Insane Hydraulics

Site theme image

Interesting Troubleshooting Case - Parker PV + Danfoss PVG32 Load Sensing System With No Movement

About a month ago, I received a Parker PV023 pump for an overhaul. It came from a scheduled equipment stop and was "at the end of its life", so I ended up selling a new pump to the client.

The unit was equipped with the standard MFC control - a single-spool Load Sensing with a built-in relief valve - a classic arrangement that works well for most LS systems. I like the PV series - these pumps have proven to be exceptional. I know for a fact that this is a very reliable pump because it outlasted all Danfoss, Rexroth, or Eaton open-loop pumps that we tried in mining conditions. Don't get me wrong - all of the pumps that I tried performed equally great when new and on the surface - but as soon as they went underground and the operation started throwing blown water coolers at them - all but the PV series "asked to resign".

Note that this comes from a guy who is "not a Parker fan" (to say the least)! Getting Parker hydraulics in this corner of the world somehow feels like biting the bullet every time I try it, not to mention the fact that getting a decent term from these guys is next to impossible, which is why I'd use Danfoss, or even Rexroth over Parker ten times out of ten, but when longevity in "less than ideal" conditions becomes the defining factor, the PV series is always my first choice.

Now you can see why I was so surprised to get a call from the client saying that "nothing worked" after they installed the new pump. In that system the PV was connected to a Danfoss PVG32 multi-section directional control valve (further on - DCV), and according to the site mechanic they had no movement on the functions, even when they controlled the spools manually, and the pressure read only 4-5 bar.

The mechanic responsible for the recommissioning was very experienced, and he, very diligently, checked the simple stuff first and then attempted to adjust both the relief valve of the DCV and the pressure compensator of the new pump, and then when nothing worked, he called me.

I never had an issue with a new PV-series pump before. So, naturally, I was very interested to see if I could discover what was wrong. In fact, since it was the tail pump of a tandem unit, I started to wonder if I had accidentally left out the splined sleeve during the assembly. The mechanic already knew me and how I worked, so when I arrived at the spot - he already had a flow tester mounted at the pump's discharge port. Beautiful! I teed another pressure gauge into the LS line, and we began the tests. But before we go into the test itself, I would like to trace the troubleshooting logic chain:

The adjustments, oil level, suction line, prime mover, and actuators had all been verified, so all I needed to determine was if it was the pump or the DCV that was causing the problem.

The easiest way to diagnose a closed center load sensing system lies in monitoring the pressures in the LS and the P lines. The logic is simple - when you move a DCV spool, it should connect the P port to the work port and the work port to the LS port, which means that pressure in the LS port should rise. Even if the pressure in the P line is low for some reason (like, for example, if the inlet section compensator is stuck in an open-to-tank position), you should be able to detect a pressure rise, which would mean that the LS signal distribution works. After that - if your pump is getting a pressure rise in the LS line and does not respond to it by increasing displacement to raise the pressure in the P line - it's because either it can't (the oil is being vented to the tank somewhere) or it won't (there's a problem in the pump itself). Thus the basic troubleshooting logic can be represented as:

do (activate directional control);

if (pressure in the LS line rose) then {DCV = good} else {DCV = bad}

In fact, even if you don't have a pressure gauge, you can often check the LS network "country-style" by unplugging the LS hose and seeing if it gushes a stream of oil when you operate a DCV spool. Just point the hose end at your face and use the following logic:

do (activate directional control);

if (I am blind now) then {DCV = good} else {DCV = bad}

But, don't do that last one! Seriously, don't!

Back to our test now. After the power pack started, I saw a reasonable 14-bar stand-by pressure at the pump outlet. Then we operated a spool using the manual control lever - and I read 0 bar in the LS line. The logic charts above would suggest that there was a problem in the DCV.

To top it up, since there already was a flow tester mounted at the pump's outlet, and the pump pressure compensator had been set to an unknown pressure - I suggested to pressure-test and re-adjust the pump, killing two rabbits with a single stone.

