Hydraulic Test Bench - Fixed!

I finally managed to get that pressure tester working on my third attempt! 

Fenwick 3654-T hydraulic pressure testing bench (credit: George Antoniou)

I'd bought it last June  - but the pump wasn't working. It wasn't building pressure. I'd found some instructions for servicing a similar LHS version of the pump and used it to tinker with mine. I'd dismantled and re-assembled it several times in the last year but couldn't find anything obviously wrong with it. I'm trying to sort through some old spheres to decide which one to P/X for new spheres so thought I'd have another go at getting the pressure tester working.

In the pump body is a check valve - a one-way valve. When you pump the pump, fluid is forced up through the body of the pump, lifting the check valve open. 

Check valve open......

While the pump is starting to build pressure, a small weak spring behind the check valve needle initially pushes it down (closed) until the built up pressure is so great that it, alone, keeps the valve closed.

Spring on the check valve.

....check valve closed

On the bottom end of the check valve assembly is a wire gauze filter.

Gauze filter on the check valve

Previously I'd flushed that filter through and it seemed fine. I'd also checked that it didn't stand proud and stop a metal-to-metal join of the valve assembly to the pump body. I'd even pushed the valve closed to see if it contacted the filter - but there were no signs of the filter moving.

On my latest attempt to fix the pump  - and having exhausted all other ideas - I winkled the filter unit out to see whether anything was trapped behind it - on it's blind side. I discovered the filter was a double-skinned lozenge. And what I found was that, on the inside face/ inner skin hidden within the check valve, there was a witness mark where the bottom tip of the check valve needle  had been contacting the gauze. In short, what was happening was that the inner skin of gauze was stopping the needle valve fully closing.

You could close the valve with your fingers (as i'd done on previous efforts to find out what was wrong), because in doing so you tended to exert as much pressure as was needed to do so. However when pumping, the spring that initially closes the check valve was weaker than the resistance of the gauze. This meant that at low pressure, the valve never really closed, allowed leak-back, and so prevented high pressure ever building up to seal the valve properly.

Once I'd worked this out, I simply squeezed the layers of gauze closer together to make the lozenge thinner and, when I rebuilt the pump, the needle valve could fully close and it very quickly built up pressure.

Pump re-assembled

As a first test for the bench, I pressure tested a couple of old spheres to decide whether I needed to buy some more at collect at the D rally next week. I'd bought these rebuilt front front spheres in about 1999 but never had the chance to fit them to my car. They should be 59 bar. Here's a video of one of those tests.

VIDEO: pressure testing a sphere

So - 25 years after bought them, they still have 56 bar of pressure. That's impressive and means they are fine as a back up pair of spheres.