optical sweep generator

What about 850? And higher orders?

I'd just want enough fixed-location photodiodes to resolve my three wavelengths. Three, maybe even two would do.

The fiber-coupled power will be a milliwatt or so, so some detection things won't work. And I want something suited to production test. I'd spend some on a spectrometer, but I haven't found one that covers my range. They all seem to go for high resolution over a narrow wavelength range.

I only want to know if we installed the wrong laser out of three.

Thorlabs sells "Laser Line" bandpass filters at around 100$ each. If you kn ow what wavelengths the diodes were when ordered, you can buy a narrow band pass, say 20 to 40 nm wide, and find out.

As for the two photon adsorption question, they are coating the CCD with a n upconversion phosphor. Not bad for 1000$. Considering used FIND-R-SCOPE S are going for nearly a grand on Ebay these days, that is not too bad. A grating and a brand new FIND-R-SCOPE with the 1550 option would cost Mr Lar kin 2900....

Gratings are easy, and with the right detector, a milliwatt is overkill. O ptical bandpass filters usually have less then 5% loss on "Laser Line" grad e units.

Thorlabs also is a good source of one off, cheap photodiodes and photocondu ctive detectors for the IR spectrum.

John, How many different wavelengths do you need to sort?

Steve

For a grating at 600 LPM, first order 800 nm comes off at 28.6 degrees and the second order comes off at 79 degrees.

Dodging the second and third order is why low cost spectrometers cover limited ranges. Even with order sorting filters, there are limits to what you can do with a single grating, or without moving the grating.

Steve

an upconversion phosphor. Not bad for 1000$. Considering used FIND-R-SCO PES are going for nearly a grand on Ebay these days, that is not too bad. A grating and a brand new FIND-R-SCOPE with the 1550 option would cost Mr L arkin 2900....

"micro lens system and special coating layer on a silicon" This is the usual CCD from SONY. If there was coverage, the CCD would not see the blue color... Frame averaging, binning only. IMHO

Three. We buy 850, 1310, and 1550 fiber-connectorized lasers.

The ideal would be a small box with a fiber in and three LEDs, for a final check on a unit before we ship it.

I don't need resolution, so filters would do. One tapered filter and a marked knob, and a single wideband detector, would work.

Or some sort of wavelength-selective beam splitter and three, maybe even two, detectors.

I can't find any instrument that will tell me what the wavelength is, over my range. It would nave to be easy to use, for production test.

Disassemble the SFP module, leave only the case with internal diagonal interference filter (1310/1550), insert two photodiodes (Ge or InGaAs).

I smell a market here for a broadband handheld box, using one moving part. I can do it without a fancy dichroic filter, as it has been done before, just the widget is off the market.

If I get furloughed in three weeks, I know what I'm working on.

Covers 330-1800, one CNCed 6061 frame, and all off the shelf parts.

Updates at say 10x a second.

Steve

You can split that milliwatt to three, use filters and InGaAs photodiodes. Or two spectrometers with a bifurcated fiber - My VIS Avantes spectrometer cost 20% of the NIR one. And used HP 71450B cost 50% of the VIS spectrometer..

Model airplane servos are the bomb for that job. With the right germanium photodiode, you can get 350-1800 nm. The trick is to make it insensitive to the alignment of the incoming beam, which is a bit of a trick even with a very very narrow slit. A bit of Scotch Magic Invisible Tape on the entrance slit helps a lot with that, at the expense of efficiency (which you don't care much about).

Running the PD open-circuit gets you a nice logarithmic output, at least above a few microamps' photocurrent, so your spectra come out nicely normalized.

To go that fast you'd probably need continuous rotation, either spinning the grating or running a cam.

Cheers

Phil Hobbs

(Who used to work on the same OPGs that you did.) ;)

I know nothing of optics but I think I heard some crystals do optical frequency doubling. Then your 830, 1310, 1550 problem shifts to 415,

piglet

Grating and mirrors Ordered, servo not ordered. Servis have their place, and I use them immensely, but Faulhaber has my love.

Steve

SHG(Doubling) sadly requires a lot more energy then he has available. Upconversion/Two Photon, using a UV or blue/green led to pump the commonly available "laser viewing card" phosphors would do that.

He really just needs a filter wheel, filters, a PD, and a voltmeter or LED at this point. Technically he needs one less filter then he has wavelengths.

I need to measure 1100-1250 for a perpetual laser development project, so I ordered a grating.

What I need to measure is just outside the range of my pile of surplus B&W Tek S110 spectrometers to detect, and I need 5 nm resolution.

Besides, the same grating can go into the BW Teks and extend the range from the current 330-640 to 300-1100 at 3 nm resolution.

At one point in time I was getting the BW Teks at 3 for 125$, surplus from a project the FDA evidently shut down. At that price I made up some kits for the local schools to use on loan.

Steve

I want one. I wouldn't mind cooperating somehow to design/manufacture/sell a wideband spectrometer or at least wavelength identifier. We do usb boxes all the time. And machining.

For crude performance, one could use a mediocre wavelength splitter, with some ambiguity, and do some software tweaks.

A tapered filter is more expensive than a grating, a lens, and multiple (or image) detectors.

I've used several Faulhaber/MicroMo motors, with good success.

The advantage of RC servos is that they have decent encoders, run closed-loop, and are super easy to control from MCU via PWM.

What are you proposing to use for an encoder?

