the mixture in the runners at low speeds. This wets the walls of the runners, weakening the mixture during transition - then when the airspeed catches up and peels the fuel off the walls it causes a rich event. Hard to compensate for. The solution? In carbureted engine practice, an accellerator pump is used to squirt extra fuel into the airflow during rapid transitions. The amount squirted is determined by both the speed, depth, and duration of the acceleration with the help of a clever "Computer" consisting of a lever, spring, check-valve and calibrated bleeds. On a CV carb this is replaced by a "damper" which acts to slow down the increase in air flow so the fuel can catch up. At best the accellerator pump or damper are a compromise - a "best guess" arrived at through lots of dyno testing, road testing, slide-rule slipping, and other Voodoo. IR systems get away from SOME of these problems - but not all. Fuel injection can aproach the problem's solution in several ways. The inertia and gravity can be delt with the same way IR systems attempt to deal with it. Shorten the path the fuel must travel with the air by employing port injectors. TBI buys you less advantage, reducing the inertial effects prior to the venturi by providing fuel under pressure, but not addressing the "gravity" or the inertia in the intake tract. To mimic the action of the fuel pump EFI can take several tacks. On the L-Jet based Nippondenso systems employed by Toyota in the eighties there was an extra pulse of fuel provided every time the throttle opened past so many degrees - basically off-idle transition. If a regular pulse at idle was 6 ms long, you got an extra 6 ms worth of fuel. Mixture was also controlled by rate of change at the airflow sensor, stretching the pulse under high airspeed acceleration. The Throttle position sensor also told the computer the engine was in one of, if I remember correctly, 4 operating modes - idle, off idle or low power, cruise or medium power, and full power At idle mixture was tailored for smooth idle and reasonable off-idle transition - read that as just rich of "stoich" -( untill emission reduction dictated otherwise - late, hot, and lean, withthe accompanying driveability problems) At off idle and low power, mixture was tailored for emissions and economy - read that as weak mixture. At cruise/medium power slightly richer for better power production - and at full power "pig rich" - as muich fuel as could be pumped in without causing a decrease in power. Like the carb, this was a "best guess" aproxemation. With todays EFI, we have MAP, or load sensors, artificial inteligence in the CPU, and closer control of other variables like ignition timing, and even intake runner sise and length, valve timing , lift, and duration, etc. A DIY system can meet many of your rquirements, but will tend to be closer to a carb in it's precision.You can have power, economy, simplicity,and driveability - pick any two. With a lot of work you can pick 3. By modifying an existing system, a dedicated, inteligent experimenter may be able to get all 4 by tuning out the compromises required to "satisfy the majority of people most of the time". Read that as give up something that does not matter to you. The application I am thinking of does not suffer much from transitions, so dirt simple and rock solid are my goals. Economy and power are important too, but driveability - the hardest to acheive- is secondary at best. Today's EFI > > > First the disclaimer. I tend to speak in a way that some people have a > problem with. I can come across as being condescending and the like. This > isn't my intention so I hope you will all take this for what it is worth. I > am putting on my flame retardant suit and placing my CSH nearby just in case > they are needed. Snippp. > I have used up enough bandwdth for now. Obviously this is a very > simplistic view of all the variables and how they interact. Hope it will > help some. > Back to lurk mode. > > Scott > > ------------------------------ > > End of DIY_EFI Digest V4 #671 > ***************************** > Ditto ------------------------------ Date: Thu, 2 Dec 1999 18:32:58 -0600 From: "Ken Kelly" Subject: ECU's and HEI Out of curiosity, I've been trying to think of ways to manage an engine by building my own ECU. Obviously, this is an arduous task, but one I'm not afraid to embark on. I started off by figuring out the inputs and outputs and determined that I could do it with a few of the new high-speed BASIC Stamps from Parallax, but lately I'm beginning to wonder if it can't be done (easier) with an analog system - i.e. transistors/diodes/resistors/capacitors and have pots for tuning? I believe this is the way Holley does most of their systems aside from their 4D and 4Di systems. My main big question, however, is HOW do the HEI electronic systems work? How does the computer adjust retard and advance? Is it a signal varied by voltage, frequency, or what? I'm at a total loss for this concept. The other parts of the engine I have figured out (injectors, MAP, TPS, O2, etc.) but I can't quite see how the timing works on the electronic versions of the HEI. (I see it has a seperate 4 wire interface?) Part of my reasoning for making my own ECU is simply for the fun of it and partly because I'm going to be making some of my own components - I'm looking into making a VFI (Venturi Fuel Injection) unit out of a Quadrajet (similar to the Barry Grant VFI unit made from a Holley carburetor body.) I'm also looking into how to control and monitor other aspects of engine management - i.e. fuel pressure. Thanks