4-Channel PID Dew Heater Controller

[rweust]

To prevent dew on the secondary mirror of my 10” CT Newton, I used a simple 2-channel PWM adjustable power controller (based on the famous NE555) to supply a NiCr heating wire behind my secondary mirror and some accessories (finder, guide scope, etc.). This solution worked of course to provide for the necessary dew-preventing temperature rise of the secondary mirror, but I always had the feeling I missed something, never knowing when to start heating and exactly how much. It was always a bit of a guess to set the right amount of heating power and always there was the wondering if I wasn’t heating too much and influence the seeing negatively by causing tube currents. The desire came up for a load sensor to measure the real temperature of the secondary mirror and compare it with the temperature of the surrounding air.

 

 

 

 

The second dew problem I rarely encountered (but occasionally still the case in wet Holland) is the main mirror becoming too cold and dewing up. It would be nice to know the temperature of the main mirror too and again being able to compare it with the surrounding air temperature. This way I could determine the moment to prevent the main mirror from cooling down to much by radiation, for example with the help of a little hat (Astrozap dust cover) around the back of the tube.

 

 

 

The cooling down of a newly set up open system like a Newton telescope shouldn’t take hours. There is even a small forced cooling fan on the back of my main mirror, but how long should it be on? Is this fan very effective? Is this cooling down time the same under all circumstances? Another reason to be able to compare the temperature of the main mirror with the surrounding air.

 

I would need a display and at least 3 temperature sensors to provide this information on my Newton. Also some power controllable Pulse Width Modulation outputs are needed. Nowadays a microcontroller is then the logical component to use. An Ardiuno module will certainly do the trick and there are a lot of software libraries to choose from, but a far more cheaper, small 18-DIL cased Microchip PIC 16F1847 can work as well and is even prepared for steering 4 PWM outputs directly from its hardware.

 

 

 

If you have a load sensor to measure the real temperature of the heated element and a microprocessor to steer a display and PWM output anyway, the best temperature controlling and most stable way to steer the power of the heating is by use of a PID controller. This PID controller can easily be programmed in the software. To show the difference between a normal on/off thermostat and a PID controlled heating, I compared the temperature reactions of a secondary mirror heated with an 2W element and set to heat 3 °C higher at 6 °C set temperature (of course too much for reality, but to give a clear picture). The red line is the temperature of the secondary mirror (°C). The blue line represents the amount of heated power (W) steered by the controller.

 

 

 

 

If there are 4 PWM outputs on this microcontroller, it is easy to make them work with each their own load sensor. The possibility to configure each output channel independently to control a heated element connected (with or without load temperature sensor) would make this an universal temperature controller. Not only useful for a Newton, but for a wide variety of telescopes that need to be temperature controlled in one way or the other or contain elements or accessories that do.

 

From these starting points this 4-Channel PID Dew Controller is born. In an earlier topic on CN and topics on Dutch/Belgium and German armature astronomy sites I showed what it can do. The PID heating controlled secondary mirror stayed within a 0,1 °C temperature window compered tot the measured surrounding air temperature. The temperature of the main mirror could be watched over closely. Together with the transparent low-emission window foil I used around my carbon fiber tube, all the temperatures of the optics and tube stay inside a 0.5 C window with the surrounding air temperature, even under fast temperature dropping circumstances that night. Not only preventing dew, but also providing wonderful low temperature differences in this open system telescope, preventing negative influences on the seeing, like tube currents. I only used one of the four PID channels in this test.

 

At 22.05 hour I started heating the fast temperature dropping secondary mirror, first to follow the surrounding air temperature exactly and later at 22.50 hour to heat with a little more offset (0,1 °C above surrounding air temperature), because of the dangerously close dew point.

 

 

 

 

I would really like to share the building of this 4-Channel PID Dew Heater Controller. The cost for building this PCB with all the components on it can be kept low, about EUR 70,-.

 

 

 

 

The specs in PDF:

 

https://www.dropbox.com/s/iqq2hk921e98oe2/4-Channel%20PID%20Dew%20Heater%20Controller%20Specs.pdf?dl=0

 

 

Robert.

 

[rweust]

The design is also simple, not just the processor.

 

 

A Schottky diode, a 4700 μF buffer capacitor and a 1 μF foil capacitor provide for an adequate decoupling of the 12V supply, preventing high frequency distortions (caused by the PWM block waves powering the dew heaters) reaching other equipment over the 12V supply lines. This Schottky diode also protects the circuit against polarity faults in the supply connection.

