Daily WTF

[AbeFM]

I'm waiting till I see what you get for jitter. I think you should average whenever you can. RPM's weight on any function of actual interest is generally small, small enough the gains in removing jitter far outweigh the problem of having RPM that's 19 ms too late (6024 rpm verses 6027?? Beats having an error of 217 because you were 3 us off in reading a tooth)

As it is, folks run miata motors on a 4-minus-1 wheel.

[Fred]

Crank speed *DOES* vary a LOT during a single cycle. That is what was causing his issues, but given that knowledge there must be a clever way to get the latest speed out of it without jitter. I'll work on that fairly soon.

[Fred]

Gold, someone I know, one of my mates, seriously asked me this today :

w): i was thinking
w): about using pvc pipes for coolant
w): cos it's so easy and cheap
w): but i'm wondering whether pvc pipes (like household type waste pipes) will be good enough
w): cos they would be very clean to install...

My response :

Fred: tip some boiling water on one and answer your own questions

When quoted to someone else I know I got a genuine :

s: coolant????
s: WTF?

At which point I knew that I had to post it here! Then the secondary discussion degraded into the merits of using coat hangers as welding wire and coke bottle gasses as shielding gas. We agreed that if you shook the bottle too much you wouldn't get much welding done...

Fred.

[Fred]

http://www.ms3efi.com/index.php

The last post was over 2 weeks ago, I wonder if the project is still even being worked on? April first I gave them a kick start by launching this site, considering there are far more of them, and far more experienced and knowledgeable than me, why is it taking so long? No bench testing the code on prototype card vids, no prelim board designs, no software spec, no hardware spec, nothing much so far really. I'm a bit disappointed in the lack of action, despite this, I'm still kinda keen to see the wonderful new MS3. I've been a MS fan for a long time after all, and old habits die hard. Come on guys, WTF is the status?

Fred.

[AbeFM]

http://video.google.com/videoplay?docid ... 1591&hl=en

This one shows the MS-II acting up - I ran the "same" test twice, got much more fuel the second time, obviously there were many more pulses. Certainly better off than I would be without it, but nowhere near as reliable as my turning the RPM knob by hand on the jimstim.

Could you use an MS-II for figuring out open times and flow rates for an injector? Probably, but it would take a lot of averaging from many runs.

[Fred]

So, as discussed over AIM earlier, when do you get started building your injector test rig based on the TA card and using my code as a base?

Fred.

[jbelanger]

Could you use an MS-II for figuring out open times and flow rates for an injector? Probably, but it would take a lot of averaging from many runs.

Yes and not necessarily. If you do a few test runs with a good number of squirts, things should be averaged out within the run. It then depends on how precise your fuel measurement is and which test fluid you use (influences how much evaporation you get).

I have done the maths before and posted it on an MS forum. Let me see if I can find it and I'll repost here.

Jean

[Fred]

Jean, his errors were huge unless he was exaggerating to me earlier :

(06:47:14) AbeFM: 800 injections. Run it five times
(06:47:46) AbeFM: 24 cc, 24.1 cc, 58cc, 23.9cc, 37 cc
(06:47:59) AbeFM: Damnit it ALMOST makes sense.
(06:48:23) Fred: LOL

Fred.

[jbelanger]

Here it is straight from the MS forum (but I'm the author anyways):

The best method, for the purpose of tuning the injector response, is to measure the fuel flow at different pulse widths and compute the dead time. You could do 1000 pulses (or whatever is practical depending on your setup because it will be factored out at the end) of a certain pulse width (say 5ms), measure the fuel injected, then do 1000 pulses or another pulse width (say 10ms), and measure the fuel injected. Then you compute the dead time. So you would have:

1000 x Inj x (PW1 - DT) = MF1
1000 x Inj x (PW2 - DT) = MF2
=> (PW1 - DT) / MF1 = (PW2 - DT) / MF2 => DT = (MF2 x PW1 - MF1 x PW2) / (M2 - MF1)
where
Inj = injector size (not important because it's factored out)
PWx = pulse width for the test
MFx = measured fuel for the test (units are not important as long as they are consistent)
DT = injector dead time (instead of calling it opening time)

This will give you the dead time. However, once you have it you can use the same measurements to compute the flow rate. The equation is left as an exercise.

Jean

p.s.: Fred, feel free to move this to a more appropriate place since it doesn't really fit here. Sorry for the threadjack.

[Fred]

It's not really thread jacking, it's on topic with regards Abe's daily wtf entry :-) If you want you could do an article on it if you think you could bulk it up a bit more.

It still doesn't explain his results if he wasn't just making up figures. I guess it could be his hardware, but my gut instinct in this case is to... well... just scratch my head really. It's hard to comprehend how it could be that wrong at something that simple to do...

Fred.