Clutch is impossible to bleed and won't release

I use my motive power bleeder more than almost any tool in the garage. 

I ended up purchasing a Motive Bleeder. What a great tool. My girlfried was relieved to learn that she will never have to help me bleed brakes again...

I hooked it up to the fluid reservoir, pumped it to 15 psi and opened the bleeder. Little bubbles came out. I removed the Motive and pressed the clutch pedal (I put the stock pushrod back in prior to bleeding) and it bottomed out before hitting the floor. I realized I forgot to put the pedal stop pad... Hopefully that didn't wreck the crank end play. Anyways it seems like the Motive bleeding accomplished something. I reinstalled the stop pad and held the pedal down with a steel tube jammed under the seat. I went to turn the engine by hand. I saw the wheels turning, but the speed wasn't very regular. I tried again with the handbrake pulled and I could easily turn the engine. So now I think it is fixed.

I wanted to start the car to try it but I found out it was out of fuel. Then I found out the oil smelled an awful lot like fuel so I drained it and there was a LOT of fuel in it, probably from all that flooding while attempting to start it with the wrong fueling settings in the Megasquirt. Hopefully I fixed the clutch and didn't mess up crank end play or didn't wash down the cylinder walls to the point of causing damage. I'll try it again during the week. Thanks for your support everyone.

In reply to Rocambolesque :

Over travelling the pressure plate will not hurt the thrust bearings of a stationary crank shaft. Not likely to hurt a running one either. What it will hurt is the diaphragm of the p.plate itself, and possibly the disc as well if the engine is running. You are good here!

Rocambolesque said:

I ended up purchasing a Motive Bleeder. What a great tool. My girlfried was relieved to learn that she will never have to help me bleed brakes again...

I hooked it up to the fluid reservoir, pumped it to 15 psi and opened the bleeder. Little bubbles came out. I removed the Motive and pressed the clutch pedal (I put the stock pushrod back in prior to bleeding) and it bottomed out before hitting the floor. I realized I forgot to put the pedal stop pad... Hopefully that didn't wreck the crank end play. Anyways it seems like the Motive bleeding accomplished something. I reinstalled the stop pad and held the pedal down with a steel tube jammed under the seat. I went to turn the engine by hand. I saw the wheels turning, but the speed wasn't very regular. I tried again with the handbrake pulled and I could easily turn the engine. So now I think it is fixed.

I wanted to start the car to try it but I found out it was out of fuel. Then I found out the oil smelled an awful lot like fuel so I drained it and there was a LOT of fuel in it, probably from all that flooding while attempting to start it with the wrong fueling settings in the Megasquirt. Hopefully I fixed the clutch and didn't mess up crank end play or didn't wash down the cylinder walls to the point of causing damage. I'll try it again during the week. Thanks for your support everyone.

Just an fyi.  I busily follow up the motive with. Couple of "traditional style pump and bleed sessions.  Like 3 or 4 pump and holds.  Gets that last little bit of feel.

Seems like I claimed victory too fast. I went back to the shop today. I started the car and once again, it's impossible to engage any gear once running. I could only try and put it in reverse but it would grind. Start the car in 1st with the clutch depressed and it rolls. 

I was super bummed out, I had my friend with me and we were looking forwards to a first drive... On one hand I'm sure the bleeding is OK with the Motive, but at the same time the pedal feels softer than before, prior to replacing the slave cylinder. My plan now is to throw in the longer slave cylinder rod (not the longest one I have, but the + 6 mm one). If I bottom out the pressure plate before the pedal hits the stop, I'll feel it. The LUK rep said that there was a slave cylinder P/N with a longer rod, but he never told me the lenght!

How much free play do you have if you wiggle the clutch arm? You need some, but not much.

A longer rod will not give you more push distance with a hydraulic system. The piston in the slave just floats in place and is moved by a distance based on the volume of fluid pushed in from the master and the diameter of the slave.

