ld drv

[phenol]

ld drv

 

[phenol]

front

 

[Benm]

At least provide some component values

I have built some opamp based drivers, but this one is a bit puzzling at best. You scale down the sense current and then compare it to a reference that is variable from Vss to Vdd... i think that might be very finicky to adjust?

 

[IgorT]

At least provide some component values

I have built some opamp based drivers, but this one is a bit puzzling at best. You scale down the sense current and then compare it to a reference that is variable from Vss to Vdd... i think that might be very finicky to adjust?

Right now the explanation is only on candlepowerforums... I'll give you the specs, pasted from there:


By Phenol:
-------------------------------------------------------------------
The current that flows thru the LD is calculated as follows:
I=[Ucc(5V)-Ur(R1+R2)/R2]/Rsense
In my case R1=10k, R2=15k, Rsense=1.8ohms
Ucc must be stabilized because the stability of I is proportional to its own.
The op amp's output must be able to swing within power supply rails. CMOS opamps have rail-to-rail capable outputs.
Ucc bypass caps not shown for clarity, however, they must be present for stability
Whatever we do, we must make sure that there are no current spikes exceeding the maximum values for the specific LD during transient events, such as toggling the circuit on/off and applying modulation...
-------------------------------------------------------------------

More coming soon (i hope)..

I hope it helps!

 

[a_pyro_is]

I should Prob. let a mod say so first, but I can't help myself. And please don't take this as a flame, just questions.

Is this Avery's Photo Hosting site? And what happened to the 10 post limit for links?

[benefit of the doubt] maybe he's planning to come back and post a few more details himself [/benefit of the doubt]

@Phenol
Please do come back and tell us a bit more. I LOVE seeing what others have made as it always gives me even more ideas. It's just disappointing to see a couple pics and no info about it other than what someone else pasted from another forum.

@Benm
You seem to know (as many do) a lot more about electronics than do I. would you be willing to start another thread describing a few of the constant current drivers you've mentioned elsewhere on the forum? ie an opamp, two transistors and a voltage divider, any others you think might be useful here.

 

[phenol]

I should Prob. let a mod say so first, but I can't help myself. And please don't take this as a flame, just questions.

Is this Avery's Photo Hosting site? And what happened to the 10 post limit for links?

[benefit of the doubt] maybe he's planning to come back and post a few more details himself [/benefit of the doubt]

@Phenol
Please do come back and tell us a bit more. I LOVE seeing what others have made as it always gives me even more ideas. It's just disappointing to see a couple pics and no info about it other than what someone else pasted from another forum.

@Benm
You seem to know (as many do) a lot more about electronics than do I. would you be willing to start another thread describing a few of the constant current drivers you've mentioned elsewhere on the forum? ie an opamp, two transistors and a voltage divider, any others you think might be useful here.

:-[
Sorry. I never thought somebody might even be remotely interested in this circuit.
R1=10k, R2=15k. The trimmer pot is a multiturn 20k one. With it adjustment is not as finical.The only benefit this circuit has is the GND potential on LD's cathode. The reason the voltage from Rsense is scaled down lies in the specs of the op amp - for relatively low currents the voltage in question is really close to Vcc, it exceeds the Vcm specification for some op amps and they no longer behav linearly. I would like to point out that while this may not be the case with LMC6482, which has a common-mode input range of at least Vdd, MCP601 /the one that I actually used in this circuit/ has Vcm=Vcc-1.2V and I was compelled to use a voltge divider.
The cap across the op amp is ceramic, 100p. The one in parallel with the LD is 1u, also ceramic. It is there to prevent transients, such as current spikes, from killing the LD. 1uF is a bit of overkill and will affect modulation response. I never actually modulated this circuit. It is just a static current source.

hope this helps....

 

[IgorT]

I should Prob. let a mod say so first, but I can't help myself. And please don't take this as a flame, just questions.

Is this Avery's Photo Hosting site? And what happened to the 10 post limit for links?

[benefit of the doubt] maybe he's planning to come back and post a few more details himself [/benefit of the doubt]

Of course more details were coming.. This is experiments and modifications threads, so this driver circuit belongs here, i think.

I have to test it a bit, cos it seems interesting.. I wonder how small this can be done?

 

[IgorT]

Sorry. I never thought somebody might even be remotely interested in this circuit.

People are always interested in driver circuits.

Can you please also give a simplified explanation of the circuit?

- How much voltage are you feeding it?
- How much comes out on the other end (on the LD)?
- What are the minimum and maximum output current ratings?


Thanks!

 

[a_pyro_is]

Sorry. I never thought somebody might even be remotely interested in this circuit.

People are always interested in driver circuits.

Can you please also give a simplified explanation of the circuit?

- How much voltage are you feeding it?
- How much comes out on the other end (on the LD)?
- What are the minimum and maximum output current ratings?


Thanks!

