Advice on Capactive fluid level sensors .

[garyd]

Hi , firstly let me say that I am not an electronics expert by any stretch of the imagination but I do have a basic grasp of the subject and have come here as it was recommended as a good place to seek advice .I have an external computer water cooling system known as a Zalman reserator xt and have been heavily modifying it I have now come up against a brick wall though as my knowledge is simply not enough , the problem I need addressing is that the reserator has a built in Aluminium reservoir for the cooling fluid connected to a water pump which I want to want to do away with and replace with 2 pumps each incorporating it's own reservoir, the reserators own reservoir has 2 water level sensors which I believe to be of the Capacitive type each has one single Yellow wire connecting it to the front control /display panel/circuit board which incorporates an alarm should the level fall too low the alarm sounds and the unit shuts down, I want to do away with the level sensors as I do not need them and they would be very difficult if not impossible to incorporate into the modded reserator , my question is can I fool the circuit board into thinking they are present when they are not or could I simply join the 2 single Yellow wires together and achieve my aim that way ? any help or comments would be most appreciated , Gary .

 

[kubeek]

My guess would be that those are not capacitive sensors, but rather plain old metal plates which use the conductivity of water to sense the level.
You could try measuring the resistance between the connector and the presumably metal end of the sensor inside the tank. If I am correct and it measures very low resistance, then simply connecting these two together or via a resistor should be enough to fool the system.

 

[cowboybob]

Welcome to ETO, garyd!

Are the sensors in direct contact with the water?

<EDIT> kubeek is correct if it's a resistive circuit (reason for the above question).

Although the single wire does suggest a cap circuit...

 

[garyd]

Thank you Kubeek and Cowboybob yes the sensors are in direct contact with the fluid one is positioned about 20mm above the other , I did a bit of research on level sensors in general and the only ones that looked very similar to mine were capacative they have a metal sleeve but what the metal is i'm afraid I do not know , it's late here now and to test the resistance i'll need to drain the system which I will do in the morning and then let you know .P.S I should have stated that I use a non conductive fluid so i'm wondering if a resistive sensor would even work and they do work so i'm not sure .

 

[cowboybob]

Just thinking out loud here:

Can you just permanently fill the reserator xt unit's reservoir (to a level that satisfies both sensors) and isolate it from the new ones, thereby eliminating an empty response from the level detector?

 

[garyd]

I should have been more clear in my original description of what I am trying to achieve Bob , the reserators own reservoir is quite large as water cooling reservoirs go and takes up a lot of the space inside of the case which prevents me from fitting larger more up to date and efficient Radiators which also sit inside it's case , there is actually no need for a liquid cooled PC to have a reservoir at all but one is usually used to help in priming pumps and evacuating air , I have tried contacting Zalman for more info but cannot even get a reply (they may no longer exist) and the only info I can find on the net that is in any way relevant is this "Waterlevel: You will need to find a water level sensor that can output 0-5V, 0-10V, 0-20ma or 4-20ma. If all you need is a water level low alarm a float and a switch on an Arduino digital pin will do. Note, 0-20, 4-20ma and 0-10V need other parts or they won't work or will damage themselves or the Arduino." which again is of very little help , I will test the resistivity later today but was wondering whether I should test across the probes themselves or across the 2 Yellow wires ?.

 

[garyd]

I have done some measurements and although they mean very little to me they may well do to you , firstly though let me make the reserators power more clear it is not in any way mains powered drawing it's 12 volt supply directly through a 2 wire molex cable connected to the PC's power supply .With the Reserator off and a multi meter set in the 2M range on the Ohm scale and the 2 yellow wires connected to the sensors i get a reading of 098 if I take the Yellow wires off and take a reading across the sensors I get a reading that climbs fast and then goes out of range fluctuating somewhat as it does so this also applies with the reserator switched on, if i go to the 20M range with the unit switched on the reading fluctuates very fast and then goes to zero , I have also noticed with the Reserator turned on if I remove the wire from the lower sensor the alarm does not sound but if I remove the upper wire it does go to alarm .I hope this helps .

 

[cowboybob]

When the wire from the lower sensor is removed, does anything happen other than no alarm, i.e., does the reserator shut down (or some other disabling action)?

If not, I'd say disconnect the lower sensor wire and be done with it.

As an aside, is the reserator controlled by an Arduino uC (does not appear to be from the pix I've seen)? Or is the quote above concerning the water level sensing only suggesting than an Arduino (of some flavor) be used for that function?

Also, do you use the recommended "anti-corrosion coolant" additive? That stuff does conduct (although, undiluted, very little - 4.4 µS/m) and leads me to conclude that the level sensing circuit might be resistive. Much easier to deal with than capacitive. We can go the jumper route for solving your problem if my first idea above doesn't hold true.

 

[garyd]

no nothing happens at all when I disconnect the lower wire even the display of the fluid level which is done by a series of illuminated bars remains unchanged , the reserator is not controlled by an arduino the quote I posted was from the only relevant discussion I could find on the net and I included it in the hope that it might help to identify what type the sensors were , I do not use the Zalman fluid as it's additives turn to something resembling Blue used chewing gum after 6 months to 1 year of use and cause no end of problems till you realise the cause, I use instead and have done for 4 years now a fluid of German origin called Fesser One which in it's unused state is non conductive but I do know that like all similar fluids it will gain conductivity over time due to it absorbing metal ions from the various water blocks and the radiators which are Copper or Nickel plated Copper as the case may be .The Reserators own reservoir although being Aluminium is heavily Anodized both inside and out and there are no signs of any corrosion which may have been caused by using mixed metals in the same cooling loop.The main reason for me wanting to do away with the reservoir and hence the level sensors is that it's large size prevents me from adding the components I want to use , I long ago relieved the Reserator of it's duty controlling the pump by using a PC fan/pump controller known as an Aquaero 6 which I use to control all the additional pumps and fans throughout the PC and the Reserator itself so in conclusion my mission is to dump the Reserator reservoir and it's level sensors but keep it's front display and it's ability to detect a loss of flow through a paddle type flow sensor which is built into the Reserator itself whilst keeping it's audible alarm .

