Barometric Compensator Failure

[jeffc280sl]

I started a thread about shims for the BC and over time the subject shifted to failed units and how to detect them.   A failed BC created a lot of extra work for me and maybe this thread will help someone else avoid the learning things the hard way.  Attached you will find a picture of a good and failed BC.  In my case and I suspect in most cases of a BC failure the bellows inside has experiened a seal failure and lost it's partial vacuum.  I've come to understand there is a spring inside the bellows and when vacuum pressure is lost the spring expands the bellows which forces the pin outward.  

The picture of the left side is the failed unit and the good one is on the right.  If you look closely at the top you will see the pin on the left unit is fully deployed.  When this happens the BC tells the fip to go very lean.  I've had my car for 15 years and had other priorities to work on until recently.  I now see that the previous owner new something was wrong with the BC influence of the fip because it had almost 5 mm worth of shims.  These shims raise the entire device and enrich the fuel mixture.  

Several months ago I bought a air fuel ratio meter and discovered that my fip seemed very lean.  My plugs always looked perfect so what gives.  Please remember that the afr varies greatly in our cars even when the engines are tuned perfectly.  The internals of the fip are constantly adjusting fuel flows to take into consideration rpm changes and engine loading. It turns out that sometimes my afr was rich, sometimes perfect but most of the time lean.  As a result my engine did not overheat and my plugs looked fine.

Not understanding the BC failure mode and knowing that replacements are NLA I setup my fip to meet the MB specs for on the road afr levels. During the process I bought a used fip to tinker with a to have as a spare. It came with a BC so I started to compare this BC to the one on my car.  I made a vacuum chamber in a Ball jar and pulled a vacuum.  I was pleased to note that the BC from the replacement pump worked according to spec.  I'm now in the process of determing how I can use the good BC on my car.  The extra work I talked about earlier is that I set up the fip to work with a failed BC.  If I had to do it again I would have rather set up the FIP with a working BC.  There are two styles of BCs.  I'm including a pic of the other version which does not include a colar at the end of the threaded section.

I suggest that each of you consider removing your bc and testing it in an diy vacuum chamber.  I suspect they work or they don't.  Atleast pull your BC out and see if it looks like the one on the left.  If it does maybe we can talk more about what to do.  I'm also attaching a pic of the vacuum chamber.

[ja17]

Very Nice Jeff,

Thanks for the work and information it is very good. I must admit I have never delved into checking these compensators. Your mini vacuum chamber is a classic and should go in the Pagoda group museum next to the "Blacklick Injection Pump Test Bench"  !

Joe Alexander
Blacklick, Ohio

[jeffc280sl]

Thanks Joe!

I'm glad the information was useful.  We only use a small portion of the BC range. 7.22 inHg is equivalent to 7,500feet.   I took the vacuum up to 25 inHg which is equivalent to 42,260 ft and the BC was at max length.  Definitely came from the aircraft industry.

[awolff280sl]

Very cool Jeff.
Do you know if the seal failure is a result of an air leak in the metal casing of the BC? And if so, can some sort of additional sealant be applied to a still functional BC to prevent failure?

[jeffc280sl]

The brass bellows lost its seal.  The bc shell is not under vacuum

[awolff280sl]

Right, I think I understand. So, the seal is around the moving pin/plunger?

[jeffc280sl]

I had a hard time with this also. The casing including the pin is not vacuum sealed. Inside the casing there is a brass bellows which is under a partial  vacuum seal.  The attached pic shows a bellows at rest with no vacuum.  It also shows a bellows which has expanded under vacuum.  The expanded bellows forces the spring loaded pin outward.  The bellows itself is creating all of the pin movement.

[awolff280sl]

So the bellows itself develops a leak? You mentioned that the bellows is "brass"? It's not some sort of softer material?

[jeffc280sl]

You have it.  When the seal leaks atmosphere air is sucked into the bellows and a spring inside the bellows (which is now not restricted by vacuum) expands the bellows.  I think this is the only failure mode for the BC and the expanded bellows causes the pin to extend which leans the mixture.  Before I took one apart I didn't know how far the pin was capable of extending.  The pin on my failed BC extended to its maximum limit which is equivalent to an altitude of 42,000 ft. 

