Post by thread:[EE]: Lasers again

But as Matt says, the rise and fall times in the transducers just kills it... ...everybody on this thread seem to be ignoring my post on Optical Ranging... ...Its simple, its cheap, its PICable, its the way to go... Just see: www.piclist.com/io/sensor/pos/opticalranging.htm and http://www.cyberg8t.com/pendragn/actlite.htm for a complete description. Am I missing something? James Newton, PICList Admin #3 spam_OUTjamesnewtonTakeThisOuTpiclist.com 1-619-652-0593 phone http://www.piclist.com {Original Message removed}

Hi James, I must really be missing something here. AFAIK, LED rise times are usually a few ns, maybe tens but not 100s. As for the receiving end, it is really governed by the amplifier you hook to the photodiode (and how you bias it, etc.) As I said, I have sent 1 MHz signals over cheap LEDs (they were actually metal can, high power IR types, but still only $3 or so) and cheap photodiodes. We aren't talking about pulsed signals here, right? Just comparing phase. I saw your page on optical ranging ,and it looks interesting, I am just curious about why people seem to insist that LEDs can't handle 1MHz signals. I would also recommend that people check out the triangulation method, which is the way many autofocus cameras do it. For example, see http://www.hobbyrobot.com/info/gp2d02/index.html Sean At 07:08 PM 2/3/01 -0800, James Newton wrote: {Quote hidden}>But as Matt says, the rise and fall times in the transducers just kills >it... > >...everybody on this thread seem to be ignoring my post on Optical >Ranging... > >...Its simple, its cheap, its PICable, its the way to go... > >Just see: >www.piclist.com/io/sensor/pos/opticalranging.htm >and >http://www.cyberg8t.com/pendragn/actlite.htm > >for a complete description. Am I missing something? > >James Newton, PICList Admin #3 >.....jamesnewtonKILLspam@spam@piclist.com >1-619-652-0593 phone >http://www.piclist.com -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.

"Sean H. Breheny" wrote: > > Hi Matt, > > Aren't you talking about modulating with a sine wave? If so, why do you > need such fast rise times? I have done >1MHz with cheap photodiodes and > LEDs (although, I wasn't doing ranging, I was experimenting with high speed > data transmission). For the XOR phase detector to work right, it needs to be full on/ full off, I'm not sure how a traditional mixer based phase detector works with two very different amplitude signals, I'm guessing that you need run an AGC, which adds some complexity that I'd rather not have. I'd rather just use the XOR gate and look at the edges, and not worry about the amplitude. For really high speed data (gigabit ethernet, for example) it is really just an on/off modulation- there are no intermediate steps, so I know it can be done, and the XOR method is nice and simple. Matt -- http://www.piclist.com#nomail Going offline? Don't AutoReply us! email listservKILLspammitvma.mit.edu with SET PICList DIGEST in the body

> I must really be missing something here. AFAIK, LED rise times are usually > a few ns, maybe tens but not 100s. As for the receiving end, it is really > governed by the amplifier you hook to the photodiode (and how you bias it, > etc.) From my experimentation, the rise/fall times in LEDs (especially the high power LEDs) are pretty poor, they're engineered for optical power, not low capacitance. The normal photodiodes I've used have way too much capacitance- more than a few pF, and you just can't get good sharp rise/fall times. I'd like to get a 10MHz signal, 50% duty cycle, with under 5ns rise/fall times, which, if I remember the equations, is about a 280 MHz bandwidth. (which is still probably easier than a 500 MHz counter) {Quote hidden}> As I said, I have sent 1 MHz signals over cheap LEDs (they were actually > metal can, high power IR types, but still only $3 or so) and cheap > photodiodes. We aren't talking about pulsed signals here, right? Just > comparing phase. > At 07:08 PM 2/3/01 -0800, James Newton wrote: > >But as Matt says, the rise and fall times in the transducers just kills > >it... > > > >...everybody on this thread seem to be ignoring my post on Optical > >Ranging... > > > >...Its simple, its cheap, its PICable, its the way to go... James, I'm not ignoring it, I'm just more interested in the phase detection method. Using triangulation, the error largely dependent upon range, and the distance between the transmitter and sensor. I'm not trying to put out a product- I'm trying to make something that intrigues me. If I absolutely had to get something out, I'd be much more interested in triangulation. Matt -- http://www.piclist.com#nomail Going offline? Don't AutoReply us! email .....listservKILLspam.....mitvma.mit.edu with SET PICList DIGEST in the body

