[FEELER] Cheap, small Laser diode Drivers

[amkdeath]

Ok, so here are my results after some testing:



VERSION 1.0 - Boost Converter

I've tested this version of the driver at various voltages; two current configurations are posted above.
Load: Same as below

So the device operates with a minimum of 2V input, and dies at a Maximum of 7V. In order to be safe, I advise running the boards with a minimum of 3.3V, and a maximum of 6.5V. Especially with the 500mA version, The driver should not be run with anything under about 3.3V when connected to a diode.

The up-swing on the switch/inductor is shown here (This is when the switch turns off, and the inductor starts driving the load). I found that by adding a 0.047uF capacitor in parallel with the output capacitor shorts one of the major ringing frequencies from before to ground. Adding a 3ohm resistor in parallel with that damps the entire thing.

The result is what you see above, and I think this much ringing on the inductor is natural given the output level. The added ground plane might reduce it even more, we'll see.

VERSION 2.0 - Single Ended Primary Induction Converter (SEPIC)

(This is the coupling capacitor voltage)
Essentially a Boost converter, followed by a buck-boost.

This little guy seems to work pretty nice. I did some pretty extensive testing, and I only managed to blow one, and that was only because I accidentally fed it 20 volts. Even then, I just swapped out the IC and it worked fine again. There seems to be some transients on the switches, I could probably increase efficiency by getting rid of those.

Here are some stats:
Load: Diode Connected Transistors, 5x 3906 with series 2W 1Ohm resistor. 5.3V drop.
Current setting: Rs=Two 0603 1Ohm surface mount resistors in parallel, expecting 0.38A output current.

Setting 1: 2.8V Input voltage
Output Voltage: 5.3V
Output Current: 338mA
Input Current: 950mA
Duty Cycle: 0.640
Efficiency: 68.1%
Peak Inductor Current Ripple: 41.3mA
Peak Output Voltage Ripple: 36mV (based on load resistance)

Setting 2: 5.9V Input voltage
Output Voltage: 5.3V
Output Current: 338mA
Input Current: 350mA
Duty Cycle: 0.0343
Efficiency: 86.8%
Peak Inductor Current Ripple: 20.9mA
Peak Output Voltage Ripple: 8.5mV (based on load resistance)

2.8Vin was the lowest setting that I could safely run the driver at without interfering with its performance. At this setting, it still had pretty decent efficiency and held the 5.3V/338mA output very steadily. The board was warm to the touch after ~2mins of continuous use, but I could not run one to failure at this current setting.

As the input voltage is increased, the efficiency goes up. As the driver has to do less boosting, the main inductor is switched less intensely, duty cycle goes down, and efficiency goes up. Due to the smaller amplitude in the inductor current, there is also a smaller amount of resulting output voltage ripple.

Vin Range: 2.8V-6V
Iout Range: ...(coming) TESTED:[338mA, ..]

Conclusion
It seems these units are ready to go. They can be used in a variety of applications, from voltage constrained applications to bigger more current intensive projects. I anticipate the maximum current I could set these to to be around 1Amp, but I'll test this later this week. I currently have 18 board left, and could possibly distribute these to early adopters after I solder them up. Or I could sell as a kit if you're good with surface mount soldering.


Pretty rock solid performance with solid current configuration, meaning there is no pot. It's set to work at a certain current, and it'll do the heck out of that. I would like to go more towards this route, in hopes of keeping the board cheap enough such that you wouldn't mind just buying another one if you need another current.


I have received Push buttons and pots, and will begin testing some new designs with those shortly after this batch.

Best,

AMK

 

[Jstr]

Awesome, thanks for the update! Do u have any pics of the driver?

If you are playing around with different types, I'd suggest possibly doing a high current (4~4.5A) buck.

 

[amkdeath]

What diodes are y'all using right now? When I left the popular reds were LD-215's and blues were 8x/6x burners.

What kind of current ranges do the diodes currently use need? I'll look into multi-amp driving as well, I'll let you know.


Best,

amk

 

[The Lightning Stalker]

Right now, it's everything between 40mA for
the 5mW reds to 4.4A for the NDB7A75.
That is why I am working on a few different
linear drivers. Some for the low range and
one for the high range. Then of course
there are the really high current IR diodes,
some of which want 15A or even more in the
case of diode bars.

 

[bdgreenb]

@amkdeath an affordable driver for our single modes (including the mm pl520-b) would be amazing. A cure all driver for every diode we use is becoming more and more unlikely with the huge current gap between the sm diodes (<1A) and the new balls to the wall 445s(4A+)

 

[bloompyle]

I'm in a weird group where I need <100mA drivers. Some of my diodes (like the 685nm) need around 35mA.

Though readily available drivers from 200-600mA every 50mA will do for most folks. After that it's the heavy hitting blues where people need 1-3A. Generally 1.8, 2.2, 2.4, 2.6, and 2.8A.

Then it's the real heavy hitters. The 5W blues that need something like 3.2A or something.

Another thing is drivers that will work for reds. We have a ton of drivers for blues, greens, and 405's. Though boost drivers kill reds. So we need good drivers for reds. Not as relevant as boost drivers though.

 

[amkdeath]

I will update the above post sometime in the immediate future with the test results from the Mark 1 Boost style driver. Bottom line is, they both work well, one is considerably cheaper than the other to make. Both designs need a couple of touches here and there to work best and can be shrunk down a bunch.

