Cooling SF800 pulse rates from T-Balancer and SpeedFan differ by a factor of 3

I couldn't find this issue addressed anywhere else on this forum, so here's my question, pulse counts from my SF800 flowmeter differ by a factor of 3 when read by T-Balancer and SpeedFan. Why, and which is correct?

I've installed a Swissflow SF800 in my watercooling system, and connected it to a T-Balancer. I'm using a D5 pump from Danger Den, set at 5 and 1/2" tubing. The T-Balancer showed a flow rate of around 2.4 liters/minute, or 144 liters/hour, and a flowmeter pulse rate of 14220 pulses/minute. I was rather disappointed with this flow rate because it is very far back on the D5 pump curve. I had hoped to be somewhere in the middle of the curve, which would be around 500 liters/hour. If this flow rate and the pump curve are correct, the head loss through my system is around 11 feet of water pressure, which sounds like a lot. The D5 motor was drawing 1.44 amps at 12 volts, or a little over 17 watts, which seemed to match the pump curve. System temps seemed okay, so I just accepted that my system was pretty restrictive.

For no particular reason, I connected the SF800 to one of my motherboard fan headers, and SpeedFan indicated 42188 pulses per minute, almost exactly 3 times what the T-Balancer showed for raw pulses per minute. If this is correct, the flow through my system is approximately 7 liters/minute, or 422 liters/hour which is close to the middle of the pump curve, and indicates that head loss through my system is a more reasonable 8.4 feet of water.

Has anyone had this experience with the SF800, and can anyone explain why the pulse rates from T-Balancer and SpeedFan differ by a factor of 3? To paraphrase Abraham Lincoln, both may be, but one must be, wrong, but which one? Any thoughts would be greatly appreciated. Thanks.

 

The SF800 generates 100 Hz per litre/min, with a peak measurement of 2000 Hz (20 litres/min). Personally, I have no experience with the T-balancer but know that Speedfan normally just reads pulses as fan RPM, thus the value it reports needs to be divided by 6000 to get actual litres/min.

However, you need to make sure that Speedfan does not multiply or divide the fan signals. There is an option for it to do so, under Configure --> Advanced tab (see: FAN1 to FAN3 "mult" and "div" values, which should all be set to 1).

Many people express some disappointment when they measure the actual flow going through their systems compared to the rather theoretical value mentioned in the specifications of their pump. Make no mistake: the average watercooling loop experiences a massive pressure drop, and flow is substantially less than what the pump is reported be capable of under rather ideal(ised) conditions (think of the max. MPH on your car speedo). This is normal. Flow is a bit overrated anyway; I get 1.73 litres/min. and my dual Opteron 250 runs just fine on a single BIX.

 

Nexxo,
Thanks for that information. I haven't used SpeedFan much, and didn't even realize it had those adjustments. I knew MBM had them, but haven't used that for a while. The fan divisor for the SF800 was set to 32 (probably from a configuration I downloaded from almico.com). When I changed the divisor to 1 the reading matched T-Balancer. Why a divisor of 32 multiplied the reading by 3 eludes me, but it isn't important.

By the way, you've probably already noticed this, but the raw rpm reading from the SF800 equals the flow rate in cc's/hour, which, if you divide by 100, gives you flow in liters/hour, which is what the D5 pump curve uses. As you noted, however, the actual flow rate isn't important, it's more a way to detect subtle changes in you system that restrict flow.


Thanks Again,
Jim

 

That's why I have one. Great way to detect blockages, pump failure or leaks (although by that time, the sparks are a definite giveaway).