5 Output Opamp based breadboarding DC power supply

[gildasd]

Hello,

Following the silliness of my previous thread:
https://www.eevblog.com/forum/beginners/beginner-opamp-questions/

I decided to not be caught short again.
This circuit is designed to set up a 3 to 5 input supply fast, without fiddling with voltage dividers:
- Run off a 12V DC 2Amp power supply that I already own.
- This is for testing simple ideas in DC.
- I only need from 0.1 to 5mA per output.
- The most important are the 0 to 6V and the 0 to -6V
- 0 to +12V is "nice to have".
- Once I've breadboarded this, it will be soldered permanently on a small test board with easy access pins.
- Based on a 339 358 and a 393 324 because I can leave my house and be back in 15 min with them.
- Total BOM is 10 euros...

Questions:
- Are 1K potentiometers ok? Would it improve usage to go to 4.7k?
- Is it OK to create a ground plane with the virtual ground?
- Do I have to protect the inputs?
- Any beginner stupidity?

[Yansi]

LOL LOL. LM393 is NOT an OPamp. Tha same for LM339.

Did you mean LM358, LM324?

[gildasd]

Well spotted - It's a mistake on my suppliers web page! me needing to read the data-sheet.
Some Comps are mixed with Opamps...

358 Ok 0.24€
324N Ok 0.5€
What about the range? I need to down to 1V...
Isn't a LM1458 LM348 combo better for this?
Or 2ex LM741 and a LM348 to be all 741?

[Yansi]

358 and 324 will go down almost to 0V. (depending on the current sinking)

1458 is a crappy old OZ, which definitely does not work around ground voltage.  See commonmode voltage range in the datasheet. Same for 348 and 741.



Then add in the circuit some supply decoupling (and/or filtering).

The 1k pots would be better 10k (smaller current consuption). 

Add resistor between ground and the noninverting input of the opamp. 100k should be fine. (if the pot fails and the wiper gets loose contact, with these pulldowns the OPAmp's output falls safely to zero voltage)

[bugs]

Be ware, Opamps do not like capacitive loads. If you power a circuit with a couple of decoupling caps it may start oscillating.
Use a couple of LM317L`s and LM337L`s instead.

[gildasd]

Thanks for the answers.

After reading the comments, and then consulting the data-sheets, I'll not be using the 358 nor the 324 because they don't go negative nor to 0...
Their range limit, is basically where I need to be (0 to 1V, positive and negative) if I read well.
But I'll get them anyhow to fiddle.
I wanted to use OP07CP and it's quad pack sibling, but I can't get them today (but I like it's specs).

Due to availability I'll be going for 2 TL084 and adding an extra 2 outputs.
I'll test the LM317L and LM337L (I wanted to do this anyway).
The opamp creating the virtual ground will get a 100K between ground and the noninverting input.

I'll be using Sockets to be able to modify quickly if I cock up.

[gildasd]

Bleuargh!
My "el cheapo" power supply is better than expected... There needs to be 50kOhms on the input to bring the amps down to the 10mA range...

[gildasd]

It's ALIVE  :
I'm getting positive AND negative!
- But the range is off, at 12v, I'm getting 1.23v to 5.4v (crap) and -4.2v to -0.9v (not bad).
- The pots have big differences in usage that are not those that are measured unloaded. So pairing will be by trial.
- OP07CP works better than LM741 that works better than LM358 (for now). But I've only got one OP07CP...
- It works only with the input at 7.5V and above...
- The 100K resistor makes the circuit much smoother and easier to adjust.
- Excess of Amps solved by wiring the pots properly...

I'll do an updated circuit layout tomorrow
I think the main problem is the virtual ground. I'll try using a separate voltage divider to get it exactly at 6V...

[dannyf]

they don't go negative nor to 0...

That became an issue only because of the design.

[gildasd]

they don't go negative nor to 0...

That became an issue only because of the design.