The Molvo was living in the exact same space you are, just barely not enough push on the clutch so the car crept in gear. The solution was a larger master cylinder. This of course comes at the expense of a harder pedal.

It sounds like there's some vagueness around what the correct pushrod length is after the interaction with Luk... Perhaps the way forward is a bit more research into what the correct part is for this application?

Unless there's wasted motion from the slave cylinder before the pushrod engages the fork, it's not too short (to address the "right part" question from a point of observation).

If the load doesn't come off the pushrod completely when it's released, it's too long.

Given that this should be a stock style installation, it seems like it shouldn't involve juggling clutch master sizes, but then we're back to Luk and assertions of incorrect parts...

The problem might be unclear because I didn't post a picture of the transmission. The slave is bolted to the side of the bellhousing, but the pushrod is hidden in there. There is no way of seeing how the clutch fork is moving (otherwise this would've been way easier!):

You can see the hardline going over the transmission in the previous picture.

Here's more too:

The slave definetly isn't at it's minimum stroke with the longer rod. I measured that when the transmission was out. 

I found a thread on another forum that was recently posted and the guy did the same swap and had the same problem. He purchased another slave cylinder and got a different pushrod length. He installed that and it worked. I messaged the guy asking him for the pushrod length. He said he didn't measure but he estimates about 1/4". So I installed the + 6 mm rod. No dice. Then I installed the longest one and bled it once again like 15 times upside down and downside up, manually, with an oil can and with the motive in every possible sequence. Now, the pedal doesn't bottom out so that's good. When it's held down and I turn the engine by hand, the wheels still turn. BUT, if I put the handbrake on and turn the engine, there is absolutely no other resistance other than the engine's compression. It feels exactly the same. However, I cannot turn the driveshaft by hand. There must still be some resistance in clutch and I'm trying to backdrive the 3.91 ratio of first gear...

Anyways, I don't know if I should call it good. It's definetly better than it was, but like someone said, there isn't 0.25" gap between pressure plate and clutch. 

I also messaged the Mercedes Classic center about the issue. I gave them the serial number of the transmission so they can verify if it is indeed off a 190E 1.8. Since only the 1.8 did not switch to dual-mass flywheel in 1988, I assumed that their 717.412 transmission was the same as our pre 1987 717.412 transmission and that it used the same clutch and throwout bearing. Maybe it isn't the same. Maybe the clutch is the wrong one. I also asked to clarify why there was 2 slave cylinder models. I hope they will respond...

So I checked on a French website à la Rockauto to see what parts were listed for a 190E 1.8. Seems like the clutch, flywheel, transmission and master were correct, BUT the slave was in fact a different P/N than what I had and what could be purchased here in North America. LUK was correct, there is indeed another cylinder P/N. I located the cylinder in the UK and had it shipped here. The rod IS longer, about 6 mm longer than the shorter one. I installed it in the car, not expecting much. Good thing because it didn't do anything. So now I have the exact setup that was OEM for the car except for the line from master to slave that is a smaller diameter and it still doesn't work. I can't figure this out, I think I officially lost at project cars 

I didn't bleed via the brake caliper, I used an oil can instead (and a Motive). It's the only thing I didn't do as the Mercedes manual described. I cannot see how it could make a difference, but at the point I am now, I might try it...

The Mercedes Classic Center won't respond, either they are closed or don't want to deal with modified cars ("Don't change the operating parameters").

How does the ID and OD of that metal line compare to a stock one? Just looking at the picture, it looks small. Most aftermarket clutch lines, from places like Jegs or Summit, are 3-AN or 4-AN size. That corresponds to IDs of 0.13" and 0.22" and ODs of 3/16 and 1/4.

The line is a 3/16 OD line, same as brake lines. The ID is about 0.125". Comparable to -3 AN, but smaller than stock I think. 

