Questions on abuse - Rohm vs. Opnext

[marconi]

Hi PN,

Well, .. what you need is a way to turn the peltier off when it gets cool enough.. but I have used a Constant Current source before but as soon as the ambient temp changes lower ..so does the peltier.
But for just experimenting you could just use a resistor or a variable supply (like a LM338) and you will see.
But you will be relying on the Dt of the peltier , its a constant. if its say 60 degrees that means when its full on its cooling 60deg below the base temp.
If the base temp gets hot so will the device you cooling so you have a thermal run-away condition. Thats what it needs to be controlled.
If the peltier goes into run-away it can be damaged.

You could probably get away with a cheap fan control circuit that uses a thermistor and an op-amp and a relay...that may work , I googled a few that look easy to make.. but I dont like full-on full off. but if you used a separate supply , that might not make much difference.

The TEC I made was a derivative of one of these circuits.

Or, You could buy a Wavelength Electronics IC that'l do all of it for you. not cheap but it'l be right. for about 100USD.

[The_Doctor]

That switch rocks, looks like something for a classic movie bomb timer, the type with big showy red digits and such. Flip the cover, toggle the switch, and see the countdown begin. Lots of interesting ticks and beeps.

Secret spy helicopter missile launcher would be a close second place. Or perhaps its ejector seat. That switch IS inspiring. It's like the Big Red Button whose state is such that murders and mayhem might ensue if it is pressed, and murders and mayhem might ensue if it is NOT pressed. Such is the fate of devices like these

Re TEC's, if you don't need very critical control of temperture, efficiency becomes a greater priority, and it's easier to control. Many are designed to work that way. I remember seeing the voltages like 14.5V and 7.8V and a small expensive company telling me how exacting the requirements are. There are manuals explaining the efficiency curves, even, with reams of technical notes. What's not as often said is how easy it can be. If you want maximum Delta T you have to heatsink VERY well, and push at the rated V max and you get either that, under very small heat load, or maximum Qmax (pumping effect) under big heat load to maintain equal hot side and cold side temperatures. When you want to cool something to a range of acceptable temperature, it can be extremely easy, as the unusual 14.5V or 7.8V ratings are really just a way of saying that the TEC is rated to run on 12V or 6V. It really IS that simple. There is an optimum point of compromise for efficiency and Delta T, and those values are set so you don't have to think of the details at all, just put it a good heatsink and run it off a standard voltage. (Fridges and CPU TEC's are mostly based on this use.)

If you do the simple method of running one on a fixed standard voltage, you will pass a strong requirement to sealing the unit against water vapour, but that can be easier than arranging precise TEC control drive. Certainly cheaper. When you're after a basic wavelength shortening and simple reliability, I'd do it this way. As Marconi says, sealing IS critical, but the additional step of sealing a dry nitrogen atmosphere in there is not needed at all. Water migrates from warm to cold. It can sublime too, ice to vapour, directly. If you look in any freezer you'll see the cumulative effects of this kind of migration. That tells you two things: 1. Assembing the device in a chest freezer is a great way to get it really dry inside, just warm the parts before putting them in there. If you're making several, it pays to cut some polycarbonate to make a cover to the freezer, and make holes in it to form a glove box. I'm looking into this, but I don't know where the glove bits will come from yet. 2. Any water molecules in the laser will migrate out of the beam path because that's the warmest part of the system. So long as the amount is small, and does not increase, thermal dynamics in there will be helpful to you.


Edit: Following up Marconi's point on thermal runaway... If all you need is some deep cooling in a vapour-proof case, the laser diode can accept greater range than ambient change, so constant voltage or current drive is fine (both amount to same thing, near enough, in non critical temperature cooling), and the thermal runaway problem won't exist. You will still need a good heatsink though, to keep the hot side temperature down, so the cold side is pushed low enough to be useful. The main thing is that if you push the diode, then switch on the TEC, the diode will die within seconds. Switch on the TEC first, then run up the diode supply cautiously until you find the right balance. Give it time after each tweak, it could take many minutes to stabilise. It's still good engineering if you can make it repeatable, and establish your safe operating ranges for heat and power. Once you've established that, it probably won't matter which gets switched on first.

