Tektronix Technical Forums are maintained by community involvement. Feel free to post questions or respond to questions by other members. Should you require a time-sensitive answer, please contact your local Tektronix support center here.

2520 Pulsed - heavy overshoot in pulsed waveform

Post Reply
Robert
Posts: 8
Joined: March 7th, 2011, 12:54 am
Country: Singapore

2520 Pulsed - heavy overshoot in pulsed waveform

Post by Robert » March 7th, 2011, 2:04 am

Hi,

We are using the 2520 for thermal resistance measurement. The 2520 is just calibrated from Keithley very recently.
We experienced heavy overshoot in pulsed waveform output of the 2520 when we used the following settings as said in the 2520 manual:

1. 500mA setting ; 10 Ohm 1/4 Watt load resistor ; 0.5us pulse width and 20us off time -> 48% overshoot (We expected it around 1%), rise/fall time ok
2. 500mA setting ; 10 Ohm 1/4 Watt load resistor ; 5us pulse width and 1000us off time -> 59% overshoot (We expected it around 0.1%), rise/fall time ok
3. 5A setting ; 1.8 Ohm 1 Watt load resistor ; 0.5us pulse width and 20us off time -> 0% overshoot but no more a square waveform (We expected it around 1%) ; rise time = 365ns, fall time =75ns
4. 5A setting ; 1.8 Ohm 1 Watt load resistor ; 5us pulse width and 1000us off time -> 12% overshoot but got damping oscillation when return to zero (We expected it around 0.1%) ; rise time = 250ns, fall time < 100ns

The waveforms are observed on an oscilloscope with probes across the load resistor.

I am looking forward to receiving help on this matter. Thank you very much.
Robert

Dale C
Keithley Applications
Keithley Applications
Posts: 2841
Joined: June 10th, 2010, 6:22 am
Country: United States
Contact:

Re: 2520 Pulsed - heavy overshoot in pulsed waveform

Post by Dale C » March 7th, 2011, 5:29 am

Are you using the supplied 15 ohm coaxial cable to connect to the load resistor?
This is very important.
Here are a few more important considerations.

When making connections to the laser diode, observe the following considerations to
avoid pulse degradation due to distributed inductance and other effects:
• Use only the supplied 15Ω coaxial cables.
• Keep cable lengths to a minimum.
• Connect the four cable shields together at the DUT.
• Carry cable shields as close to the DUT as possible. Minimize the length of
exposed (unshielded) signal lines.
• Dress the VOLTAGE SENSE cables as far away from the CURRENT OUTPUT
cables to avoid magnetic coupling. Twist the SENSE cables together to further
reduce magnetic coupling.
• Connect the VOLTAGE SENSE leads as close to the body of the DUT as possible.

Robert
Posts: 8
Joined: March 7th, 2011, 12:54 am
Country: Singapore

Re: 2520 Pulsed - heavy overshoot in pulsed waveform

Post by Robert » March 8th, 2011, 12:05 am

Thank you for your answer.

And yes, we used the 15ohm coaxial cable (minimum length) which came with the 2520. The actions that you recommend seem to only affect the rise/fall time. The rise/fall time in our tests are ok, the only thing that concerns us are the overshoots.

In addition, we got our own circuit that can transform the pulsed voltage waveform from a waveform generator to a pulsed current waveform.
Here are the images to compare the output from 2520 and our device (same cable, same load resistor, same 300mA output current, same pulse width 500ns, same duty cycle 1%, same oscilloscope).

You can see that the overshoot is heavy especially with 2520. Is there anyway to reduce it?

Thank you in advance.
Robert
Attachments
Box, 10ohm load, 200ns, 1V scale.png
Output from our device with 10ohm load, 300mA current
Box, 10ohm load, 200ns, 1V scale.png (7.73 KiB) Viewed 31203 times
Keithley, 10ohm load, 200ns, 1V scale.png
Output from 2520 with 10ohm load, 300mA current
Keithley, 10ohm load, 200ns, 1V scale.png (8.82 KiB) Viewed 31203 times

Dale C
Keithley Applications
Keithley Applications
Posts: 2841
Joined: June 10th, 2010, 6:22 am
Country: United States
Contact:

Re: 2520 Pulsed - heavy overshoot in pulsed waveform

Post by Dale C » March 21st, 2011, 12:24 pm

Not sure here...
I will check with design for help and then post the result.