Cheers

Phil Hobs

BTW an old friend of mine, Ed Yarmchuk, who invented the self-servowriting technique for hard disk drives, used to say, "Belive Mr. Newton."

Ed and his old boss, Tim Chainer, were able to get an amazing improvement in servowriting accuracy that way, without needing interferometric spin stands, which were the previous standard technique.

Cheers

Phil Hobbs

Phil,

I'm looking at motors right now. I have plenty of galvos. I have former c olleague who now makes really good galvos (175$ Each in quantity) in the U S at high production rates, I have some good titanium geared servos with g ood bearings.

If you have a stock, available, motor you'd like to suggest for a 12.5 mm square grating 6 mm thick, I'm interested. Especially if it has a good enc oder and a shaft that wont bend when you tilt the unit in your hand when sp inning. I guess I need to calculate the inertia. I agree to let Mr Newton d o his thing. I cant see needing more then 30 RPM, so geared might be accept able. I have some Portescaps laying around, with encoders, but high RPM.

I just happen to know a company that prints on plastic or glass at 22,800 D PI and and another that photo etches if I need a reference signal disk.

My goal is to have a generic core frame that can hold a galvo, motor, o r high end digital servo, a well balanced grating mount with one axis of fl exure, or for production set up a jig to just epoxy the grating right where it needs to be tilt wise.

I have a beat up old manual mill to make prototypes, so that is easy enough . I have a very good miniature lathe as well.

I ordered a TO-5 can InGaAs PD with 150 uM active area, and a couple of sto ck mirrors with a decent radius, coated for 400 nm to 20 UM, at greater th en 95%.

I'm guessing you know the wonderful piece of 1970s / 1980s technology I 'm basing this unit on. Had one at a former employer for tuning the Ti:Saph and Dye lasers. It was already 20 years old when I used it, had sat for 10 years, fired right up for me, and then a few years ago the new guy there c alled for help with determining wavelength, and I told him which shelf to l ook at, and it spun up for him. I might go look at how it counts with just a dozen or so TTL and yet compensates nicely, with a linear wavelength disp lay on the scope, and has another board with three thumbwheel switches fo r setting the marker gate and optional sample and hold. You push a button and it adds the grid and tick marks on the scope XY display.

The log amp should not be a problem for a demo.

I've got plenty of mirror mounts to prototype with. I have Ar, Hg, Kr, Ne, Cl, Al, Xe lamps for calibration. For visible ca libration I have about 20 ion laser lines quickly available and a variety o f Hene lines. I probably can dredge up a "Hot" mirror as a pre-filter.

I have a calibrated IR FO power pen for an idea of the magnitude, and if really want to get fancy I have a multiline HENE that does some IR lines an d four wave mixing lines that most people have never heard of. I know wher e there is a USB ocean optics with the IR option for verifying the near IR calibration.

Mcmaster Carr's warehouse is a 40 minute drive, so getting stock 6061 shape s from them is cheap.

I have a few slits in the junk collection and a collection of scopes.

I like the idea of this becoming a potential product, Mr. Larkin, you'll ge t a call. What connector is on the end of your fibers?

Steve

I used HiTec servos from servocity.com for rotating gratings in my transcutaneous blood glucose/alcohol sensor proto some years back. The ones I have are HS-7945TH, now discontinued. They have titanium gears, Oilite bearings, decent resolution, and lots of torque. It's super easy to control them from an MCU or a LabJack, and similar ones are still available.

Servo City also sells hubs and Meccano-style structural parts:

In my gizmo, by good luck that channel stuff was just the right size so that the surface of the grating was within 10 um of the servo's axis of rotation. (I had to cut the channel with aviation shears to get rid of the part I didn't want.)

Optical encoders get expensive fairly fast, and getting decent accuracy requires really good centration. I've used the cheapish US Digital magnetic ones, which give you a PWM output with 4096 steps per rev and are less sensitive to centration.

They're really great if you use them right, meaning measuring the duty cycle rather than the pulse width--even cold spray and a heat gun won't make them drift significantly.

Yup. Some of my favourite lasers are made like that--a small SS base plate with the optical components epoxied right on there. The great advantage of that method is that the gizmo is either (a) perfectly aligned or (b) obviously broken, nothing in between. Another advantage is that in case (b) it's obvious whose fault it was.

Active alignment + UV epoxy = a win.

InGaAs at 400 nm? Cool! They're usually useless below 800 nm on account of the very short absorption depth in the epi, which is why I'm a big fan of germanium for that use. You might get in touch with George Gasparian at GPD (formerly Germanium Power Devices, fronted by the late 'Oliver Germanium' of song and story).

My one point of reference is my Burleigh WaveMeter, which uses a corner cube Mach-Zehnder with a HeNe for calibration. I got mine for a song on eBay--some idiot had taken it apart and reassembled it ass-backwards. :) (The HeNe PSU arcs a bit for a couple of minutes when you turn it on, but nothing to worry about. Not bad for $250 plus $86 for shipping.)

Multimode HeNe's have a lot of interesting physics going on, for sure. Just shining one on a photodiode is educational--when you look at the detected signal with a spectrum analyzer, you often see a baseband spur wandering around between about 100 kHz and 1 MHz. It's a fourth-order product, coming from the interference of one mode with a third-order IM product of two others. When the frequency delta gets below 100 kHz or so, the modes jump into lock. You can get them to unlock again by cooling one end of the laser with your hand. Fun.

Good luck with your gizmo!

Cheers

Phil Hobbs