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 I was skeptical about the simple image subtraction thing, but after trying it... Wow. I used my crappy little QuickCam Express and took a photo of my whiteboard. Then I drew a line on the whiteboard, and took another photo. Imported both as layers into Paint Shop Pro, and set the blend mode to difference. White background, text, drawings - gone. Only the line remains. If you could get a laser line to show up on a black and white imager, it would be sooo easy to do with a PIC - scan one frame into some storage, kick in the laser, scan in another frame - same storage space, just subtract and take the absolute value. Simple math from there on in. - -Randy Glenn This coming from the guy with a system tray 7 icons wide... by 2 tall... ================================================= Randy_Glenn-at-tvo.org - PICxpert-at-picxpert.com PICxpert-at-yahoo.com - PICxpert-at-home.com http://www.picxpert.com/ ================================================= - {Original Message removed}

> I was skeptical about the simple image subtraction thing, but after > trying it... > > Wow. > > I used my crappy little QuickCam Express and took a photo of my > whiteboard. Then I drew a line on the whiteboard, and took another > photo. Imported both as layers into Paint Shop Pro, and set the blend > mode to difference. White background, text, drawings - gone. Only the > line remains. That sounded like such a good idea I had to try it myself. I agree with Randy. Wow! I would have included the results in this email message, but I figured that the images would exceed the PIClist line length limit. I therefore set up a web page at http://www.embedinc.com/temp for those of you that would like to see the results. ***************************************************************** Olin Lathrop, embedded systems consultant in Devens Massachusetts (978) 772-3129, EraseMEolinspam_OUTTakeThisOuTembedinc.com, http://www.embedinc.com -- http://www.piclist.com#nomail Going offline? Don't AutoReply us! email listservspam_OUTmitvma.mit.edu with SET PICList DIGEST in the body

With respect to risetime verses bandwidth: An approximation is Risetime = 1/Frequency * 0.35. This gives 70MHz bandwidth for a 5ns risetime. Not trivial but a lot easier than 280MHz. There are a lot of voltage and current feedback opamps available for these frequencies and risetimes. Not terribly expensive either. -- Rich Matt Bennett wrote: {Quote hidden}> > > I must really be missing something here. AFAIK, LED rise times are usually > > a few ns, maybe tens but not 100s. As for the receiving end, it is really > > governed by the amplifier you hook to the photodiode (and how you bias it, > > etc.) > > >From my experimentation, the rise/fall times in LEDs (especially the > high power LEDs) > are pretty poor, they're engineered for optical power, not low > capacitance. The normal photodiodes I've used have way too much > capacitance- more than a few pF, and you just can't get good sharp > rise/fall times. I'd like to get a 10MHz signal, 50% duty cycle, with > under 5ns rise/fall times, which, if I remember the equations, is about > a 280 MHz bandwidth. (which is still probably easier than a 500 MHz > counter) > > > As I said, I have sent 1 MHz signals over cheap LEDs (they were actually > > metal can, high power IR types, but still only $3 or so) and cheap > > photodiodes. We aren't talking about pulsed signals here, right? Just > > comparing phase. > > > At 07:08 PM 2/3/01 -0800, James Newton wrote: > > >But as Matt says, the rise and fall times in the transducers just kills > > >it... > > > > > >...everybody on this thread seem to be ignoring my post on Optical > > >Ranging... > > > > > >...Its simple, its cheap, its PICable, its the way to go... > > James, I'm not ignoring it, I'm just more interested in the phase > detection method. Using triangulation, the error largely dependent upon > range, and the distance between the transmitter and sensor. I'm not > trying to put out a product- I'm trying to make something that intrigues > me. If I absolutely had to get something out, I'd be much more > interested in triangulation. > > Matt > > -- > http://www.piclist.com#nomail Going offline? Don't AutoReply us! > email @spam@listservKILLspammitvma.mit.edu with SET PICList DIGEST in the body -- http://www.piclist.com#nomail Going offline? Don't AutoReply us! email KILLspamlistservKILLspammitvma.mit.edu with SET PICList DIGEST in the body

Richard Ottosen wrote: > > With respect to risetime verses bandwidth: > An approximation is Risetime = 1/Frequency * 0.35. > This gives 70MHz bandwidth for a 5ns risetime. Not trivial but a lot > easier than 280MHz. There are a lot of voltage and current feedback > opamps available for these frequencies and risetimes. Not terribly > expensive either. > Thanks- I guess I did remember the equation wrong. This also brings the frequency well within what I would expect from some fast TTL... now the big question is, how do I modulate the LED and recieve it that fast? I remember reading in a scope manual about how two consecutive rise times affect each other- something like a 10ns rise time into an amplifier with a 10ns rise time gives about a 14ns rise time? (I seem to remember a sqrt(2) in there). I guess the LED needs about a 100MHz BW from the LED to get the 5ns rise time out of a 70MHz reciever. So, how about it... anyone have circuits for transmitting/recieving light modulated at about 100MHz? Gigabit Ethernet does it at over 1GHz, and OC192 approaches 10 GHz optically, so there should be a way. Matt -- http://www.piclist.com#nomail Going offline? Don't AutoReply us! email RemoveMElistservTakeThisOuTmitvma.mit.edu with SET PICList DIGEST in the body

