Footpegs

I got started on the footpegs for the Dragon Trike. I’m making a simple T-bar that will be u-bolted to the frame. The u-bolt mount will allow it to be adjusted and removed for eventual powder coating or other modification. For a ride of any appreciable distance, I would like it to eventually be a bar with footboards on it.

I have a 4 foot long piece of 1-1/2″ square tubing and I picked up a 3 foot piece of 1″ round tubing.

I sawed a hole through one end of the square tubing and centered a 32″ length of the round tubing and welded it in place.

I placed the hole slightly off center so it would be simpler to give a chamfered look to the end. The chamfer will be about parallel to the ground.

I cut wedges out of the “wide” side:

 And folded the side down to meet them:

I then cut a rectangle to fit the opening and welded it in.

The front edge weld in the pic above is possibly the best looking weld I’ve ever done. It’s almost a shame to grind the smooth.

I cut the square tubing to length, 16 inches from the crossbar to the end.

To make the saddle, I used a 3″ x 8″ black steel pipe nipple. I cut the threaded ends off then split the remaining tube.

I had a steadier hand for the other cut 🙂

After cleaning up the edges, used a magnet, barely visible here, to hold it temporarily in place so I could tack it on the end of the square tube.

I welded to two curved straight sides. The curved sides will be carrying the most weight in service, so I cut a couple of gussets to weld in along this side.

My #0 tip just couldn’t throw enough BTUs to do a good job the the intersection of all three pieces, but I’m pretty happy with the long weld along the gusset.

I also ran out of welding rods small enough for the selfsame #0 tip, so this is how it sits at the moment. I will swing by the welding supply joint tomorrow and get a little bigger welding tip and a bunch of the smaller rods. With a bigger tip, I should be able to clean up that mess and avoid a similar mess on the other side pf this brace and the other brace.

This is the nearly completed thingy:

Decided to Blog It

I’ve started restoring a 1982 Yamaha Seca Turbo and what with the fun of bloggin my VW trike projects, I thought I should start one on the Seca.

I wanted one of these bikes the first time I saw it, which was probably in Cycle Guide magazine, to which I was subscribed in and about 1981-82. All the factory turbo bikes were cool. Probably my absolute favorite in styling is the Honda, and having had a CX500 Custom from the era, I was slightly biased. The Kawasaki was arguably the fastest of the bunch, but really didn’t look all that different from the standard GPz. However, by most accounts, the Yamaha is the more stable and reliable of the three and it’s angular look is still appealing nearly 30 years later.

I found a Craigslist ad that originally listed it for $1500, which is about the retail price on kbb.com for a nice example. This one was described as low miles (5,600) but not running, though the pic seemed to indicate all was together. The price in the ad dropped and I honestly kept busy elsewhere until the guy called me with an offer I couldn’t avoid. He was moving and needed it gone.

When I went to see it, I found it to be somewhat less assembled than the ad and our conversations seemed to indicate. The carb rack and pretty much everything else between the turbocharger and the engine was either in a box or sitting, unattached, in it’s approximate normal position on the bike. There were maybe two or three hand tight screws holding most of the body in place. My apparent reluctance, which I was not even trying to show, resulted in yet another price drop, so I happily loaded it on my trailer for $300.

As soon as I got home, I put my *other* XJ, Buzz, a 1982 Maxim 550 (with it’s own hopefully fascinating tale) on Craigslist to recover the cash outlay for the Turbo.

Had I done the degree of carburetor cleaning I have done to the Turbo’s carbs, I could probably have gotten more for Buzz than $500, and my wife has ceased pointing this out, but I think only because she doesn’t like to nag. However, the first person to actually come see Buzz bought him without quibbling over the price and I immediately had $200 in pocket for parts.

Sadly, the carburetor kits, at $42 apiece times four, ate through that pretty quickly, but look at the results:

Pay no attention to the unpainted top on carb #3; that one was in the cleaner bucket the night I was painting and I didn’t want to wait for paint to dry when I was assembling #3 last night. It will get paint, probably black. If not black, I will paint them all some color that goes with whatever color the bike is eventually painted. I’m leaning towards the period Kenny Roberts color scheme.

