In today's post:
I will describe a somewhat unexpected proportional DCV malfunction that was causing a hydraulic circuit to overheat.
I will show how the problem was diagnosed and solved without replacing parts.
I will talk about the correct and incorrect way to test classic pre-compensated proportional directional control valves.
It's Friday night, the next Tuesday is a holiday, so I decide that this time I will take a Monday off - thus gaining a four-day mini-vacation. And, of course, as usual - I get the call... Now, where did I put my red cap? - and away I fly!
The "malfunctionee" was a Putzmeister SPM 4210 concrete sprayer, which according to the operators, would occasionally fail to move the spraying arm.
The hydraulic system consisted of a 22 CC gear pump coupled to an electric motor, a six-section Hawe PSL 31 directional control valve (standard pre-compensated DCV) and basic actuators - cylinders and hydraulic motors. Nothing special, the most classic system in the world.
The valve had been overhauled by a respectable shop and, as the story went, had been thoroughly tested. I never like taking on jobs that kind of belong "by right" to someone else, but the machine was really important and we turned out to be the closest shop that would have a technician available on such short notice and on a Friday night, so... yeah... I was only hoping it would not take long.
The mechanic responsible for the sprayer and I began the tests as soon as I arrived. Naturally, when you are facing an intermittent failure, you have to keep forcing the machine until it fails in order to come to a conclusion. So, we patiently moved the levers back and forth, making the arm do a funky dance, waiting for a function to fail.
In a couple of minutes I heard the cooling fan kick in, and I knew the oil was at its operating temperature, and pretty soon the valve failed once, then it failed again, and again... Yep - the intermittent failure was there.
The good thing about these valves is that you can easily measure pressure in the load sensing gallery. Since the inlet manifold is configured for a fixed displacement pump, checking the pressure in the LS port is a good way to see how the "LS communication" works. Most of the time such "intermittent" failures are caused by contamination, very often in one of the LS shuttles, so checking this pressure a good way to confirm this.
I installed a pressure gauge in the M port (the DCV pressure inlet), another pressure gauge in the LS port and began monitoring. And indeed - when a function would fail - I saw no pressure rise in the LS gallery - which most likely meant foreign objects in one of the shuttles, and/or maybe a blocked orifice somewhere. When something like this happens the only thing you can do is disassemble the DCV and clean it, thoroughly and carefully.
But I was not worried about that. Taking apart a HAWE multi-section directional control valve and then putting it back together is a painful and time-consuming task, but it is doable. However, I discovered something else - much more concerning.
As I was working the spools, I saw that there was a really big difference between the LS and the P port (inlet) pressures. Sometimes the P would go as high as 240 bar with the LS gallery reading only about 120 bar. I even installed a pressure gauge in one of the work lines to see if the LS pressure was matching the load pressure - and it was.
Now, this was definitely something I didn't like to see. And suddenly the thought that flashed through my head half an hour before made all the sense - I remember thinking - "hmm, these fans kicked in pretty fast..."
Now I could see why - a hundred plus bar pressure drop for the sake of a pressure drop is all but healthy for a hydraulic system.
Anyways - I was in a tight spot. I came to fix one malfunction but suddenly came about a completely different one. On a recently overhauled and exhaustively tested DCV.
So - I thought to myself - "be it as it may, the intermittent failure is here, so the only thing left to do is to take the valve off of the machine, bring it over to our shop, strip it, clean it, see how it's designed (HAWE was never famous for disclosing cut views of their products), put it back together, hook it up to the test bench and see what is going on. "
And I did just that. I (very painstakingly) took the valve apart (I did find small rubber chunks that prevented the shuttles from working properly), cleaned everything, and then put it back together. Not the worst thing one could do on a Friday night at 2 o'clock in the morning, don't you agree?
Then I connected the valve to the test bench, hooked up the two pressure gauges just like I did before (one in the M port and another in the LS port), connected a needle valve to work ports of the section that had the spool with the biggest flow (so that I could induce pressure while operating with functional flow, as opposed to plugged outlets) - and began testing.
I saw a very strange pattern. Initially, I read a pretty decent pressure drop of 10-20 bar between the LS and the M ports, which could almost count as normal. But as the oil temperature rose - the pressure differential was getting bigger, and when the temperature hit 55-60 C, just like when the valve was mounted on the machine, the delta P started getting closer to the 100 bar mark. I also noticed that the pressure differential was worse at higher pressures.
Very strange... How is that even possible?
This is the moment when terminology can be of great help. Yes - you heard me right - terminology! For some reason, I never liked the word "compensator" that much. I think it's misleading. I prefer calling the spools, that are mounted in inlet sections and serve to by-pass pressure gallery to tank, pressure matching spools. The spool's function is to match the inlet pressure to the LS line pressure plus the bias spring induced differential, hence - pressure matching.
But the spool in front of me was doing all but pressure matching. Because I had 110 bar in the LS gallery and 230 bar in the P port. That's very poor pressure matching if you ask me. And I know the bias spring is not a 100 bar spring.
I thought to myself - well, I do know that the LS line pressure is on one side of the spool, and the P line pressure should be on the other side of the spool - but I am measuring the pressure in the P line (the M port), and not the side of the spool directly. How can I be sure that the pump pressure and the spool pressure are the same? So I quickly devised a fitting (thank God we have a lathe in the shop) and put it in place of the compensator plug to read the "spool-shifting pressure" directly, and - sure thing - I did see that it was significantly lower than the pressure in the P port! So the poor pressure matching spool was closing too early, causing the pressure in the inlet to rise to the relief valve setting.
Now - I am sure that If I had a nice cut view of the inlet section - I would probably come up with a correct theory earlier. But I didn't manage to find a cut view (or an exploded) view of this model Thank you so much HAWE. You should take lessons from Danfoss. Read their PVG32 catalogs! And cry!
In any case - who needs drawings when you have direct access to parts and a head on your shoulders? And if you really look into how the inlet pressure matching compensator is designed - you will immediately see what is causing this malfunction.
Check out this sketch. The compensator spool area that is "sensing" the pressure in the "P" gallery is separated from the "T" gallery by a narrow wall, and the orifice in the spool is only 0.4 mm in diameter. With sufficient wear and/or scoring of this thin sealing ring surface, leakage from the pressure matching side of the spool to the tank line can become so big that the tiny orifice will no longer be able to supply enough oil to keep the pressure equal to the P line pressure.
If you compare this design to the PVG32 inlet module, you will see that constructively it is the other way around, meaning that the P-side of the spool is in the P gallery, and thus this design will not be affected by such wear.
Quick fix? Increase the orifice diameter! And, of course, advise the client to order a new inlet module. As I increased the orifice from 0.4 mm to 1.0 mm the pressure drop returned to its expected 7-8 bar, and when I re-mounted the valve on the machine - the cooling fans took a much longer time to kick in.
So, the problem is solved but the question remains - why didn't anyone notice this seemingly apparent problem during tests?
The answer is simple. Very often, when a proportional directional control valve is tested in a shop, it is tested with the work ports plugged. This is a valid test because it allows you to adjust pressures and verify for the presence of leaks, but that's it! This test does not verify the flow function of the DCV - the very reason why such a valve was chosen over a simpler and cheaper non-compensated DCV. Why then people do such tests? For convenience! It's easier this way. Less fuss. Is there something that can be done about it? I don't think there is...
Well then - you say - you are right. The shop did a sloppy job testing the valve flow function and all... we get it. But why the hell didn't they see the intermittent malfunction? They did tell you that the valve had been tested "exhaustively", didn't they?
Well, friends - I have an answer to this question as well. As it turned out - the DCV had been tested on a smaller test bench, again - for convenience reasons. And, as it turned out with a smaller flow (the original machine worked with 35 l/min and the shop tested the valve at only 8-10 l/min) the directional control valve would still fail, but far less frequently. I actually had the chance to verify it on the actual machine, because it was equipped with a much smaller gear pump coupled to the diesel engine and connected to the same circuit via a check valve. The purpose of this pump was to serve as an emergency pressure source just to store the arm in transport position in case the main pump failed. When we tested the system with the diesel - it did fail eventually, but only a couple of times in half an hour, and with the 35 l/min supplied by the pump connected to the electric motor the valve would consistently fail every other five minutes.
I guess the two biggest bullet points I would take form this case would be:
When you test and adjust a proportional DCV in a workshop, it is very important to
a) not skip out on verifying the flow controlling function, because the simple "plug the work ports" test can easily mask a compensator-related problem, and
b) since you will be using flow for more than just opening the relief valves, make sure that you use the flow (and temperature) as close as possible to the operational ones.
This, of course, does not mean that you have to go through a lengthy test each time you work on a proportional DCV - real life, understandably, imposes real-life demands. But it is important to know that a simpler "plug'em all" test does not provide you with the same guarantee of the valve's function.