Digital Flow Meters

[daw123]

Hello guys and girls,

I'm looking to buy early next week a digital flow meter for use with the M Cubed BigNG controller. I have the following two choices:

M Cubed T-Flow. Cost: £18.50
Alphacool 17206 Stream Control SC1 Optimised for Highflow with built in sensor. Cost: £13.70

What puts me off with the M Cubed T-Flow is that the inlet and outlet are on the same side, which will be awkward to retrofit in my existing loop.

I prefer the Alphacool meter because it has a "straight through" flow configuration. The problem with the Alphacool is that it uses a stainless steel spindel on the impeller. I'm not sure if this will cause problems with my all brass/copper loop. (btw do the Swiftech pumps use stainless steel spindels?).

Do you think this will be ok, or are my concerns justified?

Edited for spelling for the pedantic (such as myself) among you

 

[Gillbot]

stainless is fine

 

[Rubycon]

316SS is fine. Best pumps are magnetic driven which completely isolate the motor from the pumped fluid and the magnet/impeller can be PTFE coated. Probably way overkill for a cooling system as there are many non intrinsically safe parts in the rest of the system!

The radial in/out of the first sensor is fine if you don't mind using a 90 on both sides. It will add a good 2' of head to the line though. (not counting flow losses of the sensor itself) You could also go ultrasonic but you probably have a budget.

 

[daw123]

Thanks Gillbot and Rubycon for the replies.

I've just placed the various orders for the following components:
M Cubed T-Balancer BigNG 4-channel Fan Controller
M Cubed T-Balancer Fan Controller - Extension Kit
Alphacool 17206 Flow Meter with Built In Sensor
2x Alphacool Thread Adapters
2x G1/4 1/2" OD EK Barbs

A nice £125 birthday present to myself for my birthday next Saturday

I think I've remembered everything; please point out any obvious (or not so obvious) cock ups.

btw its difficult in the UK to buy MCubed products. The orders I placed with the above suppliers say that the kit is in stock, but that remains to be seen (I've had this before where suppliers have lied about their stock levels).

An edit yet again for spelling - typing after coming back from the pub is awkward!!!

 

[aigomorla]

Do you really need a flow meter? :T

 

[Rubycon]

Originally posted by: aigomorla
Do you really need a flow meter? :T

Flow detection should be mandatory IMO. Having a rate is a bonus as you can get an ahead of time warning that something has changed. Think idiot light vs. gauge.

 

[LOUISSSSS]

what are the benefits of having a flow-meter vs just RPM sensing from the Pump-> MB?

besides the obvious: Extra layer of security telling you that your water is moving in your loop.
are there any other benefits?

 

[Rubycon]

Having positive indication the pump is running may be alright but that does not mean you have protection! Many factors could allow the motor to turn yet water flow to slow/stop, etc.

Positive proof of flow - even a pressure sensor may not be enough - is good to prevent irreversible damage to components in the event flow is stopped for whatever reason. A partially collapsed hose, for example; can cause a rise in operating temperature while the pump still runs. The speed may change and if you do monitor pump speed that's the first thing you need to determine what kind of pump it is! Some will drop in speed at lower head (more water flow) and some will drop in speed under higher head (less flow).

 

[daw123]

I'm using the flow meter for the reason mentioned by Rubycon; I want to know that there is liquid flowing around the loop and if not, then there is a problem. With the BigNG, I'll set it up so that if the flow rate changes for no apparent reason (i.e. pump speed hasn't changed), then it will hard shutdown my PC.

I'm hoping to set it up, so that the BigNG:
Monitors pump RPM, flow rate, and inside case temps
Controls the pump speed based on the BIOS CPU temps (or software temps)
Uses the hard shut down facilitity if there is a problem
Controls the case fan speeds using the temperature probes positioned inside the case

I understand that the BigNG can do all the above automatically, provided I setup profiles in the s/w.

 

[daw123]

I've checked out the M Cubed forum and I've found that the temp sensor data is from the analogue and digital sensors rather than from BIOS.