There was plenty of evidence that some person or persons had started a carb rebuild job and abandoned it, perhaps multiple times. Considering the problems with carb #1 described below, I’m not entirely surprised. However, as I looked at more and more on the bike, I’m beginning to think that someone had a very minor spill, maybe even just dropped it in their driveway, and just parked the bike, probably without any storage procedures.

Little things like a float in the bowl with pin in place, but not through the float hinge. Note the lovely verdigris on the brass parts. Nice for antiques, but not particularly nice for antique carburetors.

SAMSUNG

Likewise, see the emulsion tubes. This one came out of a carburetor that had been soaking in the parts dip for 6 days. That’s how much crud was *left*.


This is an emulsion tube on a carb that had not been soaked…

Once removed from the carb body and soaked directly in parts dip, the emulsion tubes came serviceably clean, and if not abused in removal, most all the brass parts can be reused. However, the carburetor kits came with almost all new brass stuff, so there was no need to preserve them.

The two biggest hiccups were the idle mixture screw on carb #1 and the bowl drain screw on carb #3.

The simpler of the two, the bowl drain screw on #3, had been chewed up long enough ago to have seriously rusted since the damage was done…


After attempting all the usual techniques with penetrating oil and my impact driver, I decided it was going to take some cutting. I did not have a handy Dremel-like tool, so I used a flat needle file and cut a deep slot into the screw. The motion required cutting a notch in the surround, but the new screws don’t fit the same way, so there will be no penalty for the notch.

As I had put this step off until just last night, I was in enough of a hurry to forget to photograph it, but the new slot and the impact driver was the magic combo.

The #1 idle mixture screw was a bigger issue.

One of the recommended steps in refurbing such carburetors is to use the proper sized tap and chase the threads for the idle mixture screw. These threads are almost universally corroded and cleaning them up will make assembly much easier and extend the life of the carb. I was having trouble getting the tap to go and, long story short (too late) I found that the old screw had been so severely seized that someone had drilled it out. The remains of the screw can be seen here as the brass “sleeve” lining the hole:

 


The hole is too larger for a screw extractor to work without damaging the threads and the hole would need to be reamed out a little for the tap to fit. If it were tapped that way and the new threads didn’t follow the old threads pretty exactly, it would be equally useless.

After much consideration, I thought it would be wise to shop for a replacement before I started, just in case I destroyed it. Found a #3, New-Old Stock, on eBay for $250. Ouch. A #3 has the fuel and vent passages drilled on both sides, whereas a #1 only has them drilled on one, so using a #3 would mean plugging the outside holes.

I found an incredibly dirty full rack for $170, which is definitely better, but might mean the inheritance of the same problem, perhaps times four.

Finally, I asked Len Chacal at XJ4Ever and sure enough, he had a couple in stock. The bare carb body with no other parts was $69. I was happy enough with that to just order it and not bother (at this time) with my old #1. It arrived in need of cleaning (as described) but in much better condition than mine, or perhaps any of mine.

So, with #1 and #3 left to finish assembling last night, I tore into the project and finished about 1:15AM.

I still need to align the carbs with a plate, check/set the float levels and bench synchronize them, but they are essentially finished.

Quick Update on Sponge Bob Square Trike

A couple of weekends ago, Sponge Bob started to experience some odd clutch-centric symptoms. During the course of the day, the clutch became less and less responsive to the pedal. Go ahead, push the pedal… the gears are going to grind anyway. At one stop, it the starter even ground a few times before it kicked off. It was beginning to look like something major was wrong. I rode the rest of the way home shifting without the clutch, shutting of the engine at each stop and starting up in gear on the green light… it was not particularly fun.

So, since I doubt the Dragon trike will be ready to take to the IMS tomorrow, I tore into Sponge Bob last night to see what the matter was.

The plan was to pull the engine out and see what was going on with the throwout bearing, my first suspect. I figured while I had it apart, I would also change out the old leaky airshocks with the nifty coilovers I ordered long enough ago for the box to be dusty. Also, when we went to look at a particular trike for sale recently, it was at the owner’s mechanic’s house. He runs a VW repair and restoration shop out of his house. Cool guy and I envy his situation. Anyway, he had some chrome pipes of the same style that are on Sponge Bob except that they still have chrome on them and Gabby bought them for $50. I’d put those on, too.

I had the bumper off and the jack under the engine when I started removing the four nuts that hold the engine and transaxle together.

