- Joined
- Dec 3, 2008
- Messages
- 70
- Points
- 0
No more updates on the experiment? :O
Welcome to Laser Pointer Forums - discuss green laser pointers, blue laser pointers, and all types of lasers
Buy Site Supporter Role (remove some ads) | LPF Donations
Links below open in new window
FrozenGate by Avery
12x Murder Experiment - Take 2 (With 12x Comparison Plots)
- Thread starter IgorT
- Start date
IgorT
0
- Joined
- Oct 24, 2007
- Messages
- 4,177
- Points
- 0
So, as i said i wanted to test an even higher power next. I felt confident enough this prime diode wouldn't just explode like the previous ones.
Luckily i was right!
But first i had a little problem to solve. Ambient temperatures started changing to the point, that i was unable to just let the diode warm up to the level it was at before.
I had to rig up a constant temperature setup. I had most of the parts already here from various old experiments, and i relied on a neat little chip from MAXIM (MAX1968) to do the peltier driving.
The chip is a high efficiency switcher that drives the TEC with PWM and can automatically change between cooling or warming to keep the temperature the same all the time. I managed to hook it up to the tiny SMD thermistor i already had thermal-epoxied to the back of the diode for temperature equalisation during earlier re-plots.
It works really nice and makes re-plotting so much easier, as i don't have to "chase" the temperature anymore.
Anyways, i decided the next step would be 600mW. I used the last step slope efficiency to calculate it would require 423mA, altho i expected a small error, because i know that slope efficiency drops a little with every step above a certain current (first it climbs, then drops).
Next came the scary part. I actually had to plot the diode to 600mW.
The little test diode reached exactly 600mW at 424mA. This is 4mA past the point where the first diode went all zombie-faced, but this one is still working fine.
First, here is the Step 4 "Zero-Hour" Plot:
The efficiency of this diode has dropped quite a bit by now, but apparently if it starts out as a good diode, it stays a good diode.
Then i let the diode cycle for 4h...
Here's the Degradation Speed Comparison @600mW:
After the initial 4h, i measured a total of 3mW Po loss at 424mA. The diode dropped to 99.5% of it's initial Po of 600mW.
But as expected, this was a minimal power loss, which didn't really change the fact, that the degradation speed comparison plots are pretty pointless, due to the much more prominent "burn-in" effect with 12x diodes.
The only interesting thing that happened is, that this time degradation was actually faster than at the step before it. You can see that the pink line of 600mW drops faster than the light-orange line of 550mW.
This probably means, that this last 50mW step up in power was harder on the diode than previous steps of equal size. I expected something like this since we are slowly nearing the diode's limits, i just didn't know how it would show.
Then came time to decide what to do next.... Problem is, all this testing sure as hell wasn't telling me.
It's a little frustrating that i have to multiply the degradation of every step with all the steps before it to get something that looks half way as if it makes sense, but means very little...
Degradation Multiplied:
This is what i was hoping to get when i saw the results of the first step, but unfortunatelly the diode won't play ball.
I thought multiplying would show me the worst-case-scenario, where degradation would come out even more pronounced than it can possibly be in reality. I thought that if i keep the last multiplied line less steep than that of the second 8x i would be safe, but at 600mW i'm just not sure anymore.
So what to do next? Well, i think that at 600mW and the diode still going strong, the decision is much simpler.
There were several wishes to go this far, since 600mW is a nice round number.
Obviously if i was asked to make a 600mW laser, i would set it a little higher, so it would stay above 600mW for the majority of it's life.
Perhaps i would even wish to set it as high as 650mW. So i did wish i could go ahead and test even higher, but since i wouldn't really go further than 650mW (which would take some ~25-26mA more or 449-450mA with the already degraded diode), i don't think it's really necessary...
If the diode tests very well at 600mW then clearly it's possible to set it above that. This we'll know from the number of hours of it's total lifetime at 424mA i'm sure.
At this point i'd much rather get "too many" hours at 600mW than too little at 650mW.