[Hemlock Mike]

PNJ - John --

It's fun reading your posts !! Glad to see you learning so much. Even I learn something from the reading.
Good work man -

Mike

[PNjunction]

Hi, Mike! Glad you're enjoying these posts, and that there's actually something to learn from discussion of my wacky visions. Thanks for the encouragement... I will definitely press on! Speaking of which, although I am learning things at a fairly furious rate, there are so many different subjects (including those not related to light, or electronics) I'm trying to learn simultaneously, that each one is taking a while. It's a very ADD approach to self-education.

So, yeah guys... I really opened up a can of worms with this TEC thing. It's gonna' make this project sooo much more complicated than it needs to be, and may not even work as well as intended. However, it just seems too cool (no pun intended) to not do it. The confidence (or ignorance... take your pick) to think it might work well, comes from that I've always been able to build anything that came to mind, without really knowing what I'm doing, and to date everything's been successful. This spirit comes from both grandfathers, however the difference being they actually knew what they were doing... lol. OK, enough about me... here's a couple details about the intended build so far.

Definitely gonna' use the Opnext. If it fries, I have a couple Rohm's as backup. The TEC will have it's own circuit and power supply, completely seperate from the LD and driver (which looks highly likely to be an EU-38-1... been talking to Roithner ). Moisture sealing will come from high temp silicone surrounding any protrusions from the box... heatsinks, switches, lens focusing ring, etc. A couple dessicant packs will be inside the box to absorb any water vapor present at time of assembly.

Marconi, and Doctor... you guys rock! Your assistance is definitely helping to steer me in the right direction. There are a lot more details to work out, and it's gonna' take a little while, but there will one day be, a very scary looking, portable death ray machine, with all kinds of mean looking things sticking out of it.

Totally ridiculous, but I'll be the only kid on the block to have one.

[The_Doctor]

I built one into a cast steel straight inspection coupling once, with a tiny red toggle switch, because it reminded me of a tiny hand held neutron bomb from a Doctor Who show from Tom Baker's time. This is not a cool confession to make. I haven't got it now, but it worked, it was based on a prebuilt module and a couple of lithium batteries. It never looked as cool as those tiny bombs, but it WAS bombproof.

Btw, if you can find a product called Tempflex from Loctite, get it, it's a low-outgassing silicone that should be ideal. It's what I'll be using if I can't find something that allows easy dissasembly.

[PNjunction]

it reminded me of a tiny hand held neutron bomb from a Doctor Who show from Tom Baker's time. This is not a cool confession to make.

Heh... maybe I should post a disclaimer?

This laser will not be used as a death ray, portable or otherwise. It will be used for educational purposes only, and not to take over the world. However, I shall not be held liable in the event the occasional flying insect "gets it".

That should cover it.

Doctor - The Tempflex sounds perfect for this in a couple of ways. I'll definitely check it out. BTW, where did you get that trapezoid shaped module mount that's attached to the TEC? It could potentially be a very nice piece to use in this project.

Now honestly, looking into the subject of wavelength vs. temperature last night had me wondering if a TEC is going to be worth the effort for this project. I think it still may be worthwhile, but the results will be slightly more marginal than originally hoped. Some research revealed the amount of cooling required to induce wavelength shifts in various diodes. Spotted info for violet and IR, but unfortunately, only one set of data for a red diode so far, and it's for 635nm at much lower output and current levels than we be talkin' 'bout.

Here it is. Improving fluorescence confocal microscopy with cryogenically-cooled diode lasers. Actually, the entire article is interesting in how diode cooling to a precise wavelength is used to optimize image resolution and saturation.

The temperature vs. wavelength curve for that reddie, shows roughly a .2nm shift per 1 degree change from rated specs at 25 C. Therefore, a 5nm shift would require a temp of approx. 0C. 10nm would be at approx. -25C, and so on. For lack of data on overdriven 658 reds, I'll use these measurements as a basic rule of thumb. IIRC, the T vs. shift curve is similar to other wavelength diodes.

