Using a DC/DC converter to power the laser

is there much advantage having the case grounded?

it must be on gnd potential if, for instance, its heatsink is not isolated from the chassis of the host and minus terminal of batts is also in contact with it.
generally i prefer to have all massive metal objects/heatsinks/ hooked to the lowest potential in a cct - normally ground/chassis of equipment for safety sakes due to risk of shorts

woop said:

is there much advantage having the case grounded?

Click to expand...

Well, if you use a flashlight host or a metal enclosure, and want to use the body as a heatsink, you need the LD negative pole at GND potential, or you short the IC..
Otherwise you have to insulate the AixiZ module from the host, and lose some heatsinking...


I don't care about this, since i'm using plastic enclosures, and a peltier for the open can.


EDIT: I see Phenol already answered, while i was writing the post.. So... What he said..

phenol said:

these are two scopeshots of what's on the output /CH1 -> AC coupled, DC component removed/. in the first one channel 1 shows the output noise almost directly across the output cap. the metal shielding barrel of the probe was touching the pcb GND as close as possible to the output cap, and the tip was also as close as possible to the 'hot' end of the same cap. moving the test point by only a millimeter or two resulted in twofold increase in noise amplitude. channel 4 shows the voltage across L2. spiking occurs during rising and falling edges.
in the second pic is one of the spikes dilated in time.
The test conditions were as follows:
Vin=2.72V
Iout=278mA
Vout=2.74V /3x1N4005+1N5819/

Click to expand...

what scope do you have?
i am thinking of getting a scope, they are so damn expensive, especially the usb ones

tek TDS3034B - 300MHz, 4 channels; it has floppy, LAN, rs232, centronics, GPIB...no USB ... digital storage scopes are a must if you intend to capture non-periodic events

Phenol... Did you manage to get the second sample order through by coilcraft?

I just recieved my second sample shipment today, but they only gave me four this time.. They gave my girlfriend 10...


But if they don't want to send you a second sample order, i could probably forward at least these four to you..


If you think the YSD needs a larger coil, i can spare these and maybe more... Let me know...

Alternativelly, i can give you an internal e-mail, of someone, who is much more willing to send samples.

thanks, igort, i havent ordered yet, these days we had an important customer audit and i have been a little too busy to my liking. im going to carefully revise my need of power coils tomorrow morning and order once again. i would encourage you to keep whatever parts uve gotten for ur ysd samples, as i can find similar coils locally or even wind smth working myself.
thanks again

phenol said:

i would encourage you to keep whatever parts uve gotten for ur ysd samples, as i can find similar coils locally or even wind smth working myself.
thanks again

Click to expand...

No problem.. Just trying to be helpfull a bit.. Especially, after bugging you with so many questions...


But you're lucky, that you can actually buy coils locally.. My engineer just told me, that i must be the first one in Slovenia, to actually have CoilCraft coils...
Our shops here don't sell anything like it.. Everything has to be ordered internationally and is very expensive..

If these same coils should also work with the YSD, then i have just the right amount of them and the ICs to build seven circuits... If not, i'll try 10uH ones.


BTW: I finally found some time, and i'm testing my circuit now...

I just successfully tested my LM3410 circuit, and it works great..

In fact, it even starts up at 2.3V and runs at this voltage with an efficiency of 65.4%, which i think is good for such a low voltage.. I'm amazed by the fact, that it would probably work from only two Ni-MH cells without modification... With it it would suck them dry.

I also tested the current stability at different voltages, and it remains constant, from 5.5V, down to 2.3V.
I didn't measure the efficiency at higher imput voltages yet, simply because i was so impressed at this low voltage.


Oh, i tested the circuit on a LED and on a dead IR LD, which actually worked for a while, without a cap on it at all.. It died completelly later, but i don't think it was the circuits fault.. It died when i was moving it's cables....


Next i'll try it with a good IR LD with a cap on and measure the efficiency at higher voltages, starting with 4.2V and down to 3V..


As it is, it will already work with a single Li-Po cell, and keep the current constant over it's entire voltage range.. It could even run with a primary lithium cell!

I must say i'm very happy with this circuit....



EDIT: Efficiency at 4.2V input 73%... I thought it would be better, but i guess it's not that bad either..

BTW: Something weird happened, when i was measuring the circuit efficiency..

I had a multimeter attached to the LED, to measure the output voltage. Then i disconnected the positive lead of the PSU from the PCB and it accidentally touched the multimeters positive lead.

I suddenly noticed a rising column of smoke, but i didn't see, where it was coming from.. I thought the IC was dead.. But when i connected everything back again, the thing worked.. It still works just like before, but i think the LED is less bright.. Not sure tho..


