Howto: simple Differential Amp + TTL ?

[Xer0]

Hello Folks!

I have a 6ch CMedia ready and want to build an quick & dirty correction board.

Could you help me?

I drawed a plan with the patient help of Nicolas (weartronics) and his great Differential Amp Design. As i cant get the ±9V DC Converter here, only 5 to ±12V, which wont get enought power from usb; i thought of taking one of the 12V wallwarts i have colleting dust, and creating ±5V from it by splitting regulated 10V. the dropout of the IC is 2V so it fits, or not?

Also, i have a only TTL modulated Laser (Pf-114) so i dont need real linear amplifying by now, it should just be able to blank the Laser with Spaghetti and HE. Maybe someone could give me a solution that uses only Transistors for the RGB chan or even raw unamped Soundcard Output.

Sorry im a total noob in Electronics. i can solder even SMD without problems but have no clue how the stuff works and how to design a circuit from scratch and what parts i need etc...

Thanks
TC

[Stoney3K]

Hint #1: Swap Voffset and your DAC inputs. Otherwise, you'd be amplifying the offset and doing something strange with the DAC input alongside it... not what you want, probably. You'll want the offset voltage to act as a ground for the op-amps... for a single amplifier stage in the chain only.

As for the power supply: Try to get some 1N4001/1N4002's and build a bridge rectifier with a few capacitors, and an LM7805/7905 combo. That way, you will get some nice stable +/-5V supply lines with the ground isolated from the adaptor's negative supply, and, more importantly, you can use a simple 9VAC stomp box adapter to power it.

If you want to do everything 'just right', add a fuse and power switch, and make sure the op-amp's power supplies are filtered near the chips with 1uF or so of capacitance to ground.

The TTL output will work fine with the correction amp, you can set the laser to TTL in HE or Spaghetti. The laser's intensity control should be 0-5V anyway, the only difference is that your laser only has on/off control instead of a continuous curve. You do not need to change the circuit because of that.

[ElektroFreak]

Hint #1: Swap Voffset and your DAC inputs. Otherwise, you'd be amplifying the offset and doing something strange with the DAC input alongside it... not what you want, probably.

I disagree. For a buffer, you apply the incoming signal to the non-inverting (+) inputs as in that drawing..

@Xero: I posted this on LPF also: A +-12V DC-DC converter should easily be able to be powered by USB. I'm using one on my board right now..

Here's the beginning of a parts list for you:

TL072 op-amp (dual op-amp, so you'll need a total of 4 according to that schematic)
+-12V dc-dc converter

(you can google search for datasheets for these, or get them from your supplier. They will show you how to hook each one up (which inputs/outputs are which).

those are the main parts. The rest are standard resistors, some capacitors on the op-amp power rails (as stoney3K mentioned), and a DB-25 connector.

I don't think you'll need the +-5V supply if you just power the DC-DC converter straight from the USB connections on your soundcard as shown in the tutorial.

[Xer0]

I disagree. For a buffer, you apply the incoming signal to the non-inverting (+) inputs as in that drawing..

whats the use of a "buffer" and why it have to be a OpAmp with ±5V supply and voltage half splitting?
just wonder why you cant use a negative constant voltage regulator set to -2.25V giving the Vref...

The TTL output will work fine with the correction amp, you can set the laser to TTL in HE or Spaghetti. The laser's intensity control should be 0-5V anyway, the only difference is that your laser only has on/off control instead of a continuous curve. You do not need to change the circuit because of that.

my intent was to reduce the board size for the TTL color channels, by using SMD Transistors or some small parts, or even the raw dac output but kompensated with the negative offest vref. would it work?

@ElektroFreak: About the ±12V DC-DC ... i can get those hre in Germany:
http://www.reichelt.de/?;ACTION=3;LA...;ARTICLE=35064
Are they ok (it's german but the important tech fact are written in english on the bottom of the site)

[ElektroFreak]

whats the use of a "buffer" and why it have to be a OpAmp with ±5V supply and voltage half splitting?
just wonder why you cant use a negative constant voltage regulator set to -2.25V giving the Vref...

The kind of circuit used in the correction amp is called a buffer. That's the only significance of the term.

Op-amps require both positive and negative voltage to operate, so a negative-only regulator wouldn't do (if I understand your question correctly). The TL072 will take up to +/-18V on its power inputs, and the DC-DC converter supplies +/-12V. This is just fine for the op amp, and will give you more voltage on the outputs and thus a wider scan angle and larger image.

You'll also need to add offsets to your modulation channels since the +/-12V will cause them to deliver more than the necessary +5V to the lasers. For the modulation channels only, use the schematic in the middle of the drawing found in this post: http://www.photonlexicon.com/forums/...5371#post55371
For the galvo channels use the circuit in your drawing. This circuit includes the necessary adjustments to set the voltage and offset. The DAC inputs from your drawing correspond to the DC audio inputs in this circuit. Then you just feed the modulation op-amps the +/-12V from the DC-DC converter just like in your drawing.

[Xer0]

I don't think you'll need the +-5V supply if you just power the DC-DC converter straight from the USB connections on your soundcard as shown in the tutorial.

