# Virtual Ground Circuit

#### GRNDPNDR

Joined Mar 1, 2012
545
I'm playing around with eagle trying to learn how to use it because I think it might be better for PCB design than ultiboard.

I wanted to build this circuit (found here) in eagle and mess around with the PCB (also, I want to use this circuit)

According to the chart, if I change the input voltage, I need to change R1 and R2 to get the desired output voltage.

What are the calculations to determine the values of R1/R2 for the "as split supply" voltages. I.E, if I had 866Ohm resistors, but was feeding it 12V, what would the "as split supply" be?

Secondly, would there be a way to automatically adjust the values of R1/R2 instead of using resistors, so that if the input voltage changed, the correct resistance would be selected.

I would have it connected to my bench PSU, which is infinitely adjustable 0-30V.

#### MrChips

Joined Oct 2, 2009
30,814
Not meaning to derail your project, but take a look at TLE2426, precision virtual ground.

#### GRNDPNDR

Joined Mar 1, 2012
545
what's the max current that thing can pass? Not sure what I'm looking for in the datasheet, but the circuit I posted above mentions being able to handle up to 1.5A, which is pretty good.

Other question relating to the circuit above, the LM336Z2.5 is a 3 pin device, that schematic only shows two pins connected.

Where does the 3rd pin connect ( A)

Last edited:

#### MrChips

Joined Oct 2, 2009
30,814
It depends on what you plan to do with the PSU.
If you want to power a bipolar opamp from a single output supply then use a virtual ground.
If you have a high power application that requires dual supply rails, then use your dual supply PSU.

#### GRNDPNDR

Joined Mar 1, 2012
545
my PSU isn't a dual supply

#### inwo

Joined Nov 7, 2013
2,419
what's the max current that thing can pass? Not sure what I'm looking for in the datasheet, but the circuit I posted above mentions being able to handle up to 1.5A, which is pretty good.

Other question relating to the circuit above, the LM336Z2.5 is a 3 pin device, that schematic only shows two pins connected.

Where does the 3rd pin connect ( A)

I just repaired a device (see other thread) using a LM336 2.5.

On pin was not connected. It was just used as a zener shunt regulator. Two pins used.

#### GRNDPNDR

Joined Mar 1, 2012
545
Good to know inwo.

I'm just messing around with making a PCB of this circuit. It looks very amateur though lol.

I see such nice boards made by people. and then there's mine.

what I need in place of those large banana plug pads, are some kind of large copper pad for soldering a wire to.

#### Attachments

• 59.5 KB Views: 31

#### magnet18

Joined Dec 22, 2010
1,227
free routing tip, use thicker traces, this isn't a logic circuit but a power circuit, bigger is better

in my opinion, also avoid routing a trace between transistor pins unless necessary, it's not exactly the end of the world, but it's a bit too close for comfort for me

edit
why is that circuit using a transistor as a diode?
why not just use a diode?

#### GRNDPNDR

Joined Mar 1, 2012
545
I just used autoroute.

didn't really think about that, but I'm not keen on the idea either.

what size should I use for the traces?

#### magnet18

Joined Dec 22, 2010
1,227
i'd go with 2 or 3 times as thick as you have
it's a bit overkill, but better overdone than underdone

#### GRNDPNDR

Joined Mar 1, 2012
545
so I set them to 24mil.

I'm playing around with grand and power planes now.... interesting stuff.

My biggest problem so far is how to choose the right footprint. a capacitor alone has tons of them, so how do I know which footprint I should use for a given value of capacitor?

Is there like some kind of footprint reference chart or something?

#### AnalogKid

Joined Aug 1, 2013
11,052
There are trace thckness calculators on the web. In MIL design there are rules about how hot a trace can get due to I^2xR losses, but in commercial work a good starting point is 0.1" of width per amp. For power work I go wider. It is a tradeoff between circuit losses and solderability. Fat traces and pads are great until you try to solder to them with a 40 W iron.

Another suggestion. Particularly when doing a prototype or first build, avoid running a trace through a pad. If a trace is going to a pad and then somewhere else, run it near the pad on the way by, and connect to the pad with a short stub. This makes no sense until you have to isolate the pad from the net; it is a single cut with the Xacto knife and no bypass jumper. Been there, done that.

ak

#### crutschow

Joined Mar 14, 2008
34,454
...................
what size should I use for the traces?
For any trace carrying power I'd make them as wide as feasible. That minimizes the trace resistance and and voltage drop from that resistance due to the current. Ground traces should also be very wide (if you aren't using a ground plane).

#### #12

Joined Nov 30, 2010
18,224
why is that circuit using a transistor as a diode?
why not just use a diode?
It's not a transistor.

#### Attachments

• 225.8 KB Views: 15

#### MrChips

Joined Oct 2, 2009
30,814
my PSU isn't a dual supply
Aren't you planning on using your new MASTECH HY3005F-3 triple power supply?

#### GRNDPNDR

Joined Mar 1, 2012
545
yes, but it doesn't have a negative rail so I was going to build this to plug my PSU into, and have dual voltages for working with opamps and whatever else may need a negative voltage.

But the biggest reason I'm playing with this is to learn the ropes of eagle and designing PCBs with it.

#### MrChips

Joined Oct 2, 2009
30,814
What do you mean your triple supply does not have a negative rail?
Every floating power supply has a negative rail.
Connect the positive terminal to GND and you have a negative supply.

#### GRNDPNDR

Joined Mar 1, 2012
545
What do you mean your triple supply does not have a negative rail?
Every floating power supply has a negative rail.
Connect the positive terminal to GND and you have a negative supply.

..... wha..?

#### bance

Joined Aug 11, 2012
315
I gave him the same advice here........

Try it with a battery..... use a dmm, put the pos to pos and neg to neg = positive voltage.
pos to neg, and neg to pos = negative voltage!!!!!

Last edited:

#### GRNDPNDR

Joined Mar 1, 2012
545
I gave him the same advice here........

Try it with a battery..... use a dmm, put the pos to pos and neg to neg = positive voltage.
pos to neg, and neg to pos = negative voltage!!!!!

I know how to do this with a battery, but my benchtop power supply does NOT have a negative voltage, and if it can produce a negative voltage I don't know how to do it.

I only have a chassis ground on my PSU.