Hello, there is a missing nut on the bolt that also holds the top of the starter. And the other top nut was backed almost off it’s stud. There was a 1/4″ gap between the engine and transaxle.

Well, that would certainly explain it. This would let the engine and transaxle intersect at an angle and explains all the symptoms.. The input shaft of the transaxle runs inside a roller bearing in the end of the crankshaft. This misalignment would put that in a bind, so it would seem that the clutch is not disengaging. The starter was not only farther from the flywheel than expected, it would also be flexing away from it on the bottom bolt.

I found a replacement for the missing nut and tightened everything up and verified that the clutch would disengage. I then went ahead with the shocks (which required a run to two parts stores to find bolts to fit the lowers) and the pipes. There are no baffles in them, so they are significantly louder than the old ones. Gabby is pleased.

Another Long Overdue Update on the Dragon Trike

The clutch issue was resolved by continuing to swap out pressure plates until I found one that worked.

The one that finally worked is one that was on the free parts shelf at the BTW clubhouse. It was coated with a fine layer of rust, but it cleaned up perfectly and more importantly, it works. I neglected to get a picture of it, but it was one similar to this.

It is still an absolute mystery why the one I bought didn’t work, but my instincts indicate that it has something to do with the mix-n-match nature of the engine and transaxle combination.

In any case, it’s streetable, though the shifter is a pain to deal with. I plan to install one of these, or one like it

I have dedicated a number of hours tuning on the EFI system. At first, I was fighting a bad O2 sensor and now I’m fighting what I can only conclude is erroneous injector specifications. All indications are that they are 245 cc/min injectors. Using that number, every iteration through the basic fuel equation returns a ReqFuel number of 22 or so, which it then divides amongst two injectors. Trouble is, the engine will hardly even idle if the number is below 30 and it doesn’t run with a decent AFR until about 44. The astute observer will note that the number that works is exactly double the number given. This leads me to wonder if I should  keep thinking that I should be able to trick it into by telling it my injectors are half the flow they are, hoping for it to come back with a 44. It apparently doesn’t work that way, for telling it 125 cc/min comes back with, like 28.

So, ignoring the set up and using 44 as the ReqFuel number gets the thing running.

I’ve been using TunerStudio instead of MegaTune. I like it better overall. The dashboard configuration is very easy to customize.

With MegaTune, I would have to swap out one gauge or another when I was working on various pieces of the tune, but with TunerStudio’s dashboard, I have all the gauges showing at once. I need to play with their layout a bit, but they are all there!

TunerStudio’s VE Analyze Live! feature is a pretty cool autotune facility, however in my case, it seems to want to lean out mixtures that are already lean, where I would expect it to richen them. Perhaps I have some settings wrong. In any case, it’s still cool to watch it change the numbers and to see the engine respond to the changes.

With continued vigilance, I hope to have this thing reliably on the street in the next few weeks.

A Solid Afternoon of Attempted Tuning

It was an eventful day, tuning wise.

First, I found that the serial cable extender had come loose regardless of the almost heroic efforts to prevent it.

Then, I found that in at the least the last of the dozen or so cycles of removing and reinstalling the engine whilst working on the clutch, I swapped the IAT and CLT sensor plugs. So, it would have been adjusting for 150 degree air on an engine that just never would warm up…

After that, I found that I was getting no usable signal from the LC-1. A bit of quick troubleshooting showed that the controller itself appeared to be working. I connected (semi-permanently) the LC-1 LED output to an LED on the speedometer/dash. It indicated error code 2, which is a bad heater circuit. I was able to locate a sensor at a nearby O’Reilly Autoparts (O’Reilly part number 17014, in case somebody needs that) for $52. It’s working correctly now.

On the way to the parts store, I suddenly remembered that I wanted to verify the actual Bosch part number against the information I had looked up. It is indeed the Bosch sensor part ending “7057”. Perhaps more importantly, though, I discovered physical damage to the sensor that I had not noticed when I removed it. Upon further investigation, the baja bumper has not yet been pinned to it’s lower mounts after all the clutch work. I think when I was towing the trike around, the bumper hit the ground and was pushed upward where it struck the sensor. So, gotta pin that bumper down so it won’t cost me $52 every time I have to tow it somewhere.