Changing the pushrod length isnt going to do anything for you because the slave will only move a distance based on the volume coming from the master. The slave will be backed off as far as there is pressure from the fork/pp when the master is released and open to the reservoir. 
 

if you need more travel out of the slave cylinder you need a bigger bore master or a smaller bore slave. 

Yes, changing the length of the pushrod will do something. The OE even did it! When I install the slave, the rod is fully extended because of the spring inside the cylinder. As the bolts are tightened, the piston travels inside  the cylinder as the other end of the pushrod rests on the clutch fork. To put it simply, lets's say the cylinder has a total stroke of 40 mm. If "rod A" is so short that it only travels inwards 15 mm during installation, you still have 25 mm of unused travel in the slave. When the pedal is depressed, the slave can only travel 15 mm until it reaches max extension. If the fork ratio is 2:1, the clutch will only be pressed 7.5 mm, that might not be enough to disengage. If a "rod B" is 20 mm longer, it will be pushed into the cylinder 35 mm (20 + 15) upon installation, leaving only 5 mm of unused travel (you need some unused travel because you don't want the hydraulic port to be blocked!). When you push the pedal (assuming the MC can displace that much fluid), you will be able to use the 35 mm of slave travel. That will net you 17.5 mm at the clutch, probably enough to disengage. 

That's how I understand it. I might be wrong or other cars might have it setup differently and we both understand the concept but didn't know similar installations. I know a lot of cars have the end of the clutch fork exposed outside the bellhousing and you see the slave movement. On this one, it's all hidden in the bellhousing, making this problem difficult to diagnose.

I am curious if there are some oversized slave cylinders or undersized master cylinders getting into the supply chain. I have seen many improperly boxed/labeled pieces over the years.  I think it is time to pull the booties back on both and measure their bores and compare to the specifications in the shop manual.

I will join this and hopefully help.

A long shot, but is #23 in the picture possibly cracked allowing air to be drawn into the system and making it impossible to bleed?  Or possibly the nipple on the reservoir is cracked allowing the same thing to happen? 

Maybe put a spring loaded rod on the clutch pedal to hold it to the floor and look for seepage.  If you made up the metal lines yourself perhaps you have a bad flare causing a problem.  Did these lines have metric bubble flares like German brake systems?  (At least like my VW's)  I am done babbling here.   

I hope you find out what the problem is.

Why not just take the entire system and unbolt it from the car (except the master) and bleed it in the "air" with the slave the highest point?

What is the stroke supposed to be?  If it is correct, I think you got something going on in the clutch.

Mercedes definitely offers different pushrods for a reason, but I strongly suspect that this is due to other things varying as well; the slave cylinder body shape, depth of its piston, shape of the clutch fork... No more than one will be correct for the combination of parts you have assembled.

I might hope, though it would be hard to be definitive, that by measuring carefully from the cylinder mounting face down into the pocket on the fork (or however the pushrod seats) you would be able to establish whether you have the correct length or not. I would think this should be pretty easy to tell if the options differ in ~6mm increments, as you neither want to be loading the throwout bearing at zero travel, nor having to close up significant slop (certainly not 6mm) before engaging the fork.

And you're not losing at project cars; this is just how the game is played!

Any reason you couldn't run a long flexible line under the transmission and avoid the loop over the bellhousing?

I somewhat read the whole thread and i don't see anything that proves that it is the clutch hydraulics at all. 

The stroke of the slave cylinder IS tied to the volume coming out of the master. All changing the pushrod length will do is remove the 'free play' in the mechanical bits between the slave cylinder piston and the pressure plate fingers. That will move your clutch pedal engagement point up, but you definitely don't want it so tight that the throwout bearing is preloaded against the pressure plate fingers at all times or you will burn up the throwout bearing in no time. It is also very possible to over-travel the pressure plate as mentioned, which can result in contact between the pressure plate fingers or the TOB and the clutch disc.

