Page 6 of 26 FirstFirst ... 234567891016 ... LastLast
Results 51 to 60 of 251

Thread: Stan_Ham diode driver (analog)

  1. #51
    Join Date
    Oct 2006
    Location
    Cleveland, Ohio
    Posts
    2,342

    Default

    For a simple protection circuit, crowbars are often used to clamp the voltage of a supply line during voltage surges to protect components downstream. They typically consist of an input zener diode which, when it conducts, triggers a thyristor. In order to protect diodes from peak current surges and high dV/dt, the crowbar circuit could be modified to that it is triggered by high inrush current or dV/dt instead of absolute voltage. If the zener were replaced by a capacitor and the cap replaced by a resistor, this could be accomplished. In fact, when the input of a thyristor experiences a high dV/dt, they tend to latch on automatically without even needing gate drive so this could be explored more. As a final piece, an r/c low pass filter network after the thyristor and before the diode may be needed to allow the thyristor time to turn on before the surge reaches the diode. The time constant of the network should be as short as possible while still allowing time for the thyristor to turn on, so that it would not affect normal operation in modulation significantly.

    This is just a simple protection circuit I came up with off the top of my head so use at your own risk!
    also it is unidirectional as described but a bidirectional implementation could be realized relatively easily

  2. #52
    Join Date
    Mar 2006
    Posts
    2,478

    Default

    Quote Originally Posted by Pangolin View Post
    In any event, the 5cm copper braid experiment has nothing to do with solder.
    .
    .
    .
    LASORB is better -- much better in fact, than a 5cm copper braid.
    Didn't say it had. That Lasorb has though, if the picture is accurate. Those are solderable wire leads, aren't they? And when you're connecting a device designed to be a very low impedance I don't need to say 'do the maths' to make it entirely clear that the device is only as good as its connections. If it really is that much better than the copper braid for protecting an unpowered diode, I have to say I think the braid wasn't connected closely or well enough. You can't argue with raw copper as a conductor. It tends to win, at atomic level. If you've really come up with something so transcendentally better, why aren't you running for the Nobel Prize instead of aiming them at a dollar profit. And if it really is out of supreme generosity to the scene, why bother patenting at all, just publish it so no-one else can.

    Re risking diodes, the point was whether or not anyone has the balls to risk theirs, wasn't it? I bet my diodes were worth a lot more in real terms to me than yours to you. That was irrecoverable loss to me at the time, and no-one was even willing to admit a reality, let alone make good the problem. I knew this before, during and after, yet I still risked it before I dared make any claim before selling a module built with one. As it happened I sold two and made damn near as much as I lost, but no-one helped me other than being willing to actually sell me diodes in the first place, or to buy a module I built. You don't have a monopoly on balls when it comes to diodes.
    Last edited by The_Doctor; 01-20-2009 at 13:10.

  3. #53
    Join Date
    Apr 2006
    Location
    Orlando, FL - USA
    Posts
    1,770

    Default

    Quote Originally Posted by The_Doctor View Post
    Didn't say it had. That Lasorb has though, if the picture is accurate. Those are solderable wire leads, aren't they?
    OK, I re-read your original post and I see your point. Yes, I suppose connections can be a factor. But in practice, I can say that they aren't as much of a factor as "lead length". The problem with the 5cm braid is the length of wire. The inductance is more than sufficient to provide higher than 50 milliohm resistance at 1GHz.

    Again, in practice normal solder will work for LASORB, since it was only normal solder used up until now -- for example as shown in our video.

    Quote Originally Posted by The_Doctor View Post
    why aren't you running for the Nobel Prize instead of aiming them at a dollar profit.
    Oh yea, I forgot. Profit is bad...


    Quote Originally Posted by The_Doctor View Post
    Re risking diodes, the point was whether or not anyone has the balls to risk theirs, wasn't it?
    That might be your point, but it wasn't mine. My point was -- laser diodes are known to be succeptable to ESD. So here we are, purposely exposing the laser diodes themselves (not a driver which could attenuate the ESD), both completely unprotected (naked), as well as protected with various schemes that are supposed to work. That takes balls because the greatest liklihood is that the diode would be destroyed (and many were).

    We are still doing it even to this very day, to prove that what we have works, and to try to determine the limits. And when diodes are destroyed, it comes out of Pangolin's pocket (which is then replenished by the profit that we are not supposed to be making...).

    Profit is the very thing which allows continual R&D. If not for profit, do you think there would be the Texas Instruments DMD (just as one example off the top of my head)?

    Quote Originally Posted by drlava View Post
    This is just a simple protection circuit I came up with off the top of my head so use at your own risk!
    also it is unidirectional as described but a bidirectional implementation could be realized relatively easily
    I have to admit Lava, you sure are smart! Want a job?


    Bill
    Last edited by Pangolin; 01-20-2009 at 20:26.

  4. #54
    Join Date
    Oct 2008
    Location
    Portland
    Posts
    1,565

    Default

    Quote Originally Posted by drlava View Post
    For a simple protection circuit, crowbars are often used to clamp the voltage of a supply line during voltage surges to protect components downstream. They typically consist of an input zener diode which, when it conducts, triggers a thyristor. In order to protect diodes from peak current surges and high dV/dt, the crowbar circuit could be modified to that it is triggered by high inrush current or dV/dt instead of absolute voltage. If the zener were replaced by a capacitor and the cap replaced by a resistor, this could be accomplished. In fact, when the input of a thyristor experiences a high dV/dt, they tend to latch on automatically without even needing gate drive so this could be explored more. As a final piece, an r/c low pass filter network after the thyristor and before the diode may be needed to allow the thyristor time to turn on before the surge reaches the diode. The time constant of the network should be as short as possible while still allowing time for the thyristor to turn on, so that it would not affect normal operation in modulation significantly.

