nubm44 driver that can be modulated with audio signal

[gattis]

Does this exist? I took a look at the blackbuck 8m which has analog modulation but from the plot on dtr's page with a 2khz square wave input, it seems it would not respond at freqs much higher than 2khz.

I don't know too much about voltage controlled current sources but I took a (perhaps naive) crack at it with just a mosfet biased into the saturation region. Attached a schematic. Any reason why this wouldn't work? I could just try it but thought I'd have some eyes on it for a sanity check so I don't bust my new diode (nubm44-81). Much appreciated.

 

[Cyparagon]

The whole point of a current source is its current remains unchanged, and you're trying to change it with external limitation. Not a good idea.

We can help you a lot better if you actually tell us what you're trying to do, rather than post a vague design spec with no context.

 

[gattis]

The whole point of a current source is its current remains unchanged, and you're trying to change it with external limitation. Not a good idea.

We can help you a lot better if you actually tell us what you're trying to do, rather than post a vague design spec with no context.

Sorry bout that, didn't want to bore but there's a couple use cases I have in mind... simplest one I want to try is to laser cut audio into the groove of a vinyl record (lets say acryllic record instead so I don't get poisoned by arsenic). I'll have a turntable running at 33 1/3 rpm and the laser is mounted above it, onto a linear actuator driven by a stepper motor at constant speed rolling from the outside of the record directly to the center at a slow speed. so the optical power of the laser module has to be modulated so that it cuts the groove along at a depth corresponding to the audio waveform.

TTL/PWM modulation won't really work without digitizing the signal first which I'd like to avoid, plus according to my calculations... if you have an audio sampling rate of say CD-quality, 44100 hz = 22676ns, and even if you can pwm with 1ns pulses (unlikely), you would be limited to log2(22676) = 14.5 bits of resolution.

I thought of replacing the turntable with another stepper motor and have it rotate at 1/10th the playback speed and cut a lot slower but there are other issues with that and stepper motor motion at slow speeds would start to leave tiny discrete jerks for your stylus to run over.

 

[gattis]

The whole point of a current source is its current remains unchanged, and you're trying to change it with external limitation. Not a good idea.

We can help you a lot better if you actually tell us what you're trying to do, rather than post a vague design spec with no context.

I think the point of a current source is that you have to "push" current across diodes in order for electrons to hop the gap, as opposed to a resistor for example where you just put a voltage across it and it will "pull" the current. Diodes certainly don't mind that current changing as a function of time. I have a laser projector for instance that scans out video frames with a tiny mems mirror and it has to modulate the current of the R,G, and B lasers at every pixel in a 720p frame and then do that at 60 frames per second, which is way faster of a modulation than mere audio waveforms.

 

[Cyparagon]

Correct, diodes have no problem being power-modulated. Current sources have a problem being modulated by an external impedance in series with your load. That's not the way to do this.

So in short you're trying to laser-cut a vinyl record. Given the spot size of a high power blue will be very large in comparison to a more traditional cutting implement, you're going to be lucky to fit anywhere near the same data. It's also going to have a heat affected zone (HAZ) that will possibly warp your disk or bleed over into an adjacent track. The smoke will also likely contaminate adjacent tracks. The groove shape may not even be playable. I don't think it'll work but it might be a fun experiment.

An off-the-shelf solution used to be the "flexmod P3", but it seems to have vanished from the marketplace. Most products that call themselves "analog laser drivers" will work. They translate a voltage input (almost always 0-5VDC) to a current that is programmable via a trimpot on the board. It'll need a DC offset, but that's simple to set up with a DC blocking cap feeding the centerpoint of a voltage divider tied to 5V and ground. You can make your own, but I would advise against it in your case. The engraving consistency is your main issue. Don't add driver design to the litany of problems you're going to have to solve.

 

[gattis]

thanks! those pointers are very helpful. with a first pass i should at least get an idea of how hard the path forward will be. most of my back-envelope calculations say this is in the realm of possibility, but will definitely have to make some major sacrifice like maybe I only get 10 minutes per side. my hope is that with the power of this diode i can rip through it at full speed and get a clean cut. i will play with various materials, maybe it is even possible to heat a thermoplastic right up to melting point with the laser and have a cutting stylus come in right behind it to do the groove and I won't need to modulate the laser.

 

[gattis]

flexmod p3 exactly what i need! 4A and 160khz modulation bandwidth. biasing and scaling the audio to 0-5v is not a problem.

 

[neodavid]

Did you find a solution?

I may be misunderstanding, but a high power voltage to current amp can be made with 3 components...


This circuit will cost you about 2 bucks... and it could easily be fed by a pulsed voltage source (160Khz as you described)... which I can also provide if you like.

Apologies if I am not following the question. I'm just getting into the laser aspect, so I have not proven this in the field. I just got my first diodes this week.

 

[gattis]

thanks for that. yeah it seems as though the flexmod p3 is no longer in existence. i wish whoever was selling it would open source the design if they don't intend to sell more but that might be a lot to ask of someone. there's also this design that would work for high current using an lm338 (pic attached) but there seems to be some controversy around its design, and personally i dont understand how it works. I agree with @Cyparagon that I probably shouldn't be adding a driver design to my list of problems, but I'm not finding any off-the-shelf solutions, at least not under $300, which is probably around the price point where I would instead get to tinkering.

your circuit seems like it would work fine if we had ideal components. I'll try to simulate it with a couple real ones and see what the freq response and distortion look like and if it will pass a smoke test at 5A.

 

[neodavid]

So why 1MHZ? That's awfully high, what's the application?

If you want to drive the mosfet to saturation you could avoid heat by varying the pulse width instead.

I try not to go over 10khz so i can use standard $1 components, but I don't have any specialized size requirements where need super tiny inductors for buckboost, etc.

For most things I just take an arduino nano and program it for whatever pulse width I want with a given input voltage, and then the mosfet always runs cool.

 

[gattis]

I personally don't need 1MHz. I am trying to modulate the laser power with an analog audio signal so it just needs to support analog modulation in the audible range of 20Hz-20KHz. However if analog is not possible (some drivers only support TTL) and we take that off the table, then I would need to approximate it by sampling and using PWM to represent each sample. If I wanted to do that relatively "hi-fi" and sample at CD-quality rate, 44100 samples per second, 16 bits per sample, I would need 2^16=65535 possible duty cycles to choose from, so my PWM "clock" would have to run at 44100Hz*16 = 705.6kHz which is up there close to 1MHz.

 

[gattis]

er no, TTL binary modulation would need to be 44100Hz * 65535 = 289MHz. So that's probably out of the picture.

 

[neodavid]

I think (ha) you are overthinking it. But I could also be remembering things incorrectly. It's late, and my brain is laughing at me.

What you are describing is 'sooooo digital'.

You won't have to sample the pulse width at an insane frequency... that's too cumbersome and will actually introduce noise and create incredible circuit complexity. You can accomplish a lot better (and easier) result with a little filtering circuitry.

If I recall, Nyquist theory says you can pull this off with about 40Khz, or double your max frequency. No one can really hear 20khz anyway, at least I can't... 10khz is SO freaking high and uncomfortable... it's only good for that 'clean' sound some audiophiles claim.

I'm missing something. My brain says there is a drawer i haven't opened.

1AM, nightly night.

 

[Cyparagon]

1MHz is the upper limit if built correctly. It doesn't mean you'll need or use all of the bandwidth. A 10GHz oscilloscope will measure a 2Hz signal just fine.

This circuit will cost you about 2 bucks...

Yeah, looks good on paper, but falls flat on its face in the real world. The scaling of the input signal, the laser diode threshold current, shunt selection, and the tendency of the system to oscillate make this a lot more complicated than you think. Trust me. I've built it.

OP, if you're still having difficulty finding a driver, I would suggest asking (or searching) at photonlexicon.com. Those folks deal with analog drivers a lot more often than the pointer crowd here.

 

[neodavid]

Yeah, looks good on paper, but falls flat on its face in the real world. The scaling of the input signal, the laser diode threshold current, shunt selection, and the tendency of the system to oscillate make this a lot more complicated than you think. Trust me. I've built it.

OP, if you're still having difficulty finding a driver, I would suggest asking (or searching) at photonlexicon.com. Those folks deal with analog drivers a lot more often than the pointer crowd here.

Interesting, I have no experience yet, but assumed a laser led would be similar to a standard led, which you can modulate easily, although frankly I've never tried doing it with a modulated constant current. I just set the threshold and combined the audio power to it.

So, applying the same concept, what I hear you saying is it would be a lot easier to set the lase current, and then pump in the audio signal if you want to drive it with AM... and to get more range, you could do the same and pump FM through it at near full current.

PWM would work this way also.

the only thing I'd add would be a current limiter to prevent overdrive...

For some reason constant current sources stuck in my head, due to my inexperience... but why bother.

Although... thinking again, I could put a diode in series and create a secondary 'protected' circuit that kept the laser just past it's threshold current, that getting rid of the circuits tendency to bounce around the junctions whack behavior.

Okay, two diodes... one each, feeding the main LED, from different power sources.

Thoughts? Mine are only theory.