Reducing temperature differences and dew in an open system

[rweust]

On a Dutch astronomy forum a discussion started  about the colour of telescope tubes. The question was if the modern telescope are more and more coloured shiny white for pure esthetical reasons or if the limitation of radiation had something to do with this. It came out that the restriction of heat loss had nothing to do with it, as most varnish and paints have a rather high emission coefficients of 0.90 till 0.96 on which colour has far less influence than might have been expected. There are low emission coatings, but they are special and expensive and not expected to be used on commercial telescopes.

_

Most of us are familiar with dew on our tubes due to heat loss by radiation to the cool, open, nightly sky. The material doesn’t really influence this radiation, as the tubes are all coated. In my case the black carbon fibre tube of my 10” Newton is always the first to moist up on the surface, even before the metal parts.

_

The calculations I made before (using air temperature and humidity to calculate the dew point) showed that the surface of the tube must be over 2 °C colder than the surrounding air temperature when condensation on the tube starts. The wall thickness is only about 2.5 mm, so the inside of the tube should be the same low temperature. The main mirror doesn’t dew up easily, so the temperature difference between tube and main mirror is 2 °C, probably even a degree or so more due to the warmer main mirror. I hate the wet surface of the telescope from a handling point of view and don’t like the possibility of tube currents occurring, so what to do?
_

With a closed system isolation can reduce the heat loss by radiation to the cool, open, nightly sky. Most widely used is insulating radiator foil. The shiny aluminium surface prevents heat loss by radiation and the foam inlayer prevents heat loss by convection. The inside of the insulated, closed system will stay behind on a higher temperature when the surrounding air cools down, but this doesn’t need to cause many problems, if the parts inside (optics, tube, etc.) differ little in temperature. But what can you do with an open system like my Newton?

_

To solve this I started by making a simple model in Excel showing one tube half directed to the sky (calculations still in Dutch, I’m sorry) and calculate the heat loss by radiation and the heat transfer by convection. Only calculated on the outside surface of the tube to get an idea about what is happening.

_

_

The following graphic shows the situation with my carbon tube cooling down by radiation. The blue line is the steadily dropping air temperature. The red line is the dew point with a humidity of 75% at 9 °C starting temperature.  The green line is the surface temperature of the tube (with an emission coefficient of 0.90). First the tube cooling down from heat loss by radiation and convection, and even cooling down below the temperature of the surrounding air due to radiation to the cold nightly sky, before a balance with the heat transfer by convection with the surrounding air occurs.

_

This shows the tube cooling down a bit more than 2 °C under the temperature of the surrounding air, representing the reality in my back yard. The tube surface will reach the dew point in a little more than 2 hours and will then become moist of dew. Sadly enough, also representing the reality in my back yard. In 5 hours the surrounding air reaches the dew point under these fast cooling circumstances causing fog or mist shortly after.
_

Now a reasonable model is set up, I can judge the influence of insulation by radiator foil and foam (heat transfer by radiation and convection reduced) . The tube will reach the temperature of the surrounding air far better and won’t dew up, but does take longer to cool down. It now cools down in about an hour. Probably quicker in reality, as in this model the heat transfer from the inside is not accounted for and the Newton tube is of course open on the front side.

_

_

But what will happen if  only the heat loss by radiation is reduced by a factor 10 (emission coefficient of the surface 0.09) and we leave the heat transfer by convection as it is?

_

This results in a tube not ending up much under the temperature of the surrounding air (ca. 0.2 °C), but still cooling down quickly. Dew on the tube surface won’t occur, but shortly before the surrounding air temperature reaches the dew point anyway, and large temperature differences between optics inside and inside wall of the tube are prevented.

_

What I was looking for is some kind of low-emission coating or foil to reduce the radiation on the outside of our tube. Aluminium foil has a low emission coefficient of 0.1 and it is cheap, but isn’t easy to apply . Not to mention the strong mirroring surface and disturbing light reflections from any surrounding objects at night. With a Newton telescope my face is close to the tube during observations, I really like a black tube and would hate to part from the carbon look.

I found a low-emission foil that is used on windows to insulate buildings. This Low-E window foil has an emission coefficient of 0.09, is almost totally clear and has a self sticking layer. It can be applied far more easily and even be removed without leaving traces on the surface. I used the Solar Gard Ecolux 70 foil  from Saint-Gobain, but 3M does provide something similar. The costs are about EUR 60,- /m2.

