# IR laser detector circuit

Discussion in 'The Projects Forum' started by mossman, Dec 31, 2013.

1. ### mossman Thread Starter Member

Aug 26, 2010
131
3
I need to design a simple IR detector circuit to measure the time delay between when a command is sent to a laser cutting machine (via a custom C# program) and when the laser actually fires. I'll be connecting channel 1 of a scope to the trigger input on the laser power supply and the detector circuit on channel 2 to determine the delay. I'm looking at the laser specs and it says the IR beam has an energy of 0.2 mJ with a pulse width of 4 ns. Is there any way to safely and accurately measure this pulse or is it too much power and/or too fast? According to my calculations, that is 50,000 W per pulse and the minimum repetition rate is 1 Hz, so that's 50,000W for 500 ms. We have some PIN diodes available (BPV10NF), but I'm not sure this is even feasible. It is a laser cutting machine after all.

Last edited: Dec 31, 2013
2. ### MikeML AAC Fanatic!

Oct 2, 2009
5,450
1,066
Just mount the IR detector a few cm off to the side; away from the beam. There will be more than enough backscatter to detect. You can change the aim of the detector while viewing the detector signal on an oscilloscope. Make your timing measurements after you have a signal amplitude which doesn't saturate the detector.

3. ### mossman Thread Starter Member

Aug 26, 2010
131
3
Wow, thanks. Why didn't I think of that?

Am I correct in saying that a laser putting out a 0.2 mJ with a 4 ns pulse is generating 50,000 W? If so, how is that possible? The power supply is only 100W. If the peak power of a pulsed laser = E/t and average power is E x repetition rate, then that gives me 50,000 watts peak power and 200 mW average power. Is it really generating 50,000 W over the 4 ns period but since it is over such a short amount of time that it barely has an effect on the sample?

Last edited: Dec 31, 2013
4. ### RichardO Well-Known Member

May 4, 2013
1,335
429
Are you sure? If this is a CO2 Laser then the output power of the laser would be only about 1 or 2 watts since the Laser is only about 1% to 2 % efficient. Most Laser cutter/etcher machines I have experience are a minimum of 25 watts.

Solid state Laser diodes are much more efficient but the wavelengths are not as good for cutting and etching plastic -- at least for inexpensive Laser diodes.

5. ### mossman Thread Starter Member

Aug 26, 2010
131
3
It's a YAG laser (EZ Laze 3) with IR and 532 nm (green) lasers.

I bread-boarded a test circuit and it seems to be working. However, I'm getting a 100 mVpp 13.9 MHz ringing signal about 2.5 µs after my trigger signal which lasts for about 1 µs, then my comparator is switching about 200 µs later. I'm not sure if the laser is actually firing after 2.5 µs or at 200 µs. What else would cause this ringing signal shortly after my trigger? I feel like it has to be the laser emissions being picked up, but why does my op-amp take so long to switch? FYI, I'm using a 741 op-amp to amplify the photodiode signal then feeding the 741 output to the inverting input of a comparator. Non-inverting input is connected to a voltage divider that sits at 2.5V (half of Vcc). The BPV10NF IR photodiode is a high speed diode and the comparator has an advertised response time of 300 ns so I don't understand why it takes so long to switch.

Last edited: Jan 3, 2014
6. ### RichardO Well-Known Member

May 4, 2013
1,335
429
The 741 op-amp is verrrrry slow. It is only useful at audio frequencies, at best. You will need a much faster amplifier for detecting pulses in the range of 1 us.

Similarly, the comparator may not be fast enough. A comparator with a 300 ns response does not have much gain and bandwidth to spare for 1 us pulses. Note that the response of a comparator is normally specified with a fairly large "overdrive" amplitude at the input. The overdrive is usually 100 mv. At larger overdrives, the comparator is faster the the spec. and at smaller overdrives the comparator is slower.

If you really need to detect a 4 ns pulse then you will need a really fast amplifier and comparator. Off the top of my head, you will need an amplifier circuit with a bandwidth of at least 100 MHz and a comparator with a response time of, maybe, 1 ns. That means an op-amp with a gain-bandwidth of several GHz and an ECL or CML comparator. This is quite challenging.

You can try looking at the Analog Devices, TI or Linear Technology web sites to find fast amplifiers and comparators.

7. ### mossman Thread Starter Member

Aug 26, 2010
131
3
...but the op-amp is detecting the 4ns pulse because its output is switching, albeit after 200 µs, but nonetheless it is switching. I'm in contact with the manufacturer trying to determine what the other, preceding signal is. I think there's a shutter that opens just prior to the laser firing that may be the cause.

Last edited: Jan 6, 2014
8. ### RichardO Well-Known Member

May 4, 2013
1,335
429
Trust me on this. The 741 has absolutely no chance of amplifying a 4 ns pulse. The 741 won't amplify a pulse shorter than a few microseconds. Look at the data sheet -- the open loop voltage gain is only about 1 at 1 MHz.