I think the bellows is made of brass but it could be another alloy.

[tel76]

Hello Jeff C,
This is just a thought ,but if you were to manufacture one of these bellows and seal it at sea level and then take the unit up to 30000ft the bellows would expand and push out the pin(the bellows are still at sea level but the pressure at 30000ft is less so the bellows would expand).
If this could be proved with your test equipment it may be possible to repair the compensator.
I have  one of my aircraft (old)altimeters and these have bellows fitted, i will strip it down and see what state the bellows are in at sea level,

[jeffc280sl]

Could be Eric.  I have read the bellows is under a partial vacuum and I believe the vacuum and spring contend with each other.  Remember when my seal gave out the spring expanded the bellows to the limits of the canister which is equivalent to 42,000 ft.  I think the bellows would be under a partial vacuum even at sea level.

[TheEngineer]

Hey Guys! Let's keep it simple! The bellows can be soldered. I don't know what's involved to open the can. And yes, it came from aircraft engines. You can fly a fuel injected engine inverted, with a carburetted engine, the best you can do is a barrel roll. Nowadays, fuel injection measures the air density. Mo betta!

[tel76]

Hello Jeff C,
As i have the FIP on the bench i removed the plate on the RHS (the one with three screws) l could then see that the BCs pin was fully deployed ,i then unscrewed the BC and was surprised how much the rack moves as you screw it out.
If we operate at sea level why do we need this unit?
If the BC is unobtainable what do you suggest to overcome this problem ?

[jeffc280sl]

Eric,

Sorry to hear about the failed BC.  As you can see the BC has a tremendous effect on the rack.  Have you driven the car with this fip installed?  I ask this because the fip can be adjusted to operate in accordance with MB specs with a failed BC. 

Options for a failed BC

1) your fip may have been adjusted to operate in spec with a failed BC.  I know it sounds unreal but that's what I did with my car before I bought a spare fip and took a BC apart to understand how it works.   That's why I ask how does your car run.
2) to compensate for a failed BC you could add shims to negate the effect of a fully deployed pin.  I have good information on the length of the pin at 50 feet and 500 feet above sea level for good BCs.  I know its a very small sample but I think I can get you in the ball park with the shims neccessary to cancel the effect of a failed BC.
3) another option is to locate a replacement BC.  I bought a spare fip on eBay for $175.00 and it happened to have a working BC.  I have using this BC on my car.
4) install a bolt with lock nut in place of the BC.  This is basically the same as option 2 because you BC is stuck fully deployed so we are talking about doing some math to set the depth of the bolt or fixed BC in the case of #2.


Only option 3 allows you to drive at various altitudes.  You need to determine if this is important.

[jeffc280sl]

As of Feb 11 I have received 3 reports concerning BCs.  Two reports indicate good BCs and one indicates a failure where the pin is fully extended.  When this data is combined with my own we have 3 good and 2 bad BCs. 

I know this is a very small sample but this data indicates a 40% failure rate. 

I'm very interested to hear more reports from members.

[66andBlue]

Hi Jeff,
thanks for providing these very informative insights on the inner working of a BC.
Some time back another member (Mrfatboy) got rid of the BC altogether but rigged it up in way that he could make IP adjustments.
See: http://www.sl113.org/forums/index.php?topic=8328.0
Perhaps with all the new data that you have now, could you make some comments on his approach?

[jeffc280sl]

Alfred,

The bolt you speak of is option 4 on my list for Eric.   It is basically a bolt with a lock nut on it.  Turn the bolt to the desired length and spin down the lock nut to hold it in place.  Getting to the correct insertion length is the hard part.  It can be estimated based on the data I've collected. If we find a failed BC with the original set of shims I think we can fairly accurately estimate the pin length when it left the factory.  If over time additional shims have been added to compensate for the failed BC, as was the case with my BC, we have to factor in another variable so the estimate is less accurate.   The correct fip setup can be verified by removing the fip and placing it on a Bosch tester. None of us has access this piece of equipment.  The other alternative is to use an air fuel tester that I describe in another post.  

