This shows the crappy union and the good one. The crappy one was very difficult to stop from leaking.
I know Loctite can be very strong. We use the stuff at work, up to 30 quid for a 20ml bottle of the more specialised types It didn’t occur to me to try that. Thanks for the suggestion.
Vic
Problem with this type of union is that both tapers, male plus female , have to be perfect not to leak fuel. This is not easy to achieve in mass production.
I can count on one hand how many times I’ve seen them leak – if the required angular tolerance were had actually a production problem I think the whole world would fall apart. Peter was pretty unlucky in my experience. If you’re trying to seal helium it’s a bigger problem than with gasoline though
In the past (20 years ago, holy cow) I used to fool around professionally with 30,000 psi 7/16 inch ID steel tubing running at -320 F (77K, liquid nitrogen temperature) and the fittings there were pretty challenging. Before pressurizing the system the first time you had to cool the system down and then walk around retorquing each fitting with very long levers, a meter long or so. After you cold cycled the system two or three times most of the fittings would then cycle hot/cold and hold pressure without retorquing. The OD on the tubes was 1 inch, so roughly 0.280 inch wall thickness (call it 7-mm wall thickness) and the seal was on a taper cut on the end of the tube. We (meaning the guys in a little shop in the middle of nowhere) cut the tapers on a lathe by hand.
Pipe threads are neat because the leak path is about a zillion times longer than the hydraulic diameter of the passage, so it takes very little to seal them. Another one of those inexpensive and very useful things developed a long time ago that we now tend to take for granted. The issue with them is angular alignment.
you mean taper threads ? I´d say you will hardly find taper threads in modern “real” hydraulics any more. This is a thing of water plumbing or very old school, well, aviation then. Basically most fittings are straight thread with an o-ring in the right place and taper seats with their high torque metal to metal contact – as you said using long levers plus retorquing the joints. The trick is to form the not so perfect fit while the components show some amount of yield. It is a bit easier with flared alu tubes to go with steel fittings. A tiny o-ring there instead is so easy for the same effect, no brute force necessary. Pairing tapered threads with the need of orientation for certain parts should qualify the designer for sacking him, having quite obviously no clue about sound engineering.
VicP.S. O-rings on the taper seem to be a DIN / metric thing, no longer so exclusive:
O-ring
I think cheap is often better than expensive, if it works OK
(You’re up early Vic )
I want to know what you were doing with liquid nitrogen at 30000psi
The issue I had was seemingly caused by a poor quality (pitted) surface on the face of the fitting.
We changed both unions and here is the job done
The LH hose (the one going into the servo) was also changed, just in case the facing surface was damaged by the rough surface on the union mating into it. These are Teflon hoses, no life limit, tested to 1500psi, yet they cost only about £50 (Saywells, Worthing) so there is absolutely no point in skimping on this stuff.
Pressure tested by running the electric pump against a closed mixture setting. No leaks, not a drop.
Looks good it’s hard to get a straight run of hose at the entry, but for what it’s worth EI doesn’t think its as important with their red cube flow transducer. Mine has a roughly 5 inch radius curve at the inlet, not too tight anyway, and it’s working fine with the factory installed K-factor.
Peter, answering your question, sort of
I want to know what you were doing with liquid nitrogen at 30000psi
The high pressure reciprocating pumps and tubing are used for cryogenic fracking, which was going on even 20 years ago. We used them for something else, but in doing so were able to contribute to their development for that purpose.
Which Loctite ref would you guys suggest for this application, ie. FF transducer fittings ?
Quick check on Loctite website – looks like 567 PST well suited to the task ?