I have absolutelly no idea what happened.. Need to look at how everything was connected, to figure out, where the current went and what made the black smoke.

Right now, i have an IR LD driven by my circuit, to see if it's safe. It has a 1uF ceramic capacitor directly on the leads..

So far so good...
Or not.. The IR LD seems to have dropped in power considerably.. I first noticed, that the faint red spot became even more faint, and it also shows up less on the camera...

Damn.. Maybe i have overdriven it, on the other hand... I don't know how much current it was supposed to take, i had it at 213mA.
Need to do more testing...

Also, with this LD the efficiency was only 63%... With the LED, at 5V and 213mA, the efficiency was 72%... This doesn't make much sense to me.



EDIT: I just tried the circuit with my sick Blue Ray LD, which needs more than 200mA to make a beam.. (I don't understand why it still works at all)
The circuit fed it 5.48V @ 213mA, and the efficiency was 74.2%...

Why is efficiency better at higher output voltages?!?


Next i'll try with a flat chip red, to see if it can take it..

as of the smoke column - touchin the + to the output of ur circuit would pass current thru the load and the shunt resistor. either, or both, may have smoked. the switch pin of LM is DC decoupled with C3, so no current should have flown there. FB pin is almost grounded by the low resistance of the shunt, so only a very high drop there would cause damage to this input.
regardin IRLD, i killed one at only 150mA, so you may have overdriven it indeed.  furthermore, i also noticed the presence of a start-up transient in the output current almost twice as high as the nominal steady state current. a rather discouraging observation   . it would probably be safe if across the load there is also a 100uF electrolytic cap and a subsquent double-check for presence of oscillations probing the output and the power input terminal with a scope. under some circumstnces i have observed sine oscillations there that went away when i placed a 100uf tantalum at  the output.
in short, this little goodie can be sonofa$%^&*.

phenol said:

i also noticed the presence of a start-up transient in the output current almost twice as high as the nominal steady state current. a rather discouraging observation   . it would probably be safe if across the load there is also a 100uF electrolytic cap and a subsquent double-check for presence of oscillations probing the output and the power input terminal with a scope. under some circumstnces i have observed sine oscillations there that went away when i placed a 100uf tantalum at  the output.
in short, this little goodie can be sonofa$%^&*.

Click to expand...

So what now? I can't use it unless i put 100uF tantalums on the LD? Damn... Or is that 100uF tantalum + 100uF electrolytic?
I was really happy when i thought those tiny ceramic SMD caps would do.. But i do have some 47uF tantalums as well.. Should i just put two of them in?

I would like to test the circuit with the flat chip red, but i kinda promissed someone i would turn it into a laser.. So i don't necessarily want to kill it.
What about tantalums on the output of the circuit for this spike? Or should i do both (the more the merrier...)?


I don't want to just accept defeat, after spending so much time on this circuit. I want it to work..

Unfortunatelly, i don't have a scope at home, and i couldn't find the time to test it at work. On the other hand, i finally have time to play with it, so i don't want to waste it.


BTW, do you have any idea, why my efficiency get's better, when the output voltage is higher (current being the same in all cases)?

I put a large ceramic SMD capacitor (value above 20uF, don't know exactly) and a 47uF tantalum directly on the leads of the flat chip 16x red LD and another 47uF tantalum directly across the circuit output..


It seems to work fine... I'm gonna try running it for a longer time, to see what happens.


Would this current ripple show it's effects immediatelly, or would it just slowly decrease the LDs lifetime?

well, the ceramic should almost completely notch out those short-lasting spikes. cycling the power on-off may push a strong startup pulse/much longer than the periodic spikes) thru the diode. the 47u cap will probably suffice in absorbing it

phenol said:

well, the ceramic should almost completely notch out those short-lasting spikes. cycling the power on-off may push a strong startup pulse/much longer than the periodic spikes) thru the diode. the 47u cap will probably suffice in absorbing it

Click to expand...

Thanks.. Do you think that the second 47uF tantalum directly on the PCB is helping any with these power on spikes, or would it be better to have both of them on the LD? Does this matter with the startup pulses?


If the swithch on spikes could be absorbed on the board itself, it would all be MUCH easier.. Soldering these to the leads directly is not much fun... Especially, since they don't fit without extending the leads.. I would rather have both tantalums on the PCB.. The small ceramic ones are much easier, when it comes to this... You think this would work?



I'm gonna let the red run for a while now.... See what happens... I really want to be able to drive my lasers from a single Li-Po or even two Ni-MHs...
It's been working for around 10 minutes now, with no drop in the brightness..




BTW: I'm getting 67.5% efficiency with the red LD.. It could be worse, but i still don't understand why it gets better at higher output voltages.. What kind of efficiency did you get, and is it constant at different output voltages?