    This is just a simple protection circuit I came up with off the top of my head so use at your own risk!
    also it is unidirectional as described but a bidirectional implementation could be realized relatively easily
    Can I expect this in the LavaDrive 3? I am not kidding, get some boards printed and let us send you money!

    -Adam
    Support your local Janitor- not solicited .

    Laser (the acronym derived from Light Amplification by Stimulated Emissions of Radiation) is a spectacular manifestation of this process. It is a source which emits a kind of light of unrivaled purity and intensity not found in any of the previously known sources of radiation. - Lasers & Non-Linear Optics, B.B. Laud.

  5. #55
    Join Date
    Oct 2006
    Location
    Cleveland, Ohio
    Posts
    2,342

    Default

    Bill, wow that's the biggest compliment I've received all year esp. coming from you. I must finish my PhD in biomedical engineering in the coming months, but let's keep in touch. I'm often interested in 'odd contract jobs'.

    Quote Originally Posted by sugeek View Post
    Can I expect this in the LavaDrive 3? I am not kidding, get some boards printed and let us send you money!

    -Adam
    If built onto the driver it won't be effective unless the driver is directly connected to the diode, and in that case its utility would be questionable. It's best left as a tiny separate circuit to be connected directly to the diode with leads that then go to the driver. We'll see.

  6. #56
    Join Date
    Jan 2008
    Location
    Belgium
    Posts
    1,009

    Default

    I know this is a bit off topic, but I've been thinking about most current driver designs, and they all make it hard to use case negative diodes

    I've drawn up some quick schematics of the basic design of most diode drivers and the one I am proposing.

    Does anyone see any reason why this setup would not work or be worse than the original one

    (I basically just moved the sense resistor and implemented a difference amp ...)

  7. #57
    Join Date
    Jul 2008
    Location
    Maryland
    Posts
    1,691

    Default

    Ill take 4 lasorbs and and 2 soft start circuits please...
    Quis custodiet ipsos custodies?
    Solid State Builders Group

  8. #58
    Join Date
    Mar 2006
    Posts
    2,478

    Default

    Quote Originally Posted by Pangolin View Post
    OK, I re-read your original post and I see your point. Yes, I suppose connections can be a factor. But in practice, I can say that they aren't as much of a factor as "lead length". The problem with the 5cm braid is the length of wire. The inductance is more than sufficient to provide higher than 50 milliohm resistance at 1GHz.

    Again, in practice normal solder will work for LASORB, since it was only normal solder used up until now -- for example as shown in our video.
    Ok. That was what I was getting at, the lead length, as well as the solder itself. People tend not to solder parts very close to their wire entry points for fear of damage, so you'll have to have specific advice about this or people might leave leads long enough to cause bother.


    Oh yea, I forgot. Profit is bad...
    Nothing wrong with profit, I just reacted to the combination of bragging and inattention. That was compounded by my scepticism that the entire electrical and electronics industry has really failed to perceive what Pangolin has perceived. A lot of people profit without having to play the patent system every time they think they have a bright idea. Most people who make a profit probably have no idea how the patent system works, but they don't let it stop them because they spend time producing and optimising work rather than aggressively and systematically defending every new idea they have about it. And you don't have a monopoly of balls with regard to risking diodes directly. What did you think retroreflection testing involved, given that no-one I could reach at that time was admitting a problem existed?

    Anyway, I thought of SCR's too, but (wrongly, it turns out) thought they weren't fast enough. I remember an SCR being triggered by fast rise times in an AC mains circuit where I needed its greater sensitivity over a triac, using it after a bridge rectifier, but my need was to prevent it, not to know its speed. Looking online earlier, I saw no shortage of patents on SCR's in various ESD protection schemes, including the main one used in CMOS IC's.

    Crowbars are interesting. Assuming an SCR is used with fast rise time being the trigger (via capacitative coupling to the gate), then arbitrary laser diode Vf can be protected, as all will be way below the SCR anode breakdown voltage, but if this is it, how to turn off the SCR later in a powered CW circuit? One way seems to be a tiny capacitance arranged (and charged by the ESD event itself, perhaps) so it puts a momentary reverse pulse into the cathode long enough to bring the SCR forward current below zero till it unlatches, but that doesn't sound likely, I read that SCR's need longer to unlatch (probably what I misremembered, causing me to think they were too slow for ESD protection, and it doesn't help that the data sheet for the last SCR I used (C106M) says next to nothing about transition timing). The cathode would have to be above ground too, which wouldn't help much with the shunting.

    For the reverse bias shunting, the Lasorb pages mention a 'zener' of sorts, but why not just an ultrafast rectifier? Not Schottky, just very fast. Clamping at 1V or so should be ok for a laser diode in reverse, I saw a max of 2V specified for the Rohm diodes, for example. The diodes I used last recover in 25 ns in either direction, and no doubt better can be found now.

  9. #59
    Join Date
    Jul 2008
    Location
    Rotterdam, Netherlands
    Posts
    789

    Default

    ok.. this topic just whent 25 steps over my head...

  10. #60
    Join Date
    Jul 2008
    Location
    Maryland
    Posts
    1,691

    Default

    Hey.. back to the original topic...
    did anyone ever post the original dual driver board?
    Quis custodiet ipsos custodies?
    Solid State Builders Group

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •