TEC and current driver chips for laser?

You could do a test board and measure the theta JA with the part soldered in place. I may be reading more into the datasheet than intended, but they mention the exposed paddle. I assume the manufacturer wants a patch of copper on the PCB to attach to the paddle? In addition, some heat leaves from the bond wires. The capacity is proportional to the square of the wire diameter, so often in these kinds of chips they spec a fatter than required bond wire. It may be possible to put a patch of copper on the PCB that connects to the power pins to improve the theta JA.

Joerg wrote in news:5Jo7h.11407$ snipped-for-privacy@newssvr29.news.prodigy.net:

What about the LM317?

It can be fast, you'll limit your slew with filtering anyway, to allow your few hundred Hz while limiting spikes. I use them for constant current LD drivers exclusively, I never use anything else. A careful choice of thermistors in its current sensing circuit can do temperature compensation if needed. That chip can be tricked into doing things it was never meant to do. By suitable control of the voltage sent to the sense input you can probably control modulation as well as set safe current limits (right down to zero with linear response and no burnt pots). I've never tried doing modulation with one yet, but it's something I want to try sometime.

LM317's ganged can also make nice TEC controllers with extremely low noise.

Those fancy new "thermally effective" packages often do require some heat to be conducted into a copper pour plane. The chips I have seen this way were within temp specs but way too hot for my taste.

Well, if there is no alternative to the ADN2830 there is always the method of using two in parallel, setting each to half the current. They are just kind of expensive for a controlled current source. North of $10 a pop. If this all goes into a larger production run it'll have to become discrete but we are far away from that point.

Joerg, my TEC might be overkill for just a laser diode, but have a look anyway.

Rene

I was thinking about that but initially we just need something off the shelf that regulates well, comes complete "in a can" and makes sure there is no glitch on power cycles. Also, we have to move really close to the max LD current. Don't want to risk $500 laser diodes.

The LM317 is indeed very versatile. Once I did an AM modulator with one and it covered the audio range with good fidelity. That way we had a voltage regulator and modulator in one package.

That can be tight here because I have to jerk the TEC around pretty good for wavelength hopping. It would have to be an H-bridge.

Hello Rene,

Wow! That ought to be the Rolls Royce of all TEC controllers. With GUI and all the bells and whistles. I wouldn't be surprised if it came with a GPS so it always knows where it is :-)

Question: Why do you charge your own countrymen about twice as much as us? Do they tax you to death over there?

One thing you might want to look at is LED drivers. They are specifically designed to be pulse modulated, and the parallel types are designed specifically to handle a few 100s milliamps of current.

Just a thought.

Cheers

PeteS

Not a Rolls Royce, but an el-cheapo model compared to those with physical user interface and a case. The competitors with case, line power supply and PC communication go for in the order of 1800$. The prices are going to drop soon.

The prices are the same for all, there is no tax included in neither price. We don't have tax listed on industrial goods, only on consumer goods. And exported goods are free of VAT anyway. Our VAT is 7.6%, being added for inland customers.

Rene

Hello Rene,

They can be a lot more expensive. We have some (bought used). But they aren't all that practical. The fancy TEC controller we have must be controlled via GPIB. Arrrgh...

And quite frankly, compared to those units yours seems to be the better deal. So if you see an order from California some day you know whom you can blame :-)

BTW, an analog input instead or in addition to the potentiometer would be nice. That's why we can't use our fancy one in the fancy box. Unless I butcher it. If you want to try etalon hopping and stuff like that you need to have an analog input.

AD also has an evaluation board for their ADN8831 for $150 but that won't come with any processing horsepower like yours does.

Well, then why is the price CHF600 but for the US it's only $461? If that is a typo maybe it should be changed.

Hello Pete,

Good point. Although laser diodes are a lot more fickle. A microsecond of excess power because of a glitch and a $500 laser diode becomes a $0.10 LED.

Joerg wrote in news:9os7h.16474$ snipped-for-privacy@newssvr27.news.prodigy.net:

This is true.. Still, I trust an LM317 with £50 laser diodes, I even used to push beyond rated power, narrowing the margin, and all failure mechanisms were accounted for, static, retroreflection, and deliberate overdrive. All duriong testing (at an expense I hope never to go through again, learning stuff the makers never tell us). No diode died from a rogue spike without notice, only from something I did to it and identified.

