Hi,

I have a solution I beleive. It’s a bit complicated, but it works well and I find It quite funny (the playground is for playing, isn’t). By the way, I’m fascinated by the vast control possibilities given by the Aquasuite and the Aquacomputer hardware.

Obviously, the idea is to control the pump with a virtual temperature sensor, using various conditions to adjust it.

One of these conditions would be an input value that you would set to 1 when you want to achieve the deaeration cycle.

The time where the pump reaches 100% would be stored. This would allow to set the 5-10 minutes delay that you want until the next ramp-up of the pump.

I join two diagrams to this message (it don't know how to insert images in the post itself). The first named “Last pump deaeration” shows how the time where the pump reaches 100% is stored in a virtual sensor named “Last pump deaeration.” Last pump deaeration.jpg The other diagram shows how the pump is controlled. Pump deaeration cycle.jpg



In the latter, the first condition is the input value named “Manual deaeration” with the “Normal/Deaeration” switch on the right. If set to 1, the rest of the diagram applies. If set to 0 it would be your ordinary way to control the pump.

A new cycle is started either if the input value is set to 1, or if 395 seconds have passed from the last pump deaeration (i.e. 5 minutes plus the time for the cycle to finish - see below). In between the pump is controlled by a fixed value (in my case 30, i.e. 30%)

If a new cycle starts, the time of the deaeration start is saved. Then the seconds from the beginning of the cycle are counted. This value is used to generate a curve that represents the virtual temperature to control the pump. It my case, the curve raises from 30% to 100% in 45 seconds, stays à 100% during 1 min. and then decreases to 30% in 45 secs. As said before, when the pump reaches 100% the sensor named “Last pump deaeration” is actualised. This will stop the cycle, but only after a delay of 95 seconds to allow the cycle to finish. The 95 seconds corresponds to the one minute where the pump is at 100% and the 45 second where it decreases to 30%

Of course, the curve could be replaced with a fixed 100 value. Then it wouldn’t be necessary to store the beginning of the deaeration and to calculate the seconds from deaeration start.

Hope it helps.

This is an interesting and elegant solution. I am still trying to fully understand the second diagram. I can't see what feeds the A input of the last X=switch {A,B].

I have a much more basic solution that sets the pump speed to 100% for 30 seconds and then to 30% for 30 seconds. I downloaded this Virtual Sensor file from an online post (I did not invent it). The logic is limited because the total time of the 2-speed cycle is always 60 seconds because the logic is based on a seconds counter that rolls over from 59 to 0. You can adjust the values in the "Switch Time" Constant, but the entire cycle will always be 60 seconds long. For example, if you set Switch Time to 10, the pump speed would be 100% for 10 seconds and 30% for 50 seconds. If you set Switch Time to 45 seconds, the pump speed would be 100% for 45 seconds and 30% for 15 seconds. Note that this Virtual Sensor will be used to control the pump speed, so the Virtual Sensor Output Units has to be set to °C because in the D5 Next pump section of Aquasuite, the pump is expecting a temperature, not a percentage or any other unit.

Once the Virtual Sensor is working, the next step is to get its output to control the pump speed. The only way I could see to do this is to go to the D5 Next Sensors section and add this Virtual Sensor. Then in the D5 Next Pump section, set the Mode to Curve Controller, then set the Data Source to the Virtual Sensor you just added. Next set up the curve so half of the curve points have temp and speed equal to the Virtual Sensor low speed value and the other half off the curve points have temp and speed set to the high speed value. This basically makes it a 2-point curve, with the low point set to the Virtual sensor low speed value and the high point set to the Virtual Sensor high speed value. As the seconds counter counts from 0 to 30 seconds, the Virtual Sensor will output 30°C (30% speed). When the seconds counter reaches 31, the output will switch to 100°C (100% speed). When the seconds counter rolls over to 0 the Virtual Sensor output will switch back to 30 °C (30% speed). I do wonder if there is a simpler way to connect the output of the Virtual Sensor to the pump speed control.

