Designing a Time to Amplitude Converter

I think the internal resistance of the fet is a bad idea. The reisistance isn't well defined. Why not build a current source, then gate it with a fet. You may want to read up on charge injection.

I'm assuming the time between pulses is very fast, else you would be building a timer.

In the dark ages, I contracted to design/build a filter for a laser based distance measuring system. As is usually the case, the design specs are written so you really don't know what the system does. [Compartmentize (sp) the info to prevent someone from walking away with the design.] However, I did get to see the set up. It was quite a show since I entered a mid-rise building (maybe 8 stories) and this floor was totally open so the laser system could have an arena in which to play.

Just eye-balling the system, I think what they were doing was modulating the laser at IIRC 10MHz. The laser beam was split optically. The phase difference of the reference laser (i.e. what was split off) and the reflected light was compared. This phase difference was related to the distance the light traveled. My filter was to take noise off the detected reflected signal.

Now I'm guessing that once you slipped a whole cycle, that a counter would remember this. Thus the distance was a function of the number of cycles that were detected, and the current phase difference.

I agree - a current source would be a much better system. I just was trying to illustrate my thought process as simply as possible.

The time that I need to measure between pulses is less than a nanosecond. I'm not aware of any good ways to measure such a short period of time using a timer, but I would love to hear of some if you know of any.

The phase difference system sounds like a solid concept. If I could measure the phase difference of just one cycle at 10Mhz I would measure anywhere from 0-30M. In fact, I would probabaly want to modulate it at a higher frequency like 60Mhz as I'm more interested in closer range than that, and I want maximum accuracy at close range. How exactly would a laser beam be modulated though? Would just the intensity be being modulated? I would think that that would be hard to keep accurate enough to measure. Also, how would you measure the phase difference? The only way I can think of to measure it is to read the signals into a DSP and do fourier transforms on them. But at 10Mhz, I don't think that is very feasible. Also, I'm unsure how you would go about counting if the phase difference was more than one phase. To me, I think you would just have to accept that that was your limited factor for your max range, but you said it's possible?

Thanks,

-Mike

You can buy lasers with a modulation input - for instance RS 564-504 claims to operate in the range 100kHz to 50MHz.

Then you measure the phase difference between transmitted and received signals at a set of frequencies. You end up with a number of distances modulo the 'wavelength' used for modulation. If you choose the frequencies carefully you can then solve for the actual distance.

To measure phases, look at how phaselock loop phase detectors work

Another approach is to connect the output of the photo detector to the laser modulation input. OTBE, the whole setup will oscillate at a frequency which (partly)depends on the path length. It is much easier to measure frequency very accurately rather than pulse delay or phase.

If your range is very short, and you like a challenge, then you could build an inteferometer and count wavelengths ...

Dave

Hi.

Uzytkownik "Mike No> The time that I need to measure between pulses is less than a

There is simple way to measure time pulses less than a nanosecon. TDC... Time to Digital Converter. ACAM produce a few diffrent TDC look at

Best Regards

Jaroslaw Grolik

Wow - I had no idea that such a chip would exist. Do you have any idea how much they cost? I have contacted their USA distributor but the guy I need to talk to is out of the office till Monday or later. Thanks!

-Mike

Thanks again for the information Jaroslaw. Out of curiosity, do you have any idea how much the TDC-GPX would cost? 10ps accuracy sounds pretty amazing (accuracy to 1.5mm!) though it'd probabaly be serious overkill. Thanks,

-Mike

Hi Dave - that definitely sounds like an interesting approach. I am somewhat worried by the cost of the laser you mentioned - at over $250 that is more than I'm hoping to spend. Are there cheaper modulateable lasers available? Thanks,

-Mike

Apart from the chip already mentioned, many ethernet type chips ive programmed have TDR capabilities wich can measure the time taken for a reflected pulse, this is to help locate where on a cable is a discontinuity.

One problem will be the low amount of reflected light from your target and the dificulty of getting both low noise and high speed signals from a low light levels via a photodetector.

Another alternative is to simply gate a constant current source (ie a resistor from on open collector/drain Q op) into an opamp integrator circuit with a slow resistive discharge, the voltage from the integrator will then be a function of the average pulse width. You dont need to discharge the C every pulse.

However the effect of noise or missing returned pulses will give rise to large errors.

