Feeler: Integrated Laser Diode Testing Environment

So, here's an idea that I've had kicking around for a while.

What I'd like to assemble is an integrated laser diode testing environment for use when a new burner diode comes out and must be murdered, or when we just want to test a laser diode. Take the (not so) recent 8x murder test for example.

As most of you know, an 8x Blu-ray burner diode was sent to IgorT to be tested... almost two years ago. He had an environment set up that would count 60-second on/off cycles and he would take the laser diode out of the environment to do P/I tests every ten hours or so.

I'd like to add a new twist to this. Here's how it would work under my model:

During the on cycle, we keep track of input current, voltage drop, and power output (efficiency will be derived from this data in real time). When the run cycle completes, that data is dumped to another microprocessor. It in turn generates an HTTP post and sends the data up to my web server while the next on cycle completes. The web server records the data into a database as a new 60-second cycle.

Here's the fun part: all data would be publicly accessible during the test. This would include full lifetime graphs of any of the data, degradation information, and complete information of the last cycle. This would allow the community to be constantly up-to-date with the progress of the test and how far the laser diode has degraded.

This system would also allow the automated collection of a P/I curve at precise intervals. If I recall correctly, this was a time consuming process for IgorT. This test would happen in a matter of seconds and the test would continue on its merry way.

----------

I think this would be a really useful tool for the community to have at its fingertips, but it's not a small amount of effort. I'd like to gauge interest in it to find out if the community at large thinks this is a good idea.

The most time consuming portion of the work is going to be in the construction of the interface that allows the complete test data to be browsed. If this doesn't sound worthwhile, I'll just cut the actual delivery out and make this a personal project.

It's also going to be kind of expensive in terms of the other equipment I'd pick up for it. Again, if it doesn't seem worthwhile, I'll simplify it and make it a little easier on my wallet.

----------

Thoughts? Criticism? Trolls?

-Trevor

Great idea! An easier way to automatically test LD's could allow us to do more detailed testing than would be practical otherwise. It could provide for easy, standardized testing for new LD's, range-testing over a # of devices, allow comparison testing of different manufacturer variants, even provide more exhaustive/modern testing of old standbys! (How far can an 815 really be pushed?)

People could even donate odd diodes they wanted tested. You might even be able to offer non-destructive "sweet spot" testing for members for free+shipping, or at a minimal cost (IIRC, IgorT offered something like this when he occasionally sold LD's for a worthy cause?)

When the run cycle completes, that data is dumped to another microprocessor

Click to expand...

Not sure a microprocessor would be the best bet to perform the actual testing? - a microcontroller would probably be better suited to this task. (I have been working with the former since the venerable 8080, but am really impressed with what some of these microcontrollers can do, particularly with regards to h/w i/f & control apps like this!)

It in turn generates an HTTP post and sends the data up to my web server while the next on cycle completes. The web server records the data into a database as a new 60-second cycle.

Click to expand...

This is in general a good approach, but might I offer a suggestion? Rather than having it act as simply an input device for the server, put initial data storage on the device itself, with periodic database updates to the server, and use the server primarily for data analysis & display.

This way, critical test data would not be lost as a result of any Internet connection glitches, server re-boots, etc. These kinds of issues are of particular concern in this application, given the extensive time-frames (weeks, or even months) of some of these tests!

A simple SD-card i/f could be devised for local data storage - this would provide more than sufficient storage capacity for this application, as well as allow easy local back-up to a PC of the raw data.

The most time consuming portion of the work is going to be in the construction of the interface that allows the complete test data to be browsed. If this doesn't sound worthwhile, I'll just cut the actual delivery out and make this a personal project.

Click to expand...

I might be willing to help-out with some of that, depending on what pgmg platform you are planning on using?

I vote YES!

I love the idea, although to be perfectly honest it is somewhat over my head in how you can implement it...

It's also too much for most casual hobbyists.

I'm sure there is interest... I know I would love to see detailed performance data for some diodes, but the logistics of this... I'm not sure how you can make it work in a way that wouldn't be too much work for you.

The other big question is what would it take to make this project work in terms of financing... I know I'd be willing to contribute $20 but would that be enough if other contributed, how many donations would it take?

Is it possible to find out from IgorT how much it cost him to set things up for his test runs? I know he's been gone due to health reasons, but I don't know the details.

Creating a standardized diode testing environment was my motivation for starting my most recent project: http://laserpointerforums.com/f55/lpm-current-voltage-temp-usb-logger-feeler-63212.html
I prefer using USB + a computer to do all of the data collection and storage, as it allows smaller and cheaper microcontrollers to be used.

You know that this is way over my head. However, I really miss the data that Igor would occasionally put out. I am willing to help with the cost

Peace,
dave

How bout Arduino? It could get data from its digital and analog pins then transmit over usb onto the computer where its graphed and uploaded.

Daniel

Problem is the programming part (for me, at least )

I made something similar times ago, when i "tortured" one of these A140 diodes ..... but mine is all-manual thing (i mean, just displays, no data-logging).

Can be a good idea to add data-logging capability, but my programming "skill" still suck a bit (read: "a lot" )

I have most of my setup connected with GPIB, if I had time I could make something similar. A dedicated setup like this would be awesome.

seoguy said:

...

