Funky Swissflow readings?

[Mathelo]

Okay, I've finally got my watercooling setup. I had to set it aside for the last 3 months because of work but I finally got it together.

I'm runnning an Opti 175 on an ASUS-SLI Deluxe and CPU / water temps are in the low 30s. No overclocking yet. Antec 180 case. The loop is a Laing D5 pump > Storm G4 > Reservoir > Swissflow > Thermochill PA160 Rad running through 7/16" hose. I'm using an mCube T-Bal to monitor temps / flows and manage the fans. In general, it looks good.

However ...

Per the T-Bal readings the Swissflow is showing less than 2,000 impulses per minute. If I understand the way the Swissflow works, that would equate to about .33 liters per minute. The Swissflow is rated at 6,100 pulses per liter per the manufacturer at
http://www.swissflow.com/en/SF800/Fl...Specifications.

With this loop I was expecting to get about 4 liters / minute, which equates to 24,000 impulses per minute.

So either:

1) I'm missing something here in the math

2) T-Bal is doing something strange with the readings

3) I have a blockage, or

4) The Swissflow is defective

I hope you guys can help me and thanks in advance.

Louis

[jaydee]

Did you set the T-Bal correctly? The divider is set to a fan's pulse rate output not the swissflows. Not sure how the T-Bal is setup but for the CF633 you just change the PPR setting to 1 instead of the standard 2 for a fan. Then whatever reading it puts out divide by 6100. I checked mine with a separate flow meter to verify.

Still though setting the divder to 1 would only give you 4000... Maybe the T-Bal is setup a different way.

Also did you put a filter in front of that Swissflow? Any little particle will ruin it. Swissflow recommends a 20micron filter.

[Mathelo]

Jaydee, thanks for your reply.

I think I did it right. The T-Bal has an input for inpulses per liter where I typed in 6,100. It also provides what would appear to be a true pulse reading without any multiplier / division effect.

Might be my problem with the lack of a filter. I dismissed it as not really needed. Where do I find one of these filters?

Louis

Originally Posted by jaydee

Did you set the T-Bal correctly? The divider is set to a fan's pulse rate output not the swissflows. Not sure how the T-Bal is setup but for the CF633 you just change the PPR setting to 1 instead of the standard 2 for a fan. Then whatever reading it puts out divide by 6100. I checked mine with a separate flow meter to verify.

Still though setting the divder to 1 would only give you 4000... Maybe the T-Bal is setup a different way.

Also did you put a filter in front of that Swissflow? Any little particle will ruin it. Swissflow recommends a 20micron filter.

[Mathelo]

If this setup was working correctly, shouldn't my temps be in the low 20s? My ambient temps are in the range of 20 to 21 C.

L

[Mathelo]

Strangly the flow has improved a bit. Now up to just under 2,600 impulses per minute from under 2,000 yesterday. Still low of course.

Is it possible that a blockage is working its way free? Maybe in the block? Maybe an air pocketin the rad? I don't think there is much air in the system but it does take a while for the radiator to give up all its air. Could that make that much difference?

I really don't know where a blockage would have come from. This is a new, clean setup ... except that I left it sit for the last 90 days after my failed first attempt. I drained it but there of course was some residual moisture left behind. I can't imagine that would have much impact.

L

[jaydee]

Originally Posted by Mathelo

Jaydee, thanks for your reply.

I think I did it right. The T-Bal has an input for inpulses per liter where I typed in 6,100. It also provides what would appear to be a true pulse reading without any multiplier / division effect.

Might be my problem with the lack of a filter. I dismissed it as not really needed. Where do I find one of these filters?

Louis

Sounds setup right. Something is a miss though. If the meter was damaged to late for a filter now though. Your temps should be around 10-15C higher than ambient under load and somwhere around 5C idle.

Also the meter itself is a huge restriction. You might want to take the meter out of the loop and hook it up with just a pump and as short amount of tubeing as possible and see what it says then. That would rule out the rad and the water block if it still reads low.

[Mathelo]

hmmm, what consitutes a load? When I wake up the computer from standby, temps are around 30C for the CPU per the ASUS PC Probe and T-Bal reports 34C. I have the analogue sensor mounted on the edge of the IHS of the CPU.

After running CPUBurn for 10 minutes both the ASUS PC Probe and T-Bal report 37C. The water temp is always within a half degree cooler. Both CPUs appear to be running at about 50% capacity per WTM.

Strangely, the readings from the Swissflow have increased to around 3,000 impulses per minute; 50% improvement. Still unexceptional but odd just the same.

Originally Posted by jaydee

Sounds setup right. Something is a miss though. If the meter was damaged to late for a filter now though. Your temps should be around 10-15C higher than ambient under load and somwhere around 5C idle.

Also the meter itself is a huge restriction. You might want to take the meter out of the loop and hook it up with just a pump and as short amount of tubeing as possible and see what it says then. That would rule out the rad and the water block if it still reads low.

[jaydee]

CPU burn is plenty good for full load test. If you are getting those temps then it is unlikely flow is a problem.

[Mathelo]

Originally Posted by jaydee

CPU burn is plenty good for full load test. If you are getting those temps then it is unlikely flow is a problem.

I guess we'll see what happens when I OC. The main reason for the Swissflow is to know when I've lost the loop and automatically shut down the computer. So the actual flow rate is less important. But it sure would be nice to get the reading right.

L

[jaydee]

Originally Posted by Mathelo

I guess we'll see what happens when I OC. The main reason for the Swissflow is to know when I've lost the loop and automatically shut down the computer. So the actual flow rate is less important. But it sure would be nice to get the reading right.

L

Yeah the right flow rate is good. Especially with a 12V adjustable pump. You can turn the pump down to lower noise and keep an eye on flow and temp to find an optimal spot.