{Quote hidden}> -----Original Message----- > From: Matt Bennett [SMTP:spamBeGonemjbspamBeGoneHAZMAT.COM] > Sent: Monday, February 05, 2001 3:22 AM > To: TakeThisOuTPICLISTEraseMEspam_OUTMITVMA.MIT.EDU > Subject: Re: [EE]: Lasers again > > Richard Ottosen wrote: > > > > With respect to risetime verses bandwidth: > > An approximation is Risetime = 1/Frequency * 0.35. > > This gives 70MHz bandwidth for a 5ns risetime. Not trivial but a lot > > easier than 280MHz. There are a lot of voltage and current feedback > > opamps available for these frequencies and risetimes. Not terribly > > expensive either. > > > > > Thanks- I guess I did remember the equation wrong. This also brings the > frequency well within what I would expect from some fast TTL... now the > big question is, how do I modulate the LED and recieve it that fast? I > remember reading in a scope manual about how two consecutive rise times > affect each other- something like a 10ns rise time into an amplifier > with a 10ns rise time gives about a 14ns rise time? (I seem to remember > a sqrt(2) in there). I guess the LED needs about a 100MHz BW from the > LED to get the 5ns rise time out of a 70MHz reciever. > > So, how about it... anyone have circuits for transmitting/recieving > light modulated at about 100MHz? Gigabit Ethernet does it at over 1GHz, > and OC192 approaches 10 GHz optically, so there should be a way. > > Matt > The likes of Maxim, Analog Devices make laser drivers, limiting amps, AGC amps and CDR IC's. The company I work for even make their own, although they are not marketed externaly AFAIK. HP make a fair range of high speed amps suitable for PIN diodes. Direct modulated 10GHz is amazingly tricky. Most 10GHz systems use a CW laser and some kind of external modulator, e.g. mach zhender or elctro absorbtion. However, getting the driver is only half of the battle. PCB design is also pretty important with such fast rise times, to avoid ringing closing the optical eye. You *may* be able to get an eval board from one of the above with some smooth talking, although don't bank on it. Mike -- http://www.piclist.com hint: To leave the PICList RemoveMEpiclist-unsubscribe-requestTakeThisOuTmitvma.mit.edu

Hi Matt, so you really want to do this? Why 100MHz? I would use a lower frequency, the commercial units do. Maybe 5MHz? What distance and resolution did you want to achieve? :o) -Roman Matt Bennett wrote: {Quote hidden}> > Richard Ottosen wrote: > > > > With respect to risetime verses bandwidth: > > An approximation is Risetime = 1/Frequency * 0.35. > > This gives 70MHz bandwidth for a 5ns risetime. Not trivial but a lot > > easier than 280MHz. There are a lot of voltage and current feedback > > opamps available for these frequencies and risetimes. Not terribly > > expensive either. > > > > Thanks- I guess I did remember the equation wrong. This also brings the > frequency well within what I would expect from some fast TTL... now the > big question is, how do I modulate the LED and recieve it that fast? I > remember reading in a scope manual about how two consecutive rise times > affect each other- something like a 10ns rise time into an amplifier > with a 10ns rise time gives about a 14ns rise time? (I seem to remember > a sqrt(2) in there). I guess the LED needs about a 100MHz BW from the > LED to get the 5ns rise time out of a 70MHz reciever. > > So, how about it... anyone have circuits for transmitting/recieving > light modulated at about 100MHz? Gigabit Ethernet does it at over 1GHz, > and OC192 approaches 10 GHz optically, so there should be a way. > > Matt > > -- > http://www.piclist.com#nomail Going offline? Don't AutoReply us! > email listservEraseME.....mitvma.mit.edu with SET PICList DIGEST in the body -- http://www.piclist.com hint: To leave the PICList EraseMEpiclist-unsubscribe-requestmitvma.mit.edu

>For the XOR phase detector to work right, it needs to be full on/ full >off, I'm not sure how a traditional mixer based phase detector works >with two very different amplitude signals, I'm guessing that you need >run an AGC, which adds some complexity that I'd rather not have. I'd >rather just use the XOR gate and look at the edges, and not worry about >the amplitude. A traditional phase detector deals just fine with signals of very different amplitude, because it is a multiplier so Uo = K * Ui1 * Ui2. Of course you need to achieve amplitude calibration, but instead of AGC I'd suggest a limiter amplifier designed for FM broadcast service. This will pass the phase information undisturbed and provide plenty of gain (and even AGC). It also contains a multiplier (product detector) that could be hijacked to do the detection. A MC3189 comes to mind, or older TBA810(??), or something much newer but more difficult to obtain. Peter -- http://www.piclist.com hint: To leave the PICList RemoveMEpiclist-unsubscribe-requestEraseMEEraseMEmitvma.mit.edu