[tel76]

Hello Jeff,
I only drove the car about 15 miles and then i embarked on rust removal,engine overhaul rear axle etc. etc.
I tried to move the knurled screw at the FIP rear end and it will not turn either way,if the pump goes to max: lean when the BC fails someone in the past may have tried to richen up the FIJ by turning the screw to max : rich.
Which way should i turn the screw to weeken it?
I do not want to put undue strain on the adjuster,or am i better removing the rear cover?

[jeffc280sl]

Eric,

Previous owner may have adjusted the fip.  The knurled screw on the rear of the pump is used to set idle co or afr.  Adjusting this screw has little or no impact on the co/afr at other rpm levels.  I suggest that you do several things to establish a baseline that you can always come back to.  Take the rear cover off the fip.  It's no big deal. You will be looking at the same thing as shown in the attached picture.

With the exception of the rack adjustment the white, black and idle screw adjust richer when turing clockwise.  The rack adjust richer when turning CCW.  All of the adjustments have spring loaded clips on the flat part of the screw head.  When turning in either direction you will feel a click or detent.  When we speak about adjustments we will be speaking about CW or CCW clicks.  With the rear cover off you will have good access to these adjustments screws.  I suggest that you count and record the CW clicks on each adjustment screw.  If you idle screw has been turned full CW or rich you will have no clicks left in the range. This may explain why you could not move it.  It may be tight but you should be able to turn the idle adjustment screw CCW if that is the case.  Adjust the white pair of screws 5 clicks at a time.  Same for the black.  What I mean is do 5 clicks on the top white then 5 clicks on the lower white.  Then back to the top then bottom.  I know the overall range of each adjustment and knowing the clicks to the CW stop will tell us where we are in the range.  Then I recomend that you return each screw to its prior location.  Should you change these screws at any point in the furture you will be able to return the pump to this position if things get messed up.


Let me know how many CW clicks you make in each screw.  If your unsure of anything just ask.  I have a spare fip here that I can take pictures that may help.   Send me a private message with your email address if you want to take some of our discussion off line.

[ja17]

Hello,

I checked the BC on nine used injection pumps here today. They were all good.  I did see that there are two different styles. I also measured the extension of the contact pin.  Here in Ohio at 770 ft above sea level, the pin  extended about 21mm from seat of the hex to the tip of the pin (all shims removed). 

[jeffc280sl]

Thanks Joe,

I understand there are two styles.  My pictures show a collar at the end of the threaded section and then the pin.  I understand the other version does not have a collar.  This makes measuring reference points very important.  My failed BC measured 24.82 mm from the seat of the hex, no shims to the end of the pin.

[ja17]

Hello Jeff,

Yes the measurement was the same on both styles.  Your elevation may be a little less, so your standard pin extension would normally be even a bit shorter (richer).

[jeffc280sl]

Given their are two styles of BCs, I think it's important for those measuring the pin length their BC to do so from the base of the hex nut to the tip of the pin.  A good BC should measure approx 21mm.  A bad BC will measure approx 24.8mm.

[jeffc280sl]

Had some time this afternoon so I took my bad BC apart and unsoldered the lid on the bellows.  Here is a picture of the components.  At 550 feet altitude the good bellows measured 22.5mm thick.  When the seal of the bad bellows brokedown outside air was sucked into the partial vacuum chamber inside the bellows.  When this occurred there was no vacuum in the bellows chamber to oppose the 26mm spring and the bellows expanded to max thickness.  This forced the pin outward to its max length.  The spring is very strong and to give you an idea I can compress it about 1.2mm between my thumb and index finger.

I believe the top piece is compressed on to the bellows section and soldered in place.  There is an extremely wide solder suface on both pieces to get a tight seal.  Notice the hole in the center of the inderside of the top.  I think the partailly assembled bellows was then taken to a chamber
with a partial vacuum.  A small round cover plate was silver soldered to the top side to seal the chamber and maintain a vacuum inside the bellows.

This is the final evidence of the failure mode for the BC.  When the vacuum seal is lost the spring inside the bellows expands forcing the pin to deploy at max length.