Thx for the partlist. I updatet my Sheme now:

But i dont think it will work so. what should i feed to the Buffer? will it still hold the around -2.25V it shoud provide when i set it with the voltage divider?

[Xer0]

Op-amps require both positive and negative voltage to operate, so a negative-only regulator wouldn't do (if I understand your question correctly)

Not fully. im confused why the "Buffer" needs an OpAmp. It's purpose is to deliver a negative Voltage offset to all Amplifier stages, right? So could i use a negative regulator to provide a voltage in the range of 2.25V?

[ElektroFreak]

that's not what I was referring to actually. Each amplifier stage is a buffer circuit.

For Vref you could simply branch a 7905 regulator off of the -12V from the DC-DC converter to get -5V and apply that to the circuit pictured in the post I linked to for all 6 channels. You would lose your differential signaling, but you gain simplicity since you can then power everything from one source. You can see the offset adjustment part of the circuit (R3). Otherwise you'd need to provide vref separately as in your circuit. I can't see a reason why you can't use a neg. regulator for vref, myself. If you refer back to the post I linked to, you can see that they accomplish a vref source using a resistor and a 5.1V zener diode. Then a pot is used to adjust the -5.1V as necessary.

EDIT: lol.. I just took a closer look at the high-res diagram you posted. Based on a comparison with the other correction amp schematic, I think Stoney3K is right. It seems that the DAC input should be placed on the (-) input of the op-amp. My fault for not realizing that earlier. It doesn't change the power requirements, though. I think you could get away with providing vref directly like you suggested using a regulator circuit, and you should be able to supply voltage to the regulator circuit from the DC-DC converter. If you add R2 and R3 from the post I linked to previously to your circuit in the same fashion and apply Vref there you should be functional.

I'm waiting for drlava to show up here and set everything 100% straight. He's the expert..

[dave]

lazy option - use a Burr Brown DRV134

[Stoney3K]

I disagree. For a buffer, you apply the incoming signal to the non-inverting (+) inputs as in that drawing..

True, but the left-hand circuit in his drawing is wrong: Voffset is applied at the negative input with 50k of forward and 150k of feedback resistance, where the DAC input is at the positive. That's an inverting amplifier with the DAC input acting as a virtual earth, and the offset being amplified 3 times.

I expect he wanted the result to be the other way round, nulling the offset and amplfiying the DAC input.

BTW, if you apply the Voffset straight to the inputs of an op-amp, it doesn't need buffering. The op-amps inputs are high-impedance already. When you use the offset voltage in a summing circuit, it's a different story.

[ElektroFreak]

You're completely correct. In post #8 I realized that you were right when I took a closer look and compared both schematics (this one and the original Dean Hammonds schematic).

[Xer0]

lazy option - use a Burr Brown DRV134

I like the Idea. How i use the Chip?
On the Vsup/Swing Graph its linear, so when i supply it with ±5V it will output 10V p-p?
And the Offset, can it be set in the IC or have to be buffered/nulled externally?


About the wrong OPA use Thing:

This is the original Design by Nicolay (weartronics) i hope i may reupload it here:

[ElektroFreak]

weartronics is at least 10X more knowledgeable about electronics than I am.. if that's what he says, that's what I'd do. That's all I'll say. Perhaps he'll see this thread and give us his $.02.

[Xer0]

Yes i hope this too. I should ask him about the Buffer thing... im not sure if it will still work when i change the Supply to ±12 V instead of the ±5V in his plans

i played with Multisim and the DRV134 a bit, but i think im going it wrong. Clipping here and there...

[Stoney3K]

weartronics is at least 10X more knowledgeable about electronics than I am.. if that's what he says, that's what I'd do. That's all I'll say. Perhaps he'll see this thread and give us his $.02.

I had a second look and the amplifier WILL work, if you treat it as a non-inverting type (with Voffset acting as ground). That's great if you need a fixed amount of gain, but an inverting amplfier is a little easier to build and allows for some more flexibility when adjusting your gain. Both work, whatever floats your boat, I guess.

[Xer0]

That's great if you need a fixed amount of gain, but an inverting amplfier is a little easier to build and allows for some more flexibility when adjusting your gain. Both work, whatever floats your boat, I guess.

I would like an adjustable gain (with fixed offset). Just tell me how to do it Symmetrical (Differential)

[Xer0]

Well...

Im simulating weartronic's Circuti now, but it works "50%" only
The Vref Spurce gives stable 2.25 like it should.
And the first OpAmp really amplifies

But the second one gives not a symmetrical negative signal, instead just an negative offest to the + one!

any clue?

[mixedgas]

Well...

Im simulating weartronic's Circuti now, but it works "50%" only
The Vref Spurce gives stable 2.25 like it should.
And the first OpAmp really amplifies

But the second one gives not a symmetrical negative signal, instead just an negative offest to the + one!

any clue?

The second opamp in the sim screen capture is NOT powered, see the Xs on the power leads? Sometimes the folks who create chip models do not always carry the power signals to where they should be, and this can cause a mess in the simulation code, because it does not have a proper bias on the device.
correct that and run it again and show us the waveforms.