That's why i think that having already reached 600mW and apparently safelly at that, the best decision is, to let the diode cycle to death at 600mW initial Po.
The diode is now nearing 10h where i'll do a re-plot. Next one will be at 20h, then i think i'll spread them out to every 20h if degradation remains slow.
Other than that, i wanted to show what a test sheet looks like, in case anyone is interested...
Test Sheet Example (Step 4):
It's a little messy, i know, but i'll explain everything.
First off, there are no currents here, because i only have them in column A, but they go in 40mA steps (0, 40, 80, 120, 160, 200, 240, 280, 320, 360, 400) except for the very last one in this case, which is 424mA.
Normally i would test to the next step and then calculate the 424mA Po, but i didn't want to push my luck, so the Diode Analyzer is set to an Imax. of 424mA. Instead of calculating, i now simply turn the Iadj. pot to the max, and read the power off the meter for analysis.
In the CN column, you can see the measured Vf values for each step, and next to that the Po.
There is some more measured data crammed in, so under the Vf title you can see the Ith reference i have there cos it's changing slightly, and to it's right the temperature during the test.
Pretty much every other value in the test sheet is automatically calculated.
The first example of calculated values is the negative Po at 0mA. It's there to make sure the graphs will cross the X axis at the right point, something that would not happen if i wrote 0mW there. It's calculated from the first two steps, to make sure the plot starts out straight.
To the right of Po you can see Pe or electrical power used by the diode, which is a multiple of current and forward voltage. It is used for Po/Pe plots and to calculate Absolute Efficiency, which again you can see to the right of Pe.
Efficiency is calculated for every step separatelly, and under it you can see the average.
To the right of the absolute efficiency, you can see "Step Size" in mW. Basically it shows me how many mW each 40mA step brought about.
As you can see it climbs a tiny bit at the start and drops after a certain point, due to the diode's heat sensitivity - which is VERY small compared to reds, but the PI plots are bent slightly if you put a ruler to them.
The step size would also warn me of any developing kinks or unusual behavior.
Under the Step Size column i have a calculation for average step and average slope efficiency (mW per mA).
Under Step 4 Percentage, you can see i have references for previous steps, to see how the latest degradation is influencing powers there.
There is a small difference to the last data i reported for Step 3. This is due to a slightly different temperature now on the TEC rig. My goal was to make it the same, but it didn't work out.
However as long as the temperature is consistent through each step (even if it's different from the last) the degradation data remains correct.
There are two strangelly marked fields saying "T. St. Sl. Eff:", which simply means "Test Step Slope Efficiency", which was used to calculate the tested powers during earlier steps.
When the test was running between 360 and 400mA, the formulas in Excel used that data to calculate the power at the tested current from the slope efficiency and the two measured powers around it.
As i mentioned, right now i just measure the power at 424mA and use that for calculating the percentage.
Anyway, this is it for now, next i have to plot 10h and 20h. I have a good feeling the diode might last past 100 tho.
Luckily i was right!
But first i had a little problem to solve. Ambient temperatures started changing to the point, that i was unable to just let the diode warm up to the level it was at before.
I had to rig up a constant temperature setup. I had most of the parts already here from various old experiments, and i relied on a neat little chip from MAXIM (MAX1968) to do the peltier driving.
The chip is a high efficiency switcher that drives the TEC with PWM and can automatically change between cooling or warming to keep the temperature the same all the time. I managed to hook it up to the tiny SMD thermistor i already had thermal-epoxied to the back of the diode for temperature equalisation during earlier re-plots.
It works really nice and makes re-plotting so much easier, as i don't have to "chase" the temperature anymore.
Anyways, i decided the next step would be 600mW. I used the last step slope efficiency to calculate it would require 423mA, altho i expected a small error, because i know that slope efficiency drops a little with every step above a certain current (first it climbs, then drops).
Next came the scary part. I actually had to plot the diode to 600mW.
The little test diode reached exactly 600mW at 424mA. This is 4mA past the point where the first diode went all zombie-faced, but this one is still working fine.