I set this up with the low speed value set to 50 and the high speed value set to 60 just to test the idea, then set up the curve controller so 8 points have temp and speed values of 50 and 8 points have temp and speed values of 60. The pump runs at 50% for 30 seconds then increases to 60% for 30 seconds, so it does work. If I wanted to change the pump low and high speeds, I can just change the values in the 2 Constants that feed into the X=switch [A,B] logic block, but I would also have to change the speed values for the 16 points in the curve controller. This project only requires a 2-point curve, but Aquasuite does not provide a way to reduce the number of curve points. I think you really only need to set up one low curve point and one high curve point with temps that are equal to the values that the Virtual Sensor outputs. The other 14 curve points really don't matter so long as none of them have a temp value equal to either of the Virtual Sensor output values. Since the Data Source is the output of the Virtual Sensor, and it can only ever have 2 values, if the rest of the curve points are set to temps that are not equal to the high or low value, they will never affect the pump speed.

I exported this Virtual Software Sensor from Aquasuite as an XML file. This forum does not allow uploading XML files but it does allow ZIP files so I put the XML fil into a ZIP file. Unzip the ZIP file, the open Virtual Software Sensors, click the Plus sign to add a new virtual sensor, then right-click on the new sensor name, select Import sensor, and select the XML file. You can select what you want for the high and low speeds, and adjust the Switch Time to control how many seconds in the 60-second cycle are high speed and how many seconds are low speed. Try it and post if it works. Maybe someone else will chime in with a better way to make the output of the Virtual Sensor control the pump speed.

Citation

This is an interesting and elegant solution. [...] I can't see what feeds the A input of the last X=switch {A,B].

Thanks ! The A input of the last switch is fed with the temperature/pump speed I want, in normal operation, i.e. when the deaeration mode is off, i.e. when the input value of "Manual deaeration" is 0.

The solution that you show doesn't have such a switch. So I guess that the solution that you show is activated by a dedicated profile, i.e. the deaeration mode is activated by switching to a deaeration profile. In my solution, there is no profile change. The deaeration mode is activated by the input value, which affects the input of the last switch (A for normal operation, B for deaeration mode).

Citation

I do wonder if there is a simpler way to connect the output of the Virtual Sensor to the pump speed control.

I don't see any other possibility than using the Curve controller mode with the Virtual sensor temperature as controller source, as you said.

For the curve, in my view, the simplest way is to set a straight curve with an input temperature going from 0° to 100°C, corresponding to an outpout power going from 0% to 100%. Only two points would be necessary for such a curve, but it doesn't matter if all the points are used. With such a curve, if the virtual temperature is 30°C, the pump will run at 30% and if the temperature is 100°C, the pump will run at 100%. The same curve will also work with any other temperature.

Quoted from "PierreV"

The solution that you show doesn't have such a switch. So I guess that the solution that you show is activated by a dedicated profile, i.e. the deaeration mode is activated by switching to a deaeration profile. In my solution, there is no profile change. The deaeration mode is activated by the input value, which affects the input of the last switch (A for normal operation, B for deaeration mode).

I had not really thought through how to activate the deaeration cycle. I normally run my pump at a fixed speed because it is generally stated, and I have confirmed through my own testing, that having the pump speed chase coolant temp or coolant minus ambient delta-T does not improve heat transfer efficiency much, if at all, outside of extremes. If I wanted to run deaeration cycles, I could just switch the pump to Curve Controller, which would already be set to use the virtual sensor as its control data source. The switch to Curve Controller could easily be tied to a profile change.

Quoted

I don't see any other possibility than using the Curve controller mode with the Virtual sensor temperature as controller source, as you said.
For the curve, in my view, the simplest way is to set a straight curve with an input temperature going from 0° to 100°C, corresponding to an output power going from 0% to 100%. Only two points would be necessary for such a curve, but it doesn't matter if all the points are used. With such a curve, if the virtual temperature is 30°C, the pump will run at 30% and if the temperature is 100°C, the pump will run at 100%. The same curve will also work with any other temperature.

