Calling All Pool Owners -Tell Me About Yours

[DaveNJ]

mas985,
Was wondering how your coming along this. Also my PSI doesn't increase over time, 2-3 weeks. Is the filter border line for my configuration.

 

[mas985]

Dave,

Sorry I missed your previous post about the vacuum gauge. You can find them at automotive stores or Sears for about $20. You may need to get an additional fitting to match the suction side drain plug of the pump. I think it is a 1/4" fitting but I am not absolutely sure. A suction measurement would definitely help the process.

Unfortunately, I have not had much time to work on this but I have found that there seems to be a large diversity in the way pools are plumbed. So to be modeled properly, more information may be necessary.

As for your pool, I did some preliminary work and found that your PSI is a bit higher than what I would have predicted. In order to get 24 PSI, I had to add 160 feet of 2.5" pipe. This could be due to several things. Pressure gauge accuracy, suction side plumbing assumptions (vacuum measurement would help here) or even something in the return side plumbing that I did not account for.

Anyway, here is what I have so far for your plumbing:

Suction side head loss: 2.4 ft
Return side head loss: 55.1 ft
Total head loss: 57.6 ft
Flow Rate: 130.6 GPM

Again, a suction side measurement would remove any error to these numbers.

 

[chem geek]

Mark,

I just took a full set of measurements on my Pentair IntelliFlo with my system in various configurations and list this below. Unfortunately, I had done a detailed list of pipe lengths, number of elbows, etc. but that all got stolen with my laptop a while ago so I'll have to reproduce all of that at some time in the future. Nevertheless, I've given some rough estimates to this below.

SOLAR ON. This has a very long pipe configuration as it must go a hundred or so feet from the pump house to the house, then down the full length of the house through 12 parallel solar panels and back again. The pipes are all 2" for this part of the system. Then, going to the pool starts out as 2" then splits into three 1.5" pipes -- one for each return. The suction side has two 1.5" pipes, one for the one skimmer and the other for two floor drains. The PSI below are pressure-side only, from my filter gauge.

[EDIT] I'm adding the implied overall pump efficiency based on GPM * PSI * 0.4347, but actual efficiency is higher since I'm ignoring suction head. [END-EDIT]

30 GPM, 2300 RPM, 715 W, 14 PSI, 25.5%
33 GPM, 2370 RPM, 785 W, 15 PSI, 27.4%
36 GPM, 2490 RPM, 895 W, 16 PSI, 28.0%
39 GPM, 2630 RPM, 1045 W, 18 PSI, 29.2%
42 GPM, 2755 RPM, 1185 W, 20 PSI, 30.8%
45 GPM, 2900 RPM, 1380 W, 22 PSI, 31.2%
48 GPM, 3000 RPM, 1530 W, 24 PSI, 32.7% <--- this is what I normally run at when the solar is on (this is 4 GPM per panel)
51 GPM, 3200 RPM, 1835 W, 27 PSI, 32.6%
54 GPM, 3325 RPM, 2075 W, 29.5 PSI, 33.3%
57 GPM, 3425 RPM, 2295 W, 31 PSI, 33.5% <--- this is close to the 3450 RPM maximum of the pump

With my SOLAR OFF, I just have the runs to/from the pool that I described above and I get the following.

15 GPM, 1185 RPM, 155 W, 3 PSI, 12.6%
18 GPM, 1325 RPM, 195 W, 4 PSI, 16.1%
24 GPM, 1460 RPM, 255 W, 4 PSI, 16.4% <--- I run close to this at 26 GPM when the solar is off
30 GPM, 1805 RPM, 420 W, 6.5 PSI, 20.2%
36 GPM, 1975 RPM, 535 W, 8 PSI, 23.4%
42 GPM, 2160 RPM, 685 W, 9.5 PSI, 25.3%
48 GPM, 2310 RPM, 835 W, 10.5 PSI, 26.2%
54 GPM, 2490 RPM, 1030 W, 12 PSI, 27.3%
60 GPM, 2710 RPM, 1305 W, 14.5 PSI, 29.0%
66 GPM, 2900 RPM, 1590 W, 17 PSI, 30.7%
72 GPM, 3080 RPM, 1895 W, 19 PSI, 31.4%
78 GPM, 3360 RPM, 2495 W, 22.5 PSI, 30.6%

I have ThePoolCleaner on a dedicated 1.5" line that I switch to for cleaning off-hours. When I disconnect the cleaner, then I get the following which shows the flow resistance solely from the 1.5" line and an outlet that has an adapter from 1.5" to 2" (since ThePoolCleaner is 2").

15 GPM, 1370 RPM, 200 W, 4 PSI, 13.0%
18 GPM, 1475 RPM, 245 W, 4.5 PSI, 14.4%
21 GPM, 1610 RPM, 300 W, 6.5 PSI, 19.8%
24 GPM, 1850 RPM, 420 W, 8 PSI, 19.9%

When I connect ThePoolCleaner it adds resistance and I get the following.

