Howto: simple Differential Amp + TTL ?

[Xer0]

Yes, that will work.

I thought you guys can get anything across EU borders duty free...

Velleman is in Belge...

Steve

Yes thats true. thats why so much China stuff is reshipped from UK

Didn't know that Vellemann is in Belgium. found their german site and Datasheet for the Kit, and it was 50% durch, i should have guessed it...

[Xer0]

Ok i now have a working simulation!!!

Dean drawed a circuit for me which works flawlessly


Now there are just 2 important (maybe) last questions: if i...

1. use R2R OPA's and a MAX768 for the -5V... can i use a common ground (connected to the USB GND too) for everything, or do i need galvanical isolation somewhere?
2. use the ±12V DC-DC converter... same question about ground and what could i do to prevent overvoltage my faul (=more than 3.8V on dac input port) - with 2x 5.1V Z Diodes shifted parallel to each Output?

[ElektroFreak]

Ok i now have a working simulation!!!

Dean drawed a circuit for me which works flawlessly


Now there are just 2 important (maybe) last questions: if i...

1. use R2R OPA's and a MAX768 for the -5V... can i use a common ground (connected to the USB GND too) for everything, or do i need galvanical isolation somewhere?
2. use the ±12V DC-DC converter... same question about ground and what could i do to prevent overvoltage my faul (=more than 3.8V on dac input port) - with 2x 5.1V Z Diodes shifted parallel to each Output?

USB ground is fine, same for the converter. Since the +/-12V is only applied to the op amp power inputs and ground, it's isolated. The only way the +/-12V could reach your galvos is through a direct short somewhere. Most galvo amps have some protection against overvoltage, but should a catastrophic short occur you never know what might happen. That's just part of the fun of building your own electronics. Just double check everything first and check the output before connecting the galvos with a scope if possible, or a DMM if that's all you have to make sure the signal looks good.

[Xer0]

Or Zener's on the Outputs, right? oh please let this be right

Besides... instead of the variable resistor between +5 and -5V i want to put a more temp stable LM336-2.5 into the Vref adjust it and invert it with a opamp. this worked fine in Multisim, will it work in real too?

Or is there any IC which by getting negative supply inverts an positive input exactly without external parts ?

[mixedgas]

Or Zener's on the Outputs, right? oh please let this be right

Besides... instead of the variable resistor between +5 and -5V i want to put a more temp stable LM336-2.5 into the Vref adjust it and invert it with a opamp. this worked fine in Multisim, will it work in real too?

Or is there any IC which by getting negative supply inverts an positive input exactly without external parts ?

No. a

[Xer0]

No a? no a what?

[mixedgas]

No a? no a what?

a= 15 character minimum post

[Xer0]

Okay, so far so good.

I have my stuff alltogether and managed to get the -2,55V Reference today.

Parts used: LM336-2.5 for stable adjustable postive Ref, ICL7660 to generate -5V, and TL071 to invert Vref.

[Xer0]

Ok, here we go:

Im currently not soldered anything to the SoundCard; im using a USB plug for power and for testing the opamp's stage behavior, im was using a 1MOhm potentiometer to adjust any signal voltage between +5 and -5V. now, after i thought the high resistance of it could be the error cause, i replaced it with 3x 10k R's at which i can get 4 different voltages for testing.

I attached my circuit plan, its exactly how i builded it. between +5V/GND and -5V/GND are a 100µF elko each for stabilising, otherwise currently no capacitors are used, except the ones for the ICL7660 negative voltage converter.


So i did a row of different measurements on different setups. Here are the results:

         1  |  2  |  3  |  4   

-5V :  -4.82 -4.80 +4.84 +4.95
 A  :  -1.55 -1.55 +0.94 +1.88
 B  :  +1.71 +1.71 -1.03 -2.06
+5V :  +4.99 +4.99 -4.74 -4.76


[ADD]

-vref + -0.94 = 3.49
-vref + +1.03 = 1.52


[SUBST]

-4.81 + -(0) = -2.41
-4.81 + -(GND) = -4.76
-4.81 + -(-vref) = -2.27
-1.55 + -(0) = -0.58
-1.55 + -(GND) = -1.16
-1.55 + -(-vref) = +1.39
+1.71 + -(0) = +0.64
+1.71 + -(GND) = +1.29
+1.71 + -(-vref) = +3.84
+4.99 + -(0) = +2.50
+4.99 + -(GND) = +4.96 
+4.99 + -(-vref) = +4.95 

+vref + -(0) = +1.28
+vref + -(GND) = +2.56
+vref + -(-vref) = +4.95

0 + -(+vref) = -2.55
GND + -(+vref) = -2.54
-vref + -(+vref) = -4.76

0 + -(0) = 0.00
GND + -(0) = 0.00
-vref + (-0) = -1.28

In the table, the 1st collumn is just measured with multimeter on the Vin 3*10k divider, withoud any load. second one is the divider voltages feed into a voltage-follower opamp circuit, 3rd one is fed into inverting-amplifier with 1x gain and 4rd finaly with 2x gain.

So this is strange. the "real" voltages on the divider are coming out unaltered with a impendance amplifier / follower; but are changing extreme when i use something with gain.
Also in the last test setup, the substractor, when i feed in 2.55V i got the half out, without any other voltage is going in which could alter it (0 in the table means unconnected cable)
On the other side, whenever i do operations with buffer stabilised negative n/o directly from LM336 reference circuit Vref ±2.55V Voltage; the output never differs from ideal theory values!


What the Fuck?!


I never get "zero-level" signal output on 2.55V "offset-level" input. SC is a CM106 so the Vref is set right. OPA's are rail2rail and able to deliver Vout = Vsupply - 0.01V !
Here's a bigger scan: http://www.abload.de/image.php?img=image0022ff8c.jpg