GMB mechanical fuel pump failures

[zKars]

I have addition news on the topic.

This week I had reason to replace a stock pump on a very old and neglected 71 that i was trying to resurrect. I was just doing a quick job to try to get it running, so I grabbed one the new GMB’s I had laying around that i was previously investigating,

Anyway, took them both apart, and slapped the top of the old Nikki (look fine) on the bottom of the new GMB and stuck it on the car. My prior guess about the GMB check valves being bad made me do that. It sucked fuel and I got the car started (another thread about Doing it for the first time in 40 years)....

That was Tuesday afternoon/evening. Maybe ran the car for 30 minutes on and off. This morning I cleaned the carbs more and tried to start the car again. Guess what. No sucky, no pumpy. You have GOT to be kidding me. 

So long story short. The failure mode on this pump is no mystery. The pivot point on the mechanical arm seized up preventing the arm from being able to oscillate.  It didn’t seize up solid, you can move it easily, but it was stiff for lack of a better term. The spring is not strong enough to push it back when it is in the up position.

I have video of me moving the arm and having it stick/stay in the min and max position. 

The spring, when compared to an OEM spring is about 1/3 the compressive strength. This is a rough feel test, but it is MUCH weaker. I also found the spring laying loose in the arm, dislodged from its end points. It might be weak enough that it falls out after a bit of use, and the pump stops working as the arm has no ability to return to the “out” position. 

No amount of lube sprayed at the pivot point changed the friction one iota. The friction is smooth, is just much too high to allow the pump to operate what that wimp of a spring. I did put the stronger OEM spring in for a test, it was not strong enough to return it. An OEM pump arm is totally loose, nearly floppy on the pivot point.

I have yet to check two other new GMB pumps I have to check if the rotation friction at the pivot is less or the same. I did initially put the pump in the vise and manually actuate it when I swamped the tops, and it felt normal and returned itself with the spring pressure.

I’m also thinking that you are also actuating the diaphragm, stretching it up and down, as the arm moves which provides some additional spring  pressure to the system. Is this enough to overcome the friction I have in this pump? No. 

Clearly without the spring in place, in this pump, it cannot and does not operate with the diaphragm spring effect alone. It worked initially, then quickly failed. 

 

OEM top, GMB bottom

Little tab at the end holding spring

The other spring perch is just a little bump. In its defense, the OEM pump uses more or less the same way to capture the spring. 

There is a geometric difference, as with the GMB spring, when you actuate the arm, the spring just BENDS as much as compresses. The OEM pump does more or less the same thing, just less bend, more compression. This contributes to the willingness to dislodge I believe.

Edited November 5, 2020 by zKars

[zKars]

More to add. I have taken apart the GMB pump to see why the arm became sticky.

I also inspected the other new GMB pumps I have and they are both free moving of course. Further, if I remove the spring, the diaphragm spring action is sufficient to return the arm after each stroke. Now clearly you need enough spring force to ensure the arm returns when the engine is running at 7000 RPM. Point is, if the spring falls off its perch, the pump may still “work” at the lower end of the RPM range. So my spring off its perch isn’t the main failure

So what happened is that the arm rubs the inside body of the pump near the pin, causing enough friction to make it sticky. I hacked off the housing where the  sealing balls are located to get the balls out. Zip zip gone. Then the pin through the arm just fell out. Loose. The arm came out with a tiny tug. The pin is just placed through two holes in the arm, and sits and rotates in the body of the pump. No bushings. Steel pin rotating in aluminum body. Oil is the only thing keeping it from wearing quickly.

Now to the issue. The U shape of the arm is simply a bit wide and is rubbing the body, just above and into the body near the pin. I went to put the arm back in and put the pin through it to start seeing what’s going on, but I couldn’t quite push the arm in far enough to get the pin started, then saw that it was rubbing the body. I can easily push a bit, put the pin back through and have the same stiff rotating arm that hangs up. 

In one picture below you can see the shiny spot where it rubs the AL body on the front side. 

Now the mystery is that it was loose to start with. How did that fairly thick, pressed into a U shape, arm “expand a little” and “become tight” with a bit of use? I do not have an answer for that yet. I can see in the other two pumps that there is room on either side, not much, but clearly enough. 

The U shaped arm is not particularly symmetrical and the hole through it is different on one side, as the sides are not perfectly parallel. When I put the pin through and attempt to hold the arm “vertical” relative to a square, the pin is not sitting quite straight through it. 

So I thought it was going to be a dragging pin in the body that was causing the friction due to oil not getting in there,, but nope, the arm expanded like a noddle when it got wet with oil I guess. Maybe it “Relaxed” from its obviously press formed shape when it got hot? I’m reaching here.....  metallurgists out there?

I do know that if I just take a BFH to that arm and make it the “right” width again (or grind a bit off, or widen the inside of the body a bit), that I will have a wonderfully performing pump. Until the pin works it way out of the body..... Anyone have any 6mm ball bearings I can put back in there? Oh dang, I’ve ground off the place where it sits. Thread it and put a very short bolt in? Silicon? Bit of epoxy?

 

Shiny spot above the hole. Bit of black Sharpy above that where I applied some to the body to see where the arm touches.

You call that square? 

Edited November 5, 2020 by zKars

[zKars]

Well, learning is SUCH a complicated business, Trial, error, more trial, more error. 

So, it turns out that the U shaped arms are not changing shape and growing wider magically. It is the outer aluminum cast case that is changing shape, ie getting narrower, causing the arm to rub and stick. And it twas “I” that was causing the squeeze. In my vise. I was holding it as you see in a couple of pics above, very gently squeezing the case just enough to hold it. Well, the wall thickness of the sides of that GMB case is SO THIN that I was actually distorting the case inward, not just momentarily reducing the clearance when its in the vise. When it failed, and I brought it in to diagnose, I put it in the vise, THEN noticed the arm was binding. No surprise now. I then assumed that was the issue and the above couple of posts resulted. 

Note that I cannot cause the same distortion and tight arm motion when clamping an OEM pump body. Much thicker stuff. Big surprise.

Thus we are back to square one. I have reassembled this pump with all its original parts and it sucks and blows just marvelously. It is not clear again what caused it to quit after my initial 30 minutes of use. Now I’m down to the diaphragm plunger getting mis-aligned in the arm hook, or the spring having fallen out of position causing poor return action. Hard to believe at cranking RPM’s that that would be an issue.  Not check valves, I had the OEM cap with their check valves in place during the on-car failure. Has to be something mechanical. 

The search continues.  

Another lesson in forensics. 

[grannyknot]

Jim, can you transfer the new rubber diaphragm to an OEM housing?

[SteveJ]

1 hour ago, grannyknot said:

Jim, can you transfer the new rubber diaphragm to an OEM housing?

I did that on a fuel pump. I transferred the diaphragm from a Delphi pump into a stock housing. I can't give a full report because I have to change out a rotten fuel line at the tank.

[zKars]

20 hours ago, grannyknot said:

Jim, can you transfer the new rubber diaphragm to an OEM housing?

No. The length of the rod and shape of the end of it are different. The oil seal is completely different as well. Pity. 

If you were brave enough or clever enough, you might be able to grind off the rivet in the center and steal the diaphragm. Just have to figure out some way to re-rivet on to the old stem. Bit tough I think.