buffo
Thanks for the input, Tocket! Glad to see you're on-board and willing to help develop this standard.Originally Posted by tocket
You're right. I alluded to it in requirement # 4, but didn't specify the parameter well enough.You left out one important issue however; linearity.
The formula you've proposed will work, but might be more difficult for the manufacturer to work with. (They'd need to take the measurements, then plot the graph, then determine the slope.) Also I'm not sure that it's realistic to expect a zero offset (m). Most analog lasers need at least .7 volts on the blanking lead before they turn on. Some don't turn on until around 1.2 volts. But yeah, there should be a spec for that turn-on point, as well as a linearity spec. Normalizing to get a consistent slope is a nice idea, but really - if the laser meets the offset requirement and is linear, then the slope will be consistent anyway... (Unless it reaches max power at significantly less than 5 volts, that is!)
How about specifying a turn on voltage spec (plus or minus a narrow band), and a separate voltage spec for when the laser reaches 100% of rated power (again plus or minus a narrow band). If we control those two end points, then we don't need to worry about the slope. (As for the linearity between those two points - that's another measurement that can be combined with the blanking speed test; see below.)
What is a happy medium for the turn-on voltage? Maybe a goal of 1 volt offset, plus or minus .5 volts? Or is that too broad a spec? What do you think? And what about the voltage for max power? 4.5 volts? 5 volts exactly? And how wide of a band do you want on the high end? (My gut reaction is to give a little more leeway on the lower turn-on voltage, but to tighten the spec on the max power voltage to something like 4.8 volts plus or minus .2)
One other problem though, is that you need to define what "turning on" consists of. Maybe it would be better to call it the voltage at which the power output is 10% of maximum...
Basically I'm trying to make the spec something that can easily be measured by a technician in the factory to see if the laser is OK or not before it ships. I am afraid that if the spec requires too many calculations, or requires graphing in order to test, then it might not be implemented at all. (Trying to keep it simple without sacrificing quality.)
That's a pretty tight spec for the slope (k). If you are going to allow for a 10% deviation at any one point (your .1 residual term), then you really need to increase the standard deviation spec, don't you think?You also need to look at the residual. Stating that the residual may at no point be larger than 0.1 and the standard deviation for k must be less than 0.05 would weed out any lasers that cannot handle analog blanking in a linear fashion.
This is what is known as jellybeaning (or blanking-induced power loss), and it's at the heart of step 4 of requirement # 4. (Specifically, the "no power fluctuations greater than 10%" part at the end of that sentence.) It's also one of the most vexing problems with DPSS lasers today. (At least from my experience.)The other related issue is that of power fluctuations due to blanking.
This is exactly what I proposed, though admittedly I didn't include the formula to calculate it. (Thanks for that!) But we're on the same page here: No more than a 10% power fluctuation at whatever rated power is selected.Normally you would expect a beam's power to be a simple function of duty cycle (D) and analog blanking on-state voltage (V) like:
P=Pmax*D*V/Vmax
<snip>
Since there are two variables in this problem it becomes more complex to analyze. A fairly simple way of doing it is to pick 10 points on the surface spanned by D and V, measure actual power at those points and compare to the ideal function above. By requiring that the measured value is within 10% of the ideal we can assure ourselves that the handles analog blanking properly.
So - for example, if you have a 100 mw analog laser that turns on at 1 volt, and reaches maximum power at 5 volts, then 3 volts on the blanking lead should = 50 mw CW. Now pulse it at 3 volts at 10Khz at a 50% duty cycle, and you should have 25 mw average power. If your measured power isn't within 10% of 25 mw, the laser fails this spec.
Also, taking a set of 10 measurements (from a low of 10% power all the way to 100% power) using this same criteria should give you enough data to predict the slope with reasonable accuracy. So if there *is* a dip in the slope, it should be apparent. This means that one test will allow you to see if the laser meets the jellybeaning requirement as well as the power linearity requirement. Or, to put it another way, if you have 10 points on the power curve that are all in line within 10%, then the slope can be interpolated from that and assumed to be within 10% as well. No need to actually plot the graph.
One thing I should probably admit here is that I'm running Pangolin, which has a very complicated color balance routine that will compensate for non-linearity in the laser's power curve. So perhaps I'm a bit less sensitive to this issue than other people might be. But I also don't want to make the spec so tight that it's unattainable. (Or, unattainable without spending a small fortune.)
What do you think? How much non-linearity is too much? And what about the turn-on point and the max power point. (For the purposes of discussion, let's call "turn on" to be 10% of rated power.) What is a reasonable voltage for that 10% power point? 1.2 volts maybe? plus or minus how much? And what about the max power point? I like 4.8 volts plus or minus .2, but what do you think?
Adam
Reply With Quote
If all CNI and Bridge can make the wish list come ture, that will be perfect.