If the clutch hose were acting as a check valve (happens all the time with brake hoses) you could check for this simply by opening the bleeder after pushing and releasing the clutch. If fluid actively spurts out even a tiny bit you may be on to something. If the fluid dribbles normally as it would in a gravity bleed situation then the hose is not causing the system to retain pressure. 

Since the car is in one piece still I would do one more check that may help narrow down the issue, which is to check for clutch release through the entire pedal travel, SLOWLY. This is to check if the clutch actually DOES fully release but then re-engages when you push the pedal further down. This is fairly easy to check if you have an unsynchronized reverse gear (most manuals do) because you can GENTLY hold the shifter just outside reverse and feel the reverse gears gnashing end-to-end slightly when the input shaft is still spinning and the output gear (reverse idler) you are moving with your hand is stopped (because the rear wheels are stopped, use e-brake or brake pedal to guarantee this). This is basically the feeling you would get before the big crunch if you tried to put it in reverse without hitting the clutch at all (engine running). So, to avoid the big crunch you just go very slowly and stop and hold it at the point where the vibration coming through the shifter is informing you that the input shaft is still spinning. Then push the clutch pedal slowly through its travel and you should feel the movement stop. You can feel the input rpm very clearly and should feel it slow to a stop at which point the shifter might try to slide into actual reverse if you let it, but you want to keep it in the same spot. As you push the pedal the rest of the way to the floor, it should not start spinning again. If you let it engage reverse when the input stops you won't feel anything if the input tries to start spinning again because the rear brakes will be holding it still, so keep the shifter in the same spot.

If it DOES start spinning again further down the pedal travel, you can try shortening the pushrod to add slack to the mechanical bits so that the pressure plate isn't pushed that far even when your clutch pedal is to the floor, or you can put an adjustable pedal stop behind the clutch pedal (which is honestly preferable assuming your clutch engagement point isn't uncomfortably high in the pedal travel), or you can take the trans out and see what is contacting in the clutch assembly and whether it needs to be corrected. If you do decide to take the trans out i would honestly start it up in gear with the e-brake on and let it run like that for a minute or so, because that will force whatever is touching to touch for a while and leave you a clearer set of 'witness marks' for when you take it apart. As long as you aren't turning the flywheel or pressure plate blue from heat, you shouldn't be costing yourself any real additional money by doing that. 

In reply to Vigo (Forum Supporter) :

I am pretty sure the OP is trying t prove his clutch with an engine that does not run.....

In reply to TurnerX19 :

The engine does runs! I'm just waiting for a fuel fitting to get here because I had a fuel leak, but the car starts good enough now. 

Vigo, I will try what you are describing when the fitting gets here and I put the fuel pump back in the car. I'm not sure I understand why the clutch would release and engage again, but it's possible.

I even tried removing the pedal stop and the carpet on the floor the other day, this enables to almost bottom out the master cylinder (maybe 1mm left). Even that won't do it.

I'm pretty sure I have the correct parts now (slave, master, clutch, transmission are matching), it just needs some tuning. 

I tried removing the cylinder from the trans and leaving the line attached to bleed it below the transmission. There is absolutely no way to move the slave to the driver's side to drop it with the line attached. The transmission might be small, but the tunnel is a very tight fit!

Thanks guys, I will find the solution one day... Hopefully during Christmas break!

 I'm not sure I understand why the clutch would release and engage again, but it's possible.

It would be either because the back side of the pressure plate fingers  or the front of the throwout bearing  is contacting the clutch disc and pushing it back against the flywheel even though the pressure plate friction surface itself is no longer touching the clutch disc.

I'm following along. My e46 auto>manual swap has run into similar issues and I haven't felt like touching it in about 3 weeks now. 

This is not over! I removed the transmission once again and took some highly scientific measurements:

First of all, this is the difference between the commonly found LUK LSC 503 (512 0040 10) slave cylinder (top) and the Europe only LUK 512 0039 10 cylinder. The rod on the 0039 is 4 mm longer. I looked on what appeared to be the French version of Rockauto and the 39 is specced for cars with single mass flywheel and 0040 is for dual mass. All other parts sources in NA list the 0040 for all cars, regardless of the flywheel type.

