Strange Artifacts of new LPM's output...

[Atomicrox]

Interesting. Seems to be about the same amount of overshoot (3-4%).

 

[djQUAN]

Soooo, I 'borrowed' a Tek scope so I can post a new set of measurements now without waiting for access to my toys at the other house. I used a 1.4W 445nm laser for this test as requested. Here are the results:

Since my laser has a momentary tact switch button which does not need a lot of force to activate, I was able to set it up fixed on the table and turn it on and off without movement. Here are the "control" data.

At turn on:

At turn off:

And I tried holding it by hand and moved the laser dot around the TEC surface at turn on, held it steady for a few seconds then made a quick move before turning it off again. (This is the same laser, I changed the vertical gain setting on the scope to more clearly capture the spikes and "ringing")

Now, back to the original graph in question, notice any similarities?

I think we can let this issue to rest by now?

 

[Trevor]

Wow - that's some thorough testing. Very well done!

Interested to see if Jerry will pop back in...

Trevor

 

[Things]

Pretty sure that overshoot is going to be the driver dropping into regulation, I don't see an issue with it.

 

[Trevor]

Pretty sure that overshoot is going to be the driver dropping into regulation, I don't see an issue with it.

It does it with my Coherent CUBE also - so I think it's most likely an attribute of the sensor.

Trevor

 

[Atomicrox]

That can be tested easily. Just cover the beam, turn the laser on and then release it. If there's no overshoot the driver is to blame, if there's overshoot the sensor is to blame,

 

[Trevor]

Well... I suppose I'll give it 'til the 3rd to make it an even month, but after that, Mr. Bauer will be earning another red bar for this silliness.

Trevor

 

[Meatball]

That can be tested easily. Just cover the beam, turn the laser on and then release it. If there's no overshoot the driver is to blame, if there's overshoot the sensor is to blame,

I was thinking about current overshoot as well - but this test would surely rule it out if any could so quickly.

Both the overshoot and ringing seems to take about 2-3 full seconds to quiet down - that's a long time. I have to go back and look at any thermal mechanisms that could account for this.

Perhaps even the ordinary thermocouple has a small response region during fast temperature changes.

You see, I've been thinking about this incorrectly. We usually say the voltage produced is a function of a temperature difference between two surfaces. But in reality, its a function of a temperature gradient. For those who've been in calc III, I'm referring to

del(voltage) = -(seebeckCoeff) * del(temperature)

not

del(voltage) = -(seebeckCoeff) * (Th-Tc)

DelTemp refers to the temperature gradient from the end of one of the thermopile's wires, to the end of the other wire, which varies depending on the length along which the temperature changes along the wiring. This is why there is cold junction compensation built into the head's amplifier. It holds one end of the thermopile's wires at a constant temp, or it will actively measure the temp on one of the wires and provide feedback into the instrumentation amp.

Since heat tends to spread relatively (compared to charge) slowly through solids (even wires), I'm thinking that our sudden whack of heat input (laser suddenly turns on), cannot be accurately reflected on a system that relies on heat flow. Makes sense to anyone else?

The overshoot could be simply reflected heat which didn't make it along the wire in less than 2 seconds. Think about re-arranging area drawn out by the spike, and squish it back over to the left of where the curve initially rises. I think it could be shown that doing so would significantly square off the rise time on the initial response.

So maybe all the power is accounted for, but it simply arrives to the amplifier "late".

But this does not explain the ringing. Ringing still say's stray capacitance to me.

 

[djQUAN]

The overshoot is part of the analog predicting circuit to speed up the 90% response time of the TEC output (that's what the 3rd pot inside the ophir head is for IIRC). The ringing has been debunked as due to the movement of the laser at turn on. Please see post #34.

 

[Meatball]

TEC? I though this was for Ophir thermopiles.

I didn't see your conclusion to the ringing posted anywhere. I just saw your posted experimental graphs.

So there IS a differentiator built into the amp?

 

[djQUAN]

Yes, Ophir "thermopiles". I couldn't remember the correct term at the time so I called it a TEC

on Post #34, I mentioned that I moved the laser at turn on and got the same graph so we basically concluded that was the cause of the "ringing" like waveform.

True ringing would look different with a constant frequency and diminishing amplitude. Pic:http://www.ridgetopgroup.com/img/sinusoid.jpg and since the supposed ringing is less than a few Hz, that should have taken a considerable size parasitic capacitance and inductance to oscillate at that low of a frequency.

I believe there is an integrator built in as the third pot inside for compensation. Adjusting it should minimize overshoot but would slow down the rise time. I haven't touched mine since it's already fine.

credit: http://laserpointerforums.com/f70/diy-ophir-lpm-78856.html#post1133802

 

[ARG]

I just refer to them as radial thermopiles. I've messed around with that 3rd pot on my test Ophir head and I wouldn't recommend touching it; they set it perfectly in the factory and it's pretty touchy so it's hard to get back into the right position

 

[Atomicrox]

What was the effect of messing the third pot?

 

[ARG]

Changed how the head would over/undershoot.

 

[Atomicrox]

Then this is solved

 

[Trevor]

Well... it's now been a month, with not even a peep from Jerry on the matter.

At this point it's pretty bleedlingly obvious that this little stunt was anticompetitive - not scientific. It's pretty funny that he'll do anything to eliminate his competition. Well, almost anything, because innovation is just too much effort.

Another red bar for your collection, Jerry.

Trevor