Monday, November 28, 2005

How to make grinding sounds vanish: lubrication!

In an earlier post, "Running, but now a noise", I noted that on the last run of the spring PCA autocross at Marina airport, Seaside, CA, I was greeted by a loud whining sound emanating from the transmission area. A whining sound that went away when I depressed the clutch. Since then I had only taken Legs out for very short runs, hoping the sound would go away. No, I didn't actually think she would heal herself (though hope springs eternal for the busy and lazy). In the summer I took an afternoon to see how quickly I could replace a throw-out bearing. The thinking was that the pressure-plate and clutch, pressing against a worn-out bearing would silence it, under load, and aligning it under pressure to better than the noisy, sloppy state with the clutch out.
Alas, this was not the problem (more on that if you stick with me here), but I did learn some things. To wit: if you get past the idea that removing your engine is a daunting task, you can R&R a throwout bearing in two hours. I did, and I even had problems with one of the engine mounting bolts (small problem really, requiring more joints in my wrist and elbow and shoulder and wrenches on the starter-side nut than I had available--but I finally got it). I didn't drop the engine completely, but just back and low enough to swap out the bearing. I surprised myself, but still don't live up to those weekend racers that do complete engine swaps between classes and other feats of motoring heroism (let's not even start on what drag racers do!). So I pretended I was in the garage at LeMans, and could still save a first-in-class over those less reliable Matra-Simcas if I could jest get back on the track and log some more laps within the next three hours.
When I took Legs with her new T/O bearing down to the grocers, my heart sank as the sound was still there. No worse, but no better. I skipped the AutoX the next day, didn't make the required 4/6 events in order to score class points for a trophy, and never got to go heads-up against the only other Class Ap entrant of the year.

I was sure that the noise was then coming from the input shaft main bearing. The shop manual illustrations showed that this was a likely candidate for the noise. It fit the symptoms. I ordered a replacement bearing at modest cost. Then went to the shop manuals again and started reading the procedures. Perhaps I should stop trying to educate (and scare) myself before starting any project of this kind. Because it stopped me cold. I started investigating purchasing a new (used) transmission. Opportunity for upgrade, it would be a 741 trans, I would take it to Ron's Transaxles up in San Pablo and have them go through it, I would swap out the stock 716 nose cover (to match the mounting points and A-shifter) and I'd be better than new. But that was a serious $$ commitment, and the $$s didn't fall into my pocket over the summer.

Eager to make the Coastal Driving School's December event at Laguna Seca, I was ready to dive into the removal of the transmission during my T'Givings vacation. Luckily, the lightbulb went off in my head before I did that major surgery (done once before in 2000 to replace a broken hockey stick, but that's a different story). I thought, "check the fluid in the transaxle before you do the Full Monty".

I did, on Wed. Nov. 23, and was rewarded (kinda, but considering the alternative, you know) in only a pint or so of transaxle fluid falling out of the drain hole. The shop manual confirmed a 3.5Liter (3.17qt) normal operational volume, so I was sliiiightly low. Drips on the bottom of the trans case at the median line were observed, and one of the swing axle boots had a tear. Enough to lose 2.75 quarts? Over 4 years, and hard driving, yes. Nary a spot under the car (observed, anyway, there may have been some on the gravel drive) means that the loss was under pressure and probably mostly on the freeway at high gear speeds. The exit point was likely the boot and also at the nosecone-to-trans-case junction. Where, in 2000, I neglected to reinstall the new paper gasket. There's a metal-metal contact there, which shouldn't be, but which is almost good enough, but not quite. So I will be R&Ring the transaxle to fix (rather than patch with a transfusion) this problem, but not until 2006 and warmer/drier weather. And maybe I'll find a good deal on a 741 transaxle when I'm not shopping under a deadline.
The Porsche Gods smiled upon me, when after pumping (and pumping, and pumping--something like 150 strokes of the hand-pump per quart) in the requisite volume of Valvoline multigrade 85-140 gear oil (with the leaks I wouldn't trust a synthetic to stay in), and firing up Legs, the whining vanished like a toddler placed on Santa's lap.

Sunday, July 10, 2005

Weights of the pistons and conrods

This is for the new motor, I should probably start a new blog for that and keep the running car entries separate. For now, I just want to get this in.

On 10/31/2004 I did big-end journal sizes of one rod, to get some practice in on using my caliper (analog, thimble-type). I concluded that it was 2.244, which is about 4 thou above high-limit for a std rod.

On 7/3/2005 I measured by weight the rods and the new JE pistons I got. The rods have numbers on the outside of their caps (so you don't mix/match them--why two of mine have the same number(!!) is beyond me.

700 -479g
742 -480.5g
204 -481g
204 -480.5g

Because there are two 204s, but their weights are different, I'm referring to one as "204(h)" (for "heavy"). It is a pig, well, really, the 700 is 1.5g too light and shouldn't be in this set, but what are you going to do?

The pistons are a little close than the rods, only 0.5g off, rather than a full gram (unacceptable).