_

_

My carbon fibre tube regretfully isn’t a perfect cylinder, its reinforced around the focuser and at the back around the main mirror. The foil is a very wear resistant polyester (probably PET) and has to be cut in on some places to follow the tube. With an aluminium tube this problem would be far less. But still it’s sticking on nicely and with the use of some water and (dish cleaning) soap can be corrected where needed for some time.

_

_

In reality this foil proves to make a big difference in temperature and dew problems. I even made a PID dew controller to control the temperature of my secondary mirror accurately. This also gave me the possibility to measure temperatures of the main mirror and tube insides accurate by contact sensors. All resulting in an open system in which all temperatures end up within a 0.5 °C window relative to the temperature of the surrounding air. I will post these results later on.

_

Robert.

[ChrisLX200]

Keep us informed on how well the mod works!

ChrisH

[rweust]

Thanks Chris for the kind reply,

To control the temperature of my secondary mirror I modified my 10” carbon tube GSO Newton with a heating spiral of NiCr wire, as have many done before me.

_

_

I reduced the heating power with the use of a home build 2-chanel PWM power controller based on the  famous NE555 integrated circuit, as also many have done before me.

_

_

This was of course working against dew on the secondary mirror, but I always had the feeling I missed something. I never knew the exact temperatures, never knew when to start the heating (sometimes to late) and couldn’t be sure if I wasn’t heating far too much and influenced the seeing negatively.

_

To make a long story shorter. I decided to build another dew controller, but now one I can measure the temperatures with, so sensors and a LCD display are wished for. This automatically means the use of a microprocessor. If we use a microprocessor anyway and have load sensors measuring the temperature of each channel,  the temperature control can be a lot more accurate with a far more stable power output  by the use of a PID temperature control instead of thermostats or differential temperature control by opamps.

_

_

This kind of dew controller can of course be build with the use of an Arduino board, but I choose to use a PIC 16F1847 microcontroller. It’s extremely cheap and can handle 4 PWM outputs directly in its hardware. For the six temperature sensors I used the DS1820, with a 9-bit digital temperature data transfer they normally are not so accurate (0.5 °C), but by use of the also readable slope and remain bytes, this accuracy can easily being improved to 0.1 °C. They are, as digital temperature sensors, more linear than analogue sensors, not that expensive, easy to build in and can all be wired parallel on one input or cable (Dallas one-wire protocol)

_

_

Dew controller might not even be the right name for this device, as you can control four temperature zones in an accurate way on the telescope and read all relevant temperatures out on the display, including surroundings temperature and main mirror or tube. In this way keeping the temperatures between the different parts of the telescope as low in difference as possible, which is of course more than just preventing dew.

_

_

I will probably start another topic on this dew controller, if there is enough interest. In this topic I used it to measure the temperatures of the surrounding air, main mirror, secondary mirror and inside of the tube (bottom side and top) simultaneously and constantly by using the DS1820 as contact temperature sensors.

_

For more interested, I provide a link to the specifications and manual:

_

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

_

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

_

Measuring results will follow.

_

Robert.

[Gina]

Very good   BTW the DS18B20 can be set to 12bit resolutuion giving 0.0625°C.

[rweust]

Praxis test with the low-emission window foil around the tube.

_

_

Sensors S1 is assembled in the secondary mirror holder, with direct contact to the glass and fixated from above with (insulating) foam tape and a polystyrene cover.

_

_

The main mirror sensor S5 is also assembled with direct glass contact, fixated/isolated with foam tape and a polystyrene bend plate.

_

_

Inside the tube I temporarily taped two sensors to measure the wall temperature inside the tube, both about half way inside. One sensor S3 direct to the top side of the tube, the side radiating fully to the open sky, and one sensor S4 directed more to the bottom.

_

_

The surrounding air temperature is measured by sensor S0 on top of the case of the dew controller. This sensor is inside a gold plated Cinch socket, hopefully preventing to much heat loss by radiation (gold has a very low emission coefficient) and measuring the right air temperature.

_

_

The telescope was placed outside at 19.15 hours from a room with a 20.7 °C temperature. At  19.35 the dew controller was connected, the main mirror was already cooled down till 12.4 °C, the temperature outside was 5.6 °C.

_

At 19.45 the cool down cycle started on this telescope. The blower on the back of the main mirror is powered by the dew controller. I wrote down the temperatures every 10 minutes and let this run for 1,5 hour. This resulted in the following graphic:

_

_

We can conclude a few things from this measurements:

_

* The measured outside temperature is about 0.3 °C lower as the real air temperature. The gold plated layer on the Cinch socket works for outside temperature sensor S0, but not fully against heat loss by radiation to the open sky. The surroundings temperature (dark blue line) has to be corrected, in fact it is 0.3 °C higher than in this graphic.