Other than taking the meter and pump out and testing seperatly I have no others suggestions.

[starbuck3733t]

I had the same problem - I thought it was reading strangely in SpeedFan, but in fact my Storm had clogged up and I really as running that low of a flow. I cleaned out the storm and am back at about 2 liters per minute.

Edit: blow into the swissflow gently and make sure the rotor spins. Check your electrical connections too, though that's less likely.

[Mathelo]

Originally Posted by starbuck3733t

I had the same problem - I thought it was reading strangely in SpeedFan, but in fact my Storm had clogged up and I really as running that low of a flow. I cleaned out the storm and am back at about 2 liters per minute.

Edit: blow into the swissflow gently and make sure the rotor spins. Check your electrical connections too, though that's less likely.

Hmmm, I'm going to break down the block this weekend and see if there is a problem there. I'll do that before pulling out the Swissflow.

L

[starbuck3733t]

Good luck. mine just saved me a lot of headache.

[Mathelo]

I disassembled the Storm this weekend and sure enough - it was plugged pretty bad. After cleaning it up and reinstalling, the flow improved considerably to 10,000 impulses per minutes (from 3,500). However, this is still low. I'm expecting 24,000+ pulses.

Temps after cleaning running CPUBurn for an hour, no oc:

Ambient: 22C
CPU: 40C
Water: 39C

This is a Opti 175
Loop = Laing D5>Storm>reservoir>Swissflow>PA160

Thoughts?

[BillA]

expectations, lol
what makes you think that is not (approximatly) your flow rate ?
you have a flow restrictor in the ststem eh ?

[jaydee]

If there was large enough particles to clog the storm then the meter may have been damaged to. I would definatly get a filter. A couple options: http://forums.procooling.com/vbb/showthread.php?t=12311

Water temp looks way to high.

[bigben2k]

Jaydee's right; that water temp (39C) is too high. Do you have any fans running?

Otherwise, I can tell you that you have a very restricted loop, and getting to 1 gpm (4 lpm) might be very difficult. Your main restrictions are the water block and the flow meter.

You need a filter, but it's going to add another restriction.

[Mathelo]

Originally Posted by BillA

expectations, lol
what makes you think that is not (approximatly) your flow rate ?
you have a flow restrictor in the ststem eh ?

Not saying it isn't but my non-scientific guess based on a loop with a Laing D5, Storm, and PA160 should get me about 4 liters/minute. The Swissflow is sold as "low restriction" but that doesn't apper to be the case. I'm also wondering if there are any limitations in the T-Balance software that might be causing a misreading. I know this may sound off but point in fact, the software does have threshold limits for setting the alarm off. Can't set it higher than about 250 impulses per second. If I was getting the readings I expected, I'd be double this.

So without more sophisticated equipment I'm left with process of elimination and guessing.

[Mathelo]

Originally Posted by jaydee

If there was large enough particles to clog the storm then the meter may have been damaged to. I would definatly get a filter. A couple options: http://forums.procooling.com/vbb/showthread.php?t=12311

Water temp looks way to high.

I've not checked the Swissflow so you may be right but the deposits in the Storm were not particles so much as gunk (if that makes any real difference). I had this loop setup over 90 days ago but it developed a leak in a poorly designed reservoir (my design ) so I set it aside. I did not completely drain it. I'm "guessing" the deposits were from sitting idle and having a small amount of water in it - not 100% drained.

In any case, adding a filter to the system is just going to add another restriction. If this low flow is a result of restriction at the flowmeter, I'm tempted to remove it and find another means to monitor flow. My main reasons for adding it in priority order were:

1) Safety - let me know when flow had dropped too low
2) Report flow for performance reasons

The flowmeter has already served both purposes but may also be part of the problem. Further testing (and disassembly) will be needed.

I believe the the temps are high because I've got this system setup for "quiet," at least at the moment. Everything is installed in a P180, all fans are controlled with the T-Bal and its closed up pretty tight. It is very quiet. I'll run a test with fans at max and improved ventilation and report the numbers back.

Louis

[Mathelo]

I've spent a considerable amount of time with the beast today. First things first. These foil temp sensor needed to be calibrated. They were way off. I don't have the greatest equipment for calibration but I do have a lab grade mercury thermometer and a digital thermometer we use for cooking. The two were within 1 degree of each other and I was able to use the two to get most of my temps set. Wish I could say that for the CPU temp. Just a tough place to get.

In any case, this is where I ended up. Temps are “better” than I had previously reported but strange. Remember, this is still without any oc but case is opened up for better airflow:

At Idle:
Mobo CPU reading = 32c
IHS (analogue sensor)= 32c
Water = 31c
Ambient = 23c

After running CPUBurn = K7 high for 60 minutes:
Mobo CPU reading = 39c
IHS = 34.5c
Water = 32.5c
Ambient = 22.5c

Here’s what I previously reported:
Mobo CPU/IHS: 40C
Water: 39C
Ambient: 22C

So the water temp has improved considerably and I have a high degree of confidence in their accuracy. I can’t say the same about the CPU or IHS temps. I calibrated the IHS temp to match the Mobo CPU temps at idle but if that spread as the CPU heats up is accurate, it strikes me that the IHS is acting more as an insulator than a conductor. I used Artic Silver 5 for mounting the Storm. I understand this paste need some time to develop maximum efficiency.

Curious what you guys think of these numbers?

The Swissflow is still showing 170 to 175 pulses per second. My next step is to plug it into a mobo fan header and take a reading there and see what it tells me. That will either confirm the T-Bal reading or tell me there is something off in the T-Bal software.

Louis

EDIT
I went back and read some of pHaestus's water block reviews and now see that the CPU - Water Delta is about right (6.5C). If anything it is low, but I'm measuring the temp on the exit side of the block.