The mirror has finally arrived.

I created the thread "12.5 inch Truss Tube Dobsonian" , for the purposes of documenting the construction of my new telescope. However, there is one part of the design that must be finalized before construction can begin: the dimensions and layout of the mirror box. One factor driving the layout of the mirror box is how, exactly, I should cool the mirror.

There is a whole lot of documentation and arguments concerning how to properly and efficiently cool the primary mirror. Most people on this forum will know what I'm talking about, but for those of you readers who don't understand why, it has to do with air turbulence over the primary mirror, due to the difference in temperature between the mirror and the air directly above it. Immediately after the sun goes down, the air gets cool. Throughout the night the temperature continues dropping. A large mirror has a decent amount of mass to prevent it from reaching ambient air temperature quickly. Without this ability, objects will blur, suffering from the "wavey" look you see across hot desert highways during the heat of the day.

I decided to start this thread to document this testing separately, since I believe it's going to be rather rigorous. Once this thread (and tests) are complete, I can finish the design and get back to building.....

First of all, I have in mind, after much reading and discussions, my own idea of how the mirror box should be shaped. The dimensions will adhere to the following requirements:

1) The box must be large enough to house the primary mirror, the cooling fans, and the upper truss assembly (UTA). I fully intend to carry the UTA inside the box, making the system more compact than the average model I'm used to seeing.

2) The fans used will cool the bottom and the top (face) of the mirror simultaneously. This will remove the "boundary layer" of air, which is otherwise difficult to get rid of. For this purpose, two fans, side by side, will be positioned such that approximately 1/3 of the upper portion of the fans will blow across the face, and the remaining 2/3 of the fans' areas will be aimed at the bottom portion of the mirror. The fans will blow from the side - NOT the BOTTOM! The bottom of the box will be sealed, except for the protruding collimation bolts for mirror alignment. Therefore, the box must allow these fans to be mounted on the inside, facing the mirror.

3) Two additional fans will be mounted on the opposite side of the box to blow the air back out. Despite the fact that this is an open truss system, and the top of the mirror box will be open, I have read enough horror stories of people trying to get the heat out of the mirror box without exit fans, and I am sold that this is a requirement, due to turbulence and eddies, trapping warm air inside the box, while cool air blows right on through. Most of the news related to entrance and exit fans together have proven the pair as a more reliable system.

4) The mirror box will still remain the smallest dimensions possible to allow carrying all these items, minus the side bearings, the side bearing support, the azimuth bearing (or foot), & truss tubes.

In the past, people have used 2" thick mirrors. These mirrors have usually been cooled directly from the bottom. But these conventional mirrors are now being replaced by thinner, sturdier mirrors, called conical mirrors. Conical mirrors have thinner sides (about 1/2"), and thicker middles (1 3/4"). The mirror is thinner all the way around, than the conventional types. Furthermore, a special "tailgate" for the mirror is no longer required. Not only that, but these mirrors can be easily cooled from the side (or so I believe). I will explain in detail, the mirror cooling technique, with many pictures, as the test continues.

The test requires the following:

* A model (actual size) of the mirror box will be constructed out of cheap wood or even cardboard. The actual fans will be mounted in the box over the positions required. The bottom and sides of the box will otherwise be sealed with box tape, so that it is hermetically sealed (air-tight).

* The model will be placed on a table in the middle of a temperature controlled room.

* A backplate will be attached to the mirror, allowing handling of it without touching its face. This will also allow safer heating of the mirror, since the backplate will be closest to the heating coils, rather than the mirror's surface.

* A thermocouple will be attached to the center of the face of the mirror (where the collimation dot will later be placed). The manufacturer's tissue will cover the mirror face, then 3 1/2" of fiberglass insulation will be used to cover the face of the mirror, insulating it from ambient air.

* The mirror will be placed in a 110 degree oven and allowed to be saturated in the heat until its face reaches the oven's temperature.

* Once equilibrium is well-established, the mirror will be removed as a stopwatch is set. The mirror will be placed in the box with the fans running. The time and temperature will be tracked until ambient temperature is reached.

* This process will be completed three times for an average of the results.

* A second set of three tests will be repeated once a heat sink is placed between the backplate and the rear of the primary mirror.