 

All the DS18S20 temperature sensors are on one input of the microcontroller. These one-wire devices can all be connected parallel. So in theory you can jump from sensor to sensor on your telescope with only one cable to connect for all the temperature sensors.
I used the DS18S20 temperature sensor which normally is a 9-bit temperature value sensor, so the accuracy is 0.5 °C, but this sensor also sends back a remain and slope byte that can be used to calculate a decimal more and increase the accuracy to 0.1 °C. The 12-bit DS18B20 can also be used (with a small software adaption) providing the same accuracy of 0.1 °C.

 

The push buttons are connected to another input of the microcontroller with a normal resistor-capacitor network. The LCD display is connected the common HD44780 interface way. Leaving 4 hardware PWM outputs directly powering the 4 heating channels by IRLZ44N logic level Mosfets. The fifth Mosfet output is an on/off power output for cooling blowers or fans.

 

 

No rocket science here, all parts can be assembled on a single layer Euro format (100x160 mm) PCB and no SMD’s are used. Simple and cheap to build.

 

Robert.

 

[rweust]

A few guys in Germany are currently building this controller. I wrote a manual for them.

 

https://www.dropbox.com/s/mtfuew6pz06q72b/Manual%20PID%20Dew%20Controller.pdf?dl=0

 

 

 

 

Robert.

 

[rweust]

And of course a building instruction. I programmed the microcontroller for them and etched a PCB

 

https://www.dropbox.com/s/upsocthmrl5gxal/Building%204-Chanal%20PID%20Dew%20Controller.pdf?dl=0

 

 

 

 

Robert.

 

[Gina]

WOW   Proper job as we say   Great.

[Astroboffin]

Superb, thanks for sharing

AB

[Physopto]

Nice !

Great project for some spare time, (when it is cloudy or raining, which is most of the time these days!)

Derek

[rweust]

I might be able to help there, getting you through these rainy days. I spent a Saturday before making some home-etched (but nicely tinned, see earlier picture) PCB’s for my fellow Dutchman and German neighbors, who are all building the temperature controller right now. I now have only one PCB left from this series.

I provided them all with a pre-programmed PIC 16F1847 microcontroller with the (unique) software addresses of the six DS18S20 temperature sensors (also included in the package) already in. When addressing the sensors in the software, it is little extra work to calibrate them too. I completed the kit for them with all the components on the PCB, including LCD display, push buttons, heat-sinks and connectors.

BTW, I’m not doing this professional, nor for the money. I really like the interest in this dew controller and also like helping the less experienced with microcontrollers. I do take responsibility for the design though, if anything is wrong, it will of course be solved.

Robert.

 

[barkis]

Robert. I have edited your proposal to sell  PCB's   The forum, other than the  Classifieds, must not be used as a vehicle for selling.
I do like the project above, and it will glean a lot of interest I'm sure.

I wish you well with it.

Barkis.
Mod Team.

[rweust]

I don’t want to make the impression I’m a seller. If I did that, I’m very sorry. With this project in Holland and Germany there are a few problems to be solved, which I did the following way:

 

1^th.

Not everybody is able to program a microcontroller. With this 1-wire protocol the unique addresses of the sensors have to be implemented in the software. Furthermore with these accurate temperature measurements the sensors each have to be calibrated, also done in the software. It is not possible to simply load the software and run, but it has to be adapted to the sensors. This is why I helped people with this, by providing the programmed microcontroller and associated sensors. I can’t do it totally for free, but don’t want to make any money either.

2^nd.

It is a one layer PCB, easily made, but again not something everybody can do at home. I made an amount now for others, but setting up etching equipment and chemicals for a few PCB’s does cost something also. Therefore I suggested to let them produce in Asia in a small series of 10. I offered to organize this, but do need some interest of course

3^rd.

The design is free, as are the PCB Gerber files and is the source code software (Proton Basic). All available for non-commercial use on request. I’m willing to help everybody, but not everybody is experienced with electronic components. To prevent problems with enthusiast beginners I assembled a small kit.

 

I started this from a  hobby and it has to stay this way. Hope to have cleared that I have absolutely no commercial interest in this.

 

Robert.

[rweust]

Others building this dew heater controller sent me pictures.

Robert.