There are other ways of measuring distance with light (LIDAR) I'm using a hetrodyne aproach wich modulates the laser at 1ghz+ and measures the phase difference with a quadrature demodulator. Frequency is scanned to cope with distances of multiple wavelenghts.

Another aproach is to use a dual cavity correlating interferometer, resolution of absolute distance down to a few nm.

Colin =^.^=

Dave wrote: ...

And in the intermediate range do not forget triangulation. A laser diode or a LED and a position sensitive photodiode and some lenses will do the job.

Uzytkownik "Mike Noone" napisal w wiadomosci news: snipped-for-privacy@m73g2000cwd.googlegroups.com...

It is quite cheap comparing to its capabilities .

Best Regards

Jaroslaw Grolik

Uzytkownik "Mike Noone" napisal w wiadomosci news: snipped-for-privacy@m79g2000cwm.googlegroups.com...

Unfortunately I do not know the price of TDC-GPX ... TDC-GP2 costs 15 Euro / pcs. (Above 100 pcs 10 Euro ). The TDC have many modes of operation so check that 65 ps would be enough for you. TDC-GP2 has that accuracy .

Best Regards.

Jaroslaw Grolik

What sort of accuracy can such Ethernet chips obtain? Any particular models you can reccomend? It seems to me that that might be a more cost effective solution as I expect such chips are made in much higher quantities.

You think the signal from the photodetector won't be clean enough to trigger a Ethernet chip or an Acam chip? I was thinking if I put a filter in front of it that blocked out everything but the wavelength of the laser it would be OK. If the laser were modulated and I blocked out everything but that frequency coming out of the photodetector that would work alot better I should think, though that would mean finding a modulateable laser which appears to be an expensive task. I have never experimented with this though, so that is truly a guess. I think the reflection may be brighter than what is typically experienced in range finders as I am looking for a very low range range finder, whereas most rangefinders that I have seen are designed for much longer ranges.

Your method using an op-amp integrator sounds interesting. Wouldn't it suffer from the same problem that my circuit suffers from in that it couldn't handle high enough speeds? Note that I mean the measured time

- the time between measurements can be as long as is necessary. (within reason)

Would you care to share your work on the LIDAR? I'm guessing no as it sounds like it may be with a company, but it never hurts to ask :)

Thanks!

-Mike

I see the short time pulse question has been answered. Regarding phase comparison, this can lead to some confusion. There are optical phase comparisons based on the wavelength of the light. In my case, the phase was relative to the modulation signal. That is, you amplitude modulate the laser.

It might make sense to prototype this design with a LED and photodector to perfect the "demodulation" (phase comparison). You wouldn't bother to beam split the light as it is just a LED, so use the modualtion signal. It may be that the beam splitting was needed in the setup I saw since the laser was a helium neon that might have significant time delay relative to the modulation source. [I did this work over 20 years ago.] I would just set up a scope and look at the reflected signal first before designing the phase detector.

In the laser version, you would optically filter the light falling on the photodector. That would reduce the ambient noise.

discontinuity.

and

low

circuit

then

phase

with

well 10mbit ethernet was clocked at 10mhz so resolution was limited to this clockrate, however you can get 100mbit and 1gbit ethernet these days but ive not used these.

It all depends on the returned power, consider a black plastic surface at an angle, then consider the capacitance of the photodiode and amplifier and how weak your electrical signal will then be, then compare that to the amplifier noise etc.

short range is probably fairly easy, ive seen tank range finder laser diodes wich are pulsed at 1kw. narrow band light filters are available and work well, if a little expensive.

If you modulate the laser with a carrier wave and then pulse that it means you need a very high frequency carrier, wich isnt easy, you might as well then use the phase delay of the carrier.

You can use the laser diode from cheap $5 laser pointers, and drive it with whatever modulation you need. Its advisable to use optical power sensing feedback if the diode has an inbuilt photodiode, otherwise a constant current might be too low for it to work yet might also be too high and destroy it.

It can handle high speed as long as you have a fast op amp and pulse detector logic, as it relies on continous stream of pulses at a constant rate to generate an average. but as I said any noise or mising pulses might cuase exagerated errors.

measuring individual pulses then averaging them might be better as you can discard ones that are way out.

another alternative is to use an old HP 5328 timer counter, it can average pulse timings down to very high resolution.

Ive discussed most of it here and there anyway, Yannick is also working on similar lines. you might be able to find more if u search for "heterodyne lidar" The difficult part for mine is getting it to work >1ghz, if you stick with say