Discussion:
Exhaust gas analyzer have any value in tuning a motorcycle carb?
(too old to reply)
mike
2017-08-14 09:43:58 UTC
Permalink
I'm still tweaking this Kawasaki Eliminator 125
to get rid of the detonation.

Somewhere in the attic, I have an ancient exhaust gas analyzer.
IIRC, it works by passing an infrared light thru the exhaust
and divines air/fuel ratio from that.
Was made back when there was lead in the gas.

Does that method have any value with current engine/fuel
designs?
Mark Olson
2017-08-14 17:45:47 UTC
Permalink
Post by mike
I'm still tweaking this Kawasaki Eliminator 125
to get rid of the detonation.
Somewhere in the attic, I have an ancient exhaust gas analyzer.
IIRC, it works by passing an infrared light thru the exhaust
and divines air/fuel ratio from that.
Was made back when there was lead in the gas.
Does that method have any value with current engine/fuel
designs?
Sure. Assuming it still works, it has no way of knowing whether you
are using a carb or fuel injection.

Are you sure your detonation problem is definitely down to the mixture
being too lean? Ignition timing would be the first thing I would
check. And before anyone says if the timing is right at idle it must
be OK at higher RPMs, I would not assume that.

I had a weird issue with a 1995 Kawasaki EX500 ignition module (K calls
it the "IC Igniter") that had a timing problem between cylinders (one
was OK the other one was advanced by 6 degrees) even though the engine
only has a single ignition trigger coil, and the magnetic reluctor
pole pieces on the rotor were exactly as they should have been.
Replacing the IC Igniter did in fact fix the problem (verified with
an O-scope).

If this was my bike I would be checking ignition timing at various
RPMs to make sure it's correct.
mike
2017-08-14 20:57:13 UTC
Permalink
Post by Mark Olson
Post by mike
I'm still tweaking this Kawasaki Eliminator 125
to get rid of the detonation.
Somewhere in the attic, I have an ancient exhaust gas analyzer.
IIRC, it works by passing an infrared light thru the exhaust
and divines air/fuel ratio from that.
Was made back when there was lead in the gas.
Does that method have any value with current engine/fuel
designs?
Sure. Assuming it still works, it has no way of knowing whether you
are using a carb or fuel injection.
Are you sure your detonation problem is definitely down to the mixture
being too lean? Ignition timing would be the first thing I would
check. And before anyone says if the timing is right at idle it must
be OK at higher RPMs, I would not assume that.
I had a weird issue with a 1995 Kawasaki EX500 ignition module (K calls
it the "IC Igniter") that had a timing problem between cylinders (one
was OK the other one was advanced by 6 degrees) even though the engine
only has a single ignition trigger coil, and the magnetic reluctor
pole pieces on the rotor were exactly as they should have been.
Replacing the IC Igniter did in fact fix the problem (verified with
an O-scope).
Can you give more details on how you set up the scope to make that
measurement?
I have an inductive pickup that clamps on the spark plug wire to
read the spark. How did you determine the TDC position?
I could intercept the crank position sensor, but that doesn't
tell me whether that signal is in the correct relationship to TDC.
If the timing light says the spark is at the T position on the
cam sprocket, there isn't much one can do.

The service manual has a detailed write-up on how the ignition works.
Has a nice graph of the advance curve, but it has no numbers about
the RPM of each inflection point. I'll take a crack at the timing
this afternoon.

And the bike has no tach. I have a timing light with a tach, but
it's for 4-8 cylinders and the resolution is useless for one cylinder.
Post by Mark Olson
If this was my bike I would be checking ignition timing at various
RPMs to make sure it's correct.
Thanks, I'll take all the advice I can get.

I've concentrated on carburetor because:
There's little written about this bike, but most of what exists
complains about it running dangerously lean.
The spark plug insulator was bone-chilling white.
It was detonating at speed ranges much above idle under acceleration.
Cruising was not a problem for detonation, but surging at high speed
was an issue.

It's a 2007 model with 423 miles on it. Apparently been in
storage for 5 years.
A treatment with SeaFoam and backing out the pilot screw fixed
the hanging idle, but it still detonates. Cranked the pilot screw
out 4 turns and still not much improvement in ping, but it feels/sounds rich
just off idle.

