Insane Hydraulics
Bold and audacious blog about fluid power
Today I want to talk about the viscosity of mineral hydraulic oil and the importance of understanding how much it changes with temperature and how this affects your turbine flow meter.
Regular readers of this blog already know that for the last year I have been obsessed with designing and building my very own wireless monitoring system, which started with pressure sensors and now is transitioning into the "flow meter stage".
Since I enjoy the "amateur-scientific approach", I have been testing all the flow sensors that I managed to "put my hands on", within the limits of available testing gear, of course, and I want to share the result of one particular test that I ran on every flow-meter and consistently got the same result.
I always suspected that a turbine flow meter should show a lower reading with cold oil because a more viscous fluid causes more drag on the turbine, but I wasn't sure by how much. So, I devised an ingenious test, which consisted of connecting a turbine flow meter in series with an oval-gear volumetric flow meter/counter (which should be affected by the oil viscosity to a smaller extent), and compare the readings for different temperatures.
A small aside now - these meters, that are normally used for lubricants or diesel, are great! I use them a lot. They boast 0.5% accuracy, they are volumetric, which means that they can work with incredibly low flows and, as I said before, are less affected by the fluid viscosity. This particular model can go up to 100 l/min. And they show you both the flow rate and the amount of fluid that passed through. The flow rate only gives a usable reading when the flow is above 10-15 liters per minute, but the volumetric count is invaluable because you can do the good old bucket/stop-watch test without actually needing the bucket and open lines, which means you can very accurately measure flows even below one liter per minute.
I once put this flow meter vertically on a bench, took a carefully measured liter of hydraulic oil, and poured it slowly through the meter. When the oil ran out - it read 999 CC. I think it's precise enough for me...
The only disadvantage of such meters is their inability to reliably measure hot liquids. The pick-up is most likely magnetic (I haven't opened one yet, but I suspect a ferrite magnet in one of the gears that trips a reed-switch) and as magnets loose performance with temperature, at about 80-ish C these meters stop counting, which is very unfortunate because such a meter would be an ideal precise instrument to evaluate a case drain of a hydraulic motor in the field, however more often than not, a drain line is also the hottest line in a hydraulic system, and I have seen situation were these meters fail to operate. It's a pity, really... You can get them quite cheap. One day I will gut one of these to see if I can replace the magnet with one with a higher temp. rating.
And, of course, such meters can't take high pressures. But, I guess, this is apparent.
Anyhow - that's not important. What is important is the result, which you can see in the table below. The oil is VG46, and the table contains readings from the oval gear meter when the turbine one (Parker SCFT-600) read exactly 25, 50 and 75 l/min at different temperatures.
| Oil Temperature, C | Oil Viscosity, cSt | Oval Gear Reading (Turbine - 25 l/min) | Oval Gear Reading (Turbine - 50 l/min) | Oval Gear Reading (Turbine - 75 l/min) |
| 16-19 | 100 | 31.4 | 58.3 | 84.0 |
| 42-46 | 40 | 26.4 | 52.8 | 79.0 |
| 58-60 | 20 | 25.8 | 51.5 | 78.8 |
As you can see, at room temperature, the turbine flow meter falls way behind the oval gear one.
Just a few days ago I had a Hydac flow meter in my shop. I didn't have much time with it - but I did connect it in series with the Parker and the oval gear flow-meters, and this one showed an even worse discrepancy. With oil temperature of about 25C, it was falling 5-8 l/min behind the Parker SCFT -600 (which, in its turn, was already "behind"). As the temperature rose, however, the Hydac's reading gradually started to "catch up", and if I recall correctly at about 50C was almost equal to Parker.
Unfortunately, I was short on time that day, so I didn't manage to conduct the test to the end and register it properly - but still - the tendency is clear (even if absolute flow readings may be off because none of the meters that I used were certified or calibrated,) - a turbine flow meter can show 10-15% less flow with a more viscous fluid, which is a lot.
Now, it's important to realize that it doesn't take much for a common hydraulic oil to become "more viscous". Here's a chart that shows a typical relationship of viscosity to temperature for common mineral hydraulic oil grades:
| Temp, C | ISO32 | ISO46 | ISO68 |
| 0 | 309 | 527 | 894 |
| 10 | 150 | 244 | 393 |
| 20 | 82 | 127 | 196 |
| 30 | 49 | 73 | 107 |
| 40 | 32 | 46 | 68 |
| 50 | 21 | 30 | 41 |
| 60 | 15 | 20 | 27 |
| 70 | 11 | 15 | 19 |
| 80 | 8 | 11 | 14 |
| 90 | 6 | 9 | 11 |
| 100 | 5 | 7 | 9 |
Below 20 C the viscosity increases very fast with each degree. Much faster than in the ranges above 50 (which is the point when pipes become too hot to touch).
This is just another (very good) reason to conduct all flow tests (and adjustments) only when a hydraulic system has reached its operating temperature!