Vacuum tank testing/proof of function?

[m-mman]

The car 1929 Cadillac. I have been filling the vacuum tank with the electric pump and having it gravity feed into the carb to get it running after the rebuild. However it has poor performance when test driving and I suspect it was running out of fuel. So its now time to test/prove/fix the vacuum tank. 

 

Put the brake bleeder vacuum hand pump to the fuel line and it fills the bottle easily. So gas tank & lines are good. I Remove the vacuum tank from the car for further testing.

Connect the vacuum tank to the vacuum inlet on my running 71 Cad (good reliable 15" of vacuum)  Then plumb a vacuum gauge into the fill port on the Vac tank. The valve in the glass bowl filter at the bottom is shut off. Tank being held vertically. 

 

Cover the fuel inlet port and no vacuum develops(!) UNTIL I plug up the atmospheric vent port Then it starts to rise S-L-O-W-L-Y ultimately reaching 10"

Open the vent port and it drops quickly. 

 

Cover vent port and it sometimes rises and then doesn't. Play with the fuel inlet valve and the cracks in the lid seem to be leaking. 

Cover the cracks with JB weld (currently setting up) and now I have some questions. 

 

The vent valve seems to be failing. How do you prove that this is the problem besides plugging it up as I have?  

Then if it has failed how do you fix it? 

 

If the flap on bottom of the inner tank is leaking, will that keep it from sealing and developing vacuum? (the filter valve is closed) 

It just hangs there and swings easily BUT how do you know/test that it is really sealing well under a vacuum situation?

 

The fuel inlet fitting, how does it seal? The threads seem to be straight (not tapered pipe) and there is a tapered area on the fitting and in the lid. BUT it also seems like there is a space for an O-ring(?) is there supposed to be one? 

 

While the outside of the vacuum inlet fitting is big, it has only  a small hole to draw vacuum from the tank. I would think that this would delay building vacuum in the tank and slow the whole process. Should it be enlarged?

In testing I have been using a large diameter brass fitting that would accept the rubber line from the 71 and it still draws down very slow. I would think drawing through the small hole it would be even slower. 

 

Thanks for any thoughts & suggestions

[Spinneyhill]

Have you searched these fora? There have been some very good documents posted on servicing vacuum tanks, esp. Stewart Warner tanks. Here is one thread about it...

http://forums.aaca.org/topic/306192-1929-desoto-stewart-warner-vacuum-fuel-pump/?tab=comments#comment-1709537

 

Here is another, which includes an article on the SW tank.

http://forums.aaca.org/topic/309809-vacuum-tank-woes/?tab=comments#comment-1740040

[m-mman]

Yes I had seen the discussion regarding the White truck, and the Stewart Warner  PDF was worthwhile. 

It is a systematic series of tests that can prove each component that seems to be scarce. 

I'll keep working on it 

[wmsue]

There are a few people that visit this forum regularly that are far more knowledgeable than I on vacuum tanks. They have helped me for sure. 

I'm sure they will read your post and respond.

I can check my Dyke's manual and see if there is troubleshooting info for your model.

What model # SW tank do you have?

 

 

Bill

[cahartley]

By 1929 your tank SHOULD have the inner tank which has a VERY narrow flange.

That one requires one CORK gasket which goes on top of the tank flange.

If you used paper it's undoubtedly not sealing.

 

 

[old car fan]

Make sure your brass seats are tight in the top,can be deceiving.A yes flapper valve will cause problems.Also check your top for flatness,sounds like you need a top.

Edited July 14, 2018 by old car fan (see edit history)

[m-mman]

I bought a 'repair kit' several years ago as I was starting the project and didn't know much (I still don't) It came with a cork gasket. 

When I took it apart this time I reassembled it with some permatex to give a better seal. 

 

The JB weld I put around the fuel inlet hole hardened over night and it seems to have helped. when I connected it to the 1971 vacuum source today - and when I sealed up the filter, the vent and the fuel inlet, it quickly rose to a full 15" that the 1971 was putting out.  ? 

BUT

open the valve on the filter and the vacuum drops quickly. I think this proves that the flapper valve is NOT holding/sealing

OR

Open the vent port and the vacuum drops quickly also. I think this proves that the vent port is not sealing. 

 

I have a couple of old parts tanks I got at a swap meet. I removed each of the inner tanks (same as mine) and I dipped them into a bucket of water. Fill the tank and each drains easily However, on one the water will seep in and will fill up the tank. If you put your finger and touch the flap momentarily it will seal and not leak water. 

