Log from 2019-10-27:
[01:04:09] *** Joins: physkets (~physkets@unaffiliated/physkets)
[09:25:52] *** Quits: physkets (~physkets@unaffiliated/physkets) (Quit: physkets)
[10:03:41] <ljrbot> New news from bzr: [0.4-armagetronad] r3113 Merge branch 0.2.8 into 0.4 || [0.4-armagetronad] Merge branch '0.2.8.3' into 0.2.8... || [0.2.8-armagetronad-work] r1525 Merge branch '0.2.8.3' into 0.2.8...
[10:33:44] <ljrbot> New news from bzr: [0.2.8-armagetronad-work] r1526 Merge branch '0.2.8.3' into 0.2.8 || [0.2.8.3-armagetronad-work] r1228 Explicitly request python2 for sortresources.... || [0.2.8.3-armagetronad-work] r1227 Add .gitignore || [0.2.8.3-armagetronad-work] r1226 Cherry pick merge from 0.2.8 of build fixes...
[11:03:50] <ljrbot> New news from bzr: [0.2.8.3-armagetronad-work] r1230 Make binreloc work again... || [0.2.8.3-armagetronad-work] r1229 Disable some GCC warnings
[11:50:00] *** Joins: Armanelgtron (~AmarokNel@unaffiliated/amaroknelg)
[11:50:01] *** Server sets mode: +cnrt
[11:50:04] *** Server sets mode: +cnrt
[11:55:10] <Lucifer_arma> seems to me that you can make a fuel injection system using only a throttle position sensor and an oxygen sensor
[12:16:11] <Z-Man> Are there oxygen sensors fast, reliable and cheap enough for that?
[12:18:40] <Z-Man> And durable. Must be able to still work after being exploded repeatedly, essentially.
[12:19:40] <Z-Man> (I admit that I don't even know how an oxygen sensor would even work, the only thing I can think of would be spectral analysis.)
[12:19:57] <Z-Man> Which at least deals with the 'fast' part.
[12:25:59] <Lucifer_arma> yes, they put them on cars all the time :)
[12:26:05] <Lucifer_arma> ]math calc 360/8
[12:26:05] <ljrbot> Lucifer_arma: 45
[12:26:51] <Lucifer_arma> they have a 1V drop when your fuel/air mixture is pretty close to stoichometric. Higher and it's rich, lower and it's lean
[12:27:36] <Lucifer_arma> so, you adjust the time the fuel injector is open to change the fuel, and, in my thinking, you'd use the throttle position sensor + a model of the cross section in the throttle to determine how much air is coming in
[12:28:12] <Lucifer_arma> what I'm finding on scooters and motorcycles, though, and other small engines, is that they use a manifold air pressure sensor and an intake air temperature sensor to determine the mass of air going in
[12:28:36] <Lucifer_arma> but it seems to me that you can take the volume from a decent model of the throttle and tweak it based on the oxygen sensor readings
[12:29:29] <Lucifer_arma> so, you'd have, say injectorinterval(angle) = (some coefficient) * (calculation of cross section) * angle
[12:29:55] <Lucifer_arma> angle would be given by the throttle position sensor, and you'd adjust "some coefficient" based on the output from the oxygen sensor
[12:30:37] <Lucifer_arma> the calculation of cross section is a function I have to derive, but it'll have the area of a circle in it somewhere
[12:30:47] <Lucifer_arma> or ellipse, something like that
[12:31:31] <Lucifer_arma> because inside the throttle body, you have a disc that rotate when you open the throttle. When it's closed, it cuts off 99.99% of the air intake, and there's a little passage to allow air to pass so the engine can idle
[12:32:24] <Lucifer_arma> so, as you rotate that disc, you get an ellipse of increasing size - area of the disc that represents the area that air can flow through to the engine
[12:32:50] <Lucifer_arma> you also have the crankshaft position sensor, which is already on the engine I'm talking about, so you know not only what position the crankshaft is in, but also how fast it's turning
[12:32:59] <Z-Man> Neat geometric fact for you there: the shape of the disk does not matter.
