When this bike arrived, it needed a little bit of everything, including tuning. It arrived with 150 mains, which are huge for a stock engine. After some trial and error, seat-of-the-pants tuning, and a little cheating with an AFR meter (that never read correctly because of air leaks in the exhaust), I arrived at 130 mains, #3 clip position on the needles, 40 pilots, and 1.75 turns on the air screw. Floats were set and all carbs were synchronized. The ignition was set statically and the points were gapped according to the shop manual. This made the bike run relatively smoothly, with good throttle response through all positions except a stumble around 3/4 throttle. It did not like any quick crack of the throttle. The bike would never start cold on the kick starter and required several rotations of the engine to begin roughly idling for a moment. Idle at full operating temperature was also a bit rough for my liking. For the time, it was sufficient to move on to other work.
After the trip to Gettysburg, I decided I wanted to see what I could improve in the tuning. While it started and ran ok, it was certainly not a cold kickstart bike, the idle seemed choppy, and its mannerisms did not exude confidence or reliability. Honestly though, most CB750 chops I see running around seem to be about the same, but I know they could be better. I guess it just depends on the performance you're willing to accept.
Ignition Precision (actually, accuracy, but that didn't sound as good)
In the past, I've always set my points gap with a feeler gauge and set the timing statically. I rarely yanked out the timing light to check operational timing and full advance. This was good enough, but clearly there are more accurate ways of achieving the proper settings. I was curious how much there was to gain from setting dwell time instead of gap and setting timing dynamically. Spoiler: setting dwell time instead of the gap on the points made the biggest difference, but using the timing light to ensure your mechanical advance is operating properly is a good idea.
Setting the dwell time (or dwell angle) on the points is much easier (IMO) and more accurate than using a feeler gauge to set the gap on used points. Used points have pits all over the surface and this affects the true break point of the electrical connection. You may set your gap properly, but the points actually electrically short at a different time than is expected based on the gap you set. Setting the dwell angle directly sets the time that the points are in electrical contact and the time that the coils will charge.
To set the dwell angle on the CB750, which has a set of points for each pair of cylinders (1/4, 2/3), you must set your dwell meter to 8 cylinder. It is a wasted spark system with 2 sets of points, so you cant use the 4 cylinder setting. The 4 cylinder setting on dwell meters is based on a single points set and no wasted spark. Connect your leads to each side of the points so that it measures across the points. Start the engine and set idle around 850-900 rpm. When I checked the dwell on the Waffle, it measured 19 degrees with the points freshly set with my feeler gauges. The proper dwell angle should be 24 degrees. This means that my coils were charging for 20% less time than they should. I adjusted the points until I achieved a dwell angle of 24 degrees and then set the timing back to factory specs.
The timing was set using a timing light, then checked statically to see how much of a difference there was between the methods. The difference between static timing and using the timing light was negligible, but I preferred using the timing light. While I had the light hooked up, I spun the engine up in rpm and verified a smooth advance that ended between the two hash marks on the advancer. Base timing, advance, and dwell were all now within spec.
With the ignition system set properly, the bike idled a lot smoother, both at operating temp and cold start. While riding, the throttle response seemed more crisp as well. Cold start was still about the same. Meaning that the bike did not want to kickstart cold, but would fire up on the starter after several rotations. Overall, it seemed to start easier, but it still wasn't a kicker. On top of that, the stumble at 3/4 throttle was still present
It looks like I'll be doing a few things differently from now on. This includes how I set float height, which I discuss below. Improvements were made by setting the ignition system via dwell angle and a timing light, but I still suspected fueling issues.
Float Height Quandary (Using float height for fine tuning)
I felt that my main jet was sized properly and any change in the needle clip position from #3 resulted in a worsening of the throttle characteristics in other areas where it was previously good. At #3 on the needle clip, the bike still exhibited what seemed to be fuel starving around 3/4 throttle.
While doing some research on float heights for these carbs (Keihin 657A/B), I noticed a discrepancy in where the height was measured from. The shop manual isn't explicit on where to measure the height from. The most agreed upon method is to measure 26 mm from the lower level of the flange on the carb body. I also found a wet fuel height, which at some point must have been determined by some racers back in the day because Honda never specified a wet fuel height.
|Lower flange on carb body (picture rights: Hondachopper.com).|
The wet fuel height is measured using what some call "the clear tube method." This method simply uses a clear tube connected to the drain of the float bowl and routed up against the body so that the opening of the tube is above the float bowl gasket surface. This allows you to see the height of the fuel in the bowl. The ideal height is claimed to be 3-4 mm from the gasket surface.
After setting all other variables (synchronization, main jet, needle height, pilot, air screws), changing the fuel height in the bowl can offer some fine tuning... or, another way to put it, compensation for not quite getting the other variables correct.
In an attempt to remedy the lean condition off idle and at 3/4 throttle, I moved the floats up toward the carb body 2 mm to wet the emulsion tube more. This richens the mixture slightly throughout the throttle range, which is similar to a slight increase in jet size. This mostly eliminated both issues. Off idle manners were perfect. Although it still seemed very slightly lean at 3/4 throttle, I didn't want to bugger with it because 95 percent of my riding is going to be low in the throttle range. Fixing the slightest bit of perceived power loss at 3/4 throttle might sacrifice the off idle performance.
The final float height was 24 mm from the lower part of the flange. The bike kickstarts well, has a good cold start idle, good off-idle throttle response/cruising behavior, and great response with quick throttle changes. When the engine is cold, the off idle response is a bit sluggish, but this isn't unexpected because the intake tract is still cold and fuel is dropping out of suspension. Even though there might be a barely noticeable loss in power at 3/4 throttle, the hard stumble is gone. Overall, response and power throughout the throttle range is very good. Using the float height to fine tune the fuel mixture worked well in this instance.
Final Carburetor specs for the Waffle:
Stock engine with breadbox air filter and 4-4 drag pipes
Altitude: Sea Level
Needle Clip Position: #3
Air Screw: 1.5
Float height: 24 mm from the gasket surface (not the lower flange on the side)
Other useful notes:
- When setting your float height, ensure that the float tang is just barely touching the float needle. I set my float height with the carbs turned sideways. This way, I can set my calipers to the desired height and push the floats in so the tang contacts the float needle. I can vary the pressure to verify that the contact is present, but not compressing the spring.
- The float needle spring should compress completely under the weight of the floats when the carbs are held upside down. If this is not the case (which can occur if aftermarket float needles are used), you may need to set your float height a few millimeters higher to achieve the wet fuel height of 3-4 mm.
- In the event that you need to swap points or an entire points plate roadside, some spare wire and a lightbulb can be useful for setting your timing. You can carry wire and rob a lightbulb from your bike temporarily or carry a couple LED's with integrated resistors spec-ed for 12 volts in your tool bag. I purchased 20- 3mm LED's with integrated resistors for this purpose. Simply disconnect the blue or yellow ground wire from the points and the capacitor, clamp the short lead of the LED to the point under the nut and run a wire from the battery positive to the long lead on the LED. Now you can set your points based on when the light turns off. Below is a picture to help clarify this procedure.