I cleaned the carb best I could. There was no gunk in it at all.
I was amazed at how clean it looked.
I wasn't too aggressive because
the manual warns about non-removable internal plastic parts.
They don't say where, but I decided not to use carb cleaner with
acetone or toluene. Stuck with simple green and water in an ultrasonic
cleaner.
I couldn't get the pilot jet out. Twisted as hard as I dared and decided
to come back another day. Low speed was ok.
I did hook up a vacuum pump and sucked fluid back thru the jets.
All the tiny holes in the venturi seemed to be open.

External fuel level measurement showed it to be at the low end of
the 4mm spec range. I moved it up to the top end. I was unable
to remove the float pin. It's pressed in and resisted careful
attempts to press it out.

At that point the performance was markedly improved. Spark plug
now has a little color. Still pinging
around 1/4-1/2 throttle accelerating thru the gears, but ok at cruise.

I raised the main jet needle .020", then .035". Both improved the
situation. I'll try one more increment to .050" before I give up.

Raised the main jet to 130. Improved the pinging. It's now
sounding rich at top end. Putting the fuel level back to
midrange might help that.

The performance of the 125 is crap, but probably the best I can
expect from a small engine. My concern is the pinging destroying
the engine. The pinging increases as the engine warms up.

I've been all over it with propane an detected no air leaks.
I experimented with blocking part of the air filter.
With 1/3 of it blocked, the bike would hardly run. Much greater
effect than I expected. I can't see how you'd ever keep the bike
in tune if it's that sensitive.

I'd agree that delaying the spark advance to a higher
RPM might fix the problem. I just don't think that's possible.
I'll take a timing light to it as soon as I get the tach built.
I'll drag out the gas analyzer and see what it tells me.
I have a Snap-ON automotive oscilloscope, but the screen resolution
is too low to see two sparks on screen to measure RPM. Probably
works great on an 8-cylinder car.

Probably time to twist on the pilot jet until it comes out or breaks.

I've tried a lot of stuff, but it should be obvious that I have no idea
what I'm doing. I've been driving the same Honda Shadow VT-500c for
over 30 years. Never had any issues with the CV carbs in that one.

Any additional advice welcome.
mike
Mark Olson
2017-08-14 22:51:23 UTC
Permalink
Post by mike
Post by Mark Olson
Post by mike
I'm still tweaking this Kawasaki Eliminator 125
to get rid of the detonation.
Somewhere in the attic, I have an ancient exhaust gas analyzer.
IIRC, it works by passing an infrared light thru the exhaust
and divines air/fuel ratio from that.
Was made back when there was lead in the gas.
Does that method have any value with current engine/fuel
designs?
Sure. Assuming it still works, it has no way of knowing whether you
are using a carb or fuel injection.
Are you sure your detonation problem is definitely down to the mixture
being too lean? Ignition timing would be the first thing I would
check. And before anyone says if the timing is right at idle it must
be OK at higher RPMs, I would not assume that.
I had a weird issue with a 1995 Kawasaki EX500 ignition module (K calls
it the "IC Igniter") that had a timing problem between cylinders (one
was OK the other one was advanced by 6 degrees) even though the engine
only has a single ignition trigger coil, and the magnetic reluctor
pole pieces on the rotor were exactly as they should have been.
Replacing the IC Igniter did in fact fix the problem (verified with
an O-scope).
Can you give more details on how you set up the scope to make that
measurement?
I "made" a scope by constructing a resistive voltage divider network
and connected it to the primary side of the coil, running it into the
sound card of an old laptop. Recording the "sound" of the waveform let
me look at the relationship of the spark pulse vs. the pulse from the
trigger coil (stereo input of the sound card = 2 channel scope).