Steve

[mixedgas]

Well...

Im simulating weartronic's Circuti now, but it works "50%" only
The Vref Spurce gives stable 2.25 like it should.
And the first OpAmp really amplifies

But the second one gives not a symmetrical negative signal, instead just an negative offest to the + one!

any clue?

The second opamp in the sim is NOT powered, see the Xs on the power leads? Sometimes the folks who create chip models do not always carry the power signals to where they should be, and this can cause a mess in the simulation code, because it does not have a proper bias on the device.
correct that and run it again and show us the waveforms. Without the power its gonna act as a mess of diodes, and perhaps clip like your drawing shows..

Also you don't need R6-100 ohms in series with the 47K.

Steve

[Xer0]

OK i put a new TL072 at the other ones place and rewired the connections bevore/after R6 correctly

but its still not what i need...

[mixedgas]

, tl07x needs +/- 7V to turn on and swing that high.. its not a rail to rail opamp. according to the data sheet I just looked at. 9V batteries would be a good start, plus a pair of 7805/7905 for the offset pot. 7805 and 7905 need at least 7 volts and -7v to turn on...

May I suggest "Art of Electronics" by Horrowitz and Hill . Google books has it as a limited preview. Also "Opamps" by David Terrell

Rules of the opamp:

The inputs will always try to drive the output so that there is 0 volts between them.
The inputs in theory will have a impedance of infinity and almost no current will ever flow in them
The output in theory will have almost no impedance and is a current source.
Unless its a rail to rail opamp designed for low voltages, opamps need 2-3 volts higher rails then the desired output voltage.
Inputs should not exceed their common mode range, nor be greater then the rails, see data sheet
Inverting amplifiers cannot have gains less then one



Steve

[ElektroFreak]

Unless its a rail to rail opamp designed for low voltages, opamps need 2-3 volts higher rails then the desired output voltage.
Steve

@Xer0: That's what I was referring to when I described powering the op-amps with the +/-12V converter. You just get the 5V to power that from USB.

[Xer0]

Ok, so whats a good R2R opa i could use for?
The only one i know, [IMG]file:///C:/DOKUME%7E1/TC/LOKALE%7E1/Temp/moz-screenshot.png[/IMG]TS912, istn avaiable in Multisim.

My idea was to make -5V with a Max768 boost inverter. Is this possbible or cant i use the common Gnd?

Meh... the ±12V thing creeps me. in need some kind of security that under no circumstances a higher voltage than allowed will pass the x/y output, othwerwiese the galvos go "clonk", 5.1 z diode or so...

[ElektroFreak]

Ok, so whats a good R2R opa i could use for?
The only one i know, [IMG]file:///C:/DOKUME%7E1/TC/LOKALE%7E1/Temp/moz-screenshot.png[/IMG]TS912, istn avaiable in Multisim.

My idea was to make -5V with a Max768 boost inverter. Is this possbible or cant i use the common Gnd?

Meh... the ±12V thing creeps me. in need some kind of security that under no circumstances a higher voltage than allowed will pass the x/y output, othwerwiese the galvos go "clonk", 5.1 z diode or so...

+/-12V is used by both myself and 300EVIL, and it's necessary IMO. Drlavas version comes with a +/-9V converter. +/-5V will not even power the op-amps as mentioned above. You need at least 7V to the op-amps, which will in turn give +/-4V max to the galvos. And no, your galvos won't go pop if you use a +/-12V converter. I use one every day, and my galvos don't even get warm, 300EVIL has had no problems that I'm aware of.. There's no way for +/-12V to reach the galvos as long as you build the circuit right. +/-4V to the galvos (or +/-7V to the op-amps) will only get you a 15-20deg scan angle. That will get old very fast. We're just trying to tell you how to build this circuit so that you can actually use it. If all you can get out of it is a 15-25 deg scan angle, what's the point?

I'm not sure where you got the idea that +/-5V is all the galvos can handle.. The ILDA standard is +/-10V, so they at least must be able to handle that. You won't even get +/-10V to the galvos with a +/-12V DC-DC converter. You'll get +/- 9V.

And on top of all that, why use batteries when by using a DC-DC converter you can get all your power from USB? That's impractical. I believe it's important to try to design things in the most convenient way possible. That way later, when you go to use it, you won't have to worry about dead batteries. Ever. It's all about getting the simplest system with the most functionality.

[Xer0]

I'm not sure where you got the idea that +/-5V is all the galvos can handle.. The ILDA standard is +/-10V, so they at least must be able to handle that.

The Datasheet of my Projector says that. see my first Post in the Thread. its symemtrical 10V p-p only.
So i thought, why generate 12V and then drop them down to 5 again, when i could make -5 with a MAX768 and use them directly. Besides i dont even know yet if it will work with the ±12V DC our Store has. It outputs 2x82 mA at 80% Efficiency, at 5V in this is almost all one port can deliver, so i need a microUSB Y cable too.

The strange thing is that in Multisim the circuit still not works even iv i give +12V and -12V to the OpAmps.
And i done ecactly the circuit like weartronics did. so i must missed something, or it works only in real?