It's no big deal, they cost nearly nothing and it's fun to try and see if it's possible to make it work anyway..
I'll probably use one to create a separate double 0/12V range.
I really like the way the JFET opamps work, I would have never discovered them if not for looking for the 358.

I'd like to get my hands on some Sfernice/Vishay pots...

[gildasd]

OK...I'm in a sticky place  ...

All measurements are between the "virtual ground" and the relevant Vout.
All Opamps are Ti LM741CP.
The power supply is producing 12.6V
The pots are all Omeg LinA 1K

The problem is in the 6V to 0V to -6V, and more precisely the 6V to 0V: Vout1
I cannot get it to go below 1.94V...
- I've tried various 741CP's (Ti and ST) and a OP07CP to more or less same result (max 0.2V diff)...
- Changing the pots. No difference.
- Virtual ground with separate resistors only made Vout2 min and Vout1 max go a bit lower.

The negative part (Vout2) works perfectly, -0.92v t -4v. If I tweak it I can get it to -0.2v (at that point the 741 gets confused).

- All I do on Vout1 only seems to affect it's max (6V goes up or down) but leaves it's min to within 0.2V...

I'm running out of ideas...

[gildasd]

Big improvement, it's now usable:

Essentially by doing what I told not to do 
- All the OpAmps are now connected to +12V and 0V
- The virtual ground is connected between the pots. This is not needed in theory:
    -All the Vouts are more stable: "latches".
    -There's a bit more range.
- Having the +6V via a separate 10k voltage divider works MUCH better.
- All the the pots were "burned in" after I noticed (nearly fudged up) that putting them in series on 12V 2amps for a couple of minutes reduced their variability.
- I then paired them to within 3%.
- Only one pot is still out of speck (1125 Ohms) - so 'ill keep it for "unpaired" work.

741 side:
- Vout 1: +5.86V to +4.5mV. (+5.81V usable)
- Vout 2: -4.04V to -0.7mV. (-3.97V usable)

358 side:
- The ground is used as -Vcc... I'm a rebel.
- Vout 3: +4.72V to -1.5mV. (+4.70V usable)
- Vout 4: -5.99V to -2.7mV. (-5.93V usable)

There's a few tests to run still but the 358 for the Vout 741 for the virtual ground combo is the best so far
I think the imbalance between the positive and negative is due to the slight difference between pots.
This probably can be solved via trimmer on their inputs.
 

[gildasd]

On the negative rail, when I get near the max (around -4.68), it slams into the positive (around +5.5V)...

This is the same circuit, same parameters as before...
I ran the test will all combinations using a LM741 and OP07 (for virtual ground) and get the exact same reaction at the same point...
The pots were also changed and that did nothing.
I ran the previous circuit as a reference and it's always within 2mV of yesterdays results...

I don't get it, especially for the TL084 that is supposed to be very good around O and negative...

Is this normal?

[dannyf]

I don't get it,

That's the price you pay for learning by doing.

Is this normal?

Yes.

[gildasd]

Oh golly gee, what is I to do?
There's the option of using 2 lm324's with one going 0 to +6v and the other 0 to -6...
But for expendiency's sake I'll use 2 lm358 and 2 lm741 for 6 outputs.

[gildasd]

Uses one LM741 (for the virtual ground) and 3ex LM358 (for the outputs):

- All 6 outputs now go from positive to negative.
- The 3ex 358 circuits are identical.
- On the drawing, the VoutA section is repeated 3 times.
- After mucho mucho testing (oh the joy of data generation) the LM358 gives the best range by 0.8V.
- Range is +4.75v to -6.05v with 12v in.
- The outputs are very stable.

If anybody sees any cheap and worthwhile improvements, now is the time 
I'll post the BOM and the 100x50mm protoboard layout when it's done.

[gildasd]

Trying to get everything under the lid without any shorts...

[gildasd]

6 times +4V to -6V...
The pot stalks need to be trimmed.
But it's done.

My first project, all grown up, I need to shed a tear.

Now to get some class work done