First, here is the Step 4 "Zero-Hour" Plot:
The efficiency of this diode has dropped quite a bit by now, but apparently if it starts out as a good diode, it stays a good diode.
Then i let the diode cycle for 4h...
Here's the Degradation Speed Comparison @600mW:
After the initial 4h, i measured a total of 3mW Po loss at 424mA. The diode dropped to 99.5% of it's initial Po of 600mW.
But as expected, this was a minimal power loss, which didn't really change the fact, that the degradation speed comparison plots are pretty pointless, due to the much more prominent "burn-in" effect with 12x diodes.
The only interesting thing that happened is, that this time degradation was actually faster than at the step before it. You can see that the pink line of 600mW drops faster than the light-orange line of 550mW.
This probably means, that this last 50mW step up in power was harder on the diode than previous steps of equal size. I expected something like this since we are slowly nearing the diode's limits, i just didn't know how it would show.
Then came time to decide what to do next.... Problem is, all this testing sure as hell wasn't telling me.
It's a little frustrating that i have to multiply the degradation of every step with all the steps before it to get something that looks half way as if it makes sense, but means very little...
Degradation Multiplied:
This is what i was hoping to get when i saw the results of the first step, but unfortunatelly the diode won't play ball.
I thought multiplying would show me the worst-case-scenario, where degradation would come out even more pronounced than it can possibly be in reality. I thought that if i keep the last multiplied line less steep than that of the second 8x i would be safe, but at 600mW i'm just not sure anymore.
So what to do next? Well, i think that at 600mW and the diode still going strong, the decision is much simpler.
There were several wishes to go this far, since 600mW is a nice round number.
Obviously if i was asked to make a 600mW laser, i would set it a little higher, so it would stay above 600mW for the majority of it's life.
Perhaps i would even wish to set it as high as 650mW. So i did wish i could go ahead and test even higher, but since i wouldn't really go further than 650mW (which would take some ~25-26mA more or 449-450mA with the already degraded diode), i don't think it's really necessary...
If the diode tests very well at 600mW then clearly it's possible to set it above that. This we'll know from the number of hours of it's total lifetime at 424mA i'm sure.
At this point i'd much rather get "too many" hours at 600mW than too little at 650mW.
That's why i think that having already reached 600mW and apparently safelly at that, the best decision is, to let the diode cycle to death at 600mW initial Po.
The diode is now nearing 10h where i'll do a re-plot. Next one will be at 20h, then i think i'll spread them out to every 20h if degradation remains slow.
Other than that, i wanted to show what a test sheet looks like, in case anyone is interested...
Test Sheet Example (Step 4):
It's a little messy, i know, but i'll explain everything.
First off, there are no currents here, because i only have them in column A, but they go in 40mA steps (0, 40, 80, 120, 160, 200, 240, 280, 320, 360, 400) except for the very last one in this case, which is 424mA.
Normally i would test to the next step and then calculate the 424mA Po, but i didn't want to push my luck, so the Diode Analyzer is set to an Imax. of 424mA. Instead of calculating, i now simply turn the Iadj. pot to the max, and read the power off the meter for analysis.
In the CN column, you can see the measured Vf values for each step, and next to that the Po.
There is some more measured data crammed in, so under the Vf title you can see the Ith reference i have there cos it's changing slightly, and to it's right the temperature during the test.
Pretty much every other value in the test sheet is automatically calculated.
The first example of calculated values is the negative Po at 0mA. It's there to make sure the graphs will cross the X axis at the right point, something that would not happen if i wrote 0mW there. It's calculated from the first two steps, to make sure the plot starts out straight.
To the right of Po you can see Pe or electrical power used by the diode, which is a multiple of current and forward voltage. It is used for Po/Pe plots and to calculate Absolute Efficiency, which again you can see to the right of Pe.
Efficiency is calculated for every step separatelly, and under it you can see the average.
To the right of the absolute efficiency, you can see "Step Size" in mW. Basically it shows me how many mW each 40mA step brought about.
As you can see it climbs a tiny bit at the start and drops after a certain point, due to the diode's heat sensitivity - which is VERY small compared to reds, but the PI plots are bent slightly if you put a ruler to them.
