Currently, I'm working on the narrow tunable laser and log the power value through PC. My LabView program will sample the power value as the tunable laser is sweeping. Thus, I want to know how fast the laser can sweep and how fast the power sensor module will update as the power values changed?
Besides that, what is the faster sampling time for power sensor module?
I guess you're working with the 81949A laser that you mentioned earlier? This model is equipped for stepped-sweep measurements, where the laser stops at each wavelength step for a measurement and then continues to the next step.
You can configure the sweeps with the start and stop wavelengths, the step width, and the dwell time during which the laser stays at each wavelength point.
You can probably speed up the measurement you are making by using the logging function of the power meter instead of reading each value with a separate command.
The logging function is started with a selection of the number of points in the measurement (needs to match the number of points in the TLS sweep) and the averaging time of the power meter. For the 81635A, the minimum averaging time is 100us. You will probably want to set the sweep dwell time to match the chosen averaging time.
The power meter input triggering should be set for making a "single measurement" at each trigger, and the TLS output triggering should be configured to trigger at each step. Finally the 8163B standard trigger configuration should be set for "loopback" so that the trigger from the TLS goes to the power meter.
After the measurement is completed, you can read out all of the logged power values with one command. All of this can be programmed at the SCPI command level, but would be easier using the 816x PnP driver.
Comments: 1. this is mostly to help program development, in case the previous PWM logging function is not finished. The power meter is here in slot 3. 2. this sets the mainframe trigger to â€œloopbackâ€ so the TLS step triggers are routed to the PWM 3. this sets up the PWM, the range and wavelength need to set here 4. this sets the TLS to output triggers after each wavelength step 5. this sets the PWM to use triggers for individual measurements 6. this turns the laser on 7. this sets up a sweep for a TLS in any slot (in this case #1) 8. this activates the PWM logging function with the chosen averaging time, measurements wait for the triggers 9. this starts the TLS sweep 10. this waits for the end of sweep 11. this gets the results from the PWM 12. this turns off laser 13. this sets PWM triggering back to default
Hi Mike. I really appreciate your details explanation and it did work for my work now. Thank you again.
However, I have few questions as follows. 1. How to optimize the sweeping period to scan a certain range of wavelength? Is it the only way is try and error? 2. Can I know what is the purpose of Power Range? When I run the experiment by varying the range, and it seems increase the time of sweeping.
I'm using the 8163B, 81635A, and 8194A to run an experiment and the objective is to acquire the current state power value as respect to FBG reflective wavelength. I manage acquire the power value, but I'm not sure how fast the response time of the power respect the changed of reflective wavelength changed. Is it 100us?
Hi Sychong, I'm glad this helped. To your questions:
1) I guess the main parameter you can choose, after setting the desired wavelength range and step for the required resolution, is the power meter averaging time. The dwell time should match the averaging time.
Generally shorter averaging times are used for faster measurements and longer times to reduce noise with increased averaging. When stepping the 81949A, I think the TLS uses 80-100ms to set each wavelength point. So using dwell times much less than this will not significantly reduce the total measurement time. Maybe you want to try settings in the range 10-100ms for averaging and dwell time, depending on the signal and noise level.
2) The power range sets the photodetector signal amplification to set the highest measurable power level. So a spectrum with maximum -8dBm can best be measured using the 0dBm range. Setting for example the +10dBm range can also measure this signal but the noise level is higher so the minimum measurable signal will not be as low (less dynamic range). I don't expect the power range setting to influence the TLS sweep time, but ranges with higher sensitivity do have stronger analog filtering that reduces the bandwidth. I think this addresses your question regarding "response time". The bandwidth of the 81635A is: 5kHz in the power ranges +10 to -20dBm 3.7kHz in the ranges -30 and -40dBm 1.5kHz in the ranges -50 and -60dBm
So the higher power ranges match the 100us minimum averaging time quite well, depending on how fast the signal level changes. For weaker signals, a longer averaging time may be useful to avoid distortion and in any case can match the TLS sweep time better as mentioned above.
Yes, that's right. Continuous sweep measurements are not supported with the 81949A. The 81940A, for example, can do this.
So the sweep speed is not settable. The 81949A will change wavelengths as fast as possible. The maximum speed is about 50nm/s, but for short steps most of the time is starting and stopping.
Of course you could make a sweep with only one step between start and stop, which would then move continuously, but it would be very difficult to determine the wavelength at specific times during the sweep.
Thanks for the good questions. Mike
P.S. I will try to give this topic a more specific name.
Actually, I'm trying another approach of doing the sweeping process as shown in attached picture.
In this design, I'm using Set TLS Wavelength and Read PWM Values, and then it will continuous increase the wavelength at set step sizes by using looping.
The recorded sweeping times seem faster than the suggested approached in the previous post. However, sometime, the continuous sounds is sounding from the TLS device (like relay switch on sounds). I'm not sure it will shorten the life time of the TLS or not if use for long term.
Hi Sychong, It's interesting if this is going faster. Your loop to run through the wavelength steps is pretty much what I would expect the TLS to do internally with the stepped sweep command. But maybe the stepped sweep is doing something more that increases the time.
Anyway it doesn't look like the program should do anything harmful to the TLS. You might want to make sure that a new step command is not started before the TLS has finished the last step, but I would guess that the get-wavelength call would prevent that.
The "relay" sound might be coming from the power meter if it is autoranging. You could set a fixed power range and see if the sound stops occurring.
With the two approached that we discussed in previous posts, the completed sweeping time is it always consistence? Because I can't get it right, an example, 1st few consecutive sweeping time collected are consistence, but later, the next few consecutive sweeping time will change to be very slow. From the observation, it seems for me that the completed sweeping time cannot be predicted when it will maintain the sweeping time.
This observation was found when the experiments were carrying out in the following setting: 1. Averaging time = dwell time = 100 ms 2. Averaging time = dwell time = 100 us 3. Averaging time = 100 ms, dwell time = 0 s 4. Averaging time = 100 us, dwell time = 0s for the range 1 nm (step size: 100 pm) , 800 pm (step size: 100 pm), 400 pm (step size: 20 pm) in LabVIEW platform.
I'm not sure is it because only happen in LabVIEW? or also happen in VEE or other platforms (C++, VBasic)?
The stepped measurements may not always complete in the same time, since the timing is determined by the TLS regulating itself to each wavelength position. So I don't think a variation is surprising here. But if as you say the time becomes very slow, you may need to analyze where this is happening in the program. I wouldn't expect anything special due to the programming environment. If you are using the sweep and logging routines, then the power meter will run in fixed range mode, which is good. But if you are reading each power point individually, you could have the meter set for autoranging and that can increase the time if the power is changing between points.