This will in essence, tell me if, in fact, the heat sink helps at all, and whether I should include it in the design. If there is no speed up in cooling performance, I will discard the heatsink and build the mirror box without it, about an inch shorter than otherwise needed.

In the first two pictures, the shape of the mirror becomes clear. The first picture is over the face of the mirror. The back plate is resting underneath the mirror in its proper position, although it is presently unattached. The second picture reveals the profile of the mirror. It is from this point of view that the cooling fans will be directed toward the mirror. On the opposite side, two more fans will be pulling air away from the mirror. The testing will be performed inside a box with the fans installed, so as to mimic the environment of the finished telescope.

After locating and marking the center of the mirror, a thermocouple device is attached to the mirror face. Note that the center of the mirror is where the mirror is (very near its) bulkiest. I have assumed that when this point in the mirror reaches equilibrium (when cooled strictly from the underside), the entire mirror will also be at equilibrium.

The manufacturer's tissue was placed back over the face of the mirror, with the thermocouple cable punched through and secured.

Turning the mirror once again on its face, one can see the amount of area of the back side that is exposed to the air. There is tape and tissue covering some of this back surface. One might argue that it is hardly comparable to the efficiency of the rate at which a fully exposed mirror will cool. Nevertheless, this is how the mirror will be tested. The mirror will most certainly cool faster when it is exposed. But how do we know when the interior of the bulk of the mirror has reached equilibrium? We can't know without an embedded device.

This method is designed to cool the mirror from the bottom side up, and THROUGH its bulk is the only way the face will be cooled. Therefore, we capture how long it takes for the temperature change to traverse through the glass, from bottom to top. A piece of cardboard is secured over the face and the back plate is now attached.

The primary is almost ready for testing. But it would help to take an extra measure to ensure the face will not cool except from underneath. So we add a layer of fiberglass insulation over the face.

TEST SETUP

For the first test, I wanted to test the speed of heat dissipation under constant fan flow in the box in such a way as to be a fair test. In my mind, cinching the solid wood backplate against the back of the mirror would not allow adequate airflow through the underside of the mirror. I therefore propped two cardboard spacers, approximately the same thickness as the heatsink, on either side and parallel to the air flow. See the first picture . I also cut two wedges from foam and placed them approximately 45 degrees from the air flow front, to help scoop the air under the bulk of the mirror.

The mirror box model was built from the box the mirror actually arrived in. This means, of course, nothing can go wrong with the mirror now, since I will never, ever be able to send it back!! I took too long to find the 3” fans I really wanted, so I wound up using two larger 4” fans I had on-hand – one in the front, one in the rear. The fans were placed in the optimal position for funneling air flow through the mirror and out of the box (although the inside of the box is not necessarily aerodynamically designed for efficient entrance/exit of air).

The tub! ! Here, you can see my sophisticated heating device – a space heater in a bathtub. [I earlier tested my kitchen oven for this task and discovered the temperature rising to about 260 degrees on the lowest (supposedly 150 degrees) setting. Furthermore, the temperature in the oven varied wildly - so I opted for this alternative!] I placed the heater on low to slowly heat the mirror and kept it about 6 inches from the inlet fan. I set its thermostat to max to ensure it never turned off. I turned on the bathroom vent fan to prevent the bathroom from heating up too quickly. During heating of the mirror , I ran a digital Cole Parmer thermocouple thermostat, monitoring the progress of the mirror face. I used an alternate (two-soldered wire type) thermocouple for measuring the inlet air temperature (not shown). Once the mirror reached a few degrees shy of the inlet temperature, I turned off the heater, set the timer, and quickly moved the box into its “cool room” environment.

The cool room was my living room. A ceiling fan blew overhead on high, and other fans were placed strategically throughout the house to circulate the air. Every five minutes, the temperature of the mirror was measured and recorded. The temperature of the room was also tracked.

TEST 1

My initial goal was to perform each test at least 3 times, proving its repeatability and averaging the results. As it turns out, when one heats a mirror much more than they probably should, it takes a very long time to cool it down, especially when at least half of it is insulated! After discovering that one test requires almost 6 hours (under the conditions I created), I decided two iterations was MORE than enough testing, as long as the results agree within a decent confidence level - And for me, that requirement was met. Thus two tests were run and documented. See Graph 1 and Graph 2.