I knew the timing was off simply by looking at the timing marks with a
inductive timing light from the 1980s, still works fine. I could see
the one cylinder was off by 6 degrees vs. the other one (accounting for
the fact that it's a 180 degree crank, naturally).
Post by mike
I have an inductive pickup that clamps on the spark plug wire to
read the spark. How did you determine the TDC position?
By shining the timing light on the timing marks. TDC is engraved on the
alternator rotor which is visible through the timing access port. The
correct degree position for timing at low RPM is also marked for both
cylinders.
Post by mike
I could intercept the crank position sensor, but that doesn't
tell me whether that signal is in the correct relationship to TDC.
I took a photo of the position of the timing marks when the timing
light fired and measured the angle graphically.
Post by mike
If the timing light says the spark is at the T position on the
cam sprocket, there isn't much one can do.
Sure- you can check to see how far the mark advances as you go up in
RPM. You might have to paint/engrave some of your own marks on the
flywheel to do this.
Post by mike
The service manual has a detailed write-up on how the ignition works.
Has a nice graph of the advance curve, but it has no numbers about
the RPM of each inflection point. I'll take a crack at the timing
this afternoon.
Yeah, without that info, or having some general knowledge about how
far timing should advance (I have a buddy who is an absolute expert
on this stuff) you're going to have to ask or do some research to
figure out what sort of numbers are reasonable.
Post by mike
And the bike has no tach. I have a timing light with a tach, but
it's for 4-8 cylinders and the resolution is useless for one cylinder.
Tach is easy- just measure the time interval between trigger coil
pulses and take the inverse to get Hz, then divide by 60 to get RPM.
Post by mike
Post by Mark Olson
If this was my bike I would be checking ignition timing at various
RPMs to make sure it's correct.
Thanks, I'll take all the advice I can get.
There's little written about this bike, but most of what exists
complains about it running dangerously lean.
The spark plug insulator was bone-chilling white.
It was detonating at speed ranges much above idle under acceleration.
Cruising was not a problem for detonation, but surging at high speed
was an issue.
Yep, sounds lean.
Post by mike
It's a 2007 model with 423 miles on it. Apparently been in
storage for 5 years.
A treatment with SeaFoam and backing out the pilot screw fixed
the hanging idle, but it still detonates. Cranked the pilot screw
out 4 turns and still not much improvement in ping, but it feels/sounds rich
just off idle.
I cleaned the carb best I could. There was no gunk in it at all.
I was amazed at how clean it looked.
I wasn't too aggressive because
the manual warns about non-removable internal plastic parts.
They don't say where, but I decided not to use carb cleaner with
acetone or toluene. Stuck with simple green and water in an ultrasonic
cleaner.
I couldn't get the pilot jet out. Twisted as hard as I dared and decided
to come back another day. Low speed was ok.
I did hook up a vacuum pump and sucked fluid back thru the jets.
All the tiny holes in the venturi seemed to be open.
External fuel level measurement showed it to be at the low end of
the 4mm spec range. I moved it up to the top end. I was unable
to remove the float pin. It's pressed in and resisted careful
attempts to press it out.
In my experience, motorcycle carburetor float pins aren't pressed in,
they're a slide fit and only held in place by the carb bowl.
Post by mike
At that point the performance was markedly improved. Spark plug
now has a little color. Still pinging
around 1/4-1/2 throttle accelerating thru the gears, but ok at cruise.
Fuel level has a huge effect on mixture, as you found.
Post by mike
I raised the main jet needle .020", then .035". Both improved the
situation. I'll try one more increment to .050" before I give up.
That sounds like a lot.
Post by mike
Raised the main jet to 130. Improved the pinging. It's now
sounding rich at top end. Putting the fuel level back to
midrange might help that.
The performance of the 125 is crap, but probably the best I can
expect from a small engine. My concern is the pinging destroying
the engine. The pinging increases as the engine warms up.
I've been all over it with propane an detected no air leaks.
I experimented with blocking part of the air filter.
With 1/3 of it blocked, the bike would hardly run. Much greater
effect than I expected. I can't see how you'd ever keep the bike
in tune if it's that sensitive.
I'd agree that delaying the spark advance to a higher
RPM might fix the problem. I just don't think that's possible.
What I am saying is that it is possible for your ignition unit to
be defective, but still work. Once you get your head around that,
and check to see whether or not it is the case, you can rule it in
or out. When I say defective, I mean it has the wrong advance curve,
but it still happily produces sparks at the high tension side of the
coil. Just not necessarily at the right time.

