More detail for the LFL lens from DTR:
A G71 diode is supplied by an LM317 based driver, with no modulation. Maximum current is set at just under 300 mA. Room temperature is about 20°C. Total power lost through the heatsink is slightly more than 3W with negligible temperature rise. The diode is oriented with pin 2 highest.
The beam reaches a mirror at 8.4 metres and returns to a white painted wall after a total length of 18.05 metres. Once focussed for best image at 18m, it stayed set. Acrylic wouldn't have... Fast axis width at 18m is 150mm, slow axis height is only 6mm. I don't know what the fringes under the main line are caused by, or why there are none above it, but they can be removed easily by trimming the lower tip of the vertical line at aperture.
At aperture, the beam's line is vertical, 5.5mm high. The shape appears straight by eye, but isn't. A camera shows much more detail when the beam is intercepted by a 0.1mm thick polyester diffuser as a screen. It's almost straight just above threshold current, but when driven strongly it is elliptical, and its width varies with power, about 1.5mm at about 150 mA, and about 2mm at 300mA.
DTR said this (in answer to a PM):
"It is longer than the G-lens in terms of the FL. They are not for single mode diodes. The great part about them is you get more of a square with diodes like the Mitsu reds and the 445nm diodes. The source I get them from does not have a datasheet so I have had to wing it but with good results."
While I could get a squarish beam by defocussing slightly, I wouldn't do it because the divergence is still wide in both axes. Given how fine the slow axis collimation seems to be, expansion and divergence reduction of the fast axis seems to be the best way. I don't know which would do this best, prisms or cylindrical lenses. It also looks like knife-edging is best for up to four G71 diodes, and not using a PBS unless combining whole clusters of diodes. Whether that is still true with other lenses I don't know, so far I only tried this one.
5.5mm beam width (from about 4.5mm focal length, shown in the earlier post) gets good easy results, but I will try a smaller focal length lens for a thinner beam, because it looks like slow axis divergence from a G71 will still be low. If beam correction can get the fast axis divergence anything like comparable with the slow one, the LFL lens might be very good for long distance beam shows from larger scan mirrors. Given how little light it wastes, it still collimates the slow axis very well. Most things I have read suggest that this is usually only possible by even longer focal length, which results in clipping the beam. The greatest risk of clipping with the LFL lens is in a lens retainer, so the method I used, of making a polyester film 0.1mm thick form three turns to line one of Dave's lens barrels is as good a way as any. It centres the 6mm diameter LFL lens accurately, and grips it securely enough, and allows easy refitting if necessary. And it doesn't interrupt the light. I think Dave's lens retainer has a 5.6mm bore so that might just pass the beam untouched, but adding it in this case won't help. A thin smear of glue inside the flange of the lens barrel would work better to bind the lens to the polyester and to the brass barrel. To do better, a very precise barrel would have to be made for this lens, and it might not fit any other lenses.
Reply With Quote
Originally Posted by olive13
For my eventual modules I aimed at 40mm and several tens of watts Qmax, on the assumption that low powering of a strong TEC might be the easiest way to get efficient and powerful temperature changes while running as yet uncertain heat loads on it. They haven't arrived yet, so I might yet trouble RS. They cost, but such is the price of no waiting at all. Next day, if I want it.. Can't fit a 40mm TEC between bolts with centres 34mm apart without some silly kludging, so I might do this with one of their 30mm TEC's. I'm sure I can mileage out of it in future somehow.
Got to be PT100 though. I know that small NTC thermistors are cheaper, but this is low volume, working by hand, so the difference saved would be many times lost in arranging for thermistors, probably.
