Post by thread:[EE] Charge pump as analog negative supply

Hello all, I have a project which will have both a PIC and some op-amps to amplify/filter the signal from a 40kHz ultrasonic transducer. It would be convenient to have a negative supply for the analog side, and since I've got a PIC handy I wondered about generating the negative supply with a simple capacitor/diode charge pump. That would give me roughly -3.6 (or -4.3 with Schottky diodes) to +5V which would be just fine. I would imagine I'd want to avoid running the charge pump at, say, 40kHz. The question is how much of a problem would the relatively poor regulation of the negative supply likely to cause me? The op-amp circuits will all have feedback and will be referenced to 0V. Best regards, Chris

Depending on signal characteristics and the power supply rejection ratio of the Op Amps it may give no problems. I have done this before without problem. If you need to you could use a zener to regulate the negative supply. Larry {Quote hidden}> > From: Chris Emerson <spam_OUTpicTakeThisOuTnosreme.org> > Date: 2005/11/08 Tue AM 10:05:04 EST > To: .....piclistKILLspam@spam@mit.edu > Subject: [EE] Charge pump as analog negative supply > > Hello all, > > I have a project which will have both a PIC and some op-amps to > amplify/filter the signal from a 40kHz ultrasonic transducer. > > It would be convenient to have a negative supply for the analog side, > and since I've got a PIC handy I wondered about generating the negative > supply with a simple capacitor/diode charge pump. That would give me > roughly -3.6 (or -4.3 with Schottky diodes) to +5V which would be just > fine. I would imagine I'd want to avoid running the charge pump at, > say, 40kHz. > > The question is how much of a problem would the relatively poor > regulation of the negative supply likely to cause me? The op-amp > circuits will all have feedback and will be referenced to 0V. > > Best regards, > > Chris > --

It's pretty common to do this using a max-232 or similar device to generate +/- 10 Volts. But in addition, there are plenty of good single supply op-amps available.. Neil On 11/8/05, Chris Emerson <picKILLspamnosreme.org> wrote: {Quote hidden}> Hello all, > > I have a project which will have both a PIC and some op-amps to > amplify/filter the signal from a 40kHz ultrasonic transducer. > > It would be convenient to have a negative supply for the analog side, > and since I've got a PIC handy I wondered about generating the negative > supply with a simple capacitor/diode charge pump. That would give me > roughly -3.6 (or -4.3 with Schottky diodes) to +5V which would be just > fine. I would imagine I'd want to avoid running the charge pump at, > say, 40kHz. > > The question is how much of a problem would the relatively poor > regulation of the negative supply likely to cause me? The op-amp > circuits will all have feedback and will be referenced to 0V. > > Best regards, > > Chris > -

> I have a project which will have both a PIC and some op-amps to > amplify/filter the signal from a 40kHz ultrasonic transducer. > > It would be convenient to have a negative supply for the analog side, > and since I've got a PIC handy I wondered about generating the negative > supply with a simple capacitor/diode charge pump. That would give me > roughly -3.6 (or -4.3 with Schottky diodes) to +5V which would be just > fine. I would imagine I'd want to avoid running the charge pump at, > say, 40kHz. Actually, running the charge pump at 40 kHz may not be a bad idea. If the ultrasonic and the charge pump are in phase, it'll reduce the noise during the time you're interested in- the few microseconds after the audio pulse when you're looking for the reflected signal. > The question is how much of a problem would the relatively poor > regulation of the negative supply likely to cause me? The op-amp > circuits will all have feedback and will be referenced to 0V. I've made some fairly good circuits using charge pumps to create a negative bias, although I typically use something like the LMC7660, which is an 8-pin part using two external 10 uF caps to provide a pretty good -V from a +V input. It also pays to watch the PSRR of your op-amps and be smart about bypassing. Mike H.

Chris Emerson <.....picKILLspam.....nosreme.org> wrote: > The question is how much of a problem would the relatively poor > regulation of the negative supply likely to cause me? The "regulation" will simply be the regulation of the positive supply as seen through the impedance of the charge pump. You can control the latter through the size of the pump capacitor and the switching frequency. For example, a frequency of 100 kHz and a capacitor of 1 uF will yield an effective impedance of 1/f*C = 10 ohms. Whether the variations in the negative supply are a problem for the opamps is determined by their power supply rejection ratio (PSRR), which is generally pretty good for most modern chips. -- Dave Tweed

