Insane Hydraulics
Bold and audacious blog about fluid power
Today's blog will show-and-tell a very peculiar malfunction of a Danfoss PVG120 valve PVEH actuator with active fault monitoring. I used the word "peculiar" instead of "unusual", because I am, frankly, not sure if this "malfunction" is easy to re-create, but since I've always avoided using PVE actuators with active fault monitoring (because I am "wary" of stuff that has "mind of its own") I don't have enough experience to confidently label this as a common occurrence.
I put the word "malfunction" in parenthesis for a reason - the failure that I witnessed on our test bench could easily cost our client a new actuator (these come at €500 a pop, give or take) - only in the end it didn't - and now that I said it, I reckon this post could be named either "how to save 500 bucks" or "how to make an easy 500 bucks", depending on where the moral compass of the overhauling party is pointing at. Now, joking aside, whenever I come across an anomaly that can cause unnecessary expense to an owner of hydraulic equipment - I try my best to document and publicize it because I strongly believe that unwarranted expenses, which traditionally constitute 50% of this business, should be eradicated for good. Note that I am not advocating for owners to spend less on their hydraulics, but rather for a better allocation of funds, which in the end makes equipment more productive, and our clients richer. And when our clients get richer, we get richer, too. Never forget that when our clients spend their money on unnecessary things, they end up closing their businesses, and we end up with unpaid bills on our hands.
Let me start with a very brief introduction of PVE actuators (for folks who are not used to working with Danfoss stuff). A PVE actuator is an electro-hydraulic spool positioner for Danfoss PVG directional control valves. In its proportional version - it's, essentially, a device that receives an electric (usually ratiometric) signal and then shifts the spool using an LDVT for sensing its position, and an arrangement of 4 (or 2, in simpler versions) tiny two-way valves in a bridge configuration for moving and holding the spool in the required position:
When a PVE actuator is "enhanced" with active fault monitoring, it deactivates itself whenever it encounters a fault that lasts for more than half a second and requires a reboot (power cycling) to resume operation. A fault condition can be caused by the signal/supply voltages out of range, or an error in the spool position (fun fact - this one is a bit strange for the PVEs because the actuator only considers a position of the spool to be faulty when it over-shoots the requested position or moves in the opposite direction, while any "under-travel" of the spool is not recognized as a fault), or an internal error, like a short in the LVDT or the internal frequency error. In short - just as I said - it's an actuator that has a mind of its own. Safer? Maybe. I still don't like them.
Let me set the stage now. A winch dropped its load down a mine shaft several hundred meters deep due to a faulty brake, and, as you can imagine, the hydraulic motors running the sheaves and the rope-storing drum turned into boxes filled with swarf. Stuff like this happens. So, we helped our friends with changing the design of the winch (making it safer), supplied the necessary parts, and tested all of the key components of the hydraulic system. Surprisingly, the products of the "explosive contamination" remained in the respective motor areas, and I was glad to discover that the rest of the components, like the counterbalance valves, brake valves, DCVs, etc.. survived, and it was one of these components, namely - the single section Danfoss PVG120 - that gave this story its origin:
As you can see, the actuator has part number 155G4074, which is the series 4 PVEH (proportional high) actuator with active fault monitoring. Like all PVEHs - it has a status LED, which indicates the correct operation by lighting green, and a fault condition by lighting (or blinking) red:
So, I connect the valve to our test bench, hook up the wires, power up the bench, turn on the DC power supply - and see the fault indicator lighting red. Before we go any further, let me show you the test setup. It's pretty self-explanatory, you can see our test bench with the P, T, and LS lines running to the valve, and a makeshift ratiometric control connected to a cheap DC power supply:
And now is the time to give you my first PVE-actuator-related tip - it's called "It's over when I say it's over" (accent on "I"). What this means is that quite often a malfunctioning PVE actuator is replaced because a "replacer" does not know that it can be fixed. "Malfunctioning" can be a) external leakage or b) an actuator with no or sluggish response, and you wouldn't believe me if I told you how many times I've seen these babies getting replaced because of the clogged pilot pressure screen or a leaking o-ring on one of the 2/2 valves. So, in case you didn't know, the screen, the check-valves, and the T-orifice are removable (and cleanable), and the front part actually comes off, and you can replace the o-rings if you need to:
But, as you can see in the pics above, the pilot pressure screen was pristinely clean. In fact, I didn't have to disassemble it because the status LED would turn red as soon as I powered the actuator, without any spool movement, and with the control input at neutral, so everything was pointing to the internal failure of the PVE.
I was already taking care of ordering a new actuator, but there was that nagging thought at the back of my mind that "something wasn't right"... I had all the proof in the world to "defend" my case - I had the valve on my test bench, and I could clearly show the client that his valve was powered and controlled correctly (all within the specs, and I even shot a video of the failure, and I will show it to you in a second) and it was not responding to the control signal because it was entering the fault mode as soon as it received power. Nobody in the whole world would say that this was wrong! And yet .. it felt wrong because the valve was working perfectly fine on the winch, and even though I know that stuff sometimes does break randomly, it still felt strange to me. So - I decided to investigate a bit further, just to turn my 99% certainty into a 100.
So I double-checked the voltages, connections, etc.., and all was good. Then it occurred to me that maybe there's some EMI interference from the generic VFD that our test bench uses to drive the main motor - and when I powered the actuator with the main motor of the bench stopped - I saw that the LED was green! And, as soon as would start the VFD the LED would turn red. OK, we're getting somewhere. Very strange though, because both the bench and the power supply were grounded.
Then I removed the actuator from the valve, and placed it on a separate bench like so (the aluminum block is just to keep the feedback pin in the centered position):
And then I saw another weird thing. Now the LED would stay green even when I'd spin the VFD up, but as soon as I would touch the second bench with the end of a hydraulic hose connected to the bench, or a piece of wire, connecting the two benches, the LED would instantly turn red:
The only measurable thing that I discovered with a multi-meter was about 1.5V AC between the benches when the VFD was running. And then, after I did some more testing, I discovered that when I connected the ground terminal of the power supply to the bench with an additional wire - the fault would disappear! And I finally managed to test and adjust the PVG! Here's the small video demonstrating the fault, and how the additional ground wire solved the issue:
So, what conclusions can be drawn from all this?
Unfortunately, I can't say what exactly was causing the actuator to fail like that. I didn't have time to bring over my oscilloscope to see what was happening to the power and signal rails in more detail. It had to be the VFD noise. The signal and power carrying cable of the make-shift ratiometric control was pretty short, and it was lying over a grounded metal plane, but it wasn't shielded - and I didn't have the time to test this with a shielded cable either.
Danfoss says that the series 4 PVEs possess "enhanced electromagnetic compatibility that provides protection from external interference up to 100 volts/meter". I even brought a series 7 PVE with active fault monitoring to the shop the next day, and tested it under the same conditions - and it didn't fail. Maybe Series 7 is better protected against EMI?
I guess correct grounding + shielding do matter for proportional PVEs when they are around an aggressive EMI source like a VF motor drive. I will remember that! Be it as it may - in this particular situation, I can totally see how one could be lead to pointlessly replace an expensive and perfectly functional component. If there's an electrical engineer in the audience, I would truly appreciate your opinion on the subject.