Protection circuit for Laser diode (UV)

[Blueteeth]

Hi,

Been a while since I posted here, off doing work and life etc.. But I'm back with a niggling problem. A while aog I got some UV laser diodes off ebay, some plain diodes, and one in a mount - all around 100-150mW. Inevitably, because the mount had two wires coming out of it directly to the diode - after much plugging and unplugging the diode went. Could be ESD, maybe I plugged it in wrong, or maybe just over current with me lighting candles with it

So, because the inside of the mount is rather small - 16mm long, 9mm wide inside the tube, a micro driver circuit isn't ideal. Instead I would just like to put a protection circuit in there, soldered directly to the diode, with two (maybe 3) wire-lead coming out for plugging into a half decent driver circuit. This way I can use multiple driver circuits, buck, boost, able to adjust the current and therefore output of it.

I still can't decide what would be 'enough' protection, and something to fit onto the very small PCB. I could of course just have a two-track PCB, in parallel with the diode, with a TVS diode and a cap - but I feel I could do more, and this wouldn't provide decent protection.

So, here's my tihnking. This is what can destroy a laser diode:

1.Over-voltage: ESD spikes, from having an exposed lead, over-shoot from supply, accidentally plugging into wrong supply

2. Over-current: being over zealous with current trying to maximise output, plugging into wrong supply etc..

3. Transients: These are covered by the above, but could mean very fast rise times and nanosecond events.

4. Temperature: Caused by too much current.

Ok, As I said, I'm planning on an external driver board which will control current and maximum output voltage, but would like a small proteciton circuit in there to minimise the chanecs of me popping another diode. I don't think temperature would be much of a problem, as its mounted in a brass mount (not ideal as copper, but it'll do, its only 100mW).

Aside from a TVS diode, a cap, would some form of basic current limiter in there be a good idea? I was thinking of using a shunt transistor for over-current, along with a zener for over voltage - but zeners are quite slow, and TVS diodes have a rather vague clamp voltage. Given the small space, SMT is required, but I'm confident I could fit a two-transistor circuit on there, along with diode and some passives.

Perhaps i am just being overly cautious, but as I don't want just 'wires soldered directly to the LED coming out', I'm using a small PCB in there anyway, just to solder wires to for mechanical robustness - seems a pity to waste that PCB space when these diodes are so electrically fragile. Any idea's?

Cheers!

BuriedCode.

 

[alec_t]

Could you squeeze in a SMD voltage/current regulator IC (just a thought; haven't checked what's out there)?

 

[Blueteeth]

I could! and an integrated solution would be ideal, especially like an MSOP8, or a SOT23 device. There are however some restrictions

Not only the size, but also, the current. I want to drive the diode up to 105mA. As these diodes develop up to a 4.8V drop, AND I would like to make an external driver capable of being powered by USB, that really doesn't leave much headroom at all (even without protection circuitry causing vdrop). That also leaves out large series resistors.

Efficiency isn't really a concern, except in terms of heat generation. And with size, an LM317 is out. But... as I said, I'm not really bothered about putting the 'driver' itself into the mount, just some form of protection so when I plug/unplug the diode form various different driver - it doesn't fry, either by ESD, or hot plugging.

It seems like quite a narrow field, althoughI'm sure there are specific devices for laser diodes - most of which are drivers, with set limits that aren't adjustable, thats why I'm going with an external driver

 

[alec_t]

A quick look shows SOT23 regs are typically 100mA max

 

[Blueteeth]

OK perhaps a ittle more information. The UV diode has its case connected to the cathode, which is a bit odd for laser diodes, but makes it easier because then the mount is grounded. This, of course, means that any low-side current sensing within a proteciton circuit (not a driver, just a limit) isn't ideal.

The low voltage operation (~4.8V @ 105mA) could operate an opamp, but I don't want anything to drop more than a 0.6V really. I guess transient protection can be a TVS diode, maybe some device specific for the task used in modern electronics). This leaves over-voltage protection and ESD as a primary function, with perhaps a current limit as the secondary function.