So, I have problem if I want to use the BigNG to control my pump speed, because I have no way of sticking the digital sensor to the side of the CPU heat block to meaure the CPU temperature and the readings from this sensor will also be highly inaccurate.

The other problem is that if I use something like Everest, which is compatible with the BigNG, but appears from comments made on the forum that the compatability is buggy, I am then reliant on being in the OS with these programs running for the pump control to work; i.e. what happens during POST or in BIOS or if I'm running Memtest, as examples? Since there is no temperature data, would this mean that the pump doesn't work?

The alternative is to use a temperature probe on the outlet side of the CPU block and use the temperature of the water to adjust pump speed.

Another safer bet (in case the pump gets switched off due to erroneous temp data) is to leave the pump running at 100% speed and adjust the fans on the rad instead. The benefit of this solution is that the three fans are far more noisy than the pump (which was very quiet when I bench tested it) when they are running at full speed for max. cooling. It's also safer with regard to the cooling of the CPU to have the fans not working on the rad, rather than the pump not working.

I also assume that I can't control both the pump and fans on the rad (the software would ramp up pump speed and drop fan speed or vice versa or oscillate the two speeds)

Either way, is it possible to control the fans speed on the rads OR pump speed using a temp probe in the loop? And is the best location of this probe on the outlet side of the CPU block?

Also, is the best lcoation for the flow meter on the outlet side of the pump?

Here's an edited photo of the inside of my case, showing the proposed locations in my loop for the temp probe and flow meter.

This is the temperature probe I'm thinking of using with a couple of EK G1/4 1/2OD barbs.

Thanks for reading this and any help you guys can provide is appreciated.

 

[Rubycon]

It's better to leave the pump speed constantly ON. If flow must be controlled an electronic bypass valve should be installed.

Variating speed on the fans is definitely the way to go but not both. If you control both what you find it the system will hunt and if you're using proportional integral derivative (PID) method it will eventually settle to fans somewhere near slow to slowest speed and pump fully on. (again highly dependent on system load and ambient temps, etc.)

In a simple loop system the flow will be the same anywhere in the loop. The sensor should be in a place that would get the least interference from bubble impingement. There should never be cavitation as well so always keep it on the "high side" so between the discharge of the pump and the rad is where you want it to go. Your diagram shows it on the low side for reference.

If you must use software that runs in windows make sure the pumps/fans "fail" (meaning when there's no way to control them) at full ON. This way you have maximum cooling capacity when using a foreign OS such as Linux, DOS, or in BIOS etc.

 

[daw123]

Originally posted by: Rubycon
It's better to leave the pump speed constantly ON. If flow must be controlled an electronic bypass valve should be installed.

Variating speed on the fans is definitely the way to go but not both. If you control both what you find it the system will hunt and if you're using proportional integral derivative (PID) method it will eventually settle to fans somewhere near slow to slowest speed and pump fully on. (again highly dependent on system load and ambient temps, etc.)

In a simple loop system the flow will be the same anywhere in the loop. The sensor should be in a place that would get the least interference from bubble impingement. There should never be cavitation as well so always keep it on the "high side" so between the discharge of the pump and the rad is where you want it to go. Your diagram shows it on the low side for reference.

If you must use software that runs in windows make sure the pumps/fans "fail" (meaning when there's no way to control them) at full ON. This way you have maximum cooling capacity when using a foreign OS such as Linux, DOS, or in BIOS etc.

Thanks Rubycon, I think I understand your reply.

Essentially, the flow meter can go any where in the loop since the flow rate is the same all the way around (this makes sense to me) and the temp sensor should be on the outlet side of the pump - so this could be positioned where I showed the flow meter in my linked photo.

For convenience sake; I'll place the temp probe between the pump outlet and CPU block inlet and the flow meter will go between the rad outlet and res inlet (essentially on the two long pieces of tubing in my photo).

I can then use the temp probe to control the fans on the rad, but with the controller set to fail with the fans set on max. speed. I won't bother controlling the pump - it will be left running at 100% speed.

 

[Rubycon]

Water temp sensors - tee based probe holders with the sensor directly in the flow - CAN be mounted directly downstream of a CPU/GPU so when it gets hit with a load suddenly the controller reacts faster to fan speed to prevent temps from increasing too much overall. Unless you have access to powerful simulation software (Algor?) as well as the training in using it, you may have to just play around with position to see what suits your set up best.

 

[daw123]

Ah ok. I'm going to have to experiment with the temp probe.

 

[daw123]

Sorry guys, but I've resurrected this thread to vent some steam.

I placed the order with Specialtech on the 5/4/09 for the TBalancer Analogue Extension Set and two G1/2 to G1/4 unequal female adapters (or Alphacool Stream Control Adapters as they call them).

This is what I received today from Specialtech:

> To: #####
> Subject: ####
> Date: Mon, 27 Apr 2009 10:24:42 +0000
> From: support@specialtech.co.uk
>
> Dear Sir
>
> with regards to your order we no longer have the Alphacool Stream Control adapters in stock and we have no due date for these items to come back in to stock.
> We can refund these items and send the Tbalancer or refund the whole order.
>
> Many apologies for the inconvinience
>
> Kind Regards
>
> Specialtech

This is what I sent in response:

RE: [####]: Specialtech Order ###
From: ####
Sent: 27 April 2009 11:36:52
To: support@specialtech.co.uk

Dear ####,

I am very disappointed that the Alphacool Stream Control Adapter is not in stock considering the order was placed in early April 2009 and I was told twice last week that the order had been shipped.

However, I would still like you to ship the TBalancer Analogue Extension Set (is this item in stock?) and I will obviously want a refund for the Alphacool Stream Control adapters.

This has put me in an awkward position as your shop seems to be the only one in the UK which (supposedly) sells these adapters and I have already bought and received the components which I need the adapters for. Clearly, I am going to have to source the adapters (or something similar) from elsewhere.

Kind regards,


####

I was very tempted to send something a lot harsher, however I decided against it.
I'm sick of online shops messing me about with orders (this is the fourth shop to do something like this). Why the hell do they show items in stock on their website when in actuallity its complete rubbish and the item is out of stock?

I've now got to source the adapters from somewhere else.

Do others have these problems when ordering components online? It's very annoying.

 

[WoodButcher]

I sent a PM, If you have trouble getting the parts you need and can find them in the states but they won't ship overseas just yell, I'll help. Some of the stores here are willing to source and ship. Check with Aigo, I'm sure he'll have a contact or two that do this.

 

[daw123]

Thanks for the PM Woodbutcher.

My dad thinks he may have seen a brass G1/2 internal threaded to 1/2" OD barb (presumably used for plumbing in boats) in the yacht chandlery around the corner from where he's lives. He is going to check it out and get back to me.

Thanks again and I will keep the thread updated with what's happening.

Edit: I've received a notification from Specialtech that the order for the TBalancer Extension set has been dispatched.

Funny thing is that the disptach confirmation still shows the two Alphacool Stream Control Adapters even though an earlier e-mail from Paypal confirms the refund for the two adapters. Weird.

 

[daw123]

My dad has managed to get two brass G1/2 internal threaded to 1/2" OD barbs from the local yacht chandlery, so problem solved.

The barbs don't have internal sealing O-rings, so I'm going to have to use this EDPM tape to seal between the flow meter and barbs.

Thanks to Woodbutcher for the offer of helping me get the adapter.

 

[daw123]

TBalancer Extension set has turned up, so I now have all the parts except for the temperature probe, which I will order when I place the orders for the following:
EK Supreme LGA1366 bracket
EK Supreme LGA1366 back plate
RAM
MB
CPU

Geez that's a lot of parts just to get the damn thing running.

I can also see there is going to be quite a bit of head scratching connecting the bigNG together since the instructions M Cubed provide are crap and very basic.

btw the hard shutdown function works by connecting between the case power switch and MB. I was suprised, since I though it would connect between the PSU and MB on the 24pin ATX connector. Their solution is simpler than mine though

 

[Rubycon]

That's not a hard shutdown. Use a solid state DPDT relay wired to your power supply's LINE SIDE (input).

 

[daw123]

Originally posted by: Rubycon
That's not a hard shutdown. Use a solid state DPDT relay wired to your power supply's LINE SIDE (input).

Rubycon, does a PC shutdown immediately when you press the case power button, or does it go through the Windows shutdown process. I'm not sure because I've never tried. If the latter is the case, is there a way of stopping the Windows shutdown process, so that the PC shutdowns immediately.

With regard to your idea, how would I go about doing this (specifics please for a layman).

I figure that I can use the signal (essentially a switch which is usually on) from the TBalancer Extension Set, since this signal is what would be controlling the shutdown function via the case power switch if there were a problem and if it was connected as M Cubed intended.

 

[Rubycon]

Originally posted by: daw123

Rubycon, does a PC shutdown immediately when you press the case power button, or does it go through the Windows shutdown process. I'm not sure because I've never tried. If the latter is the case, is there a way of stopping the Windows shutdown process, so that the PC shutdowns immediately.

With regard to your idea, how would I go about doing this (specifics please for a layman).

I figure that I can use the signal (essentially a switch which is usually on) from the TBalancer Extension Set, since this signal is what would be controlling the shutdown function via the case power switch if there were a problem and if it was connected as M Cubed intended.

When you ground ATXPWR on the mainboard Windows will initiate a shutdown. A hung program can stop it and if the system is hard locked it won't respond. HOWEVER if you hold that ground the system will shut down regardless of what Windows does OR if it's hard locked. At the very least you'd want it that way.

Disengaging mains power (breaking both legs hot/neutral if you're in the 'States running on 120VAC or both hots in the case of 240VAC) provides the most fail safe method to completely remove power. Electrically it's the same as you yanking the power cord from the machine but leaving the ground line intact. (A triple pole relay can be used if you desire to break the ground also but that is completely optional and MAY violate regulations depending on your local codes, etc.)

Solid state relays can be normally open or normally closed. In this case you want normally open - that is if the turn on voltage ceases to exist at the terminals the connection is broken. This is also known as fail open condition. Your monitoring circuit supplies the necessary turn on power - usually 3 to 20+ Volts either AC or DC depending on the relay type. When this power is removed the relay terminals will open. This is a solid state relay so it has no coil/armature/contacts, for example.

Finally I'd recommend a time delay with this circuit. Basically it warns you that flow has dropped/ceased and will give you time perhaps a minute (but not too long as parts heat up and if you restart the pump with hot parts the shock is not good, etc.) to correct the deficiency. This way the system is not plunged into darkness instantly without warning!

 

[Gillbot]

Originally posted by: Rubycon

Originally posted by: daw123

Rubycon, does a PC shutdown immediately when you press the case power button, or does it go through the Windows shutdown process. I'm not sure because I've never tried. If the latter is the case, is there a way of stopping the Windows shutdown process, so that the PC shutdowns immediately.

With regard to your idea, how would I go about doing this (specifics please for a layman).

I figure that I can use the signal (essentially a switch which is usually on) from the TBalancer Extension Set, since this signal is what would be controlling the shutdown function via the case power switch if there were a problem and if it was connected as M Cubed intended.

When you ground ATXPWR on the mainboard Windows will initiate a shutdown. A hung program can stop it and if the system is hard locked it won't respond. HOWEVER if you hold that ground the system will shut down regardless of what Windows does OR if it's hard locked. At the very least you'd want it that way.

Disengaging mains power (breaking both legs hot/neutral if you're in the 'States running on 120VAC or both hots in the case of 240VAC) provides the most fail safe method to completely remove power. Electrically it's the same as you yanking the power cord from the machine but leaving the ground line intact. (A triple pole relay can be used if you desire to break the ground also but that is completely optional and MAY violate regulations depending on your local codes, etc.)