15 GPM, 1955 RPM, 425 W, 10.5 PSI, 16.1%

This may not have been the final stable cleaner number since it seems to be at 2180 RPM, 540 W at other times I've measured it (so probably a higher pressure as well). At any rate, it shows that ThePoolCleaner adds about 7 PSI to pressure resistance at the 15 GPM flow rate. This flow rate, which is the lowest I can set with the Pentair Intelliflow, produces a little over a 14 wheel RPM on ThePoolCleaner which is at the high-end of its recommended 11-14 wheel RPM range, but I don't find the pressure relief valve getting triggered so it's probably still OK.

The above all implies certain equivalent pipe lengths, but I'll have to see how that matches with actual lengths later when I get a chance.

Richard

 

[mas985]

Very interesting set of data. Thanks for taking the time to do this.

A couple of things that I noticed after looking at the data. The GPM settings vs RPM do not seem to follow the pump affinity laws while the PSI values do. So if I use the top speed as a reference, then GPM should be proportional to RPM but is not. However, PSI and watts do seem to follow the affinity laws with RPM.

For No Solar:

PSI ~ 22.5 * (RPM / 3360) ^ 2
Watts ~ 2495 * (RPM / 3360) ^ 3 - Except for very low RPM.

matches pretty well but

GPM ~ 78 * (RPM / 3360)

has a lot of error for low GPM values.

I am assuming that you are setting the GPM values and the Intelliflo picks the appropriate RPM value?

If so, do you know how the pump decides what RPM value to choose given the GPM input?

I would expect that the RPM readout to be fairly accurate since it is directly related to the VFD. However, if they measure the GPM in the pump and use a control loop to set the RPM value, then they are dependent on the accuracy of the flow rate measurement. If they are using some other method to infer flow rate, the error could be much larger.

So given that the PSI measurements seem to reflect the correct RPM measurements, it looks like the actual GPM does not match the GPM setting very well. Is there any other way you can verify flow rates?

 

[chem geek]

I am assuming that you are setting the GPM values and the Intelliflo picks the appropriate RPM value?

If so, do you know how the pump decides what RPM value to choose given the GPM input?

So given that the PSI measurements seem to reflect the correct RPM measurements, it looks like the actual GPM does not match the GPM setting very well. Is there any other way you can verify flow rates?

First of all, don't forget that my PSI numbers are pressure-side only so will be lower than the actual head from both pressure and suction sides. I'll talk about that more below with some examples.

For the tests, I was in Service Mode so that I could be out at the pump during the experiments and try different settings. I entered a Set Flow and then waited for the RPM to stabilize. It is quite possible for the actual flow to be off if the RPM was still stabilizing and I was impatient, but I usually waited not only for the number to not change, but if it changed, to go between two numbers (i.e. not keep moving up or down, depending on which way I was changing the flow). So there could be error in my method by simply not waiting long enough for the pump to stabilize. Remember that the following two sets of numbers should have been representing the same steady state:

15 GPM, 1955 RPM, 425 W, 10.5 PSI
15 GPM, 2180 RPM, 540 W

where the first is what I measured when doing the experiment while the second is what I've seen at the control panel in my house after the cleaner has been on for quite a while. So that might be the sort of error I introduced by not waiting long enough, or it could just be the specifics of the state of the cleaner at the time of each snapshot. The solar on numbers matched what I see at the indoor control panel, so perhaps it is true that the higher flow rates stabilized faster so are more accurate while the lower ones have the sort of error shown above. If the error you see is more than that (i.e. more than about 10% error in RPM), then I'm not sure what the cause is.

The Pentair graphs of the IntelliFlow seemed to imply a pump equation of the following:

Head (in feet) = (RPM/350)^2 - (GPM^2)/470

If I use the above equation for the solar case at the extremes I get:

(3425/350)^2 - (57^2)/470 = 88.8 feet = 38.6 PSI whereas I measured 31 PSI
(2300/350)^2 - (30^2)/470 = 41.3 feet = 17.9 PSI whereas I measured 14 PSI

For the non-solar case at the extremes I get:

(3360/350)^2 - (78^2)/470 = 79.2 feet = 34.4 PSI whereas I measured 22.5 PSI
(1185/350)^2 - (15^2)/470 = 11.0 feet = 4.8 PSI whereas I measured 3 PSI

From the above, it seems that the suction side PSI can be estimated from a system curve that has a power of around 1.8 (which is predicted from typical pipe losses).

Suction Side PSI = ( GPM^1.8 ) / 200

which is an equivalent 1.5" pipe length of around 103 feet (actually 360 feet for each pipe since there are TWO 1.5" pipes so the flow rate is roughly half for each pipe) which sounds about right (for the length of each of two 1.5" pipes) including suction resistance from the floor drains and skimmer. There is still error at the low GPM for this, but I'll bet that my pressure side gauge is not very accurate at low PSI. It doesn't have any tick marks below 4 PSI, for example. I also could only estimate the PSI to within 0.5 PSI at best -- the tick marks were every 2 PSI. Still, you may be right that there is something else going on here.