The step size would also warn me of any developing kinks or unusual behavior.
Under the Step Size column i have a calculation for average step and average slope efficiency (mW per mA).
Under Step 4 Percentage, you can see i have references for previous steps, to see how the latest degradation is influencing powers there.
There is a small difference to the last data i reported for Step 3. This is due to a slightly different temperature now on the TEC rig. My goal was to make it the same, but it didn't work out.
However as long as the temperature is consistent through each step (even if it's different from the last) the degradation data remains correct.
There are two strangelly marked fields saying "T. St. Sl. Eff:", which simply means "Test Step Slope Efficiency", which was used to calculate the tested powers during earlier steps.
When the test was running between 360 and 400mA, the formulas in Excel used that data to calculate the power at the tested current from the slope efficiency and the two measured powers around it.
As i mentioned, right now i just measure the power at 424mA and use that for calculating the percentage.
Anyway, this is it for now, next i have to plot 10h and 20h. I have a good feeling the diode might last past 100 tho.
Attachments
jayrob
0
- Joined
- Sep 21, 2007
- Messages
- 9,862
- Points
- 113
Really nice work Igor...
So that's where you've been... working on all that!
So that's where you've been... working on all that!
- Joined
- Dec 21, 2008
- Messages
- 1,223
- Points
- 0
How about this...
IgorT
0
- Joined
- Oct 24, 2007
- Messages
- 4,177
- Points
- 0
Thanks Jay!
Yes, that too, but i wanted to post about finishing the TEC setup weeks ago. Unfortunatelly i had a "little" accident which required me to get a metal plate inserted into my hand.
While it's great that i'm going to beep on airports from now on, my hand was so swallen i couldn't do much before a day or two ago. Still hurts a little, but at least i can hold things now.
HaloBlu
0
- Joined
- Dec 27, 2009
- Messages
- 1,321
- Points
- 0
Damn, I was just going to say welcome back & glad you were just very busy...
Well I guess we can still be glad it wasn't any worse.
Mmmm 600mW.
Well I guess we can still be glad it wasn't any worse.
Mmmm 600mW.
HIMNL9
0
- Joined
- May 26, 2009
- Messages
- 5,318
- Points
- 0
Glad to see that you are feeling better, now ..... health came before anything other (yes, including lasers 
)
And be careful with airports, security guards are paranoids .....
..... now, who mumbled "blasphemy, blasphemy" in the background, when i said that lasers are less important than health ?? ..... j/k
:crackup:
)And be careful with airports, security guards are paranoids .....
..... now, who mumbled "blasphemy, blasphemy" in the background, when i said that lasers are less important than health ?? ..... j/k
:crackup:
IgorT
0
- Joined
- Oct 24, 2007
- Messages
- 4,177
- Points
- 0
The diode reached the 10h mark and everything seems just fine..
I measured the 424mA power at 595mW. That's 99.17% of initial Po.
Here's the Degradation Rate Comparison plot:
And of course degradation is at it's slowest so far. So tired of having to say that. I wanna see some action or something! The 8x's did way more changing... The only thing that happened here was, that the pink 600mW line started faster than the 551mW one, but then slowed down, while the 551mW one started slower and then sped up.
But at this point, this diode is overdriven to 3.75x it's rated CW Po...
The farthest we went with an 8x during the experiment was 3.616x their rated CW Po. It's close, but the 12x is being handled harsher...
Because of the similarity of the level of overdriving, i should probably be convinced by now, that 12x's simply act differently, instead of wondering if the slow degradation means that i could test even higher.
I think we can expect a similar total lifetime from this 12x @600mW as we got out the second 8x @451mW initial Po - just under 200h. Unless efficiency somehow comes into play again and makes it more...
I measured the 424mA power at 595mW. That's 99.17% of initial Po.
Here's the Degradation Rate Comparison plot:
And of course degradation is at it's slowest so far. So tired of having to say that. I wanna see some action or something! The 8x's did way more changing... The only thing that happened here was, that the pink 600mW line started faster than the 551mW one, but then slowed down, while the 551mW one started slower and then sped up.
