Aqua Computer Highflow Sensor tested

Printable View

06-08-2009, 12:20 PM
HESmelaugh

Aqua Computer Highflow Sensor tested

A while back you may have seen the news on a new flowmeter by Aqua Computer. They kindly provided me with a sample of this flowmeter and I did a small flow-comparison to two other flowmeters.

First of all, let's take a closer look at the AC Highflow Sensor:

http://www.abload.de/thumb/01hyx2.jpg http://www.abload.de/thumb/03eyxq.jpg http://www.abload.de/thumb/02b9hq.jpg

The flowmeter is housed inside a POM-casing which has a nice, stainless steel top cover. A cable to connect the flowmeter to an aquero, t-balancer or any other device capable of interpreting this type of signal needs to be purchased separately. A great advantage of this flowmeter is that it has 1/4" threads, so you can use it for any size tubing and with practically any fittings you like. The same can't be said for the other two flowmeters I tested:

http://www.abload.de/thumb/07g9wj.jpg http://www.abload.de/thumb/0829k7.jpg

The one on the left is a Digmesa FHKUC 70. It is one of the most popular flowmeters used in liquid cooling setups, at least where I come from. It's fairly compact and it's said to be precise and not very restrictive. On the downside, you're stuck with two 3/8" barbs, so you can't use any tubing size you want.
The one on the right is a Durchflusssensor GMR ("Flowmeter GMR"), also quite popular among LC'ers over here. It's available in POM, clear acrylic and brass versions. It has an integrated, albeit very short cable. On the downside, it has 1/2" outer threads which arent compatible to anything regularly used in PC liquid cooling I can think of. So you have to get extra adapters and O-Rings and it can sometimes be tricky to get all of this sealed properly.

Now, let's take a quick look inside the AC Highflow Sensor:

http://www.abload.de/thumb/09xmdc.jpg http://www.abload.de/thumb/10xbmr.jpg

As you can see, the electronics of this flowmeter are by Digmesa. The sensor has been given a new encasing by Aqua Computer and the inlet diameter has changed compared to the FHKUC 70, which also means that the flowmeter is recalibrated for a different impulse-rate.

Test Setup

Lacking a Manometer for this kind of thing I went for something much cheaper and much more crude: Two buckets.

Here's the basic test-setup:

http://www.abload.de/img/skienx9uv.jpg

1) Bucket #1 is the reservoir.
2) A Laing DDC with an EK X-Top.
Next comes the flowmeter (FM) to be tested.
3) Then follows a quick disconnect which serves the dual purpouses of allowing me to shut off the flow and introducing considerable restriction to the loop. Since I'm doing measurements by hand this is a good thing.
4) Bucket #2 sits on a scale. 1 litre of water equals 1 Kg in weight (metric system FTW) so I can easily use the scale and a stopwatch to determine how long it takes for X litres of water to reach the bucket.

The flow I calculate using the stopwatch and the scale is what I describe as the actual flow or "GPM actual" in the diagrams. Of course, there's human error involved here. Some preliminary testing showed that the results are very repeatable, though.

So, let's get to the data.

Maximum Flow / Restriction

http://www.abload.de/img/01b9jm.jpg

This is what we get with the pump running at full power. On the one hand, looking at the actual GPM reached, this gives an idea of how restrictive each of the flowmeters is. We also see how far apart the actual flow is from the output of each of the sensors. Here, we see that the AC Highflow Sensor does very well.

Next, let's take a look at precision of the flowmeters over a small range of flowrates. Highflow fanatics can leave now and go admire their triple Iwaki setups. Due to the restriction introduced by the quick disconnect, the flowrates are all pretty low.
I measured precision at 60 l/h, 100 l/h and 150 l/h. In GPM, the numbers don't make as much apparent sense. The measurement was done by reducing adjusting a valve until the output of the flowmeter showed 60, 100 and 150 l/h respectively. With the flow set like this, the actual flow was measured using the stopwatch and the scale.

Here are the results:

http://www.abload.de/img/02mlpo.jpg

http://www.abload.de/img/03ymon.jpg

http://www.abload.de/img/04za1r.jpg

The red line shows what the output, measured by the flowmeters was, the bars show the actual flow.
As we can see, the AC Highflow Sensor doesn't do too well at low flowrates but gets increasingly more precise as flow increases. In fact, it goes from 33% to 9% to just 1% deviation in the range tested.
The Digmesa FHKUC 70 is also quite precise and the GMR is mostly a bit off target. Though in my opinion the meters are all precise enough for practical purpouses.