So, with the EGO sensor working, I was able to get a little observable EGO correction to take place. I’m sure I need to tweak some parameters. It seems to react slowly, which is probably not a bad thing overall, but it also doesn’t seem able to dial in the target AFR. It gets pretty close, just not there. I presume there is some other enrichment factors in force and will need to find them.

I spent the most time trying to dial in some usable acceleration enrichment. While the trike has been drivable and throttle response while under way is decent, it stumbles pretty badly right off idle. It’s kind of like the infamous dead spot these engines experience when equipped with the ubiquitous Bosch 009 distributor, a unit with centrifugal advance. There are a few of people running these distributors that are very happy with them. For racing applications, the engine is either idling or running WOT, so the stumble isn’t a big deal. For the street, these distributors can be tuned with the appropriate springs so that the advance is full on at a fairly low RPM, which largely eliminates the flat spot, but it’s always at least somewhat there, ranging from minor annoyance to killing you when you try to pull out onto a busy street.

Armed with this background knowledge, I spent a couple hours today trying to tune in the spark advance table. First, I verified that the trigger offset of zero is appropriate for my EDIS ignition system with a timing light. Dead on. Then I tried scaling the table because the values in it seemed low across the board. With the 009 distributor, for example, one sets the idle timing to be about 7 degrees BTDC and by the time the engine is spinning, it’s running 28-30 degrees. So, I scaled my spark advance table to correspond with that.

Well, the first minor thing is that the nature of the EDIS module is that 10 degrees BTDC is the least advance it will allow. By having a lower number in the table, it made the *displayed* advance dither around in the 7-9 territory without affecting the actual timing at all. I manually moved all the lowest values up to 10 degrees.

After than, it seemed to be adjusting ignition timing pretty much as I would expect it to need, but the stumble was still full on, so I started looking at fuel.

Armed with a working EGO gauge, I was able to tell almost immediately that it went just about dead lean on acceleration, sometimes as high 22.

I tweaked on the acceleration wizard and was able to get a pretty nice rev out of it. I was not able to make much progress on MAP based acceleration. Seems whenever I had any percentage of control as MAP, I could not satisfy whatever basic settings it wanted and I couldn’t keep it running. So, with all TPS based numbers, I was able to rev the engine nicely, so long as I wasn’t extremely gentle.

Then I started playing with EGO correction again and somehow lost the rev. As it sits now, Accel Enrichment is pegged at 100% and never responds to TPS input. I am reading a lot….

Long overdue update

The trike has been intermittently road worthy. It spent a LONG time sitting after some clutch trauma that I thought was going to be much much worse. It’s drivable enough to do a little shakedown cruising in a big parking lot….

This shakedown was partly to have fun on the trike after all this time waiting to get it rolling, but also to run out some oldish gas. Sadly, we did a little too good a job at that and I ran out on the way home from this parking lot. My wife went on home on her trike and brought back the truck and a full gas can, but sadly happened to get about the only can that can’t reach the filler, which is under the fender. One of the many tasks left is to add an external filler cap…

I had taken the laptop, but neglected to check it’s state of charge, so I wasn’t able to log any of this driving. I am hoping to get started with that tonight…

Buzz Sold

Buzz has been sold.

But, Buzz’s sale helped finance what I’m sure will eventually be another MegaSquirt project, a 1982 Yamaha Seca Turbo.

SAMSUNG

In the mean time, I am still working on my VW Trike. I resolved a mechanical issue with the clutch, so I will have it on the road enough to seriously pursue tuning.

Denied!

Rule One: Test before you put it ALL back together.

Monday night, I got the rest of the engine wiring and hoses on, the exhaust, the bumper… Started the engine up (it cranked for about one second and lit off without touching the throttle… I love EFI)…. All looked good.

I knew I would probably need to adjust the clutch cable before taking it for a spin. That’s when I noticed that the clutch arm does not budge. At all. Well, it wiggles with the minimal play that the shaft has naturally, but otherwise, it was completely immobile.

Long story short, on Tuesday, I called AMS to consult with Steve again. He recommended that I remove all the new parts and bring them in, along with the old parts and as much other information as I could gather. Tuesday night, I fooled with it for a while, still looking for something, but pulled the parts out and got them ready to go.

This afternoon, I took everything in to AMS and we determined that everything should be all the right components. He suggested a couple of things to check concerning cable adjustment and clearances. I also consulted a couple of BTW folk. At this point, it’s a mystery to all, but my mission is to measure and test to see if I can find what’s not making up.

