SparkFun Forums

[SOI_Sentinel]

Ages and ages ago I had an ongoing discussion here about doing a laser rangefinder. I haven't had time to revisit it, but I decided to do some digging now that I'm older and wiser.

Original discussion here:

viewtopic.php?t=4147

I've got schematics for an old TI optical phase comparison rangefinder. No images right now, but here's the concept:

Take a local oscillator (say 10MHz), use it to drive a light system (say Sparkfun laser or equivalent).

We also need a reference oscillator (say, 9.999MHz).

We take two signal paths. One taps the local oscillator directly, the other is driven by an automatic gain control op-amp reading from a PIN photodiode that should be recieving the return from the laser. We might also include an analog programmable delay line to allow for zero-tuning of the system (via digital pot?).

So, each 10MHz signal is mixed with the reference signal and low pass filtered, resulting in two 1KHz signals. By converting these to zero-crossing square waves, we can feed them into an MCU with hardware capture (I prefer dsPICs) and the resulting timing difference can be converted into fairly high resolution distance. I calculated that 1LSB of a full speed 16 bit capture timer would be approximately 1mm. Now, whether we can hit this or not is a different question.

Now, the hardware idea.

The idea I have is to modernize the design, and make it completely software programmable. Everything can be changed by the controller MCU to allow different frequencies to be tested.

The driving MCU provides a clock out to one or several PLL's. We need to generate two clocks that have a fanout of 2 units. I was looking at the TI CDCE925 for this, but I can't get a 10,000:9,999 divider ratio out of it. Best I can do is 511:512, so I need another idea for this. The original TI unit used two roughly 4MHz crystals cut to be 80 Hz apart as reference.

I have a 3 op-amp DC to 25MHz mixer. http://electronicdesign.com/Articles/In ... leID=11935 Upgrade the op amps to LMH6644's and it should work fine. Analog devices has a mixer that should work fine, but I think would be overkill. AD8343

One reference for an accurate delay line I found online is two Schmidtt triggers with a low pass RC filter in between. By varying the resistance (digital pot!) you vary the delay of the signal. Much cheaper than a Maxim IC.

Before the mixers is usually a high pass filter, after the mixers a low pass filter. I'd want to use digital pots or otherwise programmable filters for this.

1MHz operation (which I could get 2KHz references from) would provide me with a 75m half phase range. As this isn't an analog summing circuit (XOR of the two signals), I could get 150m out of it. This is obviously NOT tested. I'd need to test. My original plan was to allow 25MHz (3m half phase) down to 1MHz to be programmed in, with a 1KHz default mix frequency. This would have given me sub millimeter accuracy at 25MHz and centimeter accuracy at 1MHz. Now I'm looking at 2cm accuracy as best case. Not exactly happy.

Any ideas? I'll post the schematics as soon as I get access to a scanner.

Any ideas on parts?