I brought discrete transistors into this because they are cheap, small, and could shunt excess current to ground, or limit voltage. Once agian, this isn't trying to be a 'regulator', I'm sure if I plugged it into a dodgy driver, something in the proteciton circuit would blow - but that the point, its cheaper to replace a diode or a transistor, than a laser diode Thats where the shunt idea comes in - it would short-circuit.

Google is full of miniature circuits for taking in a voltage, say 7-10V and driving the diode, but as I want full control, the driver will be external, this is just incase something goes awry when plugging it into different drivers.

 

[Mr RB]

A good cap and a series resistor is all you need locally at the laser diode, as you can control current externally at the PSU driver.

 

[Blueteeth]

A good cap and a series resistor is all you need locally at the laser diode, as you can control current externally at the PSU driver.

Good, at least that makes life easier. I think I'll still add a TVS just in case, it can't hurt and will double up as reverse bias proteciton anyway.

As always I tihnk I'm just over complicating things, because any 'over-current protection' I've been adding has essentially turned out to be a constant current regulator - almost making the external PSU redundant. Perhaps I am just paranoid having two leads (or a connector..) coming out of the diode, sometimes completely unconnected. I had that before, so I guess I blew it by connecting it wrong, or over-doing the current. With such limited space, whilst it would be 'cool' to add in maybe an opamp, or comparator to shunt excess current to ground, its over kill as you said - that will be in the PSU.

Yet another restriction is, I wanted to be able to power it from a USB port. An LM317 constant current circuit would drop too much as the diodes Vf is around 4.8V. But USB might only output 4.5V anyway, so I tihnk I'll probably end up having to make a buck/boost PSU that can cope with a wide input range above and below the diodees Vf. That is why I was hesitant to add any significant series compoennts with the diode, like a large value resistor (30ohm+).

 

[Blueteeth]

Here's a rather *over-kill* circuit I came up with. It's a classic two-trasistor current source, using a PMOS as a pass element to reduce the dropout voltage. This leaves the biggest voltage drop over the current sense resistor, at a Vbe. Considering the Vf of these laser diodes is 4.8v and can go above 5V - I guess I have to drop the idea of being able to run this off USB, so adding ~0.6 to 0.7V isn't too bad.

Essentially it's a driver in its own right, but not very precise. With the penalty of adding a maximum of 0.7V to the voltage required (when driving at 120mA, absolute max of these diodes) it provides over current protection that kicks in at the maximum, just incase the external driver decides to pump in more. With the addition of three components - two resistors and a zener, it also creates an upper limit of the input voltage required for it to run.

Sometimes the output of a driver might spike voltage, or if its say 12V, then 12V - 0.7V - Diode Vf will appear across the MOSFET as it tries to keep current constant. thats 11.3 - 4.8V = 6.5V. With the maximum current limit at 110-120mA, thats 715mW power disappation in the MOSFET. Whilst thats not big, the device is small (inside a 18mm x 9mm tube), so it could over heat (not a good thing) or blow, short circuiting, connecting the laser diode in series with the power supply via the current sense resistor. So to prevent the MOSFET disappating too much, when the input voltage rises above roughly the zener voltage + 0.7, it turns off the MOSFET - disconnecting the diode.

So, in terms of plugging this thing into an external driver, it limits current to a maximum, as well as limits voltage to a maximum. Neither of these thresholds is well defined, but should be enough as a protection measure. When the input voltage is within range, and the input current is below the maximum, the circuit effectively just passes current trhough, and is inactive.

Granted, I don't have ESD protection, but I would have thought a simple ESD device across the diode would be enough, as well as a small ceramic cap. As for reverse protection, I would have loved to have used the MOSFET for that, but its the wrong way around, and adding an extra diode in reverse across the laser wouldn't take up space.

Comments and suggests are more than welcome. I do have a terrible habit of over complicating something, and at the same time, missing the original purpose of whatever it is I'm designing (ESD, revrse voltage, sotping the diode dying). But whilst I have some discretes and passives to hand in 0603 packages, extra protection can't hurt can it?

Again, just to reitterate, this is less of a 'driver' (which should be precise, and controllable) and more of a protection measure so I can keep unplugging and plugging into drivers, without fear of any mistake I make killing it.