So, when I got home and got started, about 8PM, I first checked that the throwout bearing, fork, cable and pedal were all moving freely.

 Reassembled, same thing.

I pulled it apart again. I did some measuring. I put a straight edge across the mounting face of the transaxle and measured to the face of the throwout bearing. It was 2.68 inches. I also put the straightedge on the face of the pressure plate collar thingy and measured to the engine block face that mates with the transaxle. It was 2.81 inches. These two together should mean that, once assembled, the throwout bearing should be resting about 1/8″ from the pressure plate. This corresponds with the expected range of motion for the assembled clutch.

I verified that the clutch arm has it’s full range of motion while disassembled.

I also compared the old and new throwout bearings:

Though they are constructed in rather different ways, all the dimensions that matter appear to be identical, or at least compatible.

While I had it apart, I attempted to use a steel plate and some clamps to verify that the pressure plate can indeed move when pressed. It does. There was not a reasonable way to press the clutch and verify that the clutch disk is released when it is pressed. I have in mind a custom tool that could test that, but I was not prepared to fabricate such a tool just then. I put it all back together again, with the old throwout bearing, and verified that nothing had changed.

“Insanity: doing the same thing over and over again and expecting different results.” – Albert Einstein

While I had the camera ready, I documented the free play in the clutch arm while assembled and it corresponds with what I would expect.

So, I have decided that unless some radical new information comes to light before bedtime Thursday, I am not going to be able to drive the trike to the benefit/show this weekend.

I will, however, tow it in to the show. We are auctioning off a donated paint job and this trike has the paint from a previous auction and we hope to drum up lots of interest in the auction!

Clutch Reassembled

I dropped by Automotive Machine and Supply after work to procure parts and consult the oracle. Steve verified my concerns that there was supposed to be a collar on the pressure plate for the throwout bearing to contact and further that it’s absence is the real reason the throwout bearing jumped the tracks. He also verified that the play in the input shaft is normal, though he agrees that I should change out the seal.

So, $86 dollars later, I have a new clutch cover and pressure plate, as well as the clutch fork and such.

Besides, the old pressure plate and fingers were more damaged than they first appeared:

Note the bent fingers at the lower right. They were not really apparent with the cover installed.

The installation of clutch fork included the usual minor trauma. Installation of the clutch cover gave me a new problem that make me chuckle.

My clutch pilot tool has a ring for a handle and it was too big to fit through the collar on the pressure plate. I applied a bit of heat and a bit of sculpting and now it fits.

Otherwise there were no particular difficulties with the physical installation, other than the usual big hands in a small space issues…

After the delays, I got the engine bolted on and decided that it was late enough that I didn’t want to tear in to the wiring, fuel, exhaust and bumper. They will have to wait for Friday or Saturday…

Transaxle Work

I got all but the four bolts pulled off Tuesday night, then finished pulling the engine out last night.

It looks like there was no really serious damage done by the rogue throwout bearing. The fork was pushed too far and the throwout bearing jumped off it.

You can see the broken finger and a piece of a chewed up spring clip laying on the bottom of the bell housing. The remaining clip holding the throwout bearing has been ground against something as well. At the time of the event, the engine was still running, but in the process of towing it home, the broken finger apparently jammed between the flywheel and the bell housing, preventing me from starting the engine.

Once off the fork, it appears that the bearing landed such that the fingers on the clutch pressure plate rubbed against the bearing carrier. You can see the galling on the bearing carrier in this picture:

 It remains to be seen whether the damage to the pressure plate fingers is significant:

Some pressure plates have some kind of disc attached to the fingers, like this:

… but this one never has, at least not while I’ve had it. I’m not sure whether or not it’s missing, but since clutch operation seems reasonably normal without it, I’m going to assume it’s ok without it.

Speaking of assumptions, I took a guess that some fairly severe looking play in the input shaft of the transaxle might have been a factor in the damage done. Upon clearer thinking, I now realize that this shaft is not left hanging out there, but is rather supported by the pilot bearing in the end of the engine crankshaft. Duh! I am going to replace that seal while I’m in there anyway…

Picking up parts on the way home tonight; it’s entirely possible that I may have this thing mobile tomorrow…

Sluggy Blogs All The Subjects