for any help.. The "Other" Ken Kelly ken73@xxx.com home.houston.rr.com/ken73 ------------------------------ Date: Thu, 02 Dec 1999 22:58:18 -0500 From: goodall@xxx.com (Colin Goodall) Subject: RE: QNX Floppy Tools? ==> Linux I've been lurking away here on the list for a while now. So I thought it was time to chip in. I have a 1982 CX500 Turbo (motorcycle). It has its own EFI but its totally black-box. no info available whatsoever on how to "improve" it. So I've come to the conclusion that my best bet would be to build a new computer. > What I'm surprised at is that there's not more talk of Linux on this > list. By itself it could handle an EFI system and host it's own > analysis software. With real-time Linux (www.rtlinux.org) this is now > a certainty. Sooner or later someone will build a PC-104 (or smaller) > EFI unit. This is one option I seriously considered. right off the shelf you can buy PC104 processors anywhere from an 80386 and up to current PIII systems. There are a vast array of "I/O" cards available. including cards which contain solenoid drivers (just whats needed for firing the injectors". I figured you could build the whole thing out of off-the-shelf components and run a linux OS on it. (Or even Windows NT if your brave, or Windoze 95 if your foolish!) linux would be a great choice. Easily extensible and tons of support if you need to tweak the kernel. And lots of tools available. Their's even some companies flogging versions of linux with a real time kernel. But if you had a fast enough processor Pentium series processor it wouldn't much matter since it would just simply run more than quick enough. Another nice feature would be using a CPU card with video for very very easy debugging of the software. Once your up and runnning just use a plain vanilla CPU card with no video. (But most modern CPU cards come with EVERYTHING!, serial parallel, accelerated video, keyboard and mouse, 10baseT. Forget the internet toasted, you've got an internet internal combustion engine!) If you mount the computer inside the vehicle you could plug in oone of these new flat-screen LCD monitors and mount it on the dash. Monitor the entire system while your driving! Obviously using PC104 opens up an enormous number of possibilites. But there are 2 things to consider. 1) This setup would be a hackers paradise (for me no problem), but could be a bit much for most of the people reading this list. It strikes me that most participants on this list are more experienced on the mechanical side, while you are learning and working with the EFI you are not programmers to begin with. So the learning curve would be prohitively steep. 1) Cost. Unless you are going to buy enough to build a couple hundred units these cards can be expensive once you add up eveything you need. > This would be a great project, but by the time you're done you could > easily have bought an Electomotive or MoTec setup. Unless Lego comes > out with EFI sensors for the Mind Storm....... :) Yes, exactly! Hope this helps Colin +--------------------------------------------------------------------+ | Colin Goodall | | | goodall@xxx.com | "Sometimes you're the windshield, | | http://www.magma.ca/~goodall | Sometimes you're the bug" | | (613) 258-3831 | | +--------------------------------------------------------------------+ ------------------------------ Date: Thu, 02 Dec 1999 22:02:42 -0600 From: Tom Sharpe Subject: Re:KLUTCH!!! > Date: Thu, 02 Dec 1999 10:09:18 -0500 (EST) > From: Andris > Subject: Re:KLUTCH!!! > > > Date: Thu, 2 Dec 1999 01:50:53 EST > > From: A70Duster@xxx.com > > Subject: KLUTCH!!! > > Could one use an AC clutch to engage and disengage the impeller. At idle and > > cruise, the impeller doesn't turn, but for gettin' giggy, fire up the > > "blower" and blow past the NIGHT RIDER!!!!! > > > > Can an AC clutch "hold together" when attached to a blower??? > > Granted, it would be a cool gizmo, and all, but I really don't think > you'll gain anything by it. At the lower rpms and cruise, the blower isn't > really working too hard, so your parasitic losses would be quite small. Just use a Turbo - it's power input is exhaust heat - when crusing, very little loss, but when the throttle is down ..... TWS ------------------------------ Date: Thu, 2 Dec 1999 23:15:14 -0500 From: "Andris Skulte" Subject: Re:Transients > > First the disclaimer. I tend to speak in a way that some people have a > problem with. I can come across as being condescending and the like. I seem to have the same problem sometimes. He he he - I've been apologizing and explaining myself way too much lately :/ > control engine speed. As we go to WOT there is a transition > period equal to > the lenth of time it takes for the air, at atmosferic pressure, > to move past > the throttle and reach the cylinder. Regarding the original post - how do you know its a lean miss, rather than an over-rich condition? This is how I'm thinking of it - You're at high rpm, closed throttle, 20+ in Hg vaccuum. You go WOT, and almost instantly the manifold goes to zero vacuum, which your MAP sensor reads "instantly". Your acceleration enrichment circuit sees this, and richens it above and beyond the regular WOT conditions "instantly". There is a ton of fuel in the intake port, for the first cycle or so, while the slug of air travels down the runner. Could it be likely that your electronics systems respond faster than the physical air, and over-richen it? Andris/Z28tt ------------------------------ End of DIY_EFI Digest V4 #672 ***************************** 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".