At 10:53 AM 11/8/2005 -0500, you wrote: >Chris Emerson <EraseMEpicspam_OUTTakeThisOuTnosreme.org> wrote: > > The question is how much of a problem would the relatively poor > > regulation of the negative supply likely to cause me? > >The "regulation" will simply be the regulation of the positive supply >as seen through the impedance of the charge pump. You can control the >latter through the size of the pump capacitor and the switching >frequency. For example, a frequency of 100 kHz and a capacitor of >1 uF will yield an effective impedance of 1/f*C = 10 ohms. Which would be thoroughly swamped by the resistance of the PIC port drivers, of course. This sort of thing is most useful when the amount of current to be drawn is very small, say in the scores of uA or so, and it doesn't matter much if it is +/- 0.5V or whatever. It can come in handy where you can't control the ground reference and need to accommodate negative voltages-- otherwise it's probably best to design your circuit to work with a single solid supply. >Whether the variations in the negative supply are a problem for the >opamps is determined by their power supply rejection ratio (PSRR), >which is generally pretty good for most modern chips. Provided you don't run out of "headroom" (or is it footroom, when it's negative? ;-) ). That includes common mode range and output swing. The PSRR is also much better at DC than at high frequencies, so best to filter the supplies well (unless you're using them as outputs or something really clever like that). >Best regards, Spehro Pefhany --"it's the network..." "The Journey is the reward" speffspam_OUTinterlog.com Info for manufacturers: http://www.trexon.com Embedded software/hardware/analog Info for designers: http://www.speff.com ->> Inexpensive test equipment & parts http://search.ebay.com/_W0QQsassZspeff

Or if you're already using a MAX232 etc. you may be able to get the negative supply from that. RP On 09/11/05, Spehro Pefhany <@spam@speffKILLspaminterlog.com> wrote: {Quote hidden}> At 10:53 AM 11/8/2005 -0500, you wrote: > >Chris Emerson <KILLspampicKILLspamnosreme.org> wrote: > > > The question is how much of a problem would the relatively poor > > > regulation of the negative supply likely to cause me? > > > >The "regulation" will simply be the regulation of the positive supply > >as seen through the impedance of the charge pump. You can control the > >latter through the size of the pump capacitor and the switching > >frequency. For example, a frequency of 100 kHz and a capacitor of > >1 uF will yield an effective impedance of 1/f*C = 10 ohms. > > Which would be thoroughly swamped by the resistance of the PIC port drivers, > of course. This sort of thing is most useful when the amount of current > to be drawn is very small, say in the scores of uA or so, and it doesn't > matter much if it is +/- 0.5V or whatever. It can come in handy where you > can't control the ground reference and need to accommodate negative > voltages-- otherwise it's probably best to design your circuit to work > with a single solid supply. > > >Whether the variations in the negative supply are a problem for the > >opamps is determined by their power supply rejection ratio (PSRR), > >which is generally pretty good for most modern chips. > > Provided you don't run out of "headroom" (or is it footroom, when it's > negative? ;-) ). That includes common mode range and output swing. The > PSRR is also much better at DC than at high frequencies, so best to filter > the supplies well (unless you're using them as outputs or something really > clever like that). > > >Best regards, > > Spehro Pefhany --"it's the network..." "The Journey is the reward" > RemoveMEspeffTakeThisOuTinterlog.com Info for manufacturers: http://www.trexon.com > Embedded software/hardware/analog Info for designers: http://www.speff.com > ->> Inexpensive test equipment & parts http://search.ebay.com/_W0QQsassZspeff > > > -

PSRR at DC is usually pretty good but it falls off with frequency (check the data sheets on your favorite op amp). The real problem here is not likely to be regulation but rather ripple at the charge pump's switching frequency. By the time you get to 40 - 100kHz, PSRR on even fairly good op amps is down to 40dB or so. If you have significant ripple at your switching frequency, it'll bleed right into the signal path. You can reduce ripple by having two 180 degrees out of phase pumps charging a single supply capacitor. No theoretical reason this can't be extended to say, 4 pumps 90 degrees out of phase. It'll start to look silly though and you can do a fair bit of good with a simple filter on the output of the supply. Dave Dave Tweed wrote: {Quote hidden}>Chris Emerson <spamBeGonepicspamBeGonenosreme.org> wrote: > > >>The question is how much of a problem would the relatively poor >>regulation of the negative supply likely to cause me? >> >> > >The "regulation" will simply be the regulation of the positive supply >as seen through the impedance of the charge pump. You can control the >latter through the size of the pump capacitor and the switching >frequency. For example, a frequency of 100 kHz and a capacitor of >1 uF will yield an effective impedance of 1/f*C = 10 ohms. > >Whether the variations in the negative supply are a problem for the >opamps is determined by their power supply rejection ratio (PSRR), >which is generally pretty good for most modern chips. > >-- Dave Tweed > >

Chris Emerson wrote: > I have a project which will have both a PIC and some op-amps to > amplify/filter the signal from a 40kHz ultrasonic transducer. > > It would be convenient to have a negative supply for the analog side, > and since I've got a PIC handy I wondered about generating the > negative supply with a simple capacitor/diode charge pump. That > would give me roughly -3.6 (or -4.3 with Schottky diodes) to +5V > which would be just fine. I would imagine I'd want to avoid running > the charge pump at, say, 40kHz. > > The question is how much of a problem would the relatively poor > regulation of the negative supply likely to cause me? The op-amp > circuits will all have feedback and will be referenced to 0V. You're probably fine. I'm thinking of doing something similar in a current project. The opamp shouldn't draw that much current, so you should be able to put a resistor in series with the opamp supply and a capacitor to ground if needed. Look up the power supply rejection ratio of the opamp and figure out how much ripply you can tolerate on the supply. ****************************************************************** Embed Inc, Littleton Massachusetts, (978) 742-9014. #1 PIC consultant in 2004 program year. http://www.embedinc.com/products