But at this point, this diode is overdriven to 3.75x it's rated CW Po...
The farthest we went with an 8x during the experiment was 3.616x their rated CW Po. It's close, but the 12x is being handled harsher...
Because of the similarity of the level of overdriving, i should probably be convinced by now, that 12x's simply act differently, instead of wondering if the slow degradation means that i could test even higher.
I think we can expect a similar total lifetime from this 12x @600mW as we got out the second 8x @451mW initial Po - just under 200h. Unless efficiency somehow comes into play again and makes it more...
Attachments
IgorT
0
- Joined
- Oct 24, 2007
- Messages
- 4,177
- Points
- 0
The diode is reaching 20h seemingly very easily. Only 37 cycles to go.
If nothing continues to happen with the same intensity, i may wish to spread out replotting to 30h intervals after this... Or i could just measure the 424mA Po instead of doing a full plot, since i would like to see what happens in a higher resolution than that...
If nothing continues to happen with the same intensity, i may wish to spread out replotting to 30h intervals after this... Or i could just measure the 424mA Po instead of doing a full plot, since i would like to see what happens in a higher resolution than that...
IgorT
0
- Joined
- Oct 24, 2007
- Messages
- 4,177
- Points
- 0
Well, the diode reached 20h last night, and did something unexpected.
At 424mA i measured a Po of 597mW, which is where it was at 4h! Basically, it now produces more power than it did at 10h.
Here is the degradation rate plot:
Thinking back to 8x's, they did this too, only with more (noticable) degradation first, and perhaps slightly more pronounced (just like their entire behavior) ups and downs....
I was thinking about this even before seeing the last data, now it seems even more likelly - it's quite possible that 12x's simply do not drop nearly as much in power, during their degradation, but could live just as long as 8x's when overdriven to the same level.
If we should get little noticable degradation but a long life, that would be good news, as it would mean that 12x's will hold their power much more steady through their life. And i like my lasers keeping their power steady, instead of measuring lower every week.
The only other possibility i see is, that this diode will have an absurdly long life at this power, but this seems very unlikelly at 600mW....
However one thing is already for sure. With good, high efficiency, healthy 12x's 600mW is SAFE even if the test diode pops during the next minutes, because 20h is already a respectable age hard to reach in a pointer by hand.
Then there is the time this diode already did at lower powers - 18 hours.
This has to figure in somehow, because it did noticably degrade the diode.
Perhaps it could be interpreted, that this diode by now would have survived 30h at 600mW with the same results. Approximatlly of course, as i don't know of a way to apply the 18h at lower powers to the lifetime at 600mW....
In any case, the results are already good at this point! And i have a feeling this diode is gonna keep going for a long time..
But i have to admit i was scared of going this far, due to what happened to the two donated diodes. But apparently not all 12x's are created equal.
That's why i keep saying this data only applies to good, healthy, high efficiency 12x's.
I would highly recommend testing all 12x diodes for efficiency, before deciding what current to set them to. If you should have the bad luck to stumble on a low efficiency one, you will have no choice but to power it lower or risk losing it instantly.
I just hope i was the one with the bad luck here, and that diodes like that are the exception to the rule, and not as common place as my sample makes it seem.
P.S. I am putting together a laser with a time recorder in it, which will measure the total time the button is depressed. I will make that laser my "only" one for a while and try to use it as much as possible. This should give me an even more accurate idea of what average to heavy use really is - something to help translate the hours we get here into months and years.
I have another laser with a time recorder already in it, but it wasn't made to take the recorder out before the laser diode dies. It's only there to measure the total lifetime of the 8x diode.
I realized i needed another to create proper statistics faster, and it should come in handy when trying to put meaning to all the hours from this testing.
At 424mA i measured a Po of 597mW, which is where it was at 4h! Basically, it now produces more power than it did at 10h.
Here is the degradation rate plot:
Thinking back to 8x's, they did this too, only with more (noticable) degradation first, and perhaps slightly more pronounced (just like their entire behavior) ups and downs....
I was thinking about this even before seeing the last data, now it seems even more likelly - it's quite possible that 12x's simply do not drop nearly as much in power, during their degradation, but could live just as long as 8x's when overdriven to the same level.
If we should get little noticable degradation but a long life, that would be good news, as it would mean that 12x's will hold their power much more steady through their life. And i like my lasers keeping their power steady, instead of measuring lower every week.
The only other possibility i see is, that this diode will have an absurdly long life at this power, but this seems very unlikelly at 600mW....
However one thing is already for sure. With good, high efficiency, healthy 12x's 600mW is SAFE even if the test diode pops during the next minutes, because 20h is already a respectable age hard to reach in a pointer by hand.
Then there is the time this diode already did at lower powers - 18 hours.
This has to figure in somehow, because it did noticably degrade the diode.
Perhaps it could be interpreted, that this diode by now would have survived 30h at 600mW with the same results. Approximatlly of course, as i don't know of a way to apply the 18h at lower powers to the lifetime at 600mW....
In any case, the results are already good at this point! And i have a feeling this diode is gonna keep going for a long time..
But i have to admit i was scared of going this far, due to what happened to the two donated diodes. But apparently not all 12x's are created equal.
That's why i keep saying this data only applies to good, healthy, high efficiency 12x's.
I would highly recommend testing all 12x diodes for efficiency, before deciding what current to set them to. If you should have the bad luck to stumble on a low efficiency one, you will have no choice but to power it lower or risk losing it instantly.
I just hope i was the one with the bad luck here, and that diodes like that are the exception to the rule, and not as common place as my sample makes it seem.
P.S. I am putting together a laser with a time recorder in it, which will measure the total time the button is depressed. I will make that laser my "only" one for a while and try to use it as much as possible. This should give me an even more accurate idea of what average to heavy use really is - something to help translate the hours we get here into months and years.
I have another laser with a time recorder already in it, but it wasn't made to take the recorder out before the laser diode dies. It's only there to measure the total lifetime of the 8x diode.
I realized i needed another to create proper statistics faster, and it should come in handy when trying to put meaning to all the hours from this testing.
Attachments
Last edited:
DJNY
0
- Joined
- Nov 2, 2009
- Messages
- 5,991
- Points
- 83
jayrob
0
- Joined
- Sep 21, 2007
- Messages
- 9,862
- Points
- 113
Hope that doesn't mean it's going to die...
I seem to remember a diode doing something similar just before it died...
nospin
0
- Joined
- Feb 13, 2010
- Messages
- 182
- Points
- 0
Wow
That's 2 min a day every day for a year and a half at 600mw
Right there it sounds like getting your moneys worth!
Jim
That's 2 min a day every day for a year and a half at 600mw
Right there it sounds like getting your moneys worth!
Jim
- Joined
- Oct 11, 2007
- Messages
- 106
- Points
- 0
This is great news. At first, I was afraid that setting my 12x to 600mw was too risky. Now, the data seems to show that it's safe so far. Lets hope that the diode will make it to at least 50hrs.
IgorT
0
- Joined
- Oct 24, 2007
- Messages
- 4,177
- Points
- 0
If your 12x diode tests in the higher efficiency end, i believe it's safe to go as far as 600mW.
Right now i stopped it at 30h of 600mW for a replot. That's on top of 18h at lower powers.
Chances are good it's gonna surpass 50h and go much further at this rate, so the final result will likelly be "600mW very safe / more power possible". But i'm just guessing from what i'm seeing so far.
Last edited:
- Joined
- Dec 30, 2007
- Messages
- 573
- Points
- 0
Now that's the kind of news I like to hear... Great work Igor.
The kind of news I don't like to hear, however, is the kind where you end up in hospital again. Are you sure you're not having an affair with the triage nurse down at the emergency clinic?
Hope your hand heals well and doesn't cost you mobility that affects your mad soldering skillz. You've had enough adversity already!
Cheers, CC