TEST 2

Preparing for test two, I drilled a 5/8" hole in the heatsink (to allow for the ½” mirror bolt). The heatsink, thermal grease, and flat mirror surface can been seen in this picture . The backplate was removed and thermal grease applied . The heatsink was then pressed into to the thermal grease with firm pressure, squeezing out air voids. The backplate was replaced and the foam wedges returned to position to scoop air into the heatsink. The test was performed two more times, using the same format as the previous tests. See Graph 3 and Graph 4 for those results.

SUMMARY

No Heat Sink Present:
The average time required to drop the mirror temperature from 55 degrees above ambient temperature to ambient temperature without a heatsink was about 207 minutes; that’s 3 hours, 27 minutes.

Heat Sink Present:
The second two tests show a greater deviation. If I were to be fair, the first test would probably not be as accurate as the second test, due to interference. Activity (cooking) in the kitchen and traffic through the house likely injected erratic behavior in the air conditioning. The second test, however, was performed accurately, and I am confident the results are truthful. Therefore, foregoing the info on the first test and concentrating on the second test results, one can see that the gain in performance is at best, marginally higher.

Let’s make the most conservative comparison first:
The fastest results with no heatsink occurred during “Test 1: No Heatsink”
* 53 degree delta; 205 minutes
The slowest results with a heatsink occurred during “Test 2: With Heatsink”
* 58 degree delta; 195 minutes

Looking closely over the data, I made the following assessment:

At the greatest temperature difference, the no heatsink test showed a drop of about 4 degrees per interval, with some slight degree of variance. The with heatsink test showed a solid 5 degree drop per interval. Near the end of the tests, where the temperature difference was small, there was almost no noticeable difference between the tests.

Between the no heatsink tests, the difference between the initial temperature starting points is, in fact, 4 degrees. Conveniently, if one were to theoretically bump the first test up to 134 degrees as its starting temperature point, at the drop rate of 4 degrees per interval, one would surmise that at exactly one interval more, both tests would conclude at the same time.

The most fair comparison between the tests, based upon the data collected, would therefore be:
* 58 degree delta; 210 minutes, no heatsink
* 58 degree delta; 195 minutes, with heatsink

The difference in dissipation times is therefore, 7% better with a heatsink, for this particular test environment. For a smaller temperature change, the added performance would very likely be less.

All in all, the results are still a bit disheartening to me, as I was sure the difference would be much greater between the two methods. If I were to find a method of attaching heatsinks to the entire back surface of the mirror, the performance results would possibly be more dramatic.

. . . . . . . .

Faster cooling could possibly be obtained with active cooling systems, but adding that kind of stress to a $1500 mirror does not seem fit. The stress I put this mirror through, in fact, I would not recommend to other users, unless they had a cheaper mirror to play with. A 30 to 40 degree difference would have sufficed and kept the mirror in a safer stress zone, for example. I took a gamble for the sake of learning the truth. I will not do it again, on purpose! However, for anyone out there whose imagination I may have spurred – and who is willing to take on that risk – I am certainly interested in hearing about the ideas and/or any testing results.

CONCLUSION:

I have decided to keep the heatsink intact – not because of the marginal gain in dissipation performance, but because the presence of the heatsink allows me to conveniently use the cooling airflow method I desire: across the face as well as under its bulk. Without the heatsink, I’d have to find another method of keeping the backplate away from the mirror surface such that the air can easily pass across this bottom surface. Cardboard doesn’t quite cut it! And the heatsink is a very light aluminum, weighing in at 6.9 oz. Because the mirror is light weight (13 lbs), this should help me offset the weight of the eye pieces, later on.

And now construction may continue…. Stay tuned on the “12.5 inch truss tube dobsonian” thread!

I would had thought that the heat sink would have made a bigger difference also..................

Since you are useing the heat sink, you could incorperate a digital thermometer into you scope design, attached to the heat sink, with a digital read-out mounted on the mirror box....when turned on it would give a constent temperure of the mirror. Just a thought...

Hey, Leonard, good idea. With fans running across the heatsink, it will probably be cooler than the mirror, though. But maybe somewhere on the glass would be good.