No, I'm the one who should be sorry if I sounded angry with you Phenol. :-[

And as Igor said we're always interested. Even if it's something I don't think I'll ever use, I'll still get some new ideas from it.

 

[Benm]

You seem to know (as many do) a lot more about electronics than do I. would you be willing to start another thread describing a few of the constant current drivers you've mentioned elsewhere on the forum? ie an opamp, two transistors and a voltage divider, any others you think might be useful here.

I've posted an opamp based design before, check this thread:
http://www.laserpointerforums.com/forums/YaBB.pl?num=1189529869

Main differences are that the diode ground is not at the circuit ground, making it harder to use in flashlight-style-portables, but fine for labbies. It uses a CA3140 opamp that has a common range down to ground a even a bit below. The pot is used all the way, giving you control from zero to 300 mA over the full turn.

I'll draw up the 2-pnp driver in a bit, as it is really simple to build and has the LD common to ground.

 

[phenol]

[quote author=phenol link=1193741215/0#5 date=1193766584]Sorry. I never thought somebody might even be remotely interested in this circuit.

People are always interested in driver circuits.

Can you please also give a simplified explanation of the circuit?

- How much voltage are you feeding it?
- How much comes out on the other end (on the LD)?
- What are the minimum and maximum output current ratings?


Thanks![/quote]
It is powered from a 5V stabilized source.I use LM7805 with a 1n4001 in series with its input to protect whatever's behind from polarity inversion. Then, the supply voltage of the whole thing must be higher than 7.8Volts.Typical 7805s need a drop of about 2V to work properly. Add some 0.7-0.8 volts for the Si diode.... The 7805 needs a heatsink. It cannot dissipate more than 2W without a heatsink...

The voltage that comes out depends on the load. When no load is present the voltage would be something higher than 4 volts... In its essence this is a meaningless working condition as the current has to flow thru the air... If there is a load connected, the voltage depends on the load itself and the current that flows thru it. For resistive loads the voltage U=IR. For non-linear loads such as junctions....refer to the respective datasheet. Red LDs have forward drops anywhere between 2.4 and 3.something volts. For instance, mine develops 2.51V at 150mA. The forward drop depends on the current that flows thru the junction. The higher the current, the higher the voltage across the junction. The relationship, however, is non-linear.

The minimum current is ideally 0 and is noise-limited. For bettr accuracy the Rsense should be increased. The max current depends on certain circuit restrictions. For instance, for maximum of 3.5V across the load, the circuit will source regulated current not higher than ~320mA, for 3.2V - 485mA... If Rsense is 1ohm, the max current for 3.2V across the junction would be ~748mA. I recommend inserting a 1k resistor between the output of the opamp and the base of the transistor because at certain /marginal/ conditions the opamp will try to pull the base to 0V which results in too much current /as much as the op amp is capable of sinking/ flowing into its output. If a P-channel MOSFET is to be used, this will not be an issue.

The behaviour of this and other similar circuits can be simulated and predicted. I use OrCAD /schematics, PSpice A/D/

 

[CHP]

Don't eliminate the the capacitor in parallel with the LD because of the power up transient with this circuit. When power is first applied the op amp output voltage is zero and the darlington transistor becomes forward biased sourcing a current slug to the LD until the op amp output catches up and regulates the current. The duration of the slug of current is dependent on the slew rate of the op amp and the internal delay time. The capacitor should be sized to absorb the transient.

 

[phenol]

Don't eliminate the the capacitor in parallel with the LD because of the power up transient with this circuit.  When power is first applied the op amp output voltage is zero and the darlington transistor becomes forward biased sourcing a current slug to the LD until the op amp output catches up and regulates the current.  The duration of the slug of current is dependent on the slew rate of the op amp and the internal delay time.  The capacitor should be sized to absorb the transient.

True that. I managed to kill one IR LD because of such a current slug.
For shorter settling times high speed op amps should be used.

 

[IgorT]

It's the opposite with mine.. Usually the IR LDs don't want to die, while the red ones are more than happy to..

That (and because i don't need them) is why i use the IR ones for testing the circuit, before i attach the red ones.

 

[SenKat]

As an FYI to all that are interested - ANYONE can attach one picture to their post by using the attach erm.....picture option whitst they post ? The 10 post limit is for LINKS of any type - any URL gets whacked if you do not have 10 posts....but anyone can use the photohosting that LPF uses

 

[Benm]

True that. I managed to kill one IR LD because of such a current slug.
For shorter settling times high speed op amps should be used.

The capacitor is usually a good idea. Regardless of slew rate, while powering on the opamp briefly works off a voltage too low for reliable operation. What actually happens depends on the opamp, but usually there is a little current spike with this sort of setup. 10 uF is sufficient to surpress that to below LD threshold voltage.

I would also recommend using a 1k or so resistor across the diode, this improves turn-on behaviour as you crack the current up from zero using the pot.