The current boards I have work fine, and I am putting them up for sale before I go ahead and get the new designs ordered. Since these are the BETA units, making the purchasers "Early Adopters" (aka evangelists) I will be offering them with a 20% Discount.

So that's $8 Shipped for the Boost Units, and $12 shipped for the SEPIC units

Discount code is: "IHARTAVERY", enter at checkout. This offer lasts until 10/03/14

Link: ARK TECHNOLOGY REPOSITORY

TO DO LIST:
Update Boost driver performance specs
Explore Pushbutton functionality
Add noise-cancelling and protection elements to circuit design
Explore multi-amp currents
.....


Best,

Amk

P.S. You need to be logged into the site to use the coupon... Sorry!

 

[crazyspaz]

You have scoped the drivers output, i assume?

 

[rhd]

Scope results should be posted. I think we should, as a forum, demand scope results from all drivers being sold.

I say this as a hypocrite without a scope.

 

[amkdeath]

Scope results should be posted. I think we should, as a forum, demand scope results from all drivers being sold.

I say this as a hypocrite without a scope.

Will post as I did above. All scope readings are from the switching diode in the circuit, the current and voltage on the output don't change wildly enough to make meaningful measurements, but those values have been calculated in case you are curious.

Actually, if you look at the Web page the scope reading from the rising edge of the switch is posted there.

These drivers could and will be much, much better, it's just a matter of getting the R&D done (expensive) and working with the users (us) to make them desirable.

If anyone is willing to put down some diodes (I'm running out) to use for testing these drivers, I could give the boards to you as cheap as possible so you can hook up the prototypes as I churn them out and make sure everything is A OK... just a thought, but I think it could be helpful.

Best,

Amir

 

[rhd]

The thumbnail photo of a scope above doesn't show us anything meaningful.

You need to post a non-thumbnail image (or several) of scope results on the actual output of the driver, when connected to a test load.

I'm not a scoping expert, so maybe ARG or someone else will want to correct me on this. As far as I can tell, there's no reason to scope an earlier stage in the driver, when you could be shipping the output itself.

You may not think the current/voltage ripple is significant at the output, but that's kind of the point - to show us.

Remember, your drivers are low current drivers, that will be used on low current diodes. So ripple that might be irrelevant for a 9mm 445, could be a problem for a 50mW single mode green.

 

[rhd]

The thumbnail photo of a scope above doesn't show us anything meaningful.

You need to post a non-thumbnail image (or several) of scope results on the actual output of the driver, when connected to a test load.

I'm not a scoping expert, so maybe ARG or someone else will want to correct me on this. As far as I can tell, there's no reason to scope an earlier stage in the driver, when you could be shipping the output itself.

You may not think the current/voltage ripple is significant at the output, but that's kind of the point - to show us.

Remember, your drivers are low current drivers, that will be used on low current diodes. So ripple that might be irrelevant for a 9mm 445, could be a problem for a 50mW single mode green.

 

[amkdeath]

There is no ripple on the output, that is the point of the output capacitor. Otherwise it would not be a regulated constant current source.

The pictures (which you can click on to enlarge) provide you with the details on the only part of the circuit that an O-Scope is meaningful to use as a measurement tool.

Due to capacitor charge balance, as well as s-domain inductor-voltage balance, the current in the secondary inductor is equal to the current flowing through the laser diode. The two best ways to increase the efficiency of the driver are decreasing the forward voltage drop on the schottky diode, and preventing as much external noise from entering the switching circuitry as possible.

Since the circuit switches at 1.6MHz, traces only a couple of cm long, wires, etc can both radiate as well as pick up noise. This interferes with the operation of the driver, screws with feedback, and has all kinds of other repercussions.

Therefore, using an O-Scope, you can monitor the switching that is going on, measure duty cycles (which is the only missing variable when calculating efficiency, and can only be derived quadratically), watch for ringing, calculate losses, and change up the components. The only reason you would have a non-steady output voltage would be because the circuit is not operating properly (entering thermal shutdown, for example).

If you really want to, I can post a picture of a straight line from my O-scope measuring the DC output of the regulator, but once again, that would be using an AC measuring tool to look at the output of a DC-DC converter. If you want to enforce a tradition of posting O-Scope readings on all drivers being sold on LPF, I think it would be helpful to promote posting pictures that will actually allow us to get a feel for some of the characteristics of the driver itself (for comparison purposes), rather than just a verification of the implied.

Best,

amk
P.S. I just finished testing an upgraded 700mA variety, it works but becomes unstable as it gets hot. I need to wait for the next batch of boards to come in so I can test with proper grounding.
Spoiler alert: I think the next batch of boards will be 4 layer ones. I've added test results from the smaller driver above.

 

[crazyspaz]

One thing I immediately notice at first glance is those board seem to have a lot of wasted space. Surely you could get them down in size a fair amount, unless the other side is considerably more populated.

 

[paul1598419]

Subscribe.

 

[amkdeath]

One thing I immediately notice at first glance is those board seem to have a lot of wasted space. Surely you could get them down in size a fair amount, unless the other side is considerably more populated.

Definitely a valid observation. It would also help a lot if I moved the output capacitor and shottky diode closer to the actual chip. These are modifications I'm working into the next batch.

Best,

amk