How to create super bright laser beams

[planters]

What about the bi-directional nature of the TEC? If I'm following you then I think you would need two channels for each Fet and one channel for the thermistor or digital temp probe and the serial port as well. This exceeds the 24 pins on a typical Pic or Admel. I'm already out of my comfort zone and I'm picking (pun intended) my son's brain on this one. The metal bar might be the most practical solution, but the individual TEC controlled diodes woudl better let me discover the limits of this project. I really want to investigate this with the individual module approach as well. Can you help me more here?

[logsquared]

After posting, I realized I had over-simplified it. The diodes that need to be cooled can probably be uni-polar TEC (2 processor pins). But the heated diodes will have to be bi-polar TEC (3 chip pins). The problem with bi-polar is the need for a bridge circuit. The problem with a bridge circuit is the need for bootstrapping the fets. This increases the complexity quite a bit. I am sure there are bridge chips with all that built in.

[kecked]

Pwm to the vase of a fet or transistor will work fine. Multiplex the screen. The picaxe has built in h bridge

[andy_con]

If someone can create one of these multi tec controllers all controlled via an lcd I'll take one

[planters]

So, if the H bridge is already a component of the Picaxe, then the Fet still needs to be duplicated for each TEC channel so that the high current drive is available in each polarity. Right? Where I remain stuck and bare with me here, is that the control system is pretty evolved and I am happy to be limited to a desk top for the initial work (multiplexing a tiny screen is not necessary at this point), but the interface between the digital output or a PWM output or an analog voltage output, to simple (ready made) current/voltage switching is hard.

[logsquared]

I can't see how to configure a picaxe to do an h-bridge without an external transistor bridge of some sort.

I found this bridge chip: TLE5205-2 it can do 5A up to 40 volts. All it needs is 2 inputs. PWM on one of its pins would be heat and pwm on the other would be cool. Internal circuitry keeps the bridge from shorting if you accidentally input on both pins.

The picaxe-28X2 has 22 IO pins. Need one to drive LCD and one for a button pad input. Then 3 pins for each bi-polar drive( 1 for thermistor 2 for bridge drive, heat/cool)

[logsquared]

but the interface between the digital output or a PWM output or an analog voltage output, to simple (ready made) current/voltage switching is hard.

PWM is the only way to go here. True analog would generate too much waste heat. The chip I mentioned above is really all that is needed to amplify the computer/pic signal to the TEC. They are not cheap chips but much easier that building from discrete parts. You could mount these to a heatsink and point to point solder them without need for a PC board if you wanted. They have build in flywheel diodes so all that is needed is an external bypass cap close to the chip.

[planters]

I'm getting feed back that maybe a few pins would be required for the display? But, the concept sounds closer.

Is anyone confident enough that this would work that that they would be willing to do this. I mean, make one? If 22-a few-1 leaves enough for 6 channels then I will take one. Again, my son could/would program the front end to make it sexy and convenient and I would use it to evaluate the control of these multiplexed diodes. Meanwhile, I am moving forward with the first approach with the stainless bar. This will go a long way to proving the practicality of the system and the problems that need to be dealt with optically and thermally.

[kecked]

I'll trade a spatial filter for the multi Tec��. I'll take a crack at it. I have an encironment that I can build the whole thing vertially. It's no differen than making a robot go forward and back. In fact I could skip the bridge and do it with relays run b 4066 bilateral switches or a simple flip flop. Come to think of it you never ever hear heat or cool cool so it is always alternating so a pulse to a d flop flop to set state is all you need.

Adc reads thermocouple and determines heat or cool to start. If it's the wrong state it doesn't matter as it will quickly cycle to the right state. Have to read the slope off the adc for hysteresis to prevent overshoot.

[logsquared]

Using relays for heat cool is a great idea! Not really following your use of a bilateral switch. Did give me a thought though ... Using a analog mux for the thermisors could save pins. Also, 4 bit decoder could set the heat/cool relay for up to 16 channels.