Not sure a microprocessor would be the best bet to perform the actual testing? - a microcontroller would probably be better suited to this task. (I have been working with the former since the venerable 8080, but am really impressed with what some of these microcontrollers can do, particularly with regards to h/w i/f & control apps like this!)

This is in general a good approach, but might I offer a suggestion? Rather than having it act as simply an input device for the server, put initial data storage on the device itself, with periodic database updates to the server, and use the server primarily for data analysis & display.

This way, critical test data would not be lost as a result of any Internet connection glitches, server re-boots, etc. These kinds of issues are of particular concern in this application, given the extensive time-frames (weeks, or even months) of some of these tests!

A simple SD-card i/f could be devised for local data storage - this would provide more than sufficient storage capacity for this application, as well as allow easy local back-up to a PC of the raw data.

I might be willing to help-out with some of that, depending on what pgmg platform you are planning on using?

Click to expand...

Sorry for the confusion, I will be using a microcontroller. :tinfoil:

The data will be stored locally and posted to the server for each test.

I'll be using PHP to build the browsing interface.

InfinitusEquitas said:

I love the idea, although to be perfectly honest it is somewhat over my head in how you can implement it...

It's also too much for most casual hobbyists.

I'm sure there is interest... I know I would love to see detailed performance data for some diodes, but the logistics of this... I'm not sure how you can make it work in a way that wouldn't be too much work for you.

The other big question is what would it take to make this project work in terms of financing... I know I'd be willing to contribute $20 but would that be enough if other contributed, how many donations would it take?

Is it possible to find out from IgorT how much it cost him to set things up for his test runs? I know he's been gone due to health reasons, but I don't know the details.

Click to expand...

I've got no worries about implementation. The programming is not what I'm concerned about.

The project will be costly, yes. It'll probably end up costing me a couple hundred dollars out-of-pocket. I originally wasn't planning on asking for donations since the project is so interesting.

That being said, financial support would definitely be appreciated.

laser_freak said:

Creating a standardized diode testing environment was my motivation for starting my most recent project: http://laserpointerforums.com/f55/lpm-current-voltage-temp-usb-logger-feeler-63212.html
I prefer using USB + a computer to do all of the data collection and storage, as it allows smaller and cheaper microcontrollers to be used.

Click to expand...

I need serious power for what I'm doing. I figure if I'm doing storage via the microcontroller too, generating an HTTP post was only marginally more effort and eliminated a step from the process. Plus, I won't have to rely on the computer I'm logging to. My host has a ridiculous uptime guarantee, while my local servers do not.

daguin said:

You know that this is way over my head. However, I really miss the data that Igor would occasionally put out. I am willing to help with the cost

Peace,
dave

Click to expand...

I hope to be able to deliver more data than we could ever dream of needing. Thanks for your support.

danward said:

How bout Arduino? It could get data from its digital and analog pins then transmit over usb onto the computer where its graphed and uploaded.

Daniel

Click to expand...

Right now this is based around an Arduino Mega and Uno. Multicoring is quite useful.

----------

I still need a FlexMod P3 and a noise-free solution for powering the thermopile (I'm using an Ophir thermopile - I'm going to try and work with lasersbee or nospin on this). I might employ a datalogger module from lasersbee; I'm also working with him on this.

So yeah. Pricey.

-Trevor

An Arduino is one of the first things that came to mind when I read your thread. Heck, with one of those SD card + Ethernet shields, you could even implement a web server on the device itself if you wanted to! LOL

(Not that I would recommend that over a regular web server, but it's nice to know you've got that kind of capability at your disposal!)

Speaking of which, any particular reason you are considering the much-more-$$$ Arduino Mega over a regular Arduino?

If you're looking to be cost-conscious, it's also worth mentioning that you can get a decent Arduino Duemilanove clone for about $20. The newer Uno has both advantages & disadvantages over the less expensive Duemilanove, but I don't think any of the advantages would apply for this application?

If you are planning on using an Ophir head for sensing, then I would recommend you take a real good look at this thread -

http://laserpointerforums.com/f42/diy-laser-power-meter-26831.html

and this simple PC interface -

http://laserpointerforums.com/f42/diy-laser-power-meter-interface-module-26428.html

Note also in his PC interface design, he is simply feeding that into a PIC microcontroller - you could simply feed that signal into an Arduino instead!

This was designed for use with a Coherent sensor, and something like this should also work quite well with a real thermopile sensor like the Ophir. It is based on the same amp chip used on that Coherent board (indeed, some members extracted that chip right off that board in order to build this device!), the LTC1050 - a low noise, high-performance, zero-drift instrumentation amplifier chip. Of particular importance for this application, it utilizes a chopper-stabilized amp design to eliminate the drift that typically occurs when you attempt to direct-couple an amp to measure a DC signal like from these sensors.

For all its capabilities, this amp is actually relatively inexpensive - less than $5 from the usual sources.

This amp design was quite popular at the time, and there may even be a member here who has an old one that would be willing to donate it to the cause!

I also like the idea of using open-source hardware design (just like the Arduino itself) for a community-supported project.

The greatest benefit is that this provides you the raw data directly to the Arduino for processing / massaging. You do know something about power measurement firmware, don't you?

Since the OPHIR 20C-A thermopile has a built in amplifier, it might be easy to use an arduino with a >10 bit ADC. Alternatively you could use a SPI ADC in combination with an arduino.