Last week, we looked at gear flow dividers and how "unexpectedly mediocre" they are at dividing flow when pressures between the outlets differ by more than a hundred bar, and today I'll tell you a story of how a gear flow divider "kicked me in the rear" just a couple of weeks ago.
A client of ours had this pretty old raise boring rig crawler that started to overheat oil badly. The drive was based on a single A10VO100DFR which fed a two-section Commercial-Intertech gear flow divider, which, in its turn, fed two independent open-center proportional DCVs - one for each track. The pump was configured as fixed-displacement, with the LS port routed directly to the pump's outlet (via an electric dump valve - for emergency stop), and the compensator was adjusted above the relieves of the DCVs, which were set to 250 bar.
So, I tested the pump, and it was OK, and then when I checked the drain of the divider (quite surprisingly, it had a drain line!) I saw that it was leaking 35+ liters per minute at 200 bar, and it became immediately apparent "who" was heating the oil.
Commercial Intertech gear units are repairable, but we did not have the parts in stock, and it would take at least a couple of weeks to get them here, which was not acceptable because the crawler, like pretty much every other piece of mining equipment I ever worked on, had to be repaired yesterday. Luckily, a supplier of ours proposed a Casappa HDD30.51+51 as an option available for near-immediate delivery, and since the cost of the repair and the new divider were comparable, we opted for the replacement.
Ain't she a beauty? Let us talk numbers now. The engine was working at 1900-ish rpm, so I would get about 90 l/min for each section, which would put the divider at 1800 rpm - close to the maximum speed for the optimum performance range (2000 rpm) - so size-wise this model was more than OK, and since the catalog stated the max. continuous pressure of 310 bar for this model, I was pretty confident that it was a perfect choice, if not an overkill.
Now, what kind of flow division could one expect from it at, say, 200 bar pressure difference between the two outlets? Considering the Bucher's chart that we saw last week, I would "aim" for the 0.6 line (the theoretical 1800 is already 10% below the ideal 2000, and I'd expect the pump to lose some flow rate due to its internal leakage plus the diesel would drop rpm under load too, and also - the Casappa's external-gear HDD is less efficient than the Bucher's internal-gear QXT) - so this would put the flow division at 0.45 x 0.55. When I hear the words "gear motor" or "gear pump" - I imagine volumetric efficiency of 0.9 - 0.95, so when you are turning a 95% efficient pump with a 95% efficient motor - you are bound to get 0.95x0.95=0.9 of the theoretical flow, so the 10% flow discrepancy sounds about right.
Speaking in flow-rate language, this would mean that the 180 l/min input would result in the 100/80 l/min division at 200 bar on a single section. Or better - 95/75 - to account for some 10 l/min "lost" to the pump's internal leakage and the engine speed decreasing under load. Why do I put only one of the sections under pressure? Because it's the worst-case scenario, and it is a track drive, which pretty much guarantees that you will constantly have situations when only one of the tracks is working. Anyhow - I set my flow rate target, connected a couple of flow meters to the DCV outlets (in other words - replaced the track motors with flow testers - just saying this to make it clear where the flow testers were mounted) - and began testing.
It started great. Just look at this near-perfect flow division (with no pressure):
However, when I put the outlets at 200 bar I got very alarming figures:
What I got was:
And before you say something - no, it was not the leakage in the DCVs - it was the divider. And the most disturbing thing was the abnormal noise it began to make at a certain point.
The total flow rate under pressure was about 175 l/min, but even so, I was supposed to see something along the lines of 0.45x175=79 l/min and 0.55x175=96 l/min, and not 66/108!
I told the client that, unfortunately, the divider was malfunctioning and I would need to open it up to see what was wrong. Damn it! I hate when brand-new stuff that is supposed to be turn-key fails (miserably)!
But - you got to bite the bullet when you got to bite the bullet, right? And bite the bullet I did. And here's what I found:
Yikes! That's a lot of wear for a gear unit that ran for about an hour! One side was actually much worse than the other. You can't see this in the picture, but the worst part was the burrs on the tips of the gears which you could feel with your finger. So, why did this happen?
I'll tell you what I think about it. When I received the divider - I noticed that the tag looked like it was made with one of them shop engravers. This is not a bad thing, but it often means that a unit was assembled in-house from stocked parts. Once again - this is totally OK, but it can be an issue for gear pumps (and - by extension - gear flow dividers) - because the supplied unit may well be not run in. I recall thinking to myself. "I wonder if this divider was run in ?" Then I thought: "Na-a-a. It's probably all good. I'm sure they did it. After all, they supply this stuff all the time. I'll just be careful when I test it."