Robert
Posts: 8
Joined: March 7th, 2011, 12:54 am
Country: Singapore

Re: 2520 Pulsed - heavy overshoot in pulsed waveform

Post by Robert » March 23rd, 2011, 10:50 pm

Hi,

Thank you for your help. I will wait for your advices.

Robert

Robert
Posts: 8
Joined: March 7th, 2011, 12:54 am
Country: Singapore

Re: 2520 Pulsed - heavy overshoot in pulsed waveform

Post by Robert » June 2nd, 2011, 2:38 am

Hi,

It has been nearly 3 months. Could you please give us update? Thanks a lot

hamed.alavirad
Posts: 18
Joined: February 6th, 2015, 11:54 am
Country: Canada

Re: 2520 Pulsed - heavy overshoot in pulsed waveform

Post by hamed.alavirad » March 9th, 2015, 12:19 pm

We are using your Pulsed LIV testing system 2520 for characterizing our Laser Diodes under pulsed operation. We are applying Continuous-Pulse current (3 nm pulse width and 300 nm pulse off-duration) and trying to manually check the photocurrent of detector port coming from a photodiode placed in front of the laser.

1-How is the integrated photocurrent (or Power) evaluation done under pulsed operation in your 2520 system? Is there anyway I can measure that manually?
2- For LIV test, do I need to necessarily trigger the pulse in continuous mode?

PS: Once we trigger the pulse using the buttons Config ===> Trigger ===> COUNT ===> Continuous-Pulse , I would never be able to see the display monitor to see the detector?

Thanks

Dale C
Keithley Applications
Keithley Applications
Posts: 2841
Joined: June 10th, 2010, 6:22 am
Country: United States
Contact:

Re: 2520 Pulsed - heavy overshoot in pulsed waveform

Post by Dale C » March 10th, 2015, 7:14 am

1) Maybe the pulse time is meant to be us (micrcoseconds)?
2) The pulse photo-current is measured by taking samples near the end of the pulse. Look at the below note.
3) If the pulses are short (sub <100ms), I don’t think there’s any way to measure it with a separate instrument, if that is what it means by “measure that manually”. See #5.
4) I don’t think so. The continuous trigger mode is meant for a constant stream of pulses (all at the same settings)
5) Just doing a single pulse from the front panel should display the laser diode voltage, current and photocurrents. Also, stopping the output after a stream of pulses (using continuous pulse) should show the last set of readings.

When controlling the 2520 from a computer, I strongly suggest using the trace mode to view the waveforms (signal vs time) for the LD voltage, current and photodiode current before relying on any measurements. If the pulse shapes show ringing or significant overshoot, optimize the cabling and verify the improvements. If using a 3rd party photodiode, ensure that the response time is sufficiently fast for the desired pulse widths.



Note for item 2...
Where in the pulse is the measurement made?


The 2520 measurement algorithm was designed to prevent problems with the pulse shape causing problems in the reported measurement.

The algorithm described below is used for the laser diode V and I measurements, as well as the photodiode current measurements.

The algorithm starts at the last sample in the pulse (before the pulse fall time begins) and looks back towards the start of the pulse. Using the last A/D sample in the pulse as an “anchor,” it includes samples from the pulse top that are within a certain tolerance (I don’t recall the exact details, but it is something like +/-2-3% or a certain number of A/D counts), calculating a mean (average) for all points within that window.