On the other hand, the reason I never tried modulation with my LM317 drivers is I depended on a degree of filtering that although simple (a 'memcap' in parallel with a 10µF tantalum and a low ESR ceramic 0.1µF, and a fast low Vf diode in inverse parallel) was such that modulation on the

While I think it is beyond my experience to make in one easy stride, the ideal compromise in good filtering, and decent mod speed, I still think an LM317 is a safe bet. The more powerful diodes usually have a wider margin of error in overdive current anyway, it's only the very short duration spikes you have to stop. Stopping them on DC supply input is easy. The challenge is cheaply letting in a modulation signal while preventing them getting in that way.

Joerg, any detailed words on your AM modulator/voltage regulator with an LM317? If you're willing to share, that sounds like something fun to try, to extend my knowledge of the uses of these amazing IC's.

I don't need to modulate in the telco sense with GBit/sec data streams but I do have to move it up to a few hundred Hz. But by no more than the range of 90%-100% Imax.

Maybe I'll give that 317 a shot. It's been good to me since the beginning of my career and I trust it (in contrast to almost any LDO).

I don't have a schematic and it's been long ago. Basically it was a 317 that was set via the Vsense divider to produce about 30% of the rail at its output. The rail was unregulated but the 317 made sure the 30% side was squeaky clean. Now the modulation signal was current-fed into that Vsense pin. AFAIR that was pretty much it. That way you don't need a big old power amp for the modulator.

Joerg wrote in news:Blu7h.26702$ snipped-for-privacy@newssvr21.news.prodigy.com:

I saw.. :) I was thinking only that my own filtering would have been so limiting that you'd not even get that. The ideal place for the LD protection is right by the LD, close as possible, so any kind of modulation must absolutely compromise the protection possible, so in practise a good design around an LM317 might be no worse than the best of any other kind of design. I think most elaborate driver IC's and hybrid units just group together a range of methods that can easily (and much more cheaply) be built round the 317.

I like them that much, I think sometimes there are certain components that are like elements, even if they're IC's they are irreducible, complete and so useful they should never change, except as compatible improvements. A similar general purpose IC I really like is the LF412 (and LF422) op-amps. They directly replace 714's with dramatic improvments, and I also replaced a four-times-more-expensive Burr Brown op-amp that ran very hot, with one of these. Power consumption dropped to a tenth or less, and accuracy went up. That Burr Brown IC was supposed to be an awesome audio amp, but the basic LF412 was better, dramatically so.

The ultimate fun I had with the LM317 was the day I realised I didn't have to spend >£200 to get a Wavelength Electronics laser diode driver. Real liberation, that was.

Nice. So much gain, so little noise, and all with protection built in too. And three pins. As an audio amp, I've not tried it, as it's not efficient, but it seems to me that it might rival a class A amp and sound very good, unless there are weaknesses I haven't thought about. I found a small shortwave CW radio transmitter design on the interweb once, based on an LM317. I doubt it's run out of big surprises for me yet, either.

Typo corrections: LF411 not LF422, and 741 not 714.. I need sleep.

With a bit of perverse twist you might even get the oscillator too :-)

The input is analog, sampled. Radiometric between the the suuply rails.

Well, at 1.30 CHF/USD, 461USD = 600CHF ...

The next version might have an optional Labview protocol if find the time for it.

Rene

Fred Bartoli wrote in news:455ec929$0$30409$ snipped-for-privacy@news.free.fr:

By accident, or by design? Inquiring minds need to know...

Lostgallifreyan a écrit :

The height of perversion would be to have it by design and saying it's by accident.

Fred Bartoli wrote in news:455f252a$0$5393$ snipped-for-privacy@news.free.fr:

That is SUCH a good idea! Real potential there. Could specify a thing as a nice bit of serendipity, leaving the person we're trying to impress with the idea that we could dwarf the oppositions specs, if only we tried. >:)