We quickly connected the LS line to the pump outlet and successfully tested the pump and set it to 300 bar. Of course, the brand-new PV got through the test like a champ, and I had my proof that the component that I sold and mounted was perfectly fine. So... yet another proof that the problem is in the DCV then? It is clear that the LS pressure is not rising, so we should probably check that LS orifice in the inlet section, should we not? And probably five years ago I'd be already dismounting plugs and hoses and whatnot to get to the orifice and see if it's clogged.

Luckily, I know now that rushing diagnostics is a bad practice, and whenever I come to some conclusion - I always take a step back and look at a system again in order to see if I am missing anything and if there is a simple way to double-check my current theory.

Plus - something else was bothering me. The mechanic did mention casually that the pressure was only 4 bar, which was clearly less than the 14 bar that I read at the pump outlet. When I questioned him about it, he said that he did read about 4 bar, but he was measuring the pressure at the DCV inlet section because a pressure test fitting was already installed there.

I checked the pressure there - and... it read only 4 bar. Well - this is lower, indeed, but why? Very soon I discovered the reason - the system had previously worked with a fixed displacement pump, and the pressure line had an inline check valve with another line teed upstream. Further investigation (and by investigation I mean shoving my head in between a bunch of hoses and pipes and wondering if my ears will still be "properly aligned" when I pull it back out) revealed that the check valve had a 10-bar spring and the teed line was supplying pilot pressure for another very large directional control valve. The 10-bar check valve was installed to make sure the pilot pressure never dropped below 10 bar, and it just stayed in the line when the system was later overhauled with the variable displacement pump.

Aha - so this is why we're getting only 4 bar at the DCV inlet! And also the extra line teed to the pressure line was another possible path for the oil to be escaping to the tank, so, surely I would have to blank off that line now and then re-check the system... And, again, maybe some years ago, the wrenches "would already be flying" and I would be removing the lines "with all my heart", but, once again - I applied the good old "step back and breathe" maneuver.

I thought: "Hold on a minute here, even if the check valve in the pressure line does explain the 10-bar drop between the pump outlet and the directional valve inlet, it doesn't explain why there's no pressure rise in the LS line when a spool is moved by hand. I can understand that having such a low pressure in the inlet may be not enough for the electrical controls to work, but no pressure rise at all in the LS port when a spool is moved by hand still needs to be explained."

Then, I carefully looked at the directional valve and realized that the work sections had individual pressure adjustments, which meant that there were individual compensators in each section.

Then I thought: maybe I am just not getting enough pressure to open the compensator bias spring and the oil is simply not reaching the directional spool at all? Then another thought went through my mind - but PVG 32 is a classic pre-compensated valve, and as I said (and even wrote) before - pre-compensated systems are normally open flow limiters, so, surely the oil should be reaching the spool. But, then I thought again, and then looked at the work section cutaway view (thank god for smartphones that allow you to have a digital library at the palm of your hand!), and then I realized that the same section uses the compensator spool to perform both the "compensatory function" and the load drop check valve function, which means that the compensator is kept in the closed position by the bias spring. So, if the pressure in the P line is below the bias spring setting, which was exactly the case here, it is simply not enough to shift the compensator, and thus the oil can't even reach the spool.

This was a beautiful theory because it was very easy to test! So, I bumped the stand-by pressure up to 22 bar - and the system immediately "woke up"! No need to say that I was leaving the client's shop with another smiley in my personal file. If you wonder why I didn't simply remove the check valve - I say that it was the client's choice - too many metal pipes would have to be removed and there simply wasn't enough time for that.

The point of this post is not the hydraulic function that I just described. The case is interesting but highly unlikely to happen, so it's not the point. The point is that stopping for a second and stepping back to ponder for a while about what is happening is better than taking immediate action and rushing in to solve the problem. Human beings are not machines and may need to take several looks at the trees to notice the forest. So, taking a few moments to "think it over" is never a waste of time.

Unfortunately, clients who have their production stopped due to a hydraulic malfunction, often perceive such "action void" moments as a waste of time: "Hey, buddy, you've been looking at your smartphone for the last ten minutes and haven't touched your tools yet! What gives, man?"

Don't bother feeling bad about it. Slow is smooth, and smooth is fast!

Important points: