Positive and negative supply

I need +5V, GND and -5V. Can I just use 12V DC and put 6V as a ground and then stabilize the voltage with 7805 and 7905 stabilizers? Like this:

+12V........+6V
-.............. -
-.............. -
-...............-
-...............-
-.............. - 0V
-...............-
-...............-
-.............. -
-.............. -
0V............ -6V

Would it be better to use transformer with two secondary coils? But then I would need to rectify it first.

You need to create a ground in between the 12V supply leads. The simplest method is a 2-resistor divider, or a resistor and zener diode.

This is wasteful of power, however. If power needs are low then you may not mind.

Also take a look at the list of related threads at the bottom of this column.

What circuit do you have than needs a dual polarity supply?
Why do the voltages need to be regulated?

I need to power a op amp. But the rest of the circuit, which is Atmega 1284 and some I2C components also need to work on the same supply. Because of that voltages need regulation.

If you have a 12V DC supply, you can use a 7805 for the +5VDC, and something like a charge pump inverter (the 7660 is one I know, but it is dependent on input voltage for its output voltage) to get the -5VDC.

If you plan on using the mains AC and transforming it to a low DC voltage, one way is to use a centre-tapped transformer to get 0V, +5V and -5V, it would be a lot more components to put it together, unless that doesn't matter because you want to make it all.

Maybe easiest with ICs that I know is to look for a charge pump inverter or similar device that converts a positive voltage to a negative voltage - it saves hassle and should need few components.

Sorry, not sure if I didn't understand you, but 12VDC should be "stabilised" to begin with, otherwise I think it would be AC, unless you meant rough DC after a rectifier and filter capacitors before adding a voltage regulator.

Hope you've found a suitable method.

There could be a problem that 7660 can deliver only 400uA. For now I'm considering using center tapped transformer.

I14R10 said:

I need to power a op amp.

Click to expand...

An opamp does not need a dual polarity supply if it only has an AC (audio?) output signal. You simply bias its very low input current non-inverting input at half the supply voltage with two fairly high value (100k) resistors and a filter capacitor. Then its input, output and feedback ground resistor need coupling capacitors.
If you post your schematic then I will show you how to do it.

Hi,

As Audioguru said. . An opamp usually doesn't need dual supply. Doesn't care if 0V / 12V or -6V / 6V.
And usually it has no GND connection either.

Maybe one of the signal inputs is connected with GND ( with resistor). This needs about no current.

Klaus

This link is to a recent thread which discusses a similar project (how to create a virtual ground). Single supply 24V, wants bipolar +12 and -12V.

https://www.edaboard.com/threads/338441/

I know how to make virtual ground, but I have a WS-2300 weather station and I need TL072 as a comparator. The weather station works on it's own supply so I don't know how it would recognize the signals. Because my ground is at 6V and weather station's ground is negative pole from 2x1.5V battery. But I will try and let you know how it works.

Hi,

an OPAMP as comparator .... i don´t like this. Why not use a comparator? It´s made for it.

Are you sure the TL072 prevents phase reversal? Any comparator does...

Klaus

Add: Show us your schematic and we can better assist you.

Oh yes, I forgot to attach a schematics



Excuse me for drawing this in paint. So what am I doing here? The signals from WS-2300 are around 200mV P-P. I'm using two 2N2222 to amplify them and one 2N2222 as buffer amplifier. Without it the signal would distort when I connect it to Atmega. But the problem is that I still am not receiving correct signal into Atmega. Before trying to decode live signals from WS-2300, I first recorded them and used sound card of my computer to play them for Atmega. I amplified it to the max that sound card can amplify and it worked. Atmega would decode the signal correctly.

Then when I connected the WS.2300 trouble began. I first tried connecting it like in this schematics but that also didn't work. It seems that Atmega is not receiving all the bits in the signal (there are 52, or 48. Sometimes the station sends introduction bits as 00001001 and sometimes just as 1001).

So in the end I decided that it would be best to use comparator to compare the signal to, let's say 1V. So that brings me here trying to figure out how to connect op-amp to single positive supply.

What comparator do you suggest ? Would LM111 be good **broken link removed** ? Or LM311 https://www.fairchildsemi.com/datasheets/LM/LM311.pdf

I don't have expensive equipment to check signal, but I do have an $30 oscilloscope. I can confirm that signal is amplified correctly. But I don't know signal voltage. My oscilloscope says 15V, but that is of course nonsense, because I have 12V supply. My other instrument, regular multimeter, switched to 6V AC position shows voltage at around 2.5-3V, I don't know for sure. But I believe that they are something like that.

EDIT: What do you think about using opto-isolators (opto couplers) instead of amplifying signal like in schematics? I have some opto couplers left from other projects, but I don't know how fast they are switching.

What is the signal from a weather station, AC or DC? It must have two wires for its signal, a live and a common or a differential output.
A virtual ground is used when a circuit needs a virtual ground with a lot of current, but the input of an opamp input uses a tiny current and since the TL072 has Fet inputs then they use NO current. So two high value resistors can make the input biased correctly when there is only a positive supply. The TL072 has the problem called, "Phase Reversal" when an input voltage gets within 3V to 4V from its negative supply voltage.
Why don't you post the schematic??