I've had this discussion before with people who can't understand that
an electronic circuit can fail in unusual and unexpected ways, not just
either working or dead. If you have timing controlled by capacitors,
and the capacitors lose capacitance over time (actually quite common)
you can have weird things happen.

I'm not even saying that bad spark advance is your problem- obviously
the mixture was off quite a bit and you've sufficiently richened it
up so you might be getting to the rich side of things. If it's still
pinging, it might just be too much advance.

Does it get better with premium gas?
Post by mike
I'll take a timing light to it as soon as I get the tach built.
I'll drag out the gas analyzer and see what it tells me.
I have a Snap-ON automotive oscilloscope, but the screen resolution
is too low to see two sparks on screen to measure RPM. Probably
works great on an 8-cylinder car.
Probably time to twist on the pilot jet until it comes out or breaks.
I've tried a lot of stuff, but it should be obvious that I have no idea
what I'm doing. I've been driving the same Honda Shadow VT-500c for
over 30 years. Never had any issues with the CV carbs in that one.
Any additional advice welcome.
mike
I've pretty much exhausted what I can tell you at this point.

Good luck.
mike
2017-08-15 08:05:36 UTC
Permalink
Thanks for talking me thru this.
Post by Mark Olson
Post by mike
Post by Mark Olson
Post by mike
I'm still tweaking this Kawasaki Eliminator 125
to get rid of the detonation.
Somewhere in the attic, I have an ancient exhaust gas analyzer.
IIRC, it works by passing an infrared light thru the exhaust
and divines air/fuel ratio from that.
Was made back when there was lead in the gas.
Does that method have any value with current engine/fuel
designs?
Sure. Assuming it still works, it has no way of knowing whether you
are using a carb or fuel inject
I'd agree if it were measuring gas concentration.
The thing is a couple of thermistors in a bridge measuring
thermal conductivity of the gas...I think...
It's measuring a symptom and inferring a fuel/air ratio. I'd expect
that to be sensitive to a lot of different things.

But, I dug it out, fixed it and bolted it onto the gas tank.
I adjusted the pilot adjustment for just under 13:1 and took it
for a drive.
According to it, the bike is running seriously rich. Meter goes off
scale at 11:1.

The sequence of my modifications was dependent on the availability
of the parts, so I decided to remove the modifications to the top
of the carb. Put back the main jet needle spacers to stock.
I left the 130 main jet installed.
I had to back the pilot adjustment out to 3 turns to get the idle
mixture back near 13:1. I went for a drive. Can't tell much difference.
Was dark. I'll put the A/F meter back on it tomorrow and see
how it looks under load. I half expect a cop to pull me over
for all the equipment hanging off the bike. "Hello officer,
you need help tuning your carburetor?"