The other one seals without a problem. Both flaps look the same. Obviously one has a flaw somewhere. How might one recondition the sealing surfaces on the flapper valve? 

 

So, perhaps I can take my tank apart and substitute the 'good to seal' inner tank but that would still leave the issue with the leaking vent port. 

The little 'needle' seal. It is all riveted in place . . . . ?  How does one dismantle it? or clean up the cone seal and the seat in the lid? 

 

[m-mman]

Update on my work repairing a vacuum tank

Thank you to Tim Long for talking to me on the phone and helping me understand testing and resolve my issues. The following is what I have learned and might assist others.

 

Vacuum tank – It must have good vacuum source. Is the engine a good source (timing? Carburetor? Vacuum line to the tank?)  My 1929 Cadillac engine was drawing 15”. However, for testing you need a consistent source and the 29 wont run without fuel, so I attached a long rubber line to my 1971 Cad and used the 15” of consistent vacuum from that car.

 

Some tanks have a spare hole in the top. I guess it is for filing/priming the tank. It can also have a nipple screwed into it to connect to a vacuum gauge and measure the vacuum in the tank. I did so and with all the ports plugged, the best I could draw in the tank was 10”. (obviously leaking somewhere) I found cracks in the pot metal near the fuel inlet in the lid. I covered them up with a coating of JB weld and was then able to draw 15” into the tank with all ports sealed. <hurray> Now on to testing the valves.

 

Without getting into how a vacuum tank works just know that there are three functional ports on the top.

1. Fuel in from the tank 

2. Vacuum from the engine

3. Vent of vacuum to the outside.

There is also a flapper valve in the bottom of the inner tank.

 

Remove lid and inner tank. Examine flapper valve. Of the many that I have examined (just 3) the little flap always looks fine, so you need to test it. It opens to allow fuel out to the carb (never a problem) but it needs to seal (vacuum tight) to draw fuel in from the gas tank.

 

To test - Hold inner tank against your face and suck(!) The flapper valve should close and seal. Blow and it should open easily. I also tested it by quickly dipping the inner tank into a bucket of water to see if it would seal against a liquid. On two of the 3 tanks I have, about ¼ to ½ inch of water leaked in and it then sealed. I suspect that it took that long for the water pressure to build up against the outside of the flapper. On one inner tank the water continued to flow in until the tank was filled even with depth of the bucket.

 

The flapper is bakelite It seals a brass surface ring that is embedded in it against the brass fitting. Tim suggested using fine (320) sandpaper to clean up the surfaces, which I did but after sanding I also saw no change in the sealing capabilities of the two good inner tanks. (either sucking or dipping into water) I then pronounced them working.

 

There are two needle valves operated by a float mechanism. There are levers, pivots and springs. They do not seem to be disassembable. They are riveted together. If you have one that is bent or broken it seems that the only solution is replacement. Tim can supply new springs, and maybe he can fix the levers and arms, but you certainly will not be able to.

 

But all this float and linkage stuff does is open and close two needle valves.  One to turn on and off the vacuum from the engine and the other to vent the vacuum in and out of the inner tank. Again they are tested with your mouth. Screw in a hose barb or nipple and then attempt to move air in the proper direction and at the correct position of the float.

 

With the float dropped down you should be able to suck through the engine vacuum port and NOT be able to blow air through the vent port.

With the float up you should NOT be able to suck air through the engine vacuum port (engine vacuum shut off) and you should be able to blow air into the vent port.

 

My two needle valves passed air in both directions with the float in both directions. How to fix it?  Tim gave me the secret – polish the seats with toothpaste. Put a dab on the needles and seats and spin the shaft of the needle with your finger. It is NOT EASY to get your fingers between the levers and linkages but it can be done. However it also seems that you don’t have to spin them fast or long (I worked me for 5-10 minutes each)  I stopped a couple of times and ran water through the holes/valves and blew some compressed air through them to help displace the crud that might have been removed through this lapping process.

I was shocked(!) both valves began to seal and vent with the proper direction of the float.  <hurray>

 

I cut a new top gasket from cork (composition actually) Tim says he can supply a new one if you need it. Tim says that no additional sealer or chemicals are necessary. 

 

Assembly is tricky, getting the long pin at the bottom of the float into the hole in the inner tank takes patience (how did they do this in production?) but after that assembly is pretty straight forward.

 

I reattached it to the 1971 vacuum source (and with the fuel inlet sealed) it drew down 15” in the upright position. This is the situation when it is sucking fuel from the gas tank. You can put your finger over the fuel inlet and you should feel a vacuum.

