Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

+5v and -5v simultaneously from USB

Status
Not open for further replies.

neazoi

Advanced Member level 6
Joined
Jan 5, 2008
Messages
4,119
Helped
13
Reputation
26
Reaction score
15
Trophy points
1,318
Location
Greece
Activity points
36,918
Hi,
I want to power an AM7910 IC from the USB port of the computer. The +5v is straight forward, but the chip requires also -5V 15-25mA max (depended on the version).

So the question is how to derive this amount of current for the negative voltage, out of the USB port?
I like to do it discrete, so please do not suggest me ICs dedicated for the purpose.

Two possible circuits I have found are the next, the first is an oscillator and a negative charge pump and the second uses a multiplier instead. I think a multiplier is better, because it will allow me to regulate the voltage afterwards with a 7905 and obtain a stable voltage for the IC.
But I wonder, how about replacing the 555 with a simple 2 transistors multivibrator?
The point is, how much current could I get out of such a circuit and what are the things that define the current? (collector resistors, frequency of oscillator?)

Thank you
 

Attachments

  • NegativeVoltage.gif
    NegativeVoltage.gif
    4 KB · Views: 95
  • neg1.png
    neg1.png
    27.1 KB · Views: 104
Last edited:

Hi

Use a ICL7660 or similar voltage inverter.

Klaus
 

Just curious but a 555 timer is not a discrete circuit....

Last post needs > 5V for source, USB constrained to +5V, so maybe
thats not a workable solution. Unless USB-C PD.....


Regards, Dana.
 

Just curious but a 555 timer is not a discrete circuit....

Last post needs > 5V for source, USB constrained to +5V, so maybe
thats not a workable solution.


Regards, Dana.
Sure, please read my post, I said to replace this 555 with a discrete multivibrator. Maybe I was not clear enough though.
--- Updated ---

Although interesting, this uses +12v for the input and I have +5V from the USB port. If I had 12v, I could make it easily with a potential divider and a floating ground.

Another idea might be these "joule thieves". Just increase the voltage with the self-excited oscillator, and then rectify it at both positive and negative sides with charge pumps. I just thought that a charge pump would be easier without the need for any inductors.
 
Last edited:

I would suggest a single transistor oscillator using a transformer to produce the feedback and voltage for the -5V rectifier. Some of the negative output can feed back to the transistor base to regulate the output voltage. the transformer can be a small ferrite core.

Brian.
 
I would suggest a single transistor oscillator using a transformer to produce the feedback and voltage for the -5V rectifier. Some of the negative output can feed back to the transistor base to regulate the output voltage. the transformer can be a small ferrite core.

Brian.
Now that you mention that, I remember I have something like this in my site **broken link removed**
It uses a common mode choke though, I would really prefer to do it with a charge pump, maybe in a voltage doubler configuration. I just have no idea about the current these things can supply.
 

Attachments

  • 1.png
    1.png
    5.5 KB · Views: 97
  • 2.jpg
    2.jpg
    106.6 KB · Views: 95

One of the challenges you face is power sequencing and the loads connected
to +5 and -5. So make sure your downstream loads can tolerate the starup of
the multivibrator startup while +5 already presented to loads that need split
supply.


Regards, Dana.
 

One of the challenges you face is power sequencing and the loads connected
to +5 and -5. So make sure your downstream loads can tolerate the starup of
the multivibrator startup while +5 already presented to loads that need split
supply.


Regards, Dana.
Hm... I do not think the am7910 datasheet goes as deep as this, it just assumes the voltages. However the negative rail current is only a fraction of the positive, meaning that not all circuits inside the chip use it. Well I can test this, first feed the positive and then the negative and see how it behaves.

Another thought is to use two USB ports and connect their power pins in series (for a floating ground) although I do not like to mess with it.
 

Hi,

USB voltage may vary 4.5V ... 5.25V if I´m not mistaken.
A discrete voltage inverter will have a lot of additional voltage drop

I know you don´t want to use it. But the ICs usually use active switches and thus reduce voltage drop significantly.
You need to decide what you expect from the voltage source.
* Some circuit need the negative voltage to stay just withing given range.
* Some circuit expect the negative voltage to be symmetric to the positive voltage
* Some circuit are not critical

Klaus
 

Another thought is to use two USB ports and connect their power pins in series (for a floating ground) although I do not like to mess with it.
I think you will find any single device (computer etc.) with two USB ports has a common ground between them.
Personally, I would forget the charge pump idea. Unless you use active switching you should expect at least two diode drops lower output than input and probably even less. You could make a voltage trippler then regulate down but remember that 3 x voltage also means 3 x input current.