Solid state relays can be normally open or normally closed. In this case you want normally open - that is if the turn on voltage ceases to exist at the terminals the connection is broken. This is also known as fail open condition. Your monitoring circuit supplies the necessary turn on power - usually 3 to 20+ Volts either AC or DC depending on the relay type. When this power is removed the relay terminals will open. This is a solid state relay so it has no coil/armature/contacts, for example.

Finally I'd recommend a time delay with this circuit. Basically it warns you that flow has dropped/ceased and will give you time perhaps a minute (but not too long as parts heat up and if you restart the pump with hot parts the shock is not good, etc.) to correct the deficiency. This way the system is not plunged into darkness instantly without warning!

I believe that depends on the bios settings. Many boards have a section for ATX PWR where it states:
Instant Off
Delay
etc...

If you set it to instant off it will power down, otherwise it will initiate the power down sequence or whatever else you have it set to.

 

[daw123]

Originally posted by: Gillbot

Originally posted by: Rubycon

Originally posted by: daw123

Rubycon, does a PC shutdown immediately when you press the case power button, or does it go through the Windows shutdown process. I'm not sure because I've never tried. If the latter is the case, is there a way of stopping the Windows shutdown process, so that the PC shutdowns immediately.

With regard to your idea, how would I go about doing this (specifics please for a layman).

I figure that I can use the signal (essentially a switch which is usually on) from the TBalancer Extension Set, since this signal is what would be controlling the shutdown function via the case power switch if there were a problem and if it was connected as M Cubed intended.

When you ground ATXPWR on the mainboard Windows will initiate a shutdown. A hung program can stop it and if the system is hard locked it won't respond. HOWEVER if you hold that ground the system will shut down regardless of what Windows does OR if it's hard locked. At the very least you'd want it that way.

Disengaging mains power (breaking both legs hot/neutral if you're in the 'States running on 120VAC or both hots in the case of 240VAC) provides the most fail safe method to completely remove power. Electrically it's the same as you yanking the power cord from the machine but leaving the ground line intact. (A triple pole relay can be used if you desire to break the ground also but that is completely optional and MAY violate regulations depending on your local codes, etc.)

Solid state relays can be normally open or normally closed. In this case you want normally open - that is if the turn on voltage ceases to exist at the terminals the connection is broken. This is also known as fail open condition. Your monitoring circuit supplies the necessary turn on power - usually 3 to 20+ Volts either AC or DC depending on the relay type. When this power is removed the relay terminals will open. This is a solid state relay so it has no coil/armature/contacts, for example.

Finally I'd recommend a time delay with this circuit. Basically it warns you that flow has dropped/ceased and will give you time perhaps a minute (but not too long as parts heat up and if you restart the pump with hot parts the shock is not good, etc.) to correct the deficiency. This way the system is not plunged into darkness instantly without warning!

I believe that depends on the bios settings. Many boards have a section for ATX PWR where it states:
Instant Off
Delay
etc...

If you set it to instant off it will power down, otherwise it will initiate the power down sequence or whatever else you have it set to.

I think I understand. So if I change the ATX PWR settings in BIOS to Instant Off, when the BigNG extension set switches the case power switch to off, the PC will shutdown immediately.

Have I interpreted this correctly?

My goal is that if there is a problem with the water cooling loop, I want the PC to shutdown asap to save components from being sprayed with water.

 

[Rubycon]

I've only seen instant off outside of Windows. In either case a fail safe should remove power from the PSU just to be sure.

You can still have water leakage/spray in a catastrophic failure. Fortunately the pressures involved are very low so the risk is low. If you were connecting to a chilled water system running at 50psig it would be a completely different story. (and you certainly would NOT use any off the shelf enthusiast type waterblock.) You could however use a water-water heat exchanger in place of a radiator but this is completely off topic so...