I read the Igor's lab comparison of the High Flow Next to a Thermaltake TF-1 and a Barrowch FBFT03 flow sensor. LINK. They found that the HFN was very accurate compared to a Keyence FD-Q10C precision flow sensor, while the TT and Barrowch were wildly inaccurate. I have a test loop set up which includes 2 High Flow Next flow sensors.

D5 Next Out → alphacool rad (NexXxos ST30) → HFN #1 → - alphacool spinner flow indicator → HFN #2 → - Ball Valve → D5 Next In

This test loop is quite simple. There is no CPU or GPU block in the loop so flow restriction is very low. The 2 HFN are separated only by an alphacool spinner flow indicator. The tubing is soft PVC, 3/8"ID x 5/8"OD (~10mm/16mm) with all straight compression fittings. There are no 90° fittings. The tubing on each side of the HFN flow sensors is almost straight (there is a slight curvature). Coolant is distilled water treated with Primochill Liquid Utopia.

With the pump at 50%, HFN #1 reports the flow rate is 219.9 l/hr. HFN #2 reports the flow rate is 229.7 l/hr. This is a difference of ~10 l.hr, which is ~4.4% relative to the lower reading. Is this flow rate difference within tolerance for the HFN flow sensor? I don't expect them to report exactly the same, but I was a bit surprised that their flow rate values differed by ~10 l/Hr. Conductivity measurements differed by 1.9uS/cm. That is not very much, but it is a 6.2% variation relative to the lower value. Is this difference within tolerance of the HFN conductivity sensor?

Here in an Overview Page I made that compares the 2 HFN.

2023-0302 HFN OVP.jpg

Thank you for the suggestion. I don't think turbulence caused by the spinner is the reason because there is ~ 2-ft of tubing between HFN-#1 and the spinner and ~ 1-ft of tubing between the spinner and HFN-#2. I may be wrong so I will remove the spinner and see what happens.

The HFN is not a "lab grade" flow sensor. It may be that the difference in flow rate I am seeing is simply due to accuracy limitations of the sensor. I found the HFN spec which says it can measure flow rate from 0.7 l/min to 17 l/min +/- 2.5%. The measurements I made have a difference of 4.4% which is not within specifications. I will remove the spinner and separate the 2 HFN by 2-ft of straight tubing and see if this reduces the difference in reported flow rates.

Conductivity may vary with temperature but I doubt there is much, if any difference in coolant temp in this test loop. The only source of heat in the loop is the pump, and because there are no restrictive components in the loop, the flow rate with the pump at 50% is quite high. The HFN spec sheet says the conductivity sensor can measure from 2 uS/cm to 200 uS/cm but no accuracy range is provided. There is no way to know if the difference in reported conductivity is within spec or not.

HFN Specs.jpg

Zitat von »InfoSeeker«

One wonders if these HFNs are actually calibrated to a standard prior to release. Probably not, but then their purpose is more to show difference over time than absolute accuracy out of the box.

Agreed. I suspect they are not calibrated at the factory and perhaps that is reasonable for a flow sensor that is intended to install in a computer cooling loop and costs under $100. Perhaps I will vary the flow rate and see if they both report the same change in flow rate. Again, I am not expecting lab grade accuracy from these flow meters, but I do except them to operate within Aquacomputer's specifications, which state +/-2.5%.

I think it is rare for someone to have 2 HFN in the same loop. It took several months for me to find one in stock in the US. Sometime later, I ran into a guy on Reddit who had one on backorder, then found one somewhere else. He ended up with two and was willing to sell the extra one, brand new in box, for a very reasonable price. I bought it because they are very hard to find and I felt it may be prudent to have a spare just in case. When the second one arrived, I wanted to make sure it functioned properly so I added it to my test loop. Once it was installed, I naturally wanted to see how close they are to each other. I think 4.4% is quite reasonable in this application, but it's not within spec which is concerning. I am hoping someone from Aquacomputer will provide some insights but so far, no luck with that. Thank you again for yours.

I just updated to X.67 and .NET v7.03. After doing that (without any trouble) I checked the Flow Rate sensors on both the HFN in my test loop since the X.67 update included adjustments to the calibration of flow sensors. The values reported were about the same, and still about 10 l/hr apart. I then noticed that HFN #2 was set to Distilled but HFN #1 was set to DP Ultra. I am running distilled with Primochill Utopia and I am sure I set them both to Distilled. So either this setting somehow got changed on HFN #1, or I set it incorrectly. I'll take the blame on this one and say I must have set it incorrectly. Changing this setting from Distilled to DP Ultra lowers the reported flow rate by ~8 l/hr. Now with them both set to Distilled, they are about 1.1% apart. I am thrilled that they are so close. I attached an updated screenshot of the HFN comparison overview page. It would be nice if the state of this setting (Distilled or DP Ultra) was available in the HFN data. If it was, I could add it to the overview page.

Edit to add - I found HFN #1 set to DP Ultra again. The setting changed based on which Global Preset I selected. I tested whether the Coolant Type changes with device Profile changes and found that the behavior seems to be erratic. If I set Coolant Type to Distilled for HFN Profiles 1 and 3, and DP Ultra for Profiles 2 and 4, sometimes the setting changes with a Profile change and sometimes it doesn't. IMO, Coolant Type and Connector Type should be considered global, and not change with a Profile change. I may make a separate post about this since this thread is kind of buried, but I think I have stumbled onto a bug in the HFN firmware.

2023-0309 HFN OVP.jpg