Brian.
 

    neazoi

    Points: 2
    Helpful Answer Positive Rating
I think you will find any single device (computer etc.) with two USB ports has a common ground between them.
Personally, I would forget the charge pump idea. Unless you use active switching you should expect at least two diode drops lower output than input and probably even less. You could make a voltage trippler then regulate down but remember that 3 x voltage also means 3 x input current.

Brian.
I see...
Then if it is to go for an inductor, something like this could be a solution? https://www.romanblack.com/smps/conv.htm but with the final diode and the zener (5.1v instead) and capacitor inverted? At least it uses a single cheap molded choke. 12mA is a bit low for the max of 15mA specified for the C version of the Ic. BC337 has 800mA IC, so why does he output such a low current out of this circuit? is it the choke?
 

Attachments

  • conv_cir.gif
    conv_cir.gif
    8.5 KB · Views: 88

The current capability is a factor of several things but I doubt Roman's limit is calculated, more likely it is just the maximum he could draw before the feedback could no longer keep the voltage constant. You might find you can get a little more by letting it oscillate at its natural frequency instead of C1/R4 deciding it.

Incidentally, Q2 not only provides the feedback to make the oscillator run, it also starves the base current in Q1 to provide the regulation. If you take the feedback from a transformer winding instead and generate a negative output voltage, you can use that output voltage directly to pull bias off Q1.

Brian.
 

    neazoi

    Points: 2
    Helpful Answer Positive Rating
The current capability is a factor of several things but I doubt Roman's limit is calculated, more likely it is just the maximum he could draw before the feedback could no longer keep the voltage constant. You might find you can get a little more by letting it oscillate at its natural frequency instead of C1/R4 deciding it.

Incidentally, Q2 not only provides the feedback to make the oscillator run, it also starves the base current in Q1 to provide the regulation. If you take the feedback from a transformer winding instead and generate a negative output voltage, you can use that output voltage directly to pull bias off Q1.

Brian.
A multivibrator can be made with quite low collectors resistor values in the range of 50 ohms without heating a transistor as the 2n2222 and allowing for quite high output current, definitely something more than 15mA. A multivibrator running from 5v with such low resistor values should be able to provide almost a full swing 0-5v. How about a Modified Dickson charge pump after that, (like the one attached) but for negative polarity (just invert the diodes is enough???). All we need is something like 15-20mA and something like -8V so that we can then regulate it to -5v with a 7905
 

Attachments

  • Modified_Dickson_multiplier.svg.png
    Modified_Dickson_multiplier.svg.png
    30.7 KB · Views: 115

Hi,

You must hate me by now, anyway... From experience of actually doing these things on breadboards and simulating assorted types, I feel it's easier getting blood out of a stone than 15 to 20mA from discrete component positive or negative charge pumps, and diode/capacitor combis - forget it. Really, really hard to get anything useful out of them. Switched capacitor ones with MOSFET switches are relatively productive, such as what is the 7660 IC - homemade, not so much. Instead of asking, try breadboarding a few ideas to get the feel of how implementable what you are aiming for is or not. I'd start with the one that looks like it has the lowest output impedance and least voltage drop. I really do wish you luck but personally I think it's a fools errand, like a lot of circuits I do are - maybe interesting to explore but not very practical. AFAIK, negative voltage generators that don't rely on dual supplies and that can output the current level you are aiming for tend to contain a pre-regulator, the MOSFET-based switched capacitor charge pump and a post-regulator...
 

Hi,

The circuit shows just a 2 stage charge pump with useless (only drawbacks) last D-C stage.
The last D-C stage does nothing but reduce the output voltage by wasting energy.

The more stages the less effective, or the more power you need to input for the desired output.
A (only) 3 stage diode chargepump suffers from
* power source impedance
* driver source impedance (oscillator)
* diode voltage drop

Let´s say you have 4.5V USB voltage. An ideal driver, ideal capacitors, Si diodes with a 0.6V voltage drop.
Then each stage generated (4.5V - 2 x 0.6V) = 3.3V.
1 stage: -3.3V,
2 stages: -6.6V,
3 stages: -9.9V.

The circuit causes high peak current, and thus a lot of noise.
You need a good PCB layout. A breadboard is hardly suitable.

Try it.

Klaus
 

Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top