I then flipped it upside down (this moves the float to the full position) and it correctly stopped drawing engine vacuum and quickly vented the tank through the vent port.

 

Flip it upright and it sealed the vent port and opened the vacuum source. <hurray!>

I then installed it on the firewall but ran out of time to test it under actual working conditions. But I am hoping for the best.

 

Thank you, Tim, for your information and insight.

Edited July 24, 2018 by m-mman (see edit history)

[hchris]

Quick check to see if its drawing fuel when installed, put your hand on the brass fitting or pipe on the fuel inlet side, you will feel it go cold as fuel passes through into the inner tank; when the inner vent valve opens fuel flow will stop and you will feel a noticeable warming of the inlet line (fuel will be dumping from the inner to outer tank). This will be a cyclic event as the inner tank fills and dumps, listen closely enough to the vac tank and you will hear the click as the fill and vent valves open and close. 

[old car fan]

I also agree with Chris .I am worried about your top,already in the deterioration mode.I have tops,and the rebuild kits @ 25.00 are a great deal.Thanks for the compliments.now let's hear it run.

[m-mman]

Hurray it seems to be functioning! 

Tim you were correct the old lid was deteriorating where the fuel line screws in . When attached the tension from the line pulled out the nut and broke the casting.  ?   I almost called you but remembered that based on sound advice I picked up two parts tanks at the Bakersfield Model T swap meet a few months ago. I didn't figure I would ever use them but I would dismantle them to better understand the operation. Turns out one had the exact style of top my original tank had. 

 

So, remove it clean it, clean up the needle valves with tooth paste and begin sucking and blowing on the ports to verify their function. Reassemble the tank and the new lid, attach it to my 15" vacuum source and With the fuel intake plugged the tank easily drew down to 15" and held it until I inverted it and the float shifted the valves and it dumped its vacuum. Set it upright again and it began immediately pulling 15" of vacuum. 

 

Install it on the car. Use the electric pump to get fuel to the carb so that it will run. Start the car. I plumbed a vacuum gauge into the priming hole that was in this lid. (not all lids have this port) This allowed me to sample the presence of vacuum in the center tank. 

 

Initially there was no vacuum <yikes>  But I discovered that was because the tank was  full and it was not calling for fuel. Rev up the engine to empty the float bowl in the carb and sure enough the float in the tank drops,allows vacuum into the tank - about 5" of vacuum - (it aint much, but it means that it is sucking fuel) Then when the tank is full it shuts off the vacuum. Thus allowing atmospheric pressure into the tank and this allows fuel to flow by gravity into the carb.  ? 

 

I made a couple of short videos of this happening. I think this is the definitive way to watch a vacuum tank function. 

 

 

 

I have now driven the car around a bit, both on the flat and a hill climb and the vacuum gauge tells me that it is functioning normally sucking in fuel and shutting itself off. It did not stall going up 3-5 minutes of hill climb so it looks like it might be reliable. 

Now onto fixing other stuff to make it ready for some touring. 

 

Does anybody know the cruising (top?) speed I can expect from a 29 Cad? ?

 

Thank you everyone for your help. 

Jim

Edited August 7, 2018 by m-mman (see edit history)

[hchris]

Well done, great initiative on proofing the vac system. My only query is the apparent short time interval in which the vacuum appears to be drawing fuel, although I dont have anything to measure it by, I`m thinking that the inner tank would hold about a pint ?? and it would take a little longer to fill  before it dumps.

 

The acid test of course is on the road, and particularly on a long steady climb with lots of throttle and minimal vacuum; keep us posted I`m keen to see how it behaves on a lengthy run.

[PFitz]

Running out of fuel on a long hill with a properly functioning vac tank is usually because the driver is lugging/working the motor too hard and getting the intake vacuum down near zero.

 

At typical vacuum of 15 inches cruising on a level road, the vac tank is making 7.3 PSI in the line from the gas tank. That's a lot.   Even at pushing it to 4-5 inches of vacuum on a hill - the point at which some carb fuel enrichment (economizer/power valve circuits) kicks in - the vac tank is still making about 2 psi fuel pressure in the line from the gas tank. That's within the  normal 2-4 psi operating range of many early mechanical fuel pumps that started replacing the vac tank systems in the late 1920s.  At the carb, the more constant force of gravity is causing about 3/4 psi and plenty of volume, so when the engine load gets so great as to cause a too-big drop in vacuum, the only problem is having enough vacuum to draw fuel from the gas tank to the vac tank.

 

Pushing the throttle down so far that the vacuum is too low to pull fuel from the gas tank is not only bad for fuel delivery, it's over working the motor. Better to downshift, get the vacuum higher to be able to pull fuel, plus, take some load off the motor by gaining mechanical advantage with the transmission in a lower gear, plus, it will also increase the RPMs, thus increasing fan and water pump speeds to give better cooling of the motor.

 

Vacuum gauges are not only helpful for diagnostics, they help you see how hard your working the motor. So, if you want to see how much your working the motor when you drive, tee a vacuum gauge into the line from the intake manifold to the vac tank. A vacuum gauge works like a poor man's "load meter". The lower the readings while your driving shows that the motor is working harder. By learning to drive keeping the readings as high as possible, you'll get the best gas mileage and longer engine life.     

Paul

Edited August 7, 2018 by PFitz (see edit history)

[m-mman]

Hchris - I think the short 'filling time' is related to the fact that the tank(s) are full to start with and all the system has to do is refill what was used in the last few moments.

 

If you pull the lid off of a (downdraft) carb, that has a mechanical pump, while the engine is running, you will see that the floats do not drop to empty before they admit additional fuel. They only have to drop 1/16" (?) and fuel runs in under pump pressure. Unlike with a toilet tank, the carb floats don't drop completely & refill, rather they 'vibrate' to keep the fuel level appropriate. 

 

The same thing seems to be happening in the vacuum tank. fuel runs into the carb by gravity (so it is always full) and all the tank needs to do is keep adding another 1/4" to 1/2" of fuel into the reserve (pre carb) tank. 

 

I was surprised that it was only drawing 5" of vacuum to refill. (As describe above by Pfitz I guess that is enough?) I previously measured 15" at the manifold. Cadillac uses a piston style vacuum pump (powered by the cam) and there is a check valve in the vacuum lines to ensure that there is always a vacuum at the tank. When running, the vacuum at the line that supplies the tank is a bouncing 10"-12". (the bouncing I believe is the effect of the accessory pump)  This did not drop below 10" as I reved up the engine as you would expect to see when the gauge is attached directly to the manifold. 

 

I can easily see the advantages to having the additional vacuum pump such that your new Cadillac doesn't stall going up a long hill. It is a luxury car after all. ?

[PFitz]

Not quite. While carbs and vacuum tanks both have floats, they control fuel flow very differently.

 

Unlike the carb float needle, which it's flow rate is subject to counter acting fuel force against the force being exerted on it by the pressure of the float (which changes pressure on the needle as the fuel level changes )  in a vacuum tank the inlet and vent valves are controlled only by spring tension of the rocker mechanism. They alternate between being fully on, or fully off, and do not allow progressive flow like a carb's float needle can. So, the fuel level ranges are quite different - one controlled by engine fuel demand, the other by spring tension.

 

The fuel level in the upper tank of a vacuum tank has the same range of up/down travel no matter what the engine fuel demands are. It's only the time it takes to refill that changes with changes in engine vacuum levels.   

 

The float has to drop, or rise to the same point to flip the rocker mechanism to close one valve and open the other. So, no matter how the vacuum force changes, the fuel level in the upper tank of the vacuum goes up, or down only when the float reaches those points that cause the rocker mechanism to switch the inlet and vent valves to their other extreme of travel. You can see how much the upper tank fuel level changes when the top is off the tank by raising and lowering the float by hand to work the rocker mechanism to open/close the two valves.

 

The mechanical driven vacuum pumps were an improvement (especially for a driver with a heavy right foot ) by not being dependent on lighter engine loads to have stronger vacuum to maintain good fuel flow. As long as there was sufficient rpm, it didn't matter if it was on a hill or level ground.  

 

For majority of cars with a properly functioning vacuum tank they do fine on hills, if you don't push the motor so hard that the vacuum readings are near zero. I've driven lots of my customer's vacuum tank equipped cars on our long central NY hills and never had a fuel starvation problem. But then, I was taught to never lug a motor uphill. 

 

Paul

Edited August 7, 2018 by PFitz (see edit history)

[Grimy]

As a follow-on to PFitz's excellent explanation, the suction valve and the atmospheric valve are connected and the overcenter lever must operate sharply and decisively, so that one valve is completely closed while the other is open.  Sloppy springs can hinder that, as can whiskery corrosion on the needles and seats (use NYLON brush and carb cleaner), as can roughness on pivot pins (same treatment).  Hold the lid vertically, move float up and down while observing tripping action of the two valves, which must snap open and closed.  Loose seats can be CAREFULLY staked.