 

[cowboybob]

OK. Back to the beginning:

... the problem I need addressing is that the reserator has a built in Aluminium reservoir for the cooling fluid connected to a water pump which I want to want to do away with and replace with 2 pumps each incorporating it's own reservoir,... I want to do away with the level sensors as I do not need them and they would be very difficult if not impossible to incorporate into the modded reserator, ...

And:

no nothing happens at all when I disconnect the lower wire ...

(My emphasis).

So, best as I can tell, you've discovered to the solution to your problem.

I also recall reading somewhere about the reservoir for your cooling system that it's only needed to make up for volume lost during any air purging process. Apparently, there is little or no evaporation losses.

 

[garyd]

No it does'nt lose fluid to evaporation but I have'nt found the solution because if the upper wire ie sensor is disconnected I get the alarm and my aim is to disconnect both sensors and dump the reservoir and not have the unit go to alarm .

 

[cowboybob]

OK. Now I understand your goal. (You can call me slow, but you can't fault me for not being thorough...)

One last test. With the sensors hooked up, the Reserator powered up and with coolant in the reservoir, is there a detectable DC potential (with respect to the circuit Gnd, not the chassis Gnd)) on either of the sensors? Or across the sensors? I would expect it to be <=1V, if present (anything higher would cause unacceptable electrolysis of the cooling fluid).

If there is a DC potential, I'm going to assume that the water level sensor is monitoring conductance.

I'm confident that the following trials will not damage your Reserator's electronics.

Do you have access to a local Radio Shack (or equivalent)? If so, purchase a 5 pack of 10M ohm 1/4 W resistors (RS Model: 271-1365, they don't appear to have 1/8W available although 1/8W would be adequate).

Now, with NO coolant in the reservoir, connect all of the resistors in series (50M ohms total). Connect the two ends of the group to the sensor wires and observe the response of the Reserator when powered up.

Depending on the response, repeat the above after removing one resistor at a time until only one is left and observe the responses.

Finally, if you still haven't gotten the response you want, start adding back the resistors, one at a time, but in parallel, thus further reducing the total resistance seen by the Reserator electronics.

 

[garyd]

It's ok I knew we were getting our paths crossed so to speak as it's difficult to convey something using only words , I really appreciate your patience and your willingness to help I will test for the DC potential in the morning as it's gone midnight here now and I need my beauty sleep lol, the resistors will take a few days as i'm disabled and not able to get to my local electronics supplier so i've ordered some 1/4 watt 10 ohm resistors on ebay and they should be with me by the middle of next week I looked for 1/8 watt ones but they also don't seem to have any so i've settled for the 1/4 watt ones , when i've tested the potential in the morning i'll post the results so you can maybe start to get a better idea of what is going on and then i'm afraid the rest will have to wait till I receive the Resistors sometime next week .

 

[cowboybob]

Not 10 ohms, 10M ohms (10,000,000 ohms).

We're starting with very high resistances to emulate the very low conductance of water with few free ions.

 

[garyd]

Ok I gotcha and i've changed the ebay order to 10M Ohm .

 

[garyd]

I have tested the DC potential and across the sensors it's in the Millivolt range and fluctuating and across a single sensor it's about 0.45 of a Volt .

 

[alec_t]

Assuming the sensors are of the capacitive type you could try replacing the upper sensor by a fixed capacitor of say 470pF-1nF (a pure guestimate for the value).

 

[garyd]

Hi Alec I appreciate your input but with help from Cowboybob it's looking increasingly likely that the sensors are not Capacitive after all and are more likely to be of the Conductive variety, I think Bob has a very clear idea now of what i'm trying to achieve and he's taking me down a route he has clear in his mind , please don't read this as a rebuff because nothing could be further from the truth I welcome all Input from anyone kind enough to give it .

 

[garyd]

Hi Bob , I received the Resistors today and then realised that I could'nt carry out your instructions precisely to the letter as the Reserator would alarm if no fluid was in it when switched on as there is also a temperature sensor ( already sorted ) and the flow meter connected to the buzzer so I went through the process anyway and just diconnected the level sensors wires, starting with 5 Resistors in series nothing changed until i got to 2 resistors but then it would alarm after about 5 seconds I went to one resistor and it improved by a second or so but still always ended up on alarm , I then wired one resistor in parallel and things were a little better in that it took about 10 seconds to alarm I went to 3 resistors 30 seconds and then alarm and then on 4 resistors in parallel I finally had it cracked it has now sat switched on for over 3 hours with not a single alarm everything is working as it should and even the bars that show level on the display are lit up and saying the reservoir is full , I really cannot thank you enough you have solved a very big headache for me and you have my eternal thanks , Gary .

 

[garyd]

I'm adding an update here in case anyone reads this article and thinks that I had the problem solved as I did at first , after a day or so of running fine with 4 10M ohms resistors in parallel the system started going to alarm again so I added 1 more resistor at a time until I reached 7 in all but things got worse rather than better and I was about to seek advice again when something inside me told me to add a single 10 Ohm resistor also in parallel with the existing 4 10M Ohm resistors I was worried that i'd damage something but to my surprise it's worked beautifully for a full week now with not one single alarm the knowledgable amongst you may know the reasons why this worked and all comments are welcome , Gary .