I did post schematics, don't you see it?
https://obrazki.elektroda.pl/1161386600_1437129969.png

Hi!

Probably nothing to do with it, but maybe the oscilloscope read 15V as it was summing ground references/voltages - I know from mistakes I make connecting devices together that it happens - yesterday testing 2 x power supply voltmeters one read 18V instead of 12V as I had connected its Vin- input to the ground of the 6V supply (I'm good at mistakes like this)... And ground loops created between both circuits may be involved in your case. I only started making a power supply as I wanted a common ground reference between some devices to minimise inaccuracy due to noise in signals... (so I make my own rubbish, noisy references what a waste of time besides the learning...)
If you can tie the grounds of the WS-2300 and of your input device circuit together, might (or might not) help.

I know little about OpAmps, but maybe the 2N2222As (I like them, believe me) are pretty noisy devices, so some sort of precision signal amplifying OpAmp would possibly be a good interface instead of the two stages of transistors, it's like photocopying: the more stages, the more the original gets distorted.
And, I recently learnt that my current shunt monitor (not an OpAmp, as some-one on eda kindly pointed out to me) has a minimum unavoidable noise present of 20 to 70uV - so the ammeter always reads 3 - 5mA with no input partly due to that.

Opto-coupler I know, 4N35 says: "High speed switching. Rise and fall time: 14uS"; good on one hand, on other I found it too sensitive for what I needed as even breathing too close seems to turn them on (just my opinion, could have been my circuit design), so maybe it would need a buffer/hysteresis before the input, unless you need a highly sensitive response to the WS-2300.

If the WS-2300 has an output range of approx. 200mV, is it pointless to feed it into a 200mV calibrated ADC, even a simple one, using Vin+ for the signal and Vin- for the ground reference, then direct from the ADC to the ATMega? Possibly costs the same as a good OpAmp.

It's completely unclear which signal of your "WS-2300" you want to process with the strange amplifier circuit. The specification I have seen is talking about a PC COM port interface, which would hardly have only 200 mV level.

The 2nd and 3rd amplifier stage are missing any bias, but we should discuss it after clarifying the circuit purpose.

Hi,

do you know the ATMEL has a comparator inside? So why use an external one?

The signal from the WS2300 should be specified. With dataformat, but also with voltage levels. Maybe it is open collector and thus needs external pullup... --> Look at it´s datasheet.

Sometimes the station sends introduction bits as 00001001 and sometimes just as 1001).

Click to expand...

your circuit is AC coupled, so it will be hard to detect a DC level high to low transition (first = start signal, for your timing reference)

Optocpupler: I don´t thik it is necessary. As long as GND of ATMEL and the GND of WS2300 are tied together....


Klaus

**broken link removed** this is how the signal looks.

Of course the 2nd and 3rd transistors are missing bias since first transistor amplifies the signal enough (2V P-P) so I don't need any bias.

I know Atmega has comparator but if I use it's internal comparator then some other thing stops working. I have LCD, ethernet, EEPROM and barometer all programmed on that same atmega. It's easier to use external comparator than trying to figure why internal comparator won't work.

As I said, voltage level coming from WS-2300 is around 200mV P-P.

So, how should I couple the signal from WS-2300 to comparator, or amplifier, whatever comes first?

You need to understand there is not a datasheet. This is a commercial weather station, not intended for this use. I have to experiment with everything and find out everything by myself.

I do not know if the weather station sends sinewaves or digital pulses.

Your transistors are biased completely wrong:
1) The first transistor does not have enough base current so it is cutoff most of the time.
2) The second transistor has no base bias current whatsoever so it is almost always cutoff.
3) The third transistor is probably not needed.
4) The input impedance of the second transistor severely loads down the output of the first transistor since its base-emitter is simply a diode.

Here are simulations showing a sinewave input:

FvM said:

It's completely unclear which signal of your "WS-2300" you want to process with the strange amplifier circuit. The specification I have seen is talking about a PC COM port interface, which would hardly have only 200 mV level.

Click to expand...

No it's just regular phone jack, I don't know how it's called and I don't know how to draw it.

- - - Updated - - -

Audioguru said:

I do not know if the weather station sends sinewaves or digital pulses.

Your transistors are biased completely wrong:
1) The first transistor does not have enough base current so it is cutoff most of the time.
2) The second transistor has no base bias current whatsoever so it is almost always cutoff.
3) The third transistor is probably not needed.
4) The input impedance of the second transistor severely loads down the output of the first transistor since its base-emitter is simply a diode.

Here are simulations showing a sinewave input:

Click to expand...

It sends out digital pulses. They are pulse width modulated. So, "1" is around 0.5ms wide and "0" is 1.3ms wide. It's not a problem if the transistor clips the signal. But I agree that it's not designed properly. I will later try and design it better.