Orientations

I ran each of the flowmeters in a closed loop as well and placed the flowmeters, at different angles and orientations. This was to check whether they would provide consistant measurements, independantly of orientation. Here's what I found:

Both the FHKUC 70 and the AC Highflow (which use the same electronics) can be used in any angle and orientation. No matter how I twisted or turned these, the readings would remain the same.
The GMR doesn't work at all when it's upside down, i.e. it reads 0 GPM. In orientations where it's almost upside down, the readings go awry, but it's obvious that they are inaccurate. In any other orienatation, be it vertical or horizontal or anything in between, the readings are consistant.

Quick conclusion

Out of the three flowmeters tested, I prefer the AC Highflow Sensor simply for it's practical nature. It's the only one that I can just use my usual fittings with and thus let's me use any size tubing I want. It's also the only one of the three flowmeters that looks nice and doesn't necessarily have to be hidden away in your flashy rig.
Aqua Computer took a solid flowmeter by digmesa and did some fine-tuning on it. This results in low restriction and a good precision at higher flowrates.

This is a translation of my original article on DeXgo.

Thanks for reading.

Cheers,
Shane
06-08-2009, 12:25 PM
SNiiPE_DoGG

looks nice - you guys will be seeing this one in project flow :up:
06-08-2009, 12:50 PM
ecat

I've always been curious about flow meters, now I know.
Nice work :)
06-08-2009, 01:19 PM
Sadasius

I wonder if that wire assembly from the Koolance flow meter would work with this one and hook up to the mobo and will be able to calibrate flow accurately. May have to try that out sometime.
06-08-2009, 01:21 PM
Sparda

Thanks for the test. I'm just about to pick up Bach DFS because AC Highflow a bit pricey but doesnt need any adapter. Is there any noise in the AC Highflow with a high flow rate discharge ?
06-08-2009, 01:28 PM
Sadasius

Is it just me or does the AC unit internally look identical to the Digmesa FHKUC 70? Seriously it looks like nothing was done except make a new box for it.
06-08-2009, 08:13 PM
fox3

Thanx Shane for comparing those flowmeters. :up:

I was curious how the FHKUC 70 in my rig compared with the other two. Since I am using 3/8" tubing the barbs do not really become a problem but I do like how clean the new AC meter looks and the G1/4 thread is a win.
06-08-2009, 08:52 PM
BDW88

Thanks for the testing Hes outstanding review as usual. :up:
06-08-2009, 08:58 PM
Martinm210

Nice work!

FYI, In an open loop system like that, static pressure does have influence on the pump and flow rate.

For example, if the inlet reservoir has water about 1' higher in elevation than the pump and it's a free fall. The pump will produce a performance curve equal to it's normal pumping curve + 1' H2O. But as that 1' lowers, the performance curve will slowly lower to the normal curve.

Static pressure doesn't do anything in a closed loop because it cancels itself out coming back down and around the loop, but it does matter in open loops, so keep an eye on that.:up:

Anyhow this was an observation I made from some of my bucket pump testing I did a long time ago...bucket elevations do make a difference on flow rate for pumps in bucket tests.
06-08-2009, 09:23 PM
denice25

great info.... thanks...
06-08-2009, 11:11 PM
HESmelaugh

Quote:

Originally Posted by Sparda

Thanks for the test. I'm just about to pick up Bach DFS because AC Highflow a bit pricey but doesnt need any adapter. Is there any noise in the AC Highflow with a high flow rate discharge ?

I didn't hear any noise from the flowmeters at higher flowrates.
At very low flow, the Digmesa and the GMR produce a very slight clicking noise, but almost anything else in a system will typically be louder than that.

Quote:

Originally Posted by Martinm210

Nice work!

FYI, In an open loop system like that, static pressure does have influence on the pump and flow rate.

For example, if the inlet reservoir has water about 1' higher in elevation than the pump and it's a free fall. The pump will produce a performance curve equal to it's normal pumping curve + 1' H2O. But as that 1' lowers, the performance curve will slowly lower to the normal curve.

Static pressure doesn't do anything in a closed loop because it cancels itself out coming back down and around the loop, but it does matter in open loops, so keep an eye on that.:up:

Anyhow this was an observation I made from some of my bucket pump testing I did a long time ago...bucket elevations do make a difference on flow rate for pumps in bucket tests.

