Laser current control signal circuit design?

[3dgeo]

Hi,

building new toy – 40W CO2 laser, currently working on electronics and want to make 'elegant' solution to control/limit laser current.
Laser PSU (MYJG-40W) reads analog voltage (0-5V, filtered) and sets current accordingly. PSU capable of delivering up to 15mA(@~18kV), laser needs up to 12mA.

My goal is to have 3 stages to limit/control current:
1. Limit max current so laser will not be damaged (R29 limits signal up to max 'safe' voltage)
2. Adjust signal with potentiometer up to max 'safe' voltage
3. PWM POT output with MCU according to software needs.

So far I'm at this point:


As I'm more in to digital stuff my question would be is there something wrong with my design? Any improvements?
Also what POT value should I go with? I have 100K on hand, but I assume signal current will be way too low and laser EMI could mess it up?

P.S. I will be using shielding cables for everything.


Cheers!

[ledtester]

How does the user manual describe how the laser power control pin works? I have a lot of questions...

For instance, most analog inputs are high impedance which means they won't draw a lot of current so I'm curious how the 50 ma limit comes into play.

[phil from seattle]

Your diagram shows inputs only.  Is it safe to assume the middle input (Laser Power control) is actually an output? 

Does your laser (power supply?) power control actually do filtering or does it expect a filtered input?  Generally it does filtering but worth while verifying.

Your mosfet needs to be selected to work with your MCU voltage output level.  You might want to add a driver transistor to allow a higher gate voltage if that's an issue.

I'd look at using a 10K pot but see if the 100K works with your laser (easy enough to test). 

[Terry Bites]

You cant get 5V compliance from a linear current source with a 5V supply. You could use a switching converter eg https://www.analog.com/en/products/lt3519.html

[ajb]

The validity of this approach really depends on the input to the laser power supply.  Particularly the pot resistance really depends on the impedance of the input.  If the input has a 100k pull up or pull down, then a 100k pot isn't going to work very well (depends on what it ends up set to and if the input pulls up or down).  Likewise the bandwidth of the input matters as far as your PWM signal as phil points out.  If the input is heavily filtered and you set the PWM frequency well below that filter's cutoff then you're fine.  If not, then problems. 

One other thing to consider, when the MCU is in reset the transistor is held off via R27, which means the laser current signal will be at max voltage.  Is that what you want?  If not, consider another approach that will reliably give a minimum signal when the MCU is in reset.  Could be having the MCU gate the 5V signal to the top of the divider, or could be something else...

The most straightforward approach may be the best, especially if you don't know what the input characteristics are.  There are a fair number of MCUs out there with built-in DACs now, including some of the new ATTiny parts, so if you don't need tremendous resolution or accuracy that could be a nice way to go.  If you need a better DAC, then take your pick of external ones.  You might need an external op amp to increase the voltage range, but that's also an opportunity to adjust the max current setting by adjusting the gain (you could also adjust the reference voltage for the DAC).  This gives you a true analog output, and the output impedance can be whatever you want by placing a series resistor on the output of the op amp.  Since you don't need filtering like with a PWM signal you may be able to get better settling times although that may not matter. 

[3dgeo]

How does the user manual describe how the laser power control pin works? I have a lot of questions...

For instance, most analog inputs are high impedance which means they won't draw a lot of current so I'm curious how the 50 ma limit comes into play.

Yes, it is high impedance, 50mA is 5V rail limit, not a signal input current.

Your diagram shows inputs only.  Is it safe to assume the middle input (Laser Power control) is actually an output? 

Yes, program I used for some reason doesn't have arrow other way.

Does your laser (power supply?) power control actually do filtering or does it expect a filtered input?  Generally it does filtering but worth while verifying.

I assume it does cuz input can be PWM or analog.

Your mosfet needs to be selected to work with your MCU voltage output level.  You might want to add a driver transistor to allow a higher gate voltage if that's an issue.

Yeah, I know that stuff, I'll pic mosfets that works with 3.3V.

I'd look at using a 10K pot but see if the 100K works with your laser (easy enough to test).

I'm thinking about 5K pot to get at least 1mA to make signal stable, is it too much?

[3dgeo]

The validity of this approach really depends on the input to the laser power supply.  Particularly the pot resistance really depends on the impedance of the input.  If the input has a 100k pull up or pull down, then a 100k pot isn't going to work very well (depends on what it ends up set to and if the input pulls up or down).  Likewise the bandwidth of the input matters as far as your PWM signal as phil points out.  If the input is heavily filtered and you set the PWM frequency well below that filter's cutoff then you're fine.  If not, then problems. 

You pointed out a lot of thing that was in my head
Input is high impedance with no pull up/down. PWM – 20kHz to 50kHz.

One other thing to consider, when the MCU is in reset the transistor is held off via R27, which means the laser current signal will be at max voltage.  Is that what you want?  If not, consider another approach that will reliably give a minimum signal when the MCU is in reset.  Could be having the MCU gate the 5V signal to the top of the divider, or could be something else...

Oh yeah, good point!

The most straightforward approach may be the best, especially if you don't know what the input characteristics are.  There are a fair number of MCUs out there with built-in DACs now, including some of the new ATTiny parts, so if you don't need tremendous resolution or accuracy that could be a nice way to go.  If you need a better DAC, then take your pick of external ones.  You might need an external op amp to increase the voltage range, but that's also an opportunity to adjust the max current setting by adjusting the gain (you could also adjust the reference voltage for the DAC).  This gives you a true analog output, and the output impedance can be whatever you want by placing a series resistor on the output of the op amp.  Since you don't need filtering like with a PWM signal you may be able to get better settling times although that may not matter.

That looks a bit too overkill. I should point out that PSU itself has 5V 50mA rail output for specifically this purpose – to create 0-5V analog or PWM laser current control signal and other signals. MCU – most likely STM32F411 (no DAC).

At this point I need validation on general concept – if I'm on the right track and what recommended current should be to get stable signal. I will breadboard out mistakes like you pointed out.

[ledtester]

Maybe you can just use the MCU to generate the analog output directly:

"Using PWM to Generate an Analog Output" Microchip  TB3250

http://ww1.microchip.com/downloads/en/Appnotes/90003250A.pdf

[3dgeo]

Oh silly me – I just realised that PSU has another digital laser enable input, so I can PWM that input and use pot with limiting resistor for analog input. 
There are 3 digital inputs actually, but I thought they are used only for endstops, lid open detection and pump enable, tho I can't see reason why I can't use one for PWM.