I did check the timing. Supposed to be 10 BTDC at idle and 35 at
4500 RPM. Still don't have a tach, but the advance numbers are right on.
This is a Kawasaki BN125-A8. Versions earlier than A5 stopped the
advance at 30 BTDC.
Post by Mark Olson
Post by mike
Post by Mark Olson
Are you sure your detonation problem is definitely down to the mixture
being too lean? Ignition timing would be the first thing I would
check. And before anyone says if the timing is right at idle it must
be OK at higher RPMs, I would not assume that.
I'm not really sure of anything. My previous newest bike was/is
a 1983 Honda Shadow VT500C. What I've learned from riding and
working on it for the last 30 years doesn't seem to give much
guidance relative to the 2007 Kawasaki.
The timing appears to be right on spec.
Post by Mark Olson
Post by mike
Post by Mark Olson
I had a weird issue with a 1995 Kawasaki EX500 ignition module (K calls
it the "IC Igniter") that had a timing problem between cylinders (one
was OK the other one was advanced by 6 degrees) even though the engine
only has a single ignition trigger coil, and the magnetic reluctor
pole pieces on the rotor were exactly as they should have been.
Replacing the IC Igniter did in fact fix the problem (verified with
an O-scope).
Can you give more details on how you set up the scope to make that
measurement?
I "made" a scope by constructing a resistive voltage divider network
and connected it to the primary side of the coil, running it into the
sound card of an old laptop. Recording the "sound" of the waveform let
me look at the relationship of the spark pulse vs. the pulse from the
trigger coil (stereo input of the sound card = 2 channel scope).
Interesting.
I have a Snap-ON engine scope. I was gonna do the same as you described,
but the memory is so small that, by the time you slow the sweep
to the point you can see two sparks, the pulses are so narrow that
you can't see them.
I'm not about to put my good digital scope anywhere near a spark plug.
I did write some code for a Palm Pilot to give real-time numbers
for RPM and spark advance, but
haven't tried it. Not sure whether interpreted BASIC is fast
enough to keep up.
Post by Mark Olson
I knew the timing was off simply by looking at the timing marks with a
inductive timing light from the 1980s, still works fine. I could see
the one cylinder was off by 6 degrees vs. the other one (accounting for
the fact that it's a 180 degree crank, naturally).
Post by mike
I have an inductive pickup that clamps on the spark plug wire to
read the spark. How did you determine the TDC position?
By shining the timing light on the timing marks. TDC is engraved on the
alternator rotor which is visible through the timing access port. The
correct degree position for timing at low RPM is also marked for both
cylinders.
Post by mike
I could intercept the crank position sensor, but that doesn't
tell me whether that signal is in the correct relationship to TDC.
I took a photo of the position of the timing marks when the timing
light fired and measured the angle graphically.
Post by mike
If the timing light says the spark is at the T position on the
cam sprocket, there isn't much one can do.
Sure- you can check to see how far the mark advances as you go up in
RPM. You might have to paint/engrave some of your own marks on the
flywheel to do this.
I made a plastic cover to contain the oil, after I cleaned up
the oil that went everywhere, and marked the spec'd angles
on the plastic. Seems to be dead on.
Post by Mark Olson
Post by mike
The service manual has a detailed write-up on how the ignition works.
Has a nice graph of the advance curve, but it has no numbers about
the RPM of each inflection point. I'll take a crack at the timing
this afternoon.
Yeah, without that info, or having some general knowledge about how
far timing should advance (I have a buddy who is an absolute expert
on this stuff) you're going to have to ask or do some research to
figure out what sort of numbers are reasonable.
Post by mike
And the bike has no tach. I have a timing light with a tach, but
it's for 4-8 cylinders and the resolution is useless for one cylinder.
Tach is easy- just measure the time interval between trigger coil
pulses and take the inverse to get Hz, then divide by 60 to get RPM.
Post by mike
Post by Mark Olson
If this was my bike I would be checking ignition timing at various
RPMs to make sure it's correct.
Thanks, I'll take all the advice I can get.
There's little written about this bike, but most of what exists
complains about it running dangerously lean.
The spark plug insulator was bone-chilling white.
It was detonating at speed ranges much above idle under acceleration.
Cruising was not a problem for detonation, but surging at high speed
was an issue.
Yep, sounds lean.
Post by mike
It's a 2007 model with 423 miles on it. Apparently been in
storage for 5 years.
A treatment with SeaFoam and backing out the pilot screw fixed
the hanging idle, but it still detonates. Cranked the pilot screw
out 4 turns and still not much improvement in ping, but it feels/sounds rich
just off idle.
I cleaned the carb best I could. There was no gunk in it at all.
I was amazed at how clean it looked.
I wasn't too aggressive because
the manual warns about non-removable internal plastic parts.
They don't say where, but I decided not to use carb cleaner with
acetone or toluene. Stuck with simple green and water in an ultrasonic
cleaner.
I couldn't get the pilot jet out. Twisted as hard as I dared and decided
to come back another day. Low speed was ok.
I did hook up a vacuum pump and sucked fluid back thru the jets.
All the tiny holes in the venturi seemed to be open.
External fuel level measurement showed it to be at the low end of
the 4mm spec range. I moved it up to the top end. I was unable
to remove the float pin. It's pressed in and resisted careful
attempts to press it out.
In my experience, motorcycle carburetor float pins aren't pressed in,
they're a slide fit and only held in place by the carb bowl.
My experience agrees, but the manual for the Mikuni BS28 carb
definitely says it's pressed in and warns about breaking the
casting trying to get it out. I pushed and banged on it as hard
as I dared. It would be a lot easier if I knew which end had the
press fit so I could support that support post.
If I ever have to mess with the
float valve, I'll risk it then. All this EPA anti-modification
stuff is annoying.
Post by Mark Olson
Post by mike
At that point the performance was markedly improved. Spark plug
now has a little color. Still pinging
around 1/4-1/2 throttle accelerating thru the gears, but ok at cruise.
Fuel level has a huge effect on mixture, as you found.
I'm thinking about lowering it back to the middle of the spec range,
but it's very difficult when you can't remove the pin that holds the
float.
Post by Mark Olson
Post by mike
I raised the main jet needle .020", then .035". Both improved the
situation. I'll try one more increment to .050" before I give up.
That sounds like a lot.
Agree. I did find an independent shop that suggested the .020.
It got better, so I tried .035.
Post by Mark Olson
Post by mike
Raised the main jet to 130. Improved the pinging. It's now
sounding rich at top end. Putting the fuel level back to
midrange might help that.
The performance of the 125 is crap, but probably the best I can
expect from a small engine. My concern is the pinging destroying
the engine. The pinging increases as the engine warms up.
I've been all over it with propane and detected no air leaks.
I experimented with blocking part of the air filter.
With 1/3 of it blocked, the bike would hardly run. Much greater
effect than I expected. I can't see how you'd ever keep the bike
in tune if it's that sensitive.
I'd agree that delaying the spark advance to a higher
RPM might fix the problem. I just don't think that's possible.
What I am saying is that it is possible for your ignition unit to
be defective, but still work. Once you get your head around that,
and check to see whether or not it is the case, you can rule it in
or out. When I say defective, I mean it has the wrong advance curve,
but it still happily produces sparks at the high tension side of the
coil. Just not necessarily at the right time.
Well, I tested it and it seems to be doing what it is designed to do.
Don't yet know the exact RPM related to the curve, but it sounds about
right.
Post by Mark Olson
I've had this discussion before with people who can't understand that
an electronic circuit can fail in unusual and unexpected ways, not just
either working or dead. If you have timing controlled by capacitors,
and the capacitors lose capacitance over time (actually quite common)
you can have weird things happen.
I'm an electrical engineer, so very familiar with stuff like that.
Took me all afternoon to figger out the intermittent in the A/F meter.
Turns out that the voltage regulator can't turn on until it
is turned on.
They put a cap to pulse it on when you apply power, but my
power supply had a soft start that didn't always make it go.
Post by Mark Olson
I'm not even saying that bad spark advance is your problem- obviously
the mixture was off quite a bit and you've sufficiently richened it
up so you might be getting to the rich side of things. If it's still
pinging, it might just be too much advance.
Does it get better with premium gas?
I never run anything but premium gas. I did try three different brands.
Google did find some people claiming that premium gas is always stale
because the volume is very low.
All the gas I have has remnants of SeaFoam in it. I'll have to run thru
that and get some fresh before I finish tweaking.