On the other hand, if I try predicting pressure side PSI within a similar grouping of solar on or solar off, I get the following formulas that are decent at the high pressures:

Pressure Side PSI (Solar On) = ( GPM^1.8 ) / 45
Pressure Side PSI (Solar Off) = ( GPM^1.8 ) / 110

This implies an equivalent pipe length of 188 feet if all 1.5" pipe or 630 feet if all 2" pipe and in reality it's a bit of a mix so this sounds about right including pressure from the filter and fittings. The solar is mostly 2" pipe and implies an equivalent pipe length of 1540 feet which seems high though there is a long round-trip to the house and down the entire length of the house with a bunch of turns and zig-zags for roof hips.

So the error might be from the pump flow rates when they are low or they could be from the pressure gauge at low pressures.

Richard

 

[mas985]

First of all, don't forget that my PSI numbers are pressure-side only so will be lower than the actual head from both pressure and suction sides. I'll talk about that more below with some examples.

True but both the return side and suction side should follow the affinity equations independently so taking the return side in isolation is still a valid analysis. The head loss in plumbing, for a fixed diameter and length, is only dependent on flow rates so the PSI should move at approximately the square of the RPM which it does.

So the error might be from the pump flow rates when they are low or they could be from the pressure gauge at low pressures.

Had the PSI drop not followed the square of the RPM so closely, I would have concluded that the PSI may be in error. Remember too that the affinitiy equations state that GPM is proportional to RPM but the data you collected indicates that they are not (half the RPM should be half the GPM). So either the RPM is incorrect or the GPM. However, since the PSI tracks RPM^2 to within a few percent and that RPM is probably pretty accurate reading in the pump, the only conclusion that I can come to is that the PSI values are correct and the GPM values are not quite right. Otherwise both the PSI and RPM readings would have to be wrong by same amount for the GPM to be correct which is very unlikely.

Since the pump control system tries to match the GPM setting by varying the RPM, it must accurately measure GPM in order to set the RPM appropriately. This is where I think the Intelliflo may not be setting the correct RPM for the given GPM.

Also, if you assume that the PSI values are correct, then the equivalent length of pipe would change from high speed to low which is not correct either since it should be a constant over all speeds. So all the evidence seems to be pointing to an error in the GPM.

 

[mas985]

Read this in the manual:

To set the water temperature: Use the Up and Down arrows to select “Water
Temperature”. To set the water temperature, see step 7.
Enter the current water temperature from 68° F - 104° F (Default 75° F.) The “Water
Temp” is only for the accuracy of the flow sensor. Temperature accuracy is not
critical, just enter an approximate temperature. When the IntelliFlo is connected to an
IntelliTouch system, water and air temperature information is provided by the system
sensors. The flow reading on the IntelliFlo is –0 / +2 GPM. The closer the
temperature to the actual temperature, the more accurate the flow reading on the
IntelliFlo control panel LCD will be while it’s running.

This would indicate a fairly high accuracy of the flow measurement and that they may be using a calorimetric flowmeter. So if the temperature reading is not too far off, it should be fairly accurate.

Also, in the manual, they indicate that it takes sometime for the RPM to stabalize for the correct GPM since they use small steps in RPM.

 

[chem geek]

Yes, I see what you mean that the "power" on the affinity coefficients should have still tracked correctly even though the "factor" would be different looking only at a part of the system (pressure side only).

I'm using an IntelliTouch system and didn't notice a temperature entry on the pump itself so apparently it's getting the temperature from the sensors that are part of the IntelliTouch as described in the manual (BUT probably not when I was in Service Mode -- I'll take another look at that). So I would assume the GPM is reasonably accurate, yet the affinity equations aren't being followed for GPM. I think that I'll try a different approach of making changes at the indoor control panel instead, changing the flow rate for an override "feature" since the indoor control panel tells me the actual flow rate. The pump only lets me set the flow rate, not see the actual rate (for whatever stupid reason they have). So by setting the rate at the indoor control panel, I can verify that it has actually achieved that rate, or not. Even so, I doubt this will clear up this mystery. I'll let you know what happens. It's more of a pain since I have to go back and forth from the house to the pump shed (to get the PSI reading) so I'll just do that for a smaller sample of low, mid and high flows (with solar off).

 

[chem geek]

Mark,

Take a look at this thread on another forum and note the posts a little more than halfway down by tmoss and others that seem to imply that the IntelliFlo calculates the flow rate and doesn't actually measure it. If that's true, then that might explain the anomalous GPM results I was seeing that didn't seem to fit the affinity equations very well.

Richard

 

[mas985]

Thanks for the link. It has been a while since I visted the Garden forum.

What they are indicating makes sense. Given that there is a direct relationship between power draw, RPM and GPM, theoretically, they could just use a simple control loop to which adjusts the RPM based upon power draw and GPM setting. The accuracy of this would be fairly good if they did a multipoint calibration. However, my guess is that they probably do a linear fit to the power curve shown in their documentation.