Either the pulse is much longer in duration than 4 ns or the signal you are seeing is not coming from the 741. What is the duration of the LASER pulse you are trying to detect?

9. ### mossman Thread Starter Member

Aug 26, 2010
131
3
I believe you, I'm just trying to determine why my comparator is switching. Is it possible that the laser is so powerful that it is saturating the photodiode and there is a residual signal present for long enough to trigger the circuit? I am using the laser in single shot mode, which is a single 4 ns 0.2 mJ pulse (50MW peak power).

Last edited: Jan 7, 2014
10. ### DickCappels Moderator

Aug 21, 2008
2,753
665
RichardO is on the target about that opamp.

Your comparator might be switching because of EMI picked from the Laser's driver, possibly a flash tube or some other circuit with a very fast risetime.

Your BPV10NF diode has a rise and fall time of 2.5 ns. That should be the bottleneck of your system bandwidth, because even with that 2.5 ns, you won't really see full amplitude before the signal starts dropping again.

If you want to amplify the output of the PIN diode with a transconducatnce amplifier, you would need a gain-frequency product of several GHz, and that would get very tricky very quickly. The fewer parts you have and the shorter the signal processing chain, the better your chances of success.

If you have a fast enough scope (I guess 100 MHz minimum but 300 Mhz better) with a 50 ohm input and sensitivity a few mv/division, then you can bias your PIN diode to 10 volts or more for low capacitance, and capacitively couple the diode's current to the scope's input termination resistor using a minimum of wire for low capacitance and see the signal as the IR drop across the resistor that way.

11. ### mossman Thread Starter Member

Aug 26, 2010
131
3
I have to believe that although the pulse width of the laser is 4 ns, there are several pulses per shot, not a single pulse as the manual implies. I can see the laser spot when it fires so it has to be present for much longer than that in order for the human eye to see it. With a repetition rate of 1Hz, my guess is that the laser is really on for 500 ms, assuming a 50% duty cycle. This would equal 62.5 million pulses per cycle (500 ms/8 ns).

12. ### RichardO Well-Known Member

May 4, 2013
1,335
429
Considering the low efficiency of LASERs and the 100 Watt power supply that you mentioned earlier, I suspect that your duty cycle is way less than 50%. I am assuming that the YAG LASER is the high power one of the 3 (YAG, IR and green) you mentioned. Do you know what the efficiency of yourYAG LASER is?

I agree that from what you are seeing that it is a large number of pulses per "shot" but probably no where near 62 million. My laptop battery is getting low so I will have to do a calculation to estimate the number of pulses later...

13. ### mossman Thread Starter Member

Aug 26, 2010
131
3
The laser is a Q-switched YAG laser with green, IR, and UV options installed. We are primarily using the green laser (532 nm). I believe it is about 30% to 40% efficient. Depends on the selected wavelength and microscope objective (did I mention it is mounted to a microscope?). Whatever the duty cycle, the beam is present long enough to be seen by the naked eye so it has to be on the order of 10's of milliseconds right?

Just to be certain I am understanding things, is the following example theoretically accurate? I realize the cycles per pulse/shot is nonsensical, but I felt like including it.

Lambda = 532 nm = 536 THz, T = 0.00177 ps
Laser pulse width = 4 ns = 2,144,000 cycles per pulse
Assume 50% duty cycle pulse, T = 8 ns
Assume 30% duty cycle 1 Hz repetition rate (Ton = 300 ms, Toff = 700 ms)
Number of laser pulses per shot = 300 ms / 8 ns = 37,500,000 or 80,400,000,000,000 cycles per shot

Last edited: Jan 10, 2014
14. ### RichardO Well-Known Member

May 4, 2013
1,335
429
I think we got distracted from your original needs. I am partly to blame, sorry.

Here are some questions that I think need to be answered.

A. How small a delay do you expect after telling the LASER to fire and it actually fires.

B. How small/fast do you want this delay to be?

C. In your present setup, I wonder if the detector is integrating the LASER pulse and "storing" it long enough for the 741 to amplify it.

Other questions to help find out what is happening:

1. Is there a signal detected even with the light sensitive diode covered with an opaque mterial?

2. If a signal is detected with the sensor covered then where is the interferene getting into the circuit.

3. Once the circuit is sensing light and not interference then how fast does the detector have to be.

15. ### mossman Thread Starter Member

Aug 26, 2010
131
3
Manufacturer says 180 µs, I am measuring 200 µs.

Doesn't matter. It just needs to be consistent so it is predictable and repeatable.

That's what I was thinking too, but the laser is present long enough to be seen by the naked eye, so doesn't that mean it is present long enough for the diode and 741 to detect it?

I'll have to check again.

TBD if applicable.

Doesn't matter. We just need to know what the delay is. It's apparently between 180 µs and 200 µs.