Wow. Talk about tenacity. Good show, dude. I hope it works out
for you.


"Don Bruder" > wrote in message
...
>
> I posted a couple-few weeks ago about my '82 Mazda 626 having
> intermittent issues with losing power under load - At that
> time, figured
> that the trouble was related to a dead or dying Alternator not
> putting
> enough juice to the battery, and therefore interfering with
> spark -
> Seemed to make sense, since when the alternator died and I put
> the new
> one in, the problem seemed to have vanished.
>
> Well, the good news is that it stayed gone for at least 3-4
> weeks. The
> bad news is that it started showing up again last week -
> Symptoms same
> as ever: Hit the hill near Lakeside Market, and it starts
> chugging and
> lurching and bucking like it's right on the edge of being out
> of gas,
> even after just topping up the tank less than 10 minutes/10
> miles back.
> Well, last night, it hit with a vengance... I hit the Lakeside
> hill at
> just under 45. Half a mile up, I was down to 30, and lurching
> and
> bucking like a rodeo bronco, with another mile or so to go
> before I hit
> flat land. Another couple hundred yards along, I was down to
> 15. 10
> seconds later, I was dead on the side of the raod, and no
> amount of
> cranking would get it to start. Rolled it backwards to point
> the car
> downhill and coast back down to Lakeside Market where I could
> call a
> tow-truck. Got to the bottom of the hill, and out of sheer
> "Dammit, this
> thing SHOULD start!", I hit the key. It spun over instantly,
> humming
> along just as strong as it always has. OK, cancel the tow truck
> idea...
> Maybe I can get it home under its own power? Well, long story
> shorter, I
> did indeed manage to get it home, luurching and bucking like
> crazy every
> time I had to climb a hill, but somehow keeping it running.
> Flat or
> downhill stretches, it behaved just like normal - Namely,
> running
> strong. Uphill for more than a second or two? Chug, lurch,
> threaten to
> die. This sure sounds like a fuel issue, but everything I can
> check
> shows green and go.
>
> Problem has always been worst when travelling uphill forwards,
> with the
> occasional "on the flat" appearance when I was "getting on it"
> heavy for
> whatever reason. Could never duplicate the "dropouts" while
> standing
> still - It never happened with the car moving less than 20 MPH,
> which
> made it seriously difficult to try to troubleshoot. Part of the
> "shotgun" I fired at it was plugs, wires, rotor and cap - They
> were
> needed, at least according to the calendar (even though they
> looked fine
> to me when they got replaced) - That seemed to make a
> difference for a
> short time, but the problem returned. Couldn't find any vacuum
> leaks
> using the propane torch method. Fuel filter replaced multiple
> times.
>
> Fuel-pump test showed it was giving spec pressure, and a hair
> more than
> spec volume (Specced for 3-5 PSI, replace at 2.8PSI, Spec
> volume of 800+
> ml per minute, replace fuel pump at 750 ml/min. Actual measured
> PSI:
> 4.1. Measured volume: 825 ml/minute)
>
> Compression is rock solid, at close enough to 150Kg/cm^2 on all
> four
> cylinders that my compression tester isn't sensitive enough to
> show me
> any variation (Spec: 135-180 Kg/cm^2, < 10% variance between
> highest and
> lowest cylinder)
>
> Finally, in desperation, today I lifted the top of the carb,
> going on
> the "It's acting like it's starving to death" hunch - I've
> eliminated
> everything else that makes even a tiny bit of sense, and
> diddled with a
> few things that shouldn't have anything to do with it.
>
> Got the carb open, and I'm measuring the float setting - It's a
> plastic
> float on a metal bracket, and it's so close to the book specs
> that I
> can't measure any error in the setting. But then I get to
> taking a
> closer look...
>
> The float rides on a "slides all the way through",
> free-floating pin -
> The pin is kept in place by the fact that once the top is
> screwed down,
> the area it sits in is too narrow to allow it to back out of
> the holes
> it goes through. That part is fine...
>
> What ISN'T fine is the no-longer-round holes in the float
> bracket!
> After dinking with the float adjustment a bit, I noticed that
> dropping
> the float didn't appear to open the needle valve. Started
> wondering
> about that, and got to looking closer at how the float was hung
> on the
> posts it attaches to, and that's when I noticed that the top of
> the
> holes in the bracket have worn to the point where they're no
> longer
> circular - They're very nearly twice as tall as they are wide!
>
> Pulled apart a spare carb I have for this beast, and found that
> its
> float had nice, tight, nearly perfectly round holes at that
> location.
> And when THAT float is moved up and down, there's a very
> clearly visible
> motion of the needle. Swapped floats, buttoned everything up,
> and took
> it for a test run. Sure enough - Absolutely no trace of the
> problem now.
>
> My "working theory": The worn holes were allowing the float to
> "sink"
> without actually tilting, which kept the needle from opening,
> and in
> turn, gave every sign of the car being out of gas. Hardly
> surprising,
> since for practical purposes, having the float bowl run dry is
> ideintical to having the gas tank empty. as for why it was
> intermittent,
> I'd say that the problem could only manifest when the vehicle
> was in
> certain positions (nose up versus nose down, ferinstance) so
> that when
> the float "fell" just right to bind slightly, it would function
> as
> intended, but when it got into another orientation, the
> "extra-large"
> holes would "un-bind" and let the pivot point fall, dropping
> the float
> without opening the valve.
>
> I'll be keeping an eye on the beast, but I'm pretty sure I've
> finally
> found the actual root problem - Damn, but it has taken long
> enough!!!
>
> --
> Don Bruder - - New Email policy in effect as
> of Feb. 21, 2004.
> Short form: I'm trashing EVERY E-mail that doesn't contain a
> password in the
> subject unless it comes from a "whitelisted" (pre-approved by
> me) address.
> See <http://www.sonic.net/~dakidd/main/contact.html> for full
> details.