Piston pis. +ring










I have to do some dressing to get this gang of parts within spec. If I start by matching the 411.0 piston with the #700 rod, I can reduce that big diff between the 204(h) and the light 700 rod. I was hoping I would only have to dress two rods, to be a 1/2 gram out on the high on each throw. But the rods were two grams different high to low, and the 0.5g of the pistons doesn't cover that gap. If I took the heavy #204(h) rod down to 480.0, that would bring its total to 890.5 and within a 1/2 g of the two "middles. But that leaves the "low" pair still a gram out. Does that 1.0g put too much whip on the crank? Which journal should the low pair of rod #700 go on if there is a difference?

Mix and match table
Rod No. Rod wght Cyl. Wght. Total
204(h) 481.0 410.5 891.5
742 480.5 410.5 891.0
204 480.5 410.5 891.0
700 479.0 411.0 890.0

Saturday, June 25, 2005

Running, but now a noise

I fixed the starting problem. It was...drumroll, please...a loose condenser wire... Of all things! It was either loose where it came through the body of the 050 distributor and spade connected onto the points, or it was grounding at that point on the distributor body. The pass-through the disty body is a bit of a kludge, this condenser was not designed for this disty, so the grommet is a big square thing and needs to be a smaller round thing. I had trimmed it, but it was kind of loose, and probably when I turned the distributor to set the timing, it moved and came-loose or shorted.

What that solution, I went to the April Autocross with the LPR PCA at Marina Airport course outside Salinas. Here are a couple of pictures (1, 2) by John Reed. I also took some of the photos on the same page, further down from John's pictures. One pic made it into the June Newsletter (pdf, see p.33/44).

Sadly, I only posted four runs. I DNF'ed one, because I didn't close my door securely and it flew open on the first turn! D-oh! I couldn't close it because I had kneed the door handle up into "lock" position (with the door open). As I couldn't hold the door closed and go around turns, I drove off course before they red flagged me. I would have stopped, fixed the door, and then kept going, but looking in the rear-view (fortunately not turned down--contrary to Henry Watts' recommendations), I saw my engine decklid was open, too! What a marroon!

The last run, I started noticing a pretty loud whine from just under the rear seat tray, down low. You know, where the transmission sits. It went away with the clutch in. I did my last run of the morning and then worked my afternoon assignment then nursed it home.

Fun, but not a full day and not as fun as it could have been given the expensive-sounding noise from the tranny. Still, the engine runs good now, and she starts easily and idles well!

Saturday, April 23, 2005


After the carb reassembly I went to fire up Ms. Legs and she cranked but didn't fire. Now one always wants to go back to what last changed, and look there first. But in this case, after a physical check for the obvious (gas line not connected, coil wire off) none of which showed any faults, I considered what could have, in dissassembly and reassembly of the carb, could have caused a complete failure? The answer is nothing, if you assume the carb is delivering its dose of air/gas. A substitute shot from the ether can acted in place of that but still no start.

So, the next thing to suspect is the electrics. Sigh. Didn't we just go through all this? After more wiggling of wires, it is time to bring the meter back out. But think again: what was the last thing to change in the electrics?
I had filed the points in an operation just previous to starting this blog. The point-filing was to address some voltage bounce seen in a previous test during hand-turning of the engine while trying to set the point gap. It turned out the points had some tits on them and there was an inconsistent spark--different times, sometimes not at all!--due to the gap and uneven surface. So the points are cheap and easy to replace and may be out of spec from ham-fisted dressing anyway. You should only dress (not file, burnish--rub with xxxfine emory if you don't have a "burninshing tool") points in the field, to get back home. Because you have to ensure parallel mating of the surface. Burnishing should only removed deposit tits caused by arcing--the points tips are actually pretty hard--but enthusiastic polishing could result in points that are misaligned. They're one of the cheapest parts on the car; replace 'em.

I'm off to the parts store....

Right hand carb

I took the lid off the left hand carb after running it up to where it was stuttering, let it stutter for a while and
then killed it by pulling the coil 15+ wire. With the lid off I could see that the float level was good. The theory was that the check valve was stuck closed or the float was stuck up (previously stuck down) and the carb was being starved for gas. This wasn't the case. The level was good and the check valve was operating freely.

So I turned my attention to the other carb. I removed that one and disassembled it. I found nothing unusual nor any clogged jets. I put it back together without replacing any part--although the accelerator pump skirts on both carbs are a bit ragged. The kits I used for last winter's rebuild were not the variety that had the accelerator pumps. You can't get them separately, so that meanss new kits. Which is fine; now that I've had them apart a couple of times, I'll need some rebuild kits in the spares. I didn't measure acclerator pump output, which is probably the great sin of the weekend.

Reassembling I put the new phenolic block in, matching the other side. I used blue locktite on the new manifold studs, no gasket glue (perhaps a mistake; though isn't there an old saw about using tack where it is in wet contact and not if it is air? An intake manifold would be...which? Atomized wet?). The addition of the block raised the carb height about 1/2" (didn't measure the block thickness--gotta remember those details), which changed the necessary length of the down-link from the throttle crossbar linkage.