I hadn't considered that, but I guess it makes sense.
This test setup is one that I wouldn't be able to reproduce accurately for further tests.

EDIT: Updated the first post with a bit on the flowmeters orientations.
06-09-2009, 04:24 AM
Brodholm

Nice! I was wondering If I should buy the new AC Highflow. But I had not seen any tests on it. But it turns out its really good!. The only problem now is that the aquaero only can handle 1 flowmeter if you don't want to waste another fanheader.
12-17-2011, 12:40 PM
D749

Is there any issue with suspending this unit within a tubing run? In other words does it need to be affixed to a case or is it light enough filled with water that it isn't an issue weight wise.

Thanks.
12-17-2011, 01:22 PM
Martinm210

Suspending is fine, that's how I've got the one I have running. It's a little bit of weight with the acetal block, but not enough to be a problem I don't think. The only thing I noticed about the flow meter was initially the meter had a very light tick noise that was noticeable, but after running it a day...disappeared and fine now.

I also think any of these impeller type flow meters are going to require calibration depending on the fittings used. Not all fittings have the same ID and also an elbow before/after will probably tweak the calibration too.
12-17-2011, 01:32 PM
D749

Quote:

Originally Posted by Martinm210

Suspending is fine, that's how I've got the one I have running. It's a little bit of weight with the acetal block, but not enough to be a problem I don't think. The only thing I noticed about the flow meter was initially the meter had a very light tick noise that was noticeable, but after running it a day...disappeared and fine now.

I also think any of these impeller type flow meters are going to require calibration depending on the fittings used. Not all fittings have the same ID and also an elbow before/after will probably tweak the calibration too.

Thanks for the info. One run I'm planning is: device1 -> QDC (VL4N) fitting -> tubing -> 1/2" ID 3/4" OD compression fitting -> AC HF sensor -> 1/2" ID 3/4" OD compression fitting -> tubing -> QDC (VL4N) fitting -> device2. What would I need or how would I go about calibrating for this?
12-17-2011, 01:53 PM
Martinm210

If it's just for personal use and monitoring relatively differences, you can just pick one of the default calibration values. The AQ5 gives a "High flow", "5.6mm", and "3.3mm", but when I tried high flow with an BP 90 elbow at the outlet, I noticed it was reading about 20% low compared to my king flow meter. Since I already had the king meter, I simply went in and adjusted the pulses/liter until the two matched. If you didn't have another meter you trusted, you'd have to do some sort of flow test using a pair of 5 gallon buckets or something similar to that and measure how long it takes to transfer the water across and what the meter tells you.

But, that's only if you really wanted it accurate. I just know that barbs vary quite a bit on what their actual ID is and things like elbows would influence the result too. Not a big deal for general info and relative changes in flow rate, but I wouldn't take the number given as having absolute accuracy without calibrating it and checking.
12-17-2011, 01:58 PM
D749

Quote:

Originally Posted by Martinm210

If it's just for personal use and monitoring relatively differences, you can just pick one of the default calibration values. The AQ5 gives a "High flow", "5.6mm", and "3.3mm", but when I tried high flow with an BP 90 elbow at the outlet, I noticed it was reading about 20% low compared to my king flow meter. Since I already had the king meter, I simply went in and adjusted the pulses/liter until the two matched. If you didn't have another meter you trusted, you'd have to do some sort of flow test using a pair of 5 gallon buckets or something similar to that and measure how long it takes to transfer the water across and what the meter tells you.

But, that's only if you really wanted it accurate. I just know that barbs vary quite a bit on what their actual ID is and things like elbows would influence the result too. Not a big deal for general info, but I wouldn't take the number given as accurate without calibrating it and checking.

Any adjustments/calibration would be made in the AQ5 correct? My originally plan was to use flow meters merely to verify that there was flow. But now that I think about it since both my pumps have an RPM senor it would be redundant to use these flow meters unless I actually wanted to know more the respective flows. Thanks.
12-17-2011, 02:15 PM
Martinm210

Yes, the AQ5 has all that calibration stuff in there. Flow meters are more of a luxury item than being necessary, but I do like having some sort of relative reference when not regularly servicing a loop. I noticed in my two year setup, my flow meters very slowly went down from about 1.15GPM to 1.0 GPM or so, so it was an indicator that a little bit of crud was building in the CPU block. That's really the only benefit I see beside using one for testing purposes.