> A traditional phase detector deals just fine with signals of very > different amplitude, because it is a multiplier so Uo = K * Ui1 * Ui2. Of > course you need to achieve amplitude calibration, but instead of AGC I'd > suggest a limiter amplifier designed for FM broadcast service. This will > pass the phase information undisturbed and provide plenty of gain (and > even AGC). It also contains a multiplier (product detector) that could be > hijacked to do the detection. A MC3189 comes to mind, or older TBA810(??), > or something much newer but more difficult to obtain. Even easier would be to create digital square waves from both signals from their zero crossings, then go back to the XOR method. You might want to run each raw square wave thru a divide by two counter to make sure they are really square. In other words, only use one of the two zero crossing per cycle from each input signal. That's all you need to measure phase angle. ***************************************************************** Olin Lathrop, embedded systems consultant in Devens Massachusetts (978) 772-3129, RemoveMEolinspam_OUTKILLspamembedinc.com, http://www.embedinc.com -- http://www.piclist.com hint: To leave the PICList RemoveMEpiclist-unsubscribe-requestTakeThisOuTspammitvma.mit.edu

Just to add a note toward the end ? of the thread:- We used this method to measure the elongation of fibre optic cables during laying. A mirror was formed at the far end of the fibre and at the near (stationary) end a laser source was fitted, modulated at 200MHz . To improve resolution a number of fibres were spliced though so the actual fibre length was about 6 times the cable length (total signal path length of 12km/cable km). An optical directional coupler fed the signal reflected back from the far end of the cable to a detector (apd?) , preamp and an HP network anaylser. The other channel of the network anaylser was fed directly from the 200MHz source. By precalibrating the fibre on a tensile tester and using phase measurement we were able to resolve to about 5mm of elongation on a 1km length of fibre. This was used to calculate residual stress on the fibre and get an estimate of its likely lifetime - along with producing a direct comparison of different cable laying techniques and equipment. Richard P > A traditional phase detector deals just fine with signals of very > different amplitude, because it is a multiplier so Uo = K * Ui1 * Ui2. Of > course you need to achieve amplitude calibration, but instead of AGC I'd > suggest a limiter amplifier designed for FM broadcast service. This will > pass the phase information undisturbed and provide plenty of gain (and > even AGC). It also contains a multiplier (product detector) that could be > hijacked to do the detection. A MC3189 comes to mind, or older TBA810 (??), > or something much newer but more difficult to obtain. Even easier would be to create digital square waves from both signals from their zero crossings, then go back to the XOR method. You might want to run each raw square wave thru a divide by two counter to make sure they are really square. In other words, only use one of the two zero crossing per cycle from each input signal. That's all you need to measure phase angle. ***************************************************************** Olin Lathrop, embedded systems consultant in Devens Massachusetts (978) 772-3129, EraseMEolinspamspamBeGoneembedinc.com, http://www.embedinc.com -- http://www.piclist.com hint: To leave the PICList RemoveMEpiclist-unsubscribe-requestKILLspammitvma.mit.edu -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads

>Even easier would be to create digital square waves from both signals from >their zero crossings, then go back to the XOR method. You might want to >run each raw square wave thru a divide by two counter to make sure they >are really square. In other words, only use one of the two zero crossing >per cycle from each input signal. That's all you need to measure phase >angle. I think that I would not pick the zero crossings of a noised-over signal for anything, especially not for extracting phase. I would pick the areas with the maximum amplitude which contain most energy (and best S/N). A multiplier does precisely this, but there are other ways. To obtain squarewave signals from a low amplitude noised over signal one usually uses a high gain low noise amplifier with limiting in the last stages. The FM receiver/demodulator systems do just this and have a good S/N figure. I'd be very tempted to modify a FM radio from the junkbox and use a PIN photodiode connected to the 10.7MHz IF directly (with a polarization circuit), and a laser modulated by a 10.7MHz oscillator (using a ceramic filter probably). I'd also be tempted to transmit 38400 Bauds digital data through this system (I have done experiments with 455kHz AM and 9600 Bauds using IR LEDs before). Since I have some experience with such circuits, I say that it will be very hard to keep the transmitted signal's RF from entering the receiver and obliterating any received signal. The separation must be 120dB or better even for small distances. This is *very* hard to do in my experience, especially if the two units share a power supply or the same box (and if the box must be smaller than a Chubb safe). Peter -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads