Originally Posted by
Buffo
You are correct. Marconi was talking about this a while back. The Die4Laser driver works well, but you don't want to push the diode really close to it's limits using that design. If you do, a little current fluctuation in the modulation stage might be enough to kill it.
Actually what I meant . He didnt have any protection on the input..where overvoltage (more than 5v) on the input where someone may tie the mod line to the input supply rail (+7 to 12v) being more than 5v will cause the driver to exceed the set current. this is only true if you are using a 5v reg on board and feeding the reg outboard as most will dol with + 7 to 12volts.
Otherwise, It is the best circuit I've seen.
Originally Posted by
The_Doctor
Many diodes are run CW at strong ratings. As most DVD diodes have pulsed ratings for 240 mW and such, modulation isn't what weakens them when pushed hard, as if the CW doesn't kill them, the modulation certainly won't unless the transients are very fast. That said, there are likely stressful combinations of near-CW pulse width ratio and actual mod speed even with slower transients, but those can kill a diode quite independently of any kind of driver design. Bad drives might make it worse, but the best can't avoid this fact..
It most definitely is...because you are close to the diodes operating bounderies . The best drivers are designed to avoid
transients by using negative feedback to overcome these. The load (diode) is a resistive and capacitive load , there will be some reactance that needs tended too.. Take another look at Robins driver. He has my vote.
But you are correct, most applications lately stem from overdriving the DVD diodes.
The real risk is overcurrent due to overshoot (and brief oscillation) in any inductance or capacitance present. Any driver has this risk, even a shunt modulator
Very true
though with those the risk is less, especially as the regulator will likely respond faster than the active modulating device
not that I have seen with a scope
because the modulation speed is always restricted anyway, as part of a compromise between mod speed and transient protection for the laser diode.
Actually , why comprimise...the diode has an almost unlimited bandwidth.
As a guide, an LM317 can be a shortwave RF oscillator! That should tell you that it can more than cope with overshoot if mod speeds are limited to 30 KHz or less.
The LM317 may well oscillate at HF but that is open loop bandwidth. and not much more than several megahertz, This was done on purpose to keep the regulator from becoming an oscillator while used as a regulator. It also has poor load regulation.. about 1 percent. I do like them as simple regulators, tho.
Ive built many supplies from them, even mod'd them for below 1.25v capablity.
The devices that Robin and I use are rated at least 10x that. Mine are good to 120mhz. Good for video and any sidebands you are likely to throw at it.
But, Using faster devices require more attention to reactance. but they are more easily controllable, if you use a limited bandwidth device, you have limited control.
Dont forget, the LM317 is made up of bipolar junctions. We use mosfets. much faster. The main problem is not having negative feedback to control reactance while using the LM317. At least in any circuit Ive seen..Current yes, but not frequency compansation.
In practise, you only need ONE hard current limit if you use a shunt mod. The only other precaution (apart from the usual extremely clean supply) is to avoid inductances and capacitances that can cause over-current pulses.
I have found that pre-limiting makes it easier to control overshoots., rather than rely on only one device. Why tax the driver.
I also had problems using the shunt style with being able to react fast enough
to heavy loads. Thats why you dont see commercial manufacturers doing it this way.
Don't just take my word for it. Study that Laserbias.htm I linked to, first post. That guy gets by with mod speeds capable of video transmission, apparently. His design is a tad messy-looking, and lacking one or two finer points of detail, but don't let that fool you into thinking it's a bad design, it's not.
I have tried all of Johns circuits. Also, I know him well..as we share the same ham radio hobby. those circuits while simple..dont have enough control..they can be dangerous to the unexpected hobbiest. Just not enough control.
But I do suggest reading and making a few as a starting point.
Buffo, a close look at Robin's circuit shows it's safer than you suggest. The LD is given a 5V regulated supply, so current is dependent entirely on the assumption of a short where the MOSFET is, followed by a current limiting resistor. That's your hard limit. Then the MOSFET resistance (varied from near zero ohms to >300 Mohms) can only adjust that current downwards. It's free of problematic capacitances, the only ones being those across the diode to protect it. There are no inductances other than the tiny ones in the MOSFET and the connecting wires, and being in series with the load and a current limit resistance, they will be very well damped. I prefer a shunt mod, but that's just because the logic of it is nice.
There are some inductance / capacitance and resistive reactance cuased by the diode, while Robins circuit addresses most of this..he has made it somewhat universal..it still is diode dependant..Each diode type will react differently, It is up to the user to account for the rest.
My drivers are made to match the diode, that is why you dont see them advertised for sale.
I've been making drivers for lasers more than 5 years now...There isnt much out there that I would trust.
even some of the commercial stuff out there just dont cut the mustard for us lasershow freaks.
A scope is your best friend.
Last edited by marconi; 08-05-2007 at 09:51.
"My signature has been taken, so Insert another here"
http://repairfaq.ece.drexel.edu/sam/laserfaq.htm
*^_^* aka PhiloUHF