With the clutch still on the car, I measured the distance from the transmission mounting face (let's call this TMF) to the pilot bearing. I found out that the transmission input shaft only engages like 4 mm into the bearing. I painted that white on the picture below. You can see that in this transmission's previous life, that shaft went deeper into the bearing. It's wierd because the bearing is resting on the abutment in the crankshaft, and to set it further out would probable require a shim. Alternatively, I could probably install a second bearing since I seem to have room. Not sure if that is a good idea or if the 4 mm is OK. It's proably better to shim it.

Then, I measure the distance from TMF to clutch fingers. It's 52 mm. I'll get back to this later.

I remove the clutch and flywheel and put it on the bench. I re-assembled the kit with the throwout bearing and compressed the stack with a large bolt to measure the pressure plate total possible travel and see at which point the clutch disc gets released:

To see if the disc released, I would slip a pick between the pressure plate's fingers and try to move the disc.

Surprise surprise; the PP can be pushed in 14 mm but the disc can never be released! 

I checked the clutch disc thickness, 9.5 mm. I don't know if that is correct:

Each of the linings is 3.5 mm, which is 0.1 mm under the spec, but that could be my cheap plastic caliper...

Now remember that 52 mm TMF to clutch fingers measurement. I bolted an aluminum angle to the transmission's mounting face to transpose that measurement to the transmission:

This is how the throwout bearing sits relative to the input shaft when the clutch is clamped (TMF to clutch fingers = 52 mm):

To that, subtract the 14 mm of pressure plate travel and you get this (TMF to clutch fingers = 38 mm):

See how the bearing is right at the edge of the sleeve on the input shaft. Also, the fork is right about maxed out at this point:

But as someone mentioned in this thread, you don't want to get to that point!!

The ratio on the fork is about 2:1, meaning that in order to move the bearing 1 mm, you need to move the slave cylinder 2 mm. The bore on the master and slave cylinders is the same, so 1 mm at the master is 1 mm at the slave. The master cylinder has a stroke of 29 mm. This means that the slave can travel that same 29 mm and you can move the throwout bearing 14.5 mm. You now have to play with the slave cylinder rod length to set the slave max extension to be a little under the max stroke of the flywheel. Basically, you don't want to get to TMF to clutch fingers = 38 mm since that would mean the fork would push on the engine crankshaft directly.

Then, I measured the distance from slave cylinder mount to the back of the clutch fork:

At TMF to fingers = 52 mm (clutch fully clamped), I have 43 mm of slave cylinder rod

At TMF to fingers = 38 mm (pressure plate bottomed out), I have 73 mm of slave cylinder rod.

14 mm of travel at the bearing, 30 mm at the slave (roughly 2:1 ratio on the fork like I mentioned)

I then measured all the slave cylinders I had at max extension and full retraction. I had an old Beck-Arnley wich was exactly the same as the "american" LUK. I had 2 alternate rod lengths for that "american" LUK and finally I have the "european" LUK. The values are in the table below:

For some reason, the total stroke on the "euro" LUK is less than on the other cylinders. But it doesn't really matter.

Now I just need to get a clutch + pressure plate that release on the bench so I can find the release point height, pick the right slave cylinder combo (most likely stock of + 6 mm rod) and put the car back together! I messaged LUK, maybe they can help. Either the disc is too thick or the pressure plate defective. Another solution would be to machine the mounting surface of the flywheel, but I don't have a clue how much material to take off and that might weaken the flywheel too much!

Does the number on the pressure plate match the number on the box it came from? This looks to me like the wrong pressure plate for your flywheel. Also you do need to space the pilot bearing or get a wider one. That "202" size bearing is available in a double width unit. Go to the SKF catalog with you ID and OD and look at the options.