hamed.alavirad
Posts: 18
Joined: February 6th, 2015, 11:54 am
Country: Canada

Re: 2520 Pulsed waveform

Post by hamed.alavirad » March 10th, 2015, 9:38 am

Thanks Dale,

Some comments on your respond:
1) Pulse time is uS (microsecond), sorry for typo...
2) About this question Note for item 2...Where in the pulse is the measurement made? I myself have no idea and I am just using Keithley, but for me the important thing is ONE point for any every pulse no matter where that point is (Top of the pulse or any other points in its falling time or raising period)
5) I would be thankful if you guide me in about this topic:
- We are not going to measure in single OR Finite pulse mode. We ant to have data using Infinite chain of pulses. But you are right, if I send a single pulse, I can see some values in the scree, but in the CONTINUOUS mode I can not
- I changed my method and used INFINITE count for the pulses and I am now seeing values for voltage and photocurrent. Everything seems be fine but a new problem raised up. It seems data acquisition between detector pulses (my detector is not fast though) and Keithley is not perfect, also the value for photocurrent, being showed by Detector 1 channel, is not fixed and varies while the pulses get ON and OFF. I must find a way to sync the pulses and data acquisition of Keithley and I want to have a constant value for photocurrent in a specific time of every pulse.
I have attached two phots of my Keithley taken in two different times. As you can see one shows 1.74 mA and the other shows 3.84 mA for photocurrent.
20150310_112443.jpg
20150310_112443.jpg (1.85 MiB) Viewed 28302 times
20150310_112503.jpg
20150310_112503.jpg (1.88 MiB) Viewed 28302 times
Thank you for your considerations.
Last edited by hamed.alavirad on March 10th, 2015, 11:55 am, edited 1 time in total.

Dale C
Keithley Applications
Keithley Applications
Posts: 2841
Joined: June 10th, 2010, 6:22 am
Country: United States
Contact:

Re: 2520 Pulsed - heavy overshoot in pulsed waveform

Post by Dale C » March 10th, 2015, 11:38 am

The model 2520 cannot output or measure a pulse that fast. (3nsec width with a 300nsec pulse off time).
The minimum pulse is 500nsec(0.5usec).
That is why you not getting what you expect.

hamed.alavirad
Posts: 18
Joined: February 6th, 2015, 11:54 am
Country: Canada

Re: 2520 Pulsed - heavy overshoot in pulsed waveform

Post by hamed.alavirad » March 10th, 2015, 11:45 am

Thanks Dale...
I corrected my message, all units are on MICRO (3 uS pulse width and 300 uS pulse duration) . The photos in the last message are still valid. Now I am wondering what is wrong with my measurements and why I can not see a constant value for photocurrent.
I am thinking about some kinda synchronization between pulse and Keithley but not sure how much I am right.

Dale C
Keithley Applications
Keithley Applications
Posts: 2841
Joined: June 10th, 2010, 6:22 am
Country: United States
Contact:

Re: 2520 Pulsed - heavy overshoot in pulsed waveform

Post by Dale C » March 11th, 2015, 10:39 am

Your question about “data acquisition between detector pulses (my detector is not fast though) and Keithley is not perfect” is reasonable, but actually is not possible with the 2520.

There are separate signal paths, and dedicated A/D converters for the measurements. All measurements are synchronized by design.

Remember, the measurements are made near the end of the pulse (for all measurements).

There could be many reasons why the readings are varying. I see that the one photo you took of the “3.84mA” looks like a superposition of 2.84 and 3.84. If so, that means the photocurrent is reporting values from 1.74 to 3.84 mA.

This indicates a problem with the connection to the laser diode or the detector. And/or an inappropriate detector (one that has a response that is too slow or doesn’t settle quick enough). The problem is NOT a synchronization issue within the 2520 (assuming the 2520 is not broken).

Please follow the suggestions made in the previous email about the connections and looking at the pulse waveform (critical for these types of issues!).

When looking at the pulse waveform, it will probably show some problems like overshoot, ringing or even a lack of settling.

Is the cabling between the 2520 mainframe (box with the display) and the black testhead are fully seated. The grey cable is quite stiff and must be allowed to lay straight out from each box (at least 1 foot), with no lateral strain on the connector/box coupling. Having the cable turn quickly after exiting the testhead can cause a poor connection and results in undesirable, intermittent instrument performance.

Lastly, the infinite pulse mode is intended for troubleshooting or adjusting a setup. Although the data is displayed, it is not buffered (like a DMM/switch instrument), so there is no way to get all of the data. In the continuous mode, it would be possible to periodically get a reading (by using a single GPIB measure command, it will just take measurements that are recently made).
Hope this helps.

Post Reply

Return to “Other Keithley Instrument Products”

Who is online

Users browsing this forum: No registered users and 3 guests