Senior Member
Brass is a little firmer and threads hold up in it better. I am building some small but stout 9mm aluminum mounts with a little X/Y diode shift and aluminum screws (yes, lube mandatory to avoid galling) and I mentioned to Dave about commissioning some premium versions of his existing mounts in copper.
Senior Member
Ok, going into the right way......I'm waiting for two pieces....I'm very exited !!Originally Posted by dnar
Senior Member
My past solution to aluminium and threads has been to use Helicoils from the get go.... Not sure how they fair from a thermal transfer perspective, thats never been a requirement in the past. My grey matter tells me there should be no issue.
This space for rent.
Senior Member
Sorry, still got to do some jaywalking in this thread.. It's either that or spawn a new one and that might be worse, and if it comes up then this off-topic bit might be good practise for a clearer discussion later.
Re brass, I'm ok with it. Can't solder aluminium. Copper would be better but paying for it probably wouldn't be.The amount of conduction of heat spread through the brass mount soldered upside down is viable for cooling. The bottleneck with that relatively thin and wide bit of intervening brass is the diode clamp itself. While I have yet to solder one of these mounts, I have soldered similar scales of parts to TEC's this way, of copper and brass. Robin Bowden taught me a bit about this, I adapted his trick for soldering diode bar arrays. Plenty of flux applied to clean abraded surfaces. Doesn't have to be finely lapped, but freshly abraded clean and flat will do it. followed immediately by wiping onto it a protective layer of liquid resin flux. This is sticky and will help with initial retention of parts. Heat the substrate to within 10°C below melting point of indium based solder designed to melt at 117°C, being standard for copperclad solderable TEC's also. (A cheap block heater off eBay is what I use, as it has good thermal mass so does not veer wildly with changes in its crude thermostat.) Then heat the important bit with a 50W temperature controlled iron set to a bit higher than melting point. Perhaps 10°C more. A bit hotter or more powerful may be needed, and timing is the fine gauge needed to do it right. The solder will tin the surfaces, drawn in by capillary action, and will float the mount. So tweaking position to within the tolerances of any fine-adjustment screws is crucial, then let it set. Arctic Silver epoxy might work ok, but it won't be as good an interface, and this IS crucial, second most important to the diode clamp, and it won't allow easy re-use, repair, adjustment.. My designs will be aimed at simple dual diode arrays for this, I don't think I'd want to try it for a quad. Too many bits to maintain at once. Might be doable with a cardboard template though, with cutouts for the mount positions, and that can be lifted or torn off carefully after the solder sets.
If you look inside a Coherent C315M laser (known for high performance, and I think Sam Goldwasser likes them a great deal), you'll see that most of their optics are soldered in place on a metallised ceramic substrate. I decided if it's good enough for them, it's good enough for me.
The issue of diode rim contact is a serious one. A quick and moderately cheap way might be to make the back plate out of copper, then get heat out of that somehow. (Forced air? Clumsy thinking, but better than none.) The adjustable mounts decrease this small contact but not as much as it may appear, UNLESS you intend to solder, as I may do. In this case it compromises area too much. Given that the diode must be moved relative to the lens, why not fix the diode, and move the lens?! Of course that demands a very big change in design. I've considered ways to adapt the fixed-centering mounts in a partial answer, but not come up with anything I'd stake physical work on yet. Some of my original schemes might even work better if revisited in the right way but last time they involved tubing and some very nasty 'tolerances' that cut more slack than Bob Dobbs.
Last edited by The_Doctor; 11-21-2013 at 11:26.
Senior Member
I have "taped and screwed" a lot of aluminum. Only aluminum/aluminum presents any real risk and then only if removing the screw entirely. This is when there is a chance to introduce dirt or to jack the threads off a bit and start the galling spiral of death. A small dab of Never Seize or the equivalent and no problem for adjusting etc. To allow high clamping pressures the thread depth should be more like 4 times the diameter rather than the common minimum of 2.5 times. Personally, I would avoid the complexity of soldering or exotic metal interfaces. The mount Dave sells is fine and if out of copper I do believe this is the end of the issue. The mount I am fabricating is straightforward and anyone could lay lay it out in 10 minuets. The rule is broad, short and aluminum wherever possible.
