This is the second part of the "ideal pump test bench series". If you haven't read part one - you should definitely check it out - I explain how this whole list came to be.
No more explanations in this post. Just bullet points from now on.
In our case things were simple - there was a bell housing with a large opening bolted to the motor, a jaw coupling on the shaft, and then we had a set of intermediate flanges and couplings to accommodate various flange and shaft sizes. However, simple does not equal easy - every time I had to mount and dismount a unit resembled, pretty much, a standard industrial installation. I would always have to tighten all the bolts (sometimes a lot of them), and make sure the coupling gap is right. No matter how you look at it - it's a lot of wasted time, which is a very bad thing, especially for something you repeat every day.
An ideal test bench will tolerate none of that! And since putting a unit in and out of the bench is the one operation you can't skip - the mounting system design is the heart and soul of a pump test bench. Even you you can shave off a single minute of the mounting process by automating something - you will have saved an hour 60 tests later. I have scrubbed the internet for pictures of pump test stands in my quest for the "perfect mount" - and I can tell you that the technical solutions are plenty! But the main idea is always the same - installing a unit should be lighting fast. It should be a simple matter of choosing the right pilot and shaft adapters, and then "snapping" the pump into the unit as if it was a "giant fast coupling". Ideally - shove it in, press a "fix it" button, and you're done. None of the coupling position alignment business.
As the size of the pumps you repair goes up - so does the power requirement of your test bench. This is straightforward. We had a 70 kW four pole electric motor, that was traditionally running at 1000 rpm, so we got maybe half of the rated power at the shaft end, and very often this was not enough. Of course, since most of the time the tested units had variable displacement, we always found a way to reach or at least get pretty close to the unit's maximum pressure by reducing the displacement - but it never felt right, if you know what I mean. Like testing a sniper rifle on a 50-foot shooting range with under-loaded cartridges, and then delivering it to a client with claims that it's "fully tested and field ready".
A good test bench will have no power limitations. An ideal pump test bench, aside from packing a decent amount of "muscle" will also allow for "power recovery" - something that gave me a lot of sleepless nights, by the way. I have always been of the opinion that pump pressure tests could be "smarter". Why waste all of that energy? While never a problem with smaller units, it becomes a serious issue with big ones, because even if you have enough of the hp "in house", can you really dissipate all of the heat efficiently? OK, you can. Once. But what if you have to test ten pumps in a row? And what if you have 100 kW available but need 150? See? Being able to recover some of the power that's usually lost to heat in pressure inducing elements of your bench is a big thing!
How one would achieve that? Well, that what the sleepless nights were about. I gave it a lot of thought in the past, and I believe that several solutions are possible, but they all revolve around a variable displacement hydraulic motor, that will absorb a part of the test flow and then "give it back". I am sure someone has done this already. I would love do design and build one of these, I can promise you that!
If a hydrostatic drive was used in my ideal test bench (I believe I would definitely go for the hydrostatic solution) - I would love to see it being able to work not only as a torque "supplier" but also as a torque "receiver", so that I could use the bench to test "motors as motors". This is a great way to test motors, and in a well-designed hydraulic system, energy recovery would be possible as well.
But energy recovery is not the point. We had a second test stand, that had such a system built around a large fixed displacement Linde motor. I must tell you that being able to load a hydraulic motor as you please is a great testing experience. Of course, another way to test a hydraulic motor would be to test it like a pump, and it's valid all right, but being able to actually see how a PC of a variable displacement motor kicks in on a braking test stand is a whole different business - you get to actually see how it works, sounds, and how the pressure behaves.
In any case - with a hydraulic drive such a system is not hard to implement, so I say a perfect bench will have this "extra" as well!
Our old stand had a proportional hydraulic joystick (single axis with a friction detent) as the source of pilot pressure. I worked, but this solution was definitely not ideal. Whenever a certain pressure was needed - getting it from the friction detent would turn into what I liked to refer to as the "bump the stick nightmare", for the only way to reach a desired pilot pressure would be through a series of progressively smaller bumps applied to the lever. Say, 15 bar were needed - it would be like - bump it - 10 bar - bump it - 12,3 bar - bump it - 14,1 bar - bump it - 16 bar. "You piece of ..." - bump it - 13,5 bar - "Son of a ..." - bump it - 15,8 bar "Ah.. F# it, good enough!.. " - You get the picture. Joysticks are good for skid steers, not for test benches.
Ideally, I would love to have my pilot pressure sources to be independent knob-controlled pressure reducers, coupled to on/off valves, so that I could easily set a desired pilot pressure level, and also be able to turn it on and off without changing the setting. I believe this is much more convenient.
I've mentioned the "magical software" in my previous post, but I want to talk about it more because it is another essential part of any respectable test bench.
We had this very outdated data logger, which we'd have to wire every time a test was performed, then manually log the data points (and when I say "manually" - I literally mean - by hand - for one would have to physically press a button for each of the data points), and then take the heavy box to the office to transfer the data (via an RS-232 connector!) to the PC. Another lengthy procedure that could be improved. The logger was also pretty limited as to how many data points it could hold, so any lengthy (or fast rate) logging was out of the question.
I envisioned a PC permanently hooked up to the bench sensors and a software suite that would allow for all sorts of "goodies" - like multiple unit display, math operations between channels (something that I actually implemented in my wireless monitoring app, so I can tell first hand that this option is very handy), unlimited memory for logging (even at fast sampling rates), graphical data representation, and, of course, no fuss with carrying loggers around and uploading data to PCs and whatnot.
We had the simplest solution possible - a big needle valve with a handwheel in place of the knob. Why a hand wheel? Because an unbalanced restrictor is very hard to turn at high pressures - hence a bigger wheel. It worked for years. The main advantage of this solution was how fast one could raise the pressure when it was needed. The pressure characteristic of a restrictor is not linear, and the last half of the pressure range of a typical pump was within the last turn of the wheel, which allowed the operator to simulate both slow and fast loads.