[12:33:32] <Z-Man> your formula is just going to have a (1-cos(angle)) factor in there.
[12:33:40] <Lucifer_arma> so, I'm thinking that you can crib an air mass measurement based on the angle of the throttle, the engine speed, and the inner diameter of the throttle body
[12:34:03] <Lucifer_arma> why not just do sin(angle) ?
[12:34:26] <Lucifer_arma> the angle is going to be [0,90]
[12:34:31] <Z-Man> because that would be wrong :)
[12:34:40] <Lucifer_arma> in degrees, of course. Actually, the input coming in would be in volts :)
[12:35:06] <Z-Man> well, assuming angle 0 means its fully closed, angle 90 is fully open.
[12:35:16] <Lucifer_arma> heh, well, I'm not on the math part yet, I'm still trying to determine what I'd have to purchase just to get started :)
[12:35:50] <Z-Man> I'm definitely the wrong person to give advice there :)
[12:36:25] <Lucifer_arma> not really, the ratio we're looking for is like 14.7:1 of air:fuel
[12:36:51] <Lucifer_arma> so the question is: what sensors do I need to determine the mass of air coming in?
[12:36:59] <Lucifer_arma> I control the mass of fuel
[12:37:27] <Lucifer_arma> the basic calculation is (air mass)/14.7 = fuel mass
[12:38:06] <Lucifer_arma> but fuel mass is easy. The fuel comes from the tank pressurized to like 35psi, and I know the cross section of the fuel injector (it's a circle :) )
[12:38:26] <Lucifer_arma> so volume/time open = total volume, then volume * density = mass
[12:38:39] <Lucifer_arma> wait, volume * time open
[12:38:43] <Lucifer_arma> anyway, you get the idea :)
[12:39:13] <Lucifer_arma> typically, though, the throttle position sensor is used to determine what the driver wants, because the actual mass of air getting pulled in is dependent on engine speed
[12:39:35] <Lucifer_arma> so if I use the TPS to measure air mass, I'm only directly measuring the driver's desired air mass
[12:40:20] <Lucifer_arma> so what you do there, then, is you make the mixture slightly rich to increase engine speed when the throttle is opened, and you lean it to decrease engine speed when the throttle is closed
[12:41:16] <Lucifer_arma> so I guess the real question is, can I come to a correct air/fuel mixture using only the oxygen sensor as input?
[12:41:52] <Lucifer_arma> it'd be a lot easier to answer these questions if a simulator existed for all of this. :)
[12:42:13] <Lucifer_arma> you know, a simulator I could plug into an arduino simulator, so I can use the arduino simulator to run the code I write
[12:43:14] <Z-Man> Such a thing is bound to exist, only not affordable and certainly not OS.
[12:43:19] <Lucifer_arma> anyway, back to trying to tune my carburateur, which is what's got me thinking about building an EFI system for this thing :)
[12:43:34] <Lucifer_arma> ah, that's correct. The simulators that are out there are proprietary and owned by the engine manufacturers
[16:22:35] *** Joins: kevinh (kevinh@fsf/member/kevinh)
[17:29:24] <Lucifer_arma> ]math calc 0.03*1500
[17:29:25] <ljrbot> Lucifer_arma: 45
[17:29:38] <Lucifer_arma> $45 for a 1500 word article? Content mill. :(
[21:20:05] <Lucifer_arma> ]math calc 11*60
[21:20:05] <ljrbot> Lucifer_arma: 660
[23:00:21] *** Quits: Z-Man (~Z-Man@2002:5b32:6ed5:0:553f:d39e:1ef4:fe86) (Ping timeout: 246 seconds)
[23:05:05] *** Joins: Z-Man (~Z-Man@p4FE3EC16.dip0.t-ipconnect.de)
[23:52:12] *** Joins: physkets (~physkets@unaffiliated/physkets)
View entire month
DISCLAIMER: These logs of public chat may contain some content which may not be appropriate for all audiences. Use at your own risk.
Logs from 2006-2009 pulled from wrtlprnft
Format changes at: 2015-08-25, 2017-02-20, and 2020-03-23. Times (2015 and later) should be Eastern.