This makes sense and is similar to the curve I set up. The difference is that I set 8 points to 30% (30°C) and 8 points to 100% (100°C). I don’t think doing so was necessary because any curve point not set to 30°C or 100°C will never be sent to the pump. Regarding having to set the output of the Virtual Sensor to °C, I have had to do this before to get around the AQS limitation of only accepting values with a temperature unit. In some cases (like this one), it makes sense to send an integer RPM value or a percentage to the pump for speed control, which is what AQS does when the pump is set to Power Preset. I wish that Aquacomputer would allow data values with units other than temperature to be used in cases like this.

Quoted from "RuffHi"

Thanks all for the posts. This was my intro the playground - looks like there are a bunch of things that you can do here.

Congratulations! The Playground is a VERY powerful feature of Aquasuite. The potential is limited only by your creativity. I have seen some really complex Virtual Sensors in this forum. What @PierreV shared is a good example. Another example is calculating the amount of heat, in watts, that a radiator is dissipating. If you put temp sensors at the inlet and outlet of a radiator and you know the flow rate and the specific heat capacity of the coolant, the amount of heat the radiator dissipates can be calculated by the formula Q = ṁ x Cp x ΔT where:

Q = cooling capacity in kW
ṁ = mass rate in kg/s (which is liters per hour / 3600 for water)
Cp = specific heat capacity. Distilled water at 35C is 4.178 kJ/(K.kg)
ΔT = delta temp of radiator inlet to radiator outlet.

In this example, I put in Constants in for the temp values that would normally be coming in from the 2 temp sensors. Flow Rate is also a Constant that would normally be data coming in from a Flow Sensor. The coolant is assumed to be distilled water which has a specific heat of 4.184 joules per gram per degree Celsius (j/G°C), meaning it takes 4.184 joules of energy to raise the temperature of one gram of water by one degree Celsius. The calculation shows that if the coolant temp at the rad outlet is 3°C lower than the temp at the rad inlet, and the flow rate is 100 L/H, the rad is dissipating 348.3 watts of heat. Some people may say Who Cares but custom loop enthusiasts and engineer nerds (like me) love this kind of stuff. I attached the XML in a zip file if you want to play around with it. Just make a new Virtual Sensor, then import the XML file.



The playground can also be used to build virtual sensors that are based on real sensor data. For example, you can have your fan speeds controlled by the difference between coolant temp and ambient temp (delta-T). Many people consider this to be better than just using coolant temp because using delta-T eliminates variations in ambient temp from affecting your fan speeds. You could build on this so delta-T is used as the control data source for the fan curve but if your CPU temp exceeds some value for a period of time, the fan speeds will increase until the CPU temp comes back down. The possibilities are endless.

Congratulations to RuffHi and thanks to Speedy-VI for the comments and the interesting and very clear heat dissipation diagram.

I share with you one more diagram, which achieves an automatic deaeration on startup, at an interval of days that is defined by an input value. Auto deaeration on startup.jpg



This diagram uses the same "Last pump deaeration" virtual sensor showed in my post above. The output on the right connects to the A input of the "Manual deaeration diagram" showed in my post above. The A input of the switch on the right would be the ordinary speed of the pump (fixed speed or according to water temperature if preferred).

@Speedy_Vi : how do you insert the image directly into the post ?

Quoted from "PierreV"

@Speedy_Vi : how do you insert the image directly into the post ?

Ah the how do I add an image to my post in the Aquacomputer forum mystery.

It took me a while to figure out how to add an image to a post in this forum. I would see a post where someone had included an image in their post, then click on the Quote button so I could see what they did. Then I would do the same exact thing, and it would not work. What I finally figured out is if I click on Attachments and upload a PNG file, then click in the post box and select Add to Post, it works. If I upload the sane image as a JPG file and do the same thing, it does not work. I have no idea why Add to Post works for a PNG file but not for a JPG file. Clicking the image icon and pasting in a URL for an image uploaded to IMGUR also never worked for me, but I have seen posts where it looked like that is what the OP did and it did work.

Citation de "Speedy-VI"

Citation de "PierreV"

@Speedy_Vi : how do you insert the image directly into the post ?

I click on Attachments and upload a PNG file, then click in the post box and select Add to Post, it works.

Gosh ! Indeed ! It works. Many thanks !

Quoted from "PierreV"

Gosh ! Indeed ! It works. Many thanks !

Glad it worked for you. Strange isn't it?