Signal Mixing

[saconni]

Hi people, first post

Facts:
* I have a project to make a light-board to manage N-channels in 2 different scenes controlled by 2 masters...
* I do not know a thing about electronics, I can not even remember the Ohm thing...

Question:
I've "designed" a circuit (see attachment) to mix two signals in the next way

out = max( master_a . channel_1A, master_b . channel_1B )

mean, get the max value between the product of blah and blah

masters go from 0v to 12v and channel_1A and channel_1B are values from 0 to 1

I found many troubles in the way, and I am pretty sure that the circuit wont work (by the moment)... but I need to resolve the "get the max value" thing

I used two transistors, but there must be something else there... my simulator says that is ok... but I do not believe him.

by the way... transistors are picked randomly (except for the PNP part) from a long list of components

could any one lead me in the right path??

ps: this is GPL, anyone can use this exclusive design for its own

 

[BrownOut]

I have a feeling that this circuit has no chance of producing the output you want, if I understood what you want.

How do you design a circuit with no knowledge of electronics?
Please explain how this circuit is supposed to produce your desired result.
What is the purpose of producing this result?

 

[saconni]

mmm... to many good questions...

*the circuit "works" or need to work in the next way:
there are two **continuous** incoming signals ( master_a and master_b ), they go from 0v to 12v...
each one of these signals gets into a multiplier-configuration opamp, the gain of my configuration goes from 0 to 1... i think
then I use another opamp ( one for each signal ) to substract the value of the original signal...
hopefully, i get for each master a 0...1 multiplier

ie: if the master_a have a tension of 6v and the gain of the amplifier is 0.5 i should get a 3v out signal for this channel, in master A
master_b could be at 12v and the gain of the amp on this master could be also at 0.5 with a 6v result

I need to mix those result signals (3v and 6v) choosing the grater one

* I draw this circuit dragging everything from the net. The opamp configurations are known circuits ( not for me )

* I want to compare the result signal with a ramp signal, synchronized with the 50hz from the wall, to produce a pulse, to trigger a triac circuit

 

[saconni]

sorry, it takes me so much to answer the question that I forgot the most importante

thanks!

 

[BrownOut]

It's OK, it often takes a couple trys to understand what someone wants to do. I'm still not sure what the purpose of the system is, but... I'll try to help. The op-amps seem they should work as you describe, though the multiplier isn't quite 0-1. More like 0-1.9. But my real problem is the two transistors. Firstly, just mixing the signals doesn't really give you a MAX(...) function, at least the way I understand that function. If you mean:

output = A if A>B or
output = B if A<B

then you won't get that by mixing the outputs. Aside from that, the 2N2904's won't operate properly because you don't have any bias connected to the collectors. You just have the collectors connected to a meter, but no voltage. The collectors need a voltage to operate correctly.

Wifey says it's bedtime. We shall talk of this later.

 

[saconni]

thanks, I'll appreciate it

but I have no question about the substraction (actually I do, but) I am concerned in the MAX function

my idea of those transistors was: when one tension is grater than the other only one of the transistor will operate... and vice versa... I need to know if my though was correct, and what needs to work

 

[saconni]

damn... again....


thanks!

 

[saconni]

schema updated =)

 

[BrownOut]

Now you have the collectors connect to ground thru the resistor. That won't work either. They have to be connected to positive rail.

I'll think about your max function today. Gotta run now.

 

[BrownOut]

I think I understand how your "max" circuit works now. It makes theoritical sense, but the trick will be to get the measurement right. First of all, connect the collectors to V+ thru the resistor. That should at least get a measurement. I can't gaurentee that measurement will be accurate, however, but it will be a step in the right direction. You may have to play with the value of the resistor you're making the measurement across, and maybe need to add a base resistor.

I'll give this some more thought...

 

[saconni]

Sorry, but do not understand how a positive tension in the collector will help. ( the collector is the one with no arrow, right?? )

And how much tension? it should be less than the tension in the emitter??

Thanks for all your thought!! I really want to leave the simulation and get into the real workbench(r) =) ... but I do not want to burn my appartment

 

[BrownOut]

The collector of an NPN transistor HAS to be connected for the transistor to work. NEVER allow the collector's positive voltage go below the base potential. It won't blow up or anything, but it will not operate properly. For unity gain, you need an emitter resistor with a resistance roughly equal to the one in the collector, which is in your case, 100 Ohms. I can't do a full analysis of what is required, but try to find a "common emitter NPN amplifier with unity gain"

You've done some pretty good work for someone who doesn't know about the "Ohm thing."