But the fact that the restrictor has a very non-linear pressure/flow relation also meant that keeping a pressure at a steady lever was practically impossible without "active correction", and in certain situations, this was required. A piloted relief valve, for example, would excel in that situation.
Later the needle valve was replaced with a remotely piloted relief valve, and it was then when I began missing the old restrictor. Now I could maintain the testing pressure stable but lost my ability to induce "spiky" dynamic loads at will. Since I really like both solutions, I think that in my perfect test bench I would install both. One dynamic and one stable.
Another advantage of having two pressure-inducing systems in a test bench is the ability to test the torque summation controls of tandem pumps - something you can't do with a single pressure-inducing system.
Since we already have our software in place and are using a relief valve in one of our pressure channels - why not go a step further and introduce automation? Of course, it can't be used for all pumps, for all pumps are different, but I see that for a company that tests a lot of units of the same model, developing an automated testing routine could be a time saver.
I already mentioned this in my previous post - but there's another advantage to having the "controlled loop leakage" valve. I believe that this is useful for testing tandem closed loop pumps that are equipped with a single charge pressure pump. When you test such a tandem "one half at a time" - you are putting the pump in certain advantage, so to speak, because you are charging it with an oversized charge pump, so being able to "steal" some of the charge oil in order to simulate the losses of the second loop seems like a better test to me.
Small loggers have small screens - and when you have several readings you want to monitor at the same time - there are only so many that can fit. Another advantage of a PC-based system would be being able to display the readings on a large screen.
I would replace a pressure gauge every week. Seriously. No matter how much care we used, every now and then a 40 bar gauge was connected to a 400 bar line and boom went the gauge! Analog gauges are a nice thing to have on a test bench, but in an ideal one all of the analog gauges will be a) large and b) over-pressure protected.
Pumps are heavy so help is needed to handle them. We had a manual chain hoist on a jib crane boom, but I would love to upgrade it to electric.
These are for negative pressures or case pressure logging. "Normal range" transducers are not the best choice for precise low-pressure measurement, so I would definitely include a couple of low and negative range models. Case pressure is definitely something you don't log every day, but sometimes you do need to measure low pressures reliably.
Another option that is very simple to implement. Sometimes I was getting closed loop pumps along with their loop flushing (and relief valve) manifolds to repair and test. If a bench has a couple of extra outlets in the test loop to which you can connect an external manifold "for testing purposes" - you would be "killing two birds with one stone". I can't say that I needed that option that many times, but hey - what's an extra pair of fast couplings anyway?
This one would be used for "sucking" oil from "certain places". Sometimes you get a spill you need to clean, sometimes you have an oil-filed cavity you just opened, and want to "peek inside". I already said that an ideal test bench will have an oil dispensing gun for filling the pump cases and whatever else that needs filling. I believe that next to it should be an oil-sucking gun too. Maybe one of those compressed air waste oil suction systems? I have a large shop syringe for that, but a suction gun should be way more convenient.
Some pump (and motor) controls need an external relief valve to work (think - remote pressure controls). So, instead of using an external valve lying on the bench in a nest of pilot hoses, why not make it a part of the bench? My ideal bench would definitely carry one of these. Maybe even two...
There should be a dedicated set of tools and a decent tool cart which is exclusive for the test bench. Some pumps require especial tools for adjustment, and these should be well organized and stored.
This a "safety" feature of sorts, and may actually be a crazy idea. But I wrote it down, so I am putting it here as well. More than once I've had situations where I would doubt if it was the pump that was acting out or if there was a fault in my test bench relief valve (this can happen too, you know). Well - if one installed on of them gear type visual flow indicators in the tank line of the relief valve - there would be no doubt it the valve was opening, leaking or not. Yes, I am that paranoid.
In our case there was no way to look at the shaft seal because it was inside the bell housing, so if there was a leak - we'd only catch it when the oil would start to drip from the bell housing onto the floor. A simple web cam with a led light pointing at the shaft seal would be nice I think. Things like that are dirt cheap these days. Anyway - this is just an idea. One's got to try it to see if it's valid or not.
Data loggers are very expensive, but flow sensors, especially when acquired in bundles, are reasonably cheaper. I would love a test bench to be equipped with multiple flow sensors for various ranges - for better accuracy. A pump test bench is an instrument, and instruments must be accurate!
If you work hands-on with hydraulic equipment you'll immediately hear me! It's a known fact that 90 percent (and I am very reserved with this number - what I want to say is actually 99%) of hydraulic systems are designed and built by people who will never service them, and most likely have never serviced a piece of hydraulic equipment, which is a shame, because if they ever replaced filters on systems they designed, their hydraulic projects would instantly become much more "mechanic-friendly".
An ideal test bench is easy to service. And the basic stuff like filter replacement, oil change, and tank access is perfect. No mechanic should come to the test bench, look at the filters and say something like - "the asshole who designed this piece of carp should be here changing them elements!"
You can never underestimate how much time a well-organized test area saves. The best investment a hydraulic company that tests hydraulic pumps daily can make is an investment into the "test supporting infrastructure". I am talking about racks, shelves, and whatever is needed to store the thousand and one test fittings and accessories in an organized fashion. This comes from first-hand experience. When I started working with pump tests, we had a couple of large containers of "test fittings" for all types and sizes, and I would spent hours "fishing" for the right fittings, couplings, flanges, and everything else. Pump test bench is a production line of sorts - if you organize your tools, you immediately improve productivity!
This one came out long too, didn't it? I am at the end of my list, but I could go on for hours. I'll stop now, though, that's enough of pump benches for one day, don't you think?