And I was careful. Kind of. I let it run at idle for a minute, then I hit it with 50 bar, then with a hundred, and then I thought it was enough and hit it with 250 - and that's exactly when I heard the strange noise and then noticed the peculiar flow readings.
Anyway, I had to fix it and I had to fix it fast, and, as always, there were no parts in stock to do the proper fix. The bushings and the flanges were perfectly fine, so to fix the divider I would require two bodies, two shaft kits, and four pressure plates with seals. Unfortunately, I didn't have any, and we couldn't get them in time.
So now is the time I tell you about my "I can't believe it freaking worked" fix. Since the HDD 30 uses the same shafts as the Polaris series, I did manage to scavenge four 51cc shafts (two of which had female splines in the tail end):
My big plan for the overhaul was - cut off the tapered ends from the shafts, lap the pressure plates, and then re-assemble the divider and since the bodies weren't scratched that badly (or at least that was what I was telling myself) - I would install it on our test bench and then run it for a while, slowly building pressure one side at a time, in hopes that the new shafts would rectify the bodies into the correct shape. Desperate, I know...
My "run in protocol" was:
I would do this for each side, then say a prayer, and then remount it on the crawler and see how it would fare.
And here are the numbers that I got (in the meantime, the engine guys made some adjustments to the engine controller, and the working speed was reduced to 1800 rpm, so the total flow dropped by about 5% from my previous tests):
Side A gave 80 l/min at 108 bar and 76 l/min at 202 bar,
Side B gave 76 l/min at 110 bar and 71 l/min at 203 bar.
But most importantly the "overflow" to the low-pressure side decreased a great deal:
I don't have the pic for the other side, but it was about the same (with the same discrepancy of about 5 l/min). You can see that at 250 bar, the relief valve of the DCV was already opening - hence the flow rate on the pressurized side dropping to 50 l/min, but most importantly, the free-flowing side registered only 87 l/min, so if my flow of 166 l/min (175 l/min minus the 5% due to the lower rpm) were to be divided theoretically into 0.45 x 166 = 75 l/min and 0.55 x 166 = 91 l/min - the 87 l/min would correspond to a perfectly functional gear flow divider! At the very least, it was way better than before, and it wasn't making any strange noises. I still can't believe it worked!
Now let me tell you about another "malfunction" (you didn't think it was already the end of story, did you?)
As I was running the tests there was already a truck and an inpatient driver waiting outside to load the crawler and take it to Spain (one of the reasons the quality of the pictures is so bad - everything had to be done in a hurry). So - I registered the numbers, it was all good and dandy and then I said: "The divider is OK now, I just need a couple of minutes to see how it behaves at idle, then we'll re-connect the lines and you'll be able to load it."
I put the motor at idle, pressed the lever on the remote control to pilot the DCV, saw some "flow happen" - but (to my horror) as soon as I closed the needle valve of the flow tester the flow dropped to zero at 50-ish bar! Oups! I was expecting the divider to drop efficiency at low rpm, but not like that! This, obviously, was not OK because the crawler had to be driveable with the engine at idle as well!
Luckily - I had banged my head against such systems long enough to quickly discover what was happening. It took me a couple of minutes to figure it out, and I am secretly very proud of it.
The hydraulic motor of the cooling fan (driven by a group 2 gear pump. ) leaked badly through the shaft seal, and since there was no motor of the same size around, the mechanic replaced it with one that was double its size. Later it turned out that the replacement motor was actually about the size of the original one, and the smaller motor had been installed to increase the speed of the fan when first overheating issues appeared. The bigger motor caused the pressure to drop, especially when the engine was idling - and this same pressure was used to pilot the electric proportional valves of the DCVs (which required 30 bar to work properly). Originally, there was a 10-bar inline check valve to induce additional pressure in the fan motor line so that the sum of both would always be higher than 30 bar, but this valve was removed when the smaller (and therefore more pressure-hungry) fan motor was installed. So, when the oil hit about 50 CÂș, the bigger motor was no longer inducing enough pressure at idle for the DCV piloting system to work properly - there simply wasn't enough pressure to shift the spools of the DCVs full-travel! Installing a sequence valve upstream the fan motor instantly solved the problem. What a relief! I confess my heart sank and my mind went through a series of "less than pleasant" mental images when I saw no pressure at idle.
The crawler has already been field tested - and it is working perfectly fine, with no overheating issues, even with the slower fan.
I guess the main takeaway from all this would be that gear flow dividers, especially the larger ones with iron bodies, require running in, just like regular repairable gear pumps do. Bear this in mind should you get such a divider from a supplier that may have it assembled for you from off-the-shelf parts.