Senior Member
You already do use an exotic metal interface.Indium foil, clamped, counts. Anyway, I have no machine capable of large threads in aluminium but I remember using a lathe on it when I was 16. It's not hard to get good results with, at all. But now, I just use cold formed threads for M3 and M4 in aluminium. Stainless tap, then stainless bolts to fit the holes later. Great results.. My rule is: spread the heat. So given that the top of the mount is the easiest place to get it from, and IS fairly thin but spreads out fast, it matches your short-and-broad point. After that it's down to the interface because that's the main thing stopping it. I learned enough from TEC design guides from Marlow and Melcor (and my own experience with that gap filler and the 1U ITX cooling problem I mentioned earlier), that almost any metal will conduct better than almost any nonmetal interface. Soldering is actually easy. It was never my big problem. Far harder was trying to improvise cheap adjustable mounts using engineering modelling tube. That stuff seems to fit each N/32'' above or below the one next to it until you need short lengths, then it gets clumsy and stupid very fast.
I have even wondered if the CD player method might be viable! Not fast enough to go all the way to ideal laserists purposes, but imagine a modulatable focusser that allows realtime modification from fine line to gradient fill. I've never heard of this being done in a fast scanned pattern, and maybe there is a better way, but if not a slow one may be better than none. I have no plans to make such a thing though. Way too much like hard work..
EDIT:
I read somewhere that when using the same metal, the thread length never needs to be more than roughly the same as the diameter. That sounds too small, but if you look at most full nuts that is what you get. To prove it works, overtighten any size you care to try. The bolt will snap before the thread breaks in the nut. The real worth of long threads is in avoiding pitch and yaw (and using stainless steel hardware on aluminium substrates, where the strength does matter). This is one reason I like Dave's mounts, because that fixes a nasty problem I had with mine, which were far too short for accurate control. Whatever I do I have found that making mounts too short is a way to degrade accuracy severely, and the only way to avoid that while making them short is to make them wide! One way or another there needs to be a balance of constraint and leverage in adjustments. One way to displace this problem into a smaller overall dimension (is not a TARDIS.. ha....) is to use tiny screws. Very tiny screws. I only have part of a useful answer with those so I won't post any more here now. Too much not solved properly.
Last edited by The_Doctor; 11-21-2013 at 14:37.
Junior Member
I have a big stock of driver brand on sale here
Senior Member
Doctor,
The reason I question your focus on soldering is not because it is a poor or even a difficult method. I'm just NOT SURE if it is the right choke point to justify your efforts. Here is a thought. Set up say three brass diode mounts, dry, on the same base plate and drive some 9mm diodes to the same current. I would go for at least 2 A each to cause some serious heat flow. Then use a bead type thermistor and measure various points on each mount as well as the base plate. Finally, introduce changes in the mounts ie grease, indium film, solder, loose vs tight clamping pressure, SS vs aluminum screws and copper and aluminum substitutions for the retention plates behind the diode.
There is so much that can be learned from this experiment and it could allow others to improve their set up. Cooler diodes mean more power.
I think focus stability will be less of an issue if the mounts have more thermal stability. However, in all of my low temperature experiments I have been surprised how little if any focus shift there has been even for temperature shifts of hundreds of degrees C. I see more problems with the plasticity of the knife edge steering mirror mounts. These often need several re-visits to remove ever decreasing pointing errors over several days. Go figure!
Senior Member
The only diodes I have with 9mm size are old Philips OF4944, 5mW.
When I read the Melcor TEC guide, there was a lot of detail on clamping, thermal transfer, etc, parasitic heat loads. But the one thing they emphatically said was that soldered TEC's were by far the easiest and best way to go, if at all possible. I just took their word for it.
I agree with you that focus, using glass lenses in Dave's mounts is extremely stable. (After using acrylic aspheres this is luxury). I've just been taking pictures in a further test of DTR's LFL lens for that thread I started. I used the small TEC that dnar sent me. Over 25°C dropped and it looks like no significant change at all, in either lens or diode! This is one thing that might make me rethink the whole TEC scheme, but it's early yet, I only tested with modest constant current to the diode so far.
I've not yet begun to plan mirror mounts, but I know I'll have to because they'd be on the TEC too. I now have 19 40mm TEC's here so I have enough brute cooling to try this with eventually. But not -196°C!
Senior Member
Man, 19 x 40mm TEC's is going to create more problems than it solves... I am thinking about the huge temperature differentials and remembering a project a few years back, with 32 x CPU's on a single board, process balancing was needed because the PCB warped when a bunch of CPUs at one end ran hard and the other ran idle...
I have said it before, for me TEC's are a last resort. Their hot side has all the energy pumped from the cold side PLUS the electrical energy entering the system... Unless you want to achieve stable temperatures below the ambient conditions, I always avoid them. Stable temperatures around ambient are easy to achieve with conduction, convection and even phase change systems. Say, I am yet to see anyone here use a computer water cooling system in a projector, why not? You can easily partition the chassis to house the radiator, fans and reservoir tank....
This space for rent.
Posting Permissions
Reply With Quote