So, I get the feeling that as it applies to the plastic box o' doom, getting down to 650nm from 658/660 is gonna' be a bit outside the design parameters... lol. But if 655 is possible in ambient, nighttime summer temps, then good enough.

Off to continue figuring out what TEC to use. Been looking at Ferrotec, Melcor, etc., but I have questions... anyone mind if I ask more questions?

[The_Doctor]

That's a rectangular block. Interesting that you mention a trapezoid though, I did try that too, in copper blocks I originally made from 3/8" thick bar. More trouble than it's worth though, to do the extra grinding. I make them out of brass now, but might change to copper again if I get stock for it. A bandsaw and a pillar drill and an emery board are all you need, but a small benched disk sander helps cut the work time down a lot. With a good choice of bits, and working the drill by hand to avoid inaccuracy caused by vibration, even cheap tools can get accuracies to with a thousandth of an inch. Modellers brass tubing, and a decent digital display caliper are vital. And practise, once you select your dimensions, you need to get repeatable tolerances so that even if work is slow, the wasteage is zero. If you acheive that, you'll be able to do it more cost effectively than buying stuff, and you get full control of design.

For wavelength shifts, you're right, it isn't dramatic, that's why you are better going with a brutally simple TEC drive with no feedback, just aim for an effiecient cooling using a fixed voltage/current supply, and keep the hot side well heatsinked, and keep the laser casing sealed. There is no other way better than that. We're not after wavelength control in this case, just a large shift. A small thermistor mounted on the diode block will give adjustment to the constant current drive, but even that isn't needed, just tune it to the cold temperature, and it will protect itself at high temperatures during warmup or cooling failure by being less efficient and making slightly less power at the output mirror.

[Hemlock Mike]

PNJ - John

Retarded questions ?? I've read this thread three times and keep learning !! I never ignore free education !! Next time I might have to pay for it

Mike

[PNjunction]

Doctor - Wow... that looks so trapezoidal in the pic! Is it just the way the pic came out to make it look that way? Just to make sure we're talking about the same part, I'm referring to the chunk of metal with the hole in it that's sitting on the cold side of the TEC, and looks like it has a set screw on top for holding the diode/lens module.

This mounting block from Meredith would be absolutely perfect, except the bore is too large for an Aixiz module.

Learning how to machine to close tolerances by hand is definitely something I'd like to get into. Some purchases would have to be made, which puts a damper on things. You'd laugh though, if you saw what I presently have to work with. Not even a band saw... heh. I recently spoke to the wood/metal shop teacher at my old high school about possibly using some of their equipment for a similar heatsink project, and he didn't even think his stuff could get down to a thousandth. Actually though, it was his son substituting that day... have to go back and talk to the Dad, who was MY shop teacher in school.

you are better going with a brutally simple TEC drive with no feedback, just aim for an effiecient cooling using a fixed voltage/current supply, and keep the hot side well heatsinked, and keep the laser casing sealed. There is no other way better than that. We're not after wavelength control in this case, just a large shift. A small thermistor mounted on the diode block will give adjustment to the constant current drive, but even that isn't needed, just tune it to the cold temperature, and it will protect itself at high temperatures during warmup or cooling failure by being less efficient and making slightly less power at the output mirror.

Sounds great to me! I totally agree with brutal simplicity here, and am happy to hear there should be some margin of self-protection after tuning things to a particular range of temperature. When this thing finally gets built, I'll have some questions as to temp vs. current. Now, how do I go about creating a simple drive circuit for the TEC? Can I make my own? I've got breadboard!... lol. Or, can I even use another EU-38 diode driver? That'll make TWO Roithner boards in this thing!

Some more design ideas/solutions...

Ok, I got a bigger plastic box yesterday... lol. The one intended for this, although larger than "The Grayser" (my first laser project), just wasn't going to be big enough. Finding the correct TEC is a matter of factoring in size, voltage, and current, against the batteries being used. Likely, it'll have to be two Li-ion's. Considered three NiMH, but that only gives 3.6V. Both Li-ion, and NiMH have pretty flat discharge curves, which should reasonably simulate a steady power supply. Once a hole is cut, the TEC is going to sit flush with the top plastic surface of the box... cold side facing the inside, hot side facing out. Mounting AND sealing will be courtesy of Tempflex (Thanks, Doc!). Check out what's going to cool the hot side. Goes perfect with the scary/dangerous look.

Swiftech MCX159-CU Copper C110 Fan&Heatsinks. Haven't decided how to power the 12V fan. It may get it's own power supply, making THREE seperate supplies (one for the diode, one for the TEC, one for the fan). Or, it may be replaced by a 5V fan and run off the same supply as the TEC, or some other solution. Depending on fan vs. power, it may even get it's own little fan speed control. Really resisting overusing the devil smiley again.

More details coming soon!

Mike - Hahaha... that's just my self-critical/slightly impatient nature talkin'. As a total noob just getting into a new hobby, I somehow expect myself to know all there is to know in the first day, and feel the need to call myself semi-retarded when something confuses me, or come up with ideas that require asking questions. So don't listen to me when I say that. I tell my students to avoid that feeling, and sometimes don't listen to my own advice. Again, definitely glad you're enjoying this! Just wait until this doomsday machine-looking thing is built. Hopefully, I can repay all of you for your help with some hearty laughs!

[The_Doctor]

That's definitely a rectangle.. I think the light and the front edge perspective make it look trapezoidal, but it's not.

If you have the disk sander you might get away without the bandsaw, a hacksaw will be enough, you'd just need to cut large to allow for waste of a mm or two. You could reduce that with a cutting guide. The really important tools are the caliper and the pillar drill and the emery board and some kind of saw. And some kind of very squared bar, ideally stainless steel, as a guide on the emery board to keep angles squared.

That Meredith block is cheap, very cheap. If you can use it, do it. The only advantage to making stuff if you're competing against that low cost, is having full control, and always knowing you can have more parts, you're never dependent on the will of another maker.

For TEC drive in this case, all you need is a high current capacity, 3 to 5A as likely maximum, and good stable voltage regulation. It will be either on, or off, but that's fine so long as it's not switching repeatedly. The optimum voltage for TEC efficiency is around 0.8 of Vmax. Use the smallest TEC you can get, and mount only the diode mount on it. If you can get TEC's with metallised surfaces, you'll need a brass or copper mount, not anodised aluminium, so you can solder to them with low temperature indium based sloder, with a melting point of 117°C. TEC's internal soldering melts at 138°C, so you need to be cautious. You don't need a special iron, but you do need good timing to gauge when to apply and remove heat input. Resin flux, the kind with isopropanlo solvent, is ideal, don't get 'no-clean' types, as they leave hard-to-clean residues. That sounds wrong, but they really do.

For supplies, if you start with 12V, you can further regulate with linear regulator IC like LM317 for the diode. Either as voltage or current regulator, either is ok, so long as the diode gets the right current. LM317's are very good, better than fixed types, so protect the diode better. As voltage reg, you can easily config a soft start (see LM317 data sheet, is standard design). A small TEC will likely be below 12V, even below 5V, if you get the really small ones that are best for this. If you go with eBay you'll be taking pot luck, and might even have to test carefully by raising input volts till you get good cooling without lots of heat out of the hot side. For this, the ideal drive is an LM338 (is capable of 5A, LM317 is only 1.5A). Wire as voltage regulator and use a preset to guide the supply. This means that with a 12V 5A supply, and one LM317, and one LM338, you'll almost certainly have all the supply you'll need for fan, TEC, and diode, along with the easy adjustmment you'll need. Get a low ESR 0.1 µF ceramic capacitor and a 22 µF low ESR tantalum across the diode, do what it takes to track those down and get them, they're cheap, but might take some research to find. Don't worry if you have to buy them ten at a time, you can always use them if you're into laser diodes.

Finally, as you're running it from a battery, you can be less concerned with surge rejection, as the battery can be charged off a mains unit while the laser is safely elsewhere. If the diode survives this, you can trust the drive and protection that's in there well enough to figure our what to add if you later want mains powering direct.