Best way to do a virtual ground?

[large_ghostman]

What is the best way to provide a virtual ground for a circuit? I have access to +12V but I need to have +/- 5V + +/- 0.8V in a circuit. Peak current is 1.1A

Opamp yes, but best way to do it? Also i have very limited parts as its for an overseas project in a country with really **** parts availability. Put it this way, when Noah stopped sailing he opened this countries only electronics supplier, worse than that he still only has what he had on the big boat.

 

[ronsimpson]

Questions:
does the "virtual ground" have to be at the ground of the 12V? or can the "virtual ground" be at 1/2 of the 12V?
----edited----
If your 5V pulls 1A from 12V and you are using linear regulators that is 7 watts of heat.

 

[dknguyen]

Mid-bias the 12V supply with a resistisve and capacitive divider? But at 1.1A, probably will not be sufficient.

 

[crutschow]

Is that 1.1A going into the virtual ground?
In other words is the 1.1A load connected to virtual ground?

 

[AnalogKid]

Wally makes a good point. If the peak current out of the +12 V supply is 1.1 A, there are two return paths: through the virtual ground and directly to the power supply return. If most of the load currents are balanced between the rails, then there might be very little virtual ground current. Only you know for sure.

Separate from that, 2 x 5 V does not equal 12 V, so besides the virtual GND you also need a 10 V regulator. With only 2 V of headroom, common parts such as LM317 and 7810 (if there were such a part) will not work. 5 V +/-0.8 V x 2 equals 10 V +/-1.6 V. That is not very tight regulation, so you probably can get away with a simple "active zener" regulator made from very old part numbers. I can't do schematics right now; maybe someone else will hop in.

What is the circuit powering?
Where are you located?
Where is the circuit going?

ak

 

[ronsimpson]

there are two return paths:

I read that the +5V could be 1.1A and or the -5V could be 0 to 1.1A and the +0.8V could be up to 1A and the -0.8V could be up to 1A. (too many unknowns)

If the +5V has about the same load as the -5V then there is a solution.
What if the +5V is pulling 1A and the -5V is pulling near 0A?? Some solutions will not work.

We also need to know how good the regulation is on the 12V? How good the regulation must be on the out puts?

 

[AnalogKid]

and the +0.8V could be up to 1A and the -0.8V could be up to 1A.

Yeah, I wondered if that was unclear typing or a second output requirement. I went with typing; less work for me.

ak

 

[audioguru]

A virtual ground usually feeds a very low current bias reference voltage into the (+) input of an opamp. Then the opamp output can swing up and down.

 

[schmitt trigger]

1.1 amp Peak current TOTAL, divided among each output?
Or each output requires 1.1 amp?

 

[large_ghostman]

Thx guys, just got back i need to read all this carefully. Unfortunately you have mostly confirmed the things i was worried about, the supply can be split to 6V 1.1A is total draw.

Yes alot of heat to deal with, i have a good method of moving the heat, but its still alot of waste. I will be back to answer the questions fully in the morning. Thank you for the input, i had that horrible feeling it was going to be a PITA.

 

[large_ghostman]

Is that 1.1A going into the virtual ground?
In other words is the 1.1A load connected to virtual ground?

I am not sure! I will get back to you as i am totally exhausted, originally the virtual ground was going to be GND, but maybe there is another way?

 

[ronsimpson]

originally the virtual ground was going to be GND,

It would really help if we knew what your circuit is.
If you need -5V below the ground of the +12V then a switching power supply is needed. (need to make a negative supply)
If you can split the 12V into 6 and 6 and have "V gnd' at +6V then things can be made using linear parts.

Next the 0.8V may be a problem. What is on that supply?

 

[large_ghostman]

It would really help if we knew what your circuit is.
If you need -5V below the ground of the +12V then a switching power supply is needed. (need to make a negative supply)
If you can split the 12V into 6 and 6 and have "V gnd' at +6V then things can be made using linear parts.

Next the 0.8V may be a problem. What is on that supply?

Hi Ron, i will give more details tomorrow when i get a chance. The 0.8V is critical, i have little room on it. Its sometimes used after electroporation to condition the cells, the actual range is 0.67V - 0.81V. Anything outside that range just drops cell viability off the chart and becomes pointless to use, I will post info on it later. The rest of it is pretty much driving circuits etc to use this 0.8V, part of the cycle however goes negative -5V as part of the cycle when the cells are not in the buffer solution.

Not 100% sure what exactly this part of the cycle is supposed to do, i am following a paper and it dosnt explain what is supposed to be achieved chemically by doing this, my only guess is its something to do with conditions the electrodes before the cells are added.

I have only found 4 papers (one isnt published yet) that describes the process, so its not a common technique but it would seem if the paper is correct then this technique solves a big problem you can get with electroporation.

 

[ronsimpson]

I think you need a amplifier that outputs +/- 0.8 volts. You do not need a power supply at 0.8V.

 

[large_ghostman]

I think you need a amplifier that outputs +/- 0.8 volts. You do not need a power supply at 0.8V.

You could be right, i spent alot of time today trying to work out how much inductance etc is in each pulse.

Its a bit of a puzzle as nothing sits at one setting, might need to rethink it OR more likely stick with a single set up. Really hard to find the sweet spot frequency wise etc without cell damage, the electrodes have to be pulsed before use or you get a bacterial film on them during use. Is so little scientific info it i am working blind folded at the moment. Still i got tomorrow to look at this properly and report back, I really dont want to drop all that heat, it sounds simple enough in theory but not so easy when you break it down.

Considering its not a new technique, there is almost zero information on how it actually works or what mechanism causes the cell wall to form pores in certain areas. No idea why but i just dont think anyone has thought it important enough to look into. Whats really odd and something i can find no info on, when you first energize the electroporation machine, the cells all line up in a row at one electrode. But alter the initial charge up slightly and the cells turn slightly or flip the opposite way, this dosnt make sense as they dont have poles like a magnet or a different charge each end. But make me wonder if one way works better than the other.

Electroporation is a bit of a black art, it can take alot of attempts to get a cell batch to work. Or it can work first try with same settings, some people swear by heat shock instead and others swear by the electro method, worse still the literature is split pretty much 50/50 but heat shock dosnt work well for me personally.

 

[ronsimpson]

I think you need a amplifier that outputs +/- 0.8 volts. You do not need a power supply at 0.8V.

Now I am down to quoting my self.
It may be that a AC coupled amplifier will work with out the V Ground thing. (like a audio amp driving a speaker (big capacitor on the output))

 

[large_ghostman]

Now I am down to quoting my self.
It may be that a AC coupled amplifier will work with out the V Ground thing. (like a audio amp driving a speaker (big capacitor on the output))

Did i ever mention how much i dislike analog stuff? .

Thx Ron i will look into this, is there anything i can do at the moment to get more information that might help? I am thinking maybe i should set up and standadise the test rig for a start. Might be worth using some cells and seeing what level i can detect a change at. Then slide it back a bit and use that to condition the electrodes.

I will mix some more buffer up and have a play, but i still need a -V supply. I need to save up for a dedicated microscope camera, it would be alot easier to use a proper microscope camera and show you exactly what i am after in the cell membrane. What makes it really hard is working with oblique anaerobes.

You can buy the buffer pretreated and the electrodes, trouble is 12 hours after its made its not effective anymore. OR i wait a couple of years for some clever bstard to work it out for me , but i dont think many people are looking into it.

 

[ronsimpson]

Did i ever mention how much i dislike analog stuff? .

Did i ever mention how much I love analog stuff? .

What frequency?

 

[large_ghostman]

Did i ever mention how much I love analog stuff? .

What frequency?

it was 40k, but that is something i wanted to play with, let me stick the function gen on a tube and see what works best, i can guarantee it will be a range!

 

[AnalogKid]

Did i ever mention how much i dislike analog stuff?

No; please tell us.

ak