I'm resisting the temptation to dump it into the truck. Wouldn't be
the first time I started with one good thing and one busted thing,
only to end up with two busted things.
Post by Mark Olson
Post by mike
I'll take a timing light to it as soon as I get the tach built.
I'll drag out the gas analyzer and see what it tells me.
I have a Snap-ON automotive oscilloscope, but the screen resolution
is too low to see two sparks on screen to measure RPM. Probably
works great on an 8-cylinder car.
Probably time to twist on the pilot jet until it comes out or breaks.
I've tried a lot of stuff, but it should be obvious that I have no idea
what I'm doing. I've been driving the same Honda Shadow VT-500c for
over 30 years. Never had any issues with the CV carbs in that one.
I'm 69 years old and use the bike mostly for garage sales.
The shadow is too top heavy to manhandle around a gravel driveway
on a slope. I'm afraid I'll drop it and break a hip.
I need a smaller bike like this Kawasaki.
It's the only small affordable bike I found that lets you ride sitting up.
And it was 1/4 the cost of my next best option in a new bike.
Post by Mark Olson
Post by mike
Any additional advice welcome.
mike
I've pretty much exhausted what I can tell you at this point.
One thing I've learned is that if I keep talking about it,
more people chime in. Sometimes you even get a solution ;-)
Post by Mark Olson
Good luck.
Loading...