_

* During the cooling down cycle the dew controller shows the differential temperature DT on display between air outside and main mirror (yellow line). After about 1 hour this differential temperature doesn’t decrease anymore (stays 1.3 °C in the graphic). After 0.3 °C correction a real DT of 1.0 °C remains. The heat loss by this cooling is in balance with the heat the blower generates. This GSO blower proved not to be the most efficient, as it is assembled far too close to the main mirror.

_

* After 1 hour of cooling down the parts of the telescope have the following temperatures related to the (corrected by 0.3 °C) surroundings temperature:

        Main mirror:             + 1.0 °C

        Secondary mirror:    + 0.0 °C

        Tube:                          + 0.1 °C

_

* As there is no difference between the inside wall temperature on the top of the tube compared to the bottom side and the difference with the outside temperature is low (even after one hour), the low emission foil seems to work against radiation. There is still some forced cooling present, how inefficient this may be.

_

At 21.05 I experienced the first dew op the (black ABS plastic) casing of my dew controller itself, off all possibilities I didn’t expect this one, but no dew on my tube anymore. Without this low-E foil the tube would have been the first part to dew up in my experience. Another indication this foil works. Time to stop the cooling down cycle and start with observing. The following results:

_

_

At 21.50 the parts of the telescope have the following temperatures related to the (corrected by 0.3 °C) surroundings temperature:

        Main mirror:             + 0.6 °C

        Secondary mirror:    - 0.5 °C

        Tube:                          - 0.1 °C
_

The secondary mirror is cooling down to fast due to heat loss by radiation and nearing the dew point dangerously close. So at 22.05 the heating of the secondary mirror is switched on. Setting it to follow the measured outside temperature with an extra offset of + 0.3 °C. The PID dew controller will keep it exactly on the real temperature of the surrounding air this way. At 22.50 the outside temperature lowered me too much and I increased the offset a bit on the dew controller to +0.4 °C for the temperature of secondary mirror.

_

At 22.50 the parts of the telescope have the following temperatures related to the (corrected by 0.3 °C) surroundings temperature:

        Main mirror:             + 0.5 °C

        Secondary mirror:    + 0.2 °C

        Tube:                          - 0.1 °C

_

Really nice temperature differences for such an open system. The low emission foil really works well together with this PID heater controller on the secondary mirror. And temperature differences become even less at 23.15 they are, related to the surroundings temperature:

        Main mirror:             + 0.3 °C

        Secondary mirror:    + 0.1 °C

        Tube:                          - 0.2 °C

_

All temperature differences of the measured parts are now within a 0.5 °C window related to the surrounding air temperature. The low-E foil is doing its job perfectly preventing the tube from the 2 °C temperature drop it had before due to heat loss by radiation. The 0.2 °C temperature drop measured is exactly the value as calculated before in this topic.

_

A nice format from this dew controller is showing the outside temperature and  main mirror temperature in one screen. It’s very easy to determine the moment to give the main mirror an extra shielding hat to keep warm enough (Astro Zap dust cover, see picture below). I did this at 23.30 to keep the temperature of the main mirror a bit above the temperature of the surrounding air and preventing it from dewing up.

_

_

With correction of the surroundings temperature (+0,3 °C in reality) the temperatures of the telescope parts are as below (in Dutch, sorry, but the colors of the lines are as before) . With the heating of the secondary mirror (vangspiegel in Dutch) switched on at 22.05:

_

_

No wet tube to handle anymore, wonderful low temperature differences and visually I had absolutely no complaints this night, the seeing was rather good. The telescope performed perfectly. There was no dew on the tube at all till 1.00 at night, then the tube started to dew up slightly, long after the metal parts of mount and telescope moist up and shortly before the seeing was reduced by low cloud forming and fog anyway at 1.30 that night (too cold for this high humidity).

_

Robert.

[rweust]

On 19-1-2016 at 10:50, Gina said:

Very good   BTW the DS18B20 can be set to 12bit resolutuion giving 0.0625°C.

The DS18B20 can be used on this controller to Gina, with a small software adaption I have them also working.

Did you know that the DS18S20 sends back a 9-bit temperature value, but also a remain byte and a slope byte, which you can use to calculate the temperature almost a decimal more accurate? Providig the DS18S20 with the same accuracy of 0.0625 °C as the 12-bit temperature value DS18B20 sensor. This controller wil work with an 0.1 °C accuracy with the 9-bit DS1820, equivalent DS18S20 and (with a small software adaption) the 12-bit DS18B20 temperature sensors.

Robert.