The carb came up to the crossbar, so the link had to be shorter by that same 1/2". The linkage has a semi-heim joint at each end. It is really a ball-and-cup affair; a heim joint has a through hole for a bolt. The joints are threaded onto a rod, typically with a backing stop-nut. In this case, the rod had a stop nut welded on at the base of the threads at one end. The other end was threaded longer, but the joint's internal threads bottomed-out before the correct length could be achieved. Attempts at switching sides, going to the parts box for other joints of different body length (some shorter but none longer) didn't yield a good result. So out comes mister cut-off disk on the Dremel. The amount of thread to cut off was determined casually; I made the linkage as short as possible, put it on at one end and eyeballed how far away the other cup was with no tension on the rod (any tension on the rod can cause the accelerator linkage on the carb base to move--and an even ever so slightl opening of the butterflies can play havoc with idle response. The linkage must be against the screw stop on the carb body at rest. I added two tread turns in my head to my eyeball measure to give some additional play, cut that number of turns off the rod, dressed the edge and without some small difficulty started the cup joint back onto the rod. The rods are reverse threaded at one end. as are their matched cup joints: this allows turning of the rod while attached to the linkage to lengthen or shorten the throw in small amounts--remember the nut welded to the rod? On a rainy day I need to go through the cup joint collection in the parts bin and sort them by right-hand-thread and left-hand-thread. Some white lihium grease in the joint cups and reassembly was a hand press-fit.

Monday, April 04, 2005

Stuttering above 3000

After the compression check and plug inspection I put the sparkies back in and attached the timing light.

A new timing light, from Sears provided some interesting configuration challenges. The light is 12V, Legs is 6V. Keith was borrowing the Audi to get an alternator for his Mazda, so he graciously lent me his 12V battery out of his car to run the timing light.

You have to hook the light to the battery, then ground the 12V battery to the car(!), then put on the induction clamp. I guess the induction clamp needs the common ground loop back to the light. I didn't try running w/o it. A coil of 18g wire from the travel box did the job. I grounded it by laying the bare wire against the 3rd piece, nominally held in place by a loop around the genertor clamp. If it hadn't worked, I would have used an alligator clip or tape. But it did.

Timing was static at 5 BTDC on the Bosch 050 distributor. On the pulley, a 1.5" mark is supposed to be 30 degrees advance. This per the Distributors section of " The Maestro's Little Spec Book". From what I could see, I was on that mark, but at 3k rpms it was jumping around a lot as the engine stuttered. Of course, there was no causal information therein: was the stuttering engine causing the timing mark to jump around under the light as the motor changed speeds dramatically and momentarily, or was the timing causing the engine to stutter as the advance went all wacky for a moment or three.

Sunday, April 03, 2005

Compression check

Today I ran a compression check on Legs. First time since I've owned her. Here's what I found:

Test number: #1 #2 #3 Median

Cylinder 1: 132 135 130 132
Cylinder 2: 135 135 135 137
Cylinder 3: 109 125 120 120
Cylinder 4: 135 135 135 135

This is actually better than I was expecting. That number three was the lowest of the four also wasn't any surprise. The compression checker I used was from a box of free stuff that was on the sidewalk outside Volks Cafe in Santa Cruz about three years ago. But today was the first time I'd used it and the first time I'd looked at it closely.

A homebrew contraption, it was made of four fixtures. Down at the sensing end, there were two threaded inserts. The endmost one was probably just an extension...not sure why. It added about two inches to the length of the tester. It fortunately ended in a 18mm nut, which allowed me to thread it out of the sparkplug hole when it came disconnected from the next piece on the tester. The second piece had a check valve and was also threaded for the spark plug hole. Both this tip and the extension had O-rings at the base of the threaded ends. Nice. Its other end was fitted to a 2-ft tube that appeared to have washing machine hose fittings at both ends. Hey--works for me, they're strong clamps. Into the other end of the washer hose was a pressure relief valve. It's a press-button to relieve the air pressure in the rig. Teflon taped into that is an air pressure guage that goes to 200 lbs. That puts the typical range, 100-150 right in the middle of the gauge, helping with accuracy.

I ran the test with five "pfut!" cranks at each go. I started with three, but it wasn't gettin up to full go. Theoretically, it should, but the gauge hadn't been used in three years. At one point I put some silicone lubricant into the pressure relieve valve and re-ran cylinders 3/4. The results were more consistent after the lubrication, so those are what are recorded here. The time on the starter motor was very similar for each cylinder. So even if the numbers aren't absolutes, they're consistent and good for that inter-cylinder 10% target.

The final indication that compression in the engine is good, is that while the starter will turn over juuust fast enough to start the motor in run-ready trim, with all four plugs pulled, it will spin the motor fast enough to make the oil pressure light go out. Note to self: if you ever have to limp home (or across the finish line) on the starter motor, take out as many plugs as you can....

Some statustics to start

This is pretty much just a test to get started.
New to the blogging thing.

Car: Porsche
Year: 1959
Model: 356 (356A, T2)
VIN: 106728
Date of Manufacture: March 1959 (approximate, no Kardex info)
Engine: non-original, no 3rd piece stamp. 616/1
Color: Silver over black
Original Color: Ivory