Post by thread:[PIC]:A/D with Vref+

Hi I am trying to do an A/D with the input source to A/D channel 0 of 16F873 and I have configured ADCON1 to have 00000101 with a vref+ of 1V from volt divider and vref- to be Vss. However I get a all '1's output from my output ports even though the input volt is less than 1V. Cant figure out the problem here. __________________________________________________ Do You Yahoo!? Make a great connection at Yahoo! Personals. http://personals.yahoo.com -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics

For one thing, your Vref+ must be at least 2V higher than Vref-. Dale -- Hallo, this is Linus Torvalds and I pronounce Linux as Leennuks. Hallo, this is Bill Gates and I pronounce 'crap' as 'Windows'. On Wed, 10 Oct 2001, Keong wrote: > Hi > > I am trying to do an A/D with the input source to A/D > channel 0 of 16F873 and I have configured ADCON1 to > have 00000101 with a vref+ of 1V from volt divider and > vref- to be Vss. However I get a all '1's output from > my output ports even though the input volt is less > than 1V. Cant figure out the problem here. -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics

But then, is there a method to do an A/D on an input signal of 0V to 1V on the 16F873 using Vrefs? thanx --- Dale Botkin <spam_OUTdaleTakeThisOuTBOTKIN.ORG> wrote: {Quote hidden}> For one thing, your Vref+ must be at least 2V higher > than Vref-. > > Dale > -- > Hallo, this is Linus Torvalds and I pronounce Linux > as Leennuks. > Hallo, this is Bill Gates and I pronounce 'crap' as > 'Windows'. > > On Wed, 10 Oct 2001, Keong wrote: > > > Hi > > > > I am trying to do an A/D with the input source to > A/D > > channel 0 of 16F873 and I have configured ADCON1 > to > > have 00000101 with a vref+ of 1V from volt divider > and > > vref- to be Vss. However I get a all '1's output > from > > my output ports even though the input volt is less > > than 1V. Cant figure out the problem here. > > -- > http://www.piclist.com hint: The list server can > filter out subtopics > (like ads or off topics) for you. See > http://www.piclist.com/#topics > > __________________________________________________ Do You Yahoo!? Make a great connection at Yahoo! Personals. http://personals.yahoo.com -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics

yup, in fact all pins in porta are set to input --- Mike Mansheim <.....Michael_J_MansheimKILLspam@spam@GRACO.COM> wrote: {Quote hidden}> > I am trying to do an A/D with the input source to > A/D > > channel 0 of 16F873 and I have configured ADCON1 > to > > have 00000101 with a vref+ of 1V from volt divider > and > > vref- to be Vss. However I get a all '1's output > from > > my output ports even though the input volt is less > > than 1V. Cant figure out the problem here. > > tris set to input? > > -- > http://www.piclist.com hint: The list server can > filter out subtopics > (like ads or off topics) for you. See > http://www.piclist.com/#topics > > __________________________________________________ Do You Yahoo!? Make a great connection at Yahoo! Personals. http://personals.yahoo.com -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics

On Wed, Oct 10, 2001 at 08:49:12AM -0700, Keong wrote: > But then, is there a method to do an A/D on an input > signal of 0V to 1V on the 16F873 using Vrefs? Sure. Set Vref to exactly 2 volts and lose one bit of accuracy. Instead of 10 bits you only get 9. You can shift the 9th bit into the upper byte of the A/D and round the 10th bit in. It's an 8 bit result but pretty much has 9 bits of accuracy. I did this with fairly good results when reading a LM34 temp sensor. It's output voltage didn't get higher than 1V and I used a Vref of 5.12V. BAJ {Quote hidden}> > thanx > > --- Dale Botkin <daleKILLspamBOTKIN.ORG> wrote: > > For one thing, your Vref+ must be at least 2V higher > > than Vref-. > > > > Dale > > -- > > Hallo, this is Linus Torvalds and I pronounce Linux > > as Leennuks. > > Hallo, this is Bill Gates and I pronounce 'crap' as > > 'Windows'. > > > > On Wed, 10 Oct 2001, Keong wrote: > > > > > Hi > > > > > > I am trying to do an A/D with the input source to > > A/D > > > channel 0 of 16F873 and I have configured ADCON1 > > to > > > have 00000101 with a vref+ of 1V from volt divider > > and > > > vref- to be Vss. However I get a all '1's output > > from > > > my output ports even though the input volt is less > > > than 1V. Cant figure out the problem here. > > > > -- > > http://www.piclist.com hint: The list server can > > filter out subtopics > > (like ads or off topics) for you. See > > http://www.piclist.com/#topics > > > > > > > __________________________________________________ > Do You Yahoo!? > Make a great connection at Yahoo! Personals. > http://personals.yahoo.com > > -- > http://www.piclist.com hint: The list server can filter out subtopics > (like ads or off topics) for you. See http://www.piclist.com/#topics > -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics

On Wed, 10 Oct 2001, Keong wrote: > But then, is there a method to do an A/D on an input > signal of 0V to 1V on the 16F873 using Vrefs? Yes, as noted in the other reply you can use Vref of 2V and you'll have a 0-512 reading, or 9-bit resolution instead of 10-bit. Other than that you can use an op amp to amplify the input signal; of course then you need to divide the ADC reading by the gain of your amp - which again gets you 9-bit resolution! I don't think there's a way around this - in other words to get full 10-bit resolution on a 1V signal. I think your input signal would have to be 2V or better to get you full ADC resolution, since that's the minimum Vref. Dale -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics

> But then, is there a method to do an A/D on an input > signal of 0V to 1V on the 16F873 using Vrefs? Not if you want full range of the A/D without external parts. To get the full A/D range from a 0 to 1V signal requires an external amplifier. ******************************************************************** Olin Lathrop, embedded systems consultant in Littleton Massachusetts (978) 742-9014, .....olinKILLspam.....embedinc.com, http://www.embedinc.com -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics

> Yes, as noted in the other reply you can use Vref of 2V and you'll have a > 0-512 reading, or 9-bit resolution instead of 10-bit. Other than that you > can use an op amp to amplify the input signal; of course then you need to > divide the ADC reading by the gain of your amp - which again gets you > 9-bit resolution! Huh!? You end up with a 0 - 1023 value that maps to the full range of the input voltage, which is 0 to 1V in this case. That's what he wanted in the first case. There's an arbitrary scale factor between the EMF expressed in volts and the A/D reading. A scale factor of 512 doesn't seem more desirable than 1024. If you're doing calibration, you're going to apply an offset and scale anyway. > I don't think there's a way around this - in other > words to get full 10-bit resolution on a 1V signal. I think your input > signal would have to be 2V or better to get you full ADC resolution, since > that's the minimum Vref. ******************************************************************** Olin Lathrop, embedded systems consultant in Littleton Massachusetts (978) 742-9014, EraseMEolinspam_OUTTakeThisOuTembedinc.com, http://www.embedinc.com -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics

On Wed, 10 Oct 2001, Olin Lathrop wrote: > > Yes, as noted in the other reply you can use Vref of 2V and you'll have a > > 0-512 reading, or 9-bit resolution instead of 10-bit. Other than that you > > can use an op amp to amplify the input signal; of course then you need to > > divide the ADC reading by the gain of your amp - which again gets you > > 9-bit resolution! > > Huh!? You end up with a 0 - 1023 value that maps to the full range of the > input voltage, which is 0 to 1V in this case. Nope, if you amplify the input by two you'll get a 0 - 1023 value that maps to TWICE the full range of the (original) input voltage, and represents twice what it would normally. In other words, with a 2V reference and a 1V unamplified input signal, one ADC step equals (roughly) 1.95mV, and you won't see an ADC reading over 512. Amplify the input signal x2 and you will get ADC readings up to 1023, but now each ADC step represents 3.9mV. Same result, no? Just different numbers. That was my point. Admittedly I should have responded when I had more time to clarify, I shouldn't try to multitask when responding to posts. Dale -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics

>On Wed, 10 Oct 2001, Olin Lathrop wrote: > > > > Yes, as noted in the other reply you can use Vref of 2V and you'll have a > > > 0-512 reading, or 9-bit resolution instead of 10-bit. Other than > that you > > > can use an op amp to amplify the input signal; of course then you need to > > > divide the ADC reading by the gain of your amp - which again gets you > > > 9-bit resolution! > > > > Huh!? You end up with a 0 - 1023 value that maps to the full range of the > > input voltage, which is 0 to 1V in this case. Dale replied >Nope, if you amplify the input by two you'll get a 0 - 1023 value that >maps to TWICE the full range of the (original) input voltage, and >represents twice what it would normally. In other words, with a 2V >reference and a 1V unamplified input signal, one ADC step equals (roughly) >1.95mV, and you won't see an ADC reading over 512. Amplify the input >signal x2 and you will get ADC readings up to 1023, but now each ADC step >represents 3.9mV. Same result, no? Just different numbers. That was my >point. Admittedly I should have responded when I had more time to >clarify, I shouldn't try to multitask when responding to posts. How did the ADC step suddenly double? Vref is still 2V. Step is still 1.95mV. Amplifying the input signal up to make is closer to Vref gives better resolution. Surely measuring a 2V (max) signal with a 2V Vref uses all of the available bits (0-1023) where a 1V (max) signal only loses the MSB (0-511). The signal being amplified doesn't make it lose any resolution. The original analog range doesn't come into it at all. If you can amplify it to better match Vref, do it ! Regards... ___________________________________________ David Duffy Audio Visual Devices P/L U8, 9-11 Trade St, Cleveland 4163 Australia Ph: +61 7 38210362 Fax: +61 7 38210281 New Web: http://www.audiovisualdevices.com.au ___________________________________________ -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics

Olin, You have misinterpreted thie comment below: > > Yes, as noted in the other reply you can use Vref of 2V and you'll have a > > 0-512 reading, or 9-bit resolution instead of 10-bit. Other than that you > > can use an op amp to amplify the input signal; of course then you need to > > divide the ADC reading by the gain of your amp - which again gets you > > 9-bit resolution! The user's intent was to use a legal Vref+ value (ie: 2V with Vref- = 0V), which means that the input signal, with varies from 0 to 1V would result in a 9 bit value (0..511 or so). In effect, trading off one bit of input precision for the need to have an amplifier. Bob Ammerman RAm Systems (contract development of high performance, high function, low-level software) Olin said: > Huh!? You end up with a 0 - 1023 value that maps to the full range of the > input voltage, which is 0 to 1V in this case. That's what he wanted in the > first case. There's an arbitrary scale factor between the EMF expressed in > volts and the A/D reading. A scale factor of 512 doesn't seem more > desirable than 1024. If you're doing calibration, you're going to apply an > offset and scale anyway. > > > I don't think there's a way around this - in other > > words to get full 10-bit resolution on a 1V signal. I think your input > > signal would have to be 2V or better to get you full ADC resolution, since {Quote hidden}> > that's the minimum Vref. > > > ******************************************************************** > Olin Lathrop, embedded systems consultant in Littleton Massachusetts > (978) 742-9014, olinspam_OUTembedinc.com, http://www.embedinc.com > > -- > http://www.piclist.com hint: The list server can filter out subtopics > (like ads or off topics) for you. See http://www.piclist.com/#topics > > -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics

{Quote hidden}> > > Huh!? You end up with a 0 - 1023 value that maps to the full range of the > > > input voltage, which is 0 to 1V in this case. > > Dale replied > >Nope, if you amplify the input by two you'll get a 0 - 1023 value that > >maps to TWICE the full range of the (original) input voltage, and > >represents twice what it would normally. In other words, with a 2V > >reference and a 1V unamplified input signal, one ADC step equals (roughly) > >1.95mV, and you won't see an ADC reading over 512. Amplify the input > >signal x2 and you will get ADC readings up to 1023, but now each ADC step > >represents 3.9mV. Same result, no? Just different numbers. That was my > >point. Admittedly I should have responded when I had more time to > >clarify, I shouldn't try to multitask when responding to posts. > > How did the ADC step suddenly double? Vref is still 2V. Step is still 1.95mV. > Amplifying the input signal up to make is closer to Vref gives better > resolution. > Surely measuring a 2V (max) signal with a 2V Vref uses all of the available > bits (0-1023) where a 1V (max) signal only loses the MSB (0-511). The signal > being amplified doesn't make it lose any resolution. The original analog range > doesn't come into it at all. If you can amplify it to better match Vref, do > it ! Sorry, you're right -- the ADC step itself doesn't double, its relationship to the original signal, which is half of Vref, does. Now you're taking 1023 steps to measure twice the original signal, rather than 512 steps to measure the original signal. It's the same thing, isn't it? Dale -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics

{Quote hidden}> -----Original Message----- > From: pic microcontroller discussion list > [@spam@PICLISTKILLspamMITVMA.MIT.EDU]On Behalf Of Dale Botkin > Sent: Wednesday, October 10, 2001 18:15 > To: KILLspamPICLISTKILLspamMITVMA.MIT.EDU > Subject: Re: [PIC]:A/D with Vref+ > > > On Wed, 10 Oct 2001, Olin Lathrop wrote: > > > > Yes, as noted in the other reply you can use Vref of 2V and > you'll have a > > > 0-512 reading, or 9-bit resolution instead of 10-bit. Other > than that you > > > can use an op amp to amplify the input signal; of course then > you need to > > > divide the ADC reading by the gain of your amp - which again gets you > > > 9-bit resolution! > > > > Huh!? You end up with a 0 - 1023 value that maps to the full > range of the > > input voltage, which is 0 to 1V in this case. > > Nope, if you amplify the input by two you'll get a 0 - 1023 value that > maps to TWICE the full range of the (original) input voltage, and > represents twice what it would normally. In other words, with a 2V > reference and a 1V unamplified input signal, one ADC step equals (roughly) > 1.95mV, and you won't see an ADC reading over 512. Amplify the input > signal x2 and you will get ADC readings up to 1023, but now each ADC step > represents 3.9mV. Same result, no? Just different numbers. That was my > point. Admittedly I should have responded when I had more time to > clarify, I shouldn't try to multitask when responding to posts. Huh??? In both cases your "one ADC step" equals 1.95mV, multiplying the input signal by a gain of 2 simply brings the 1V signal to the 2V level, the "step" size doesn't change unless you change the ref. TTYL -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics

At 06:13 PM 10/10/2001 -0500, you wrote: > > > > Huh!? You end up with a 0 - 1023 value that maps to the full range > of the > > > > input voltage, which is 0 to 1V in this case. > > > > Dale replied > > >Nope, if you amplify the input by two you'll get a 0 - 1023 value that > > >maps to TWICE the full range of the (original) input voltage, and > > >represents twice what it would normally. In other words, with a 2V > > >reference and a 1V unamplified input signal, one ADC step equals (roughly) > > >1.95mV, and you won't see an ADC reading over 512. Amplify the input > > >signal x2 and you will get ADC readings up to 1023, but now each ADC step > > >represents 3.9mV. Same result, no? Just different numbers. That was my > > >point. Admittedly I should have responded when I had more time to > > >clarify, I shouldn't try to multitask when responding to posts. David: {Quote hidden}> > How did the ADC step suddenly double? Vref is still 2V. Step is still > 1.95mV. > > Amplifying the input signal up to make is closer to Vref gives better > > resolution. > > Surely measuring a 2V (max) signal with a 2V Vref uses all of the available > > bits (0-1023) where a 1V (max) signal only loses the MSB (0-511). The > signal > > being amplified doesn't make it lose any resolution. The original > analog range > > doesn't come into it at all. If you can amplify it to better match Vref, do > > it ! Dale: >Sorry, you're right -- the ADC step itself doesn't double, its >relationship to the original signal, which is half of Vref, does. Now >you're taking 1023 steps to measure twice the original signal, rather than >512 steps to measure the original signal. It's the same thing, isn't it? The fact that the input signal is doubled makes no difference to the resolution. Ignore the actual input range - it's just being scaled up to get a better result. It might as well be a 100V signal scaled down to 2V. Does this make sense? Regards... -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics

By the way, instead of using a 2 v reference, use a 2.048 v reference and then each "step" will be .001 volt. This simplifies any math that you have to do later. Other useful reference values are: 2.56 v, 4.096 v, and 5.12 v. Certain types of conversion steps can also be simplified by choosing the reference voltage carefully. In effect you are doing part of the overall conversion at the analog stage instead of the digital stage. In general, choose a reference value that makes each "step" value a "nice" number. Fr. Tom McGahee -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.

> Nope, if you amplify the input by two you'll get a 0 - 1023 value that > maps to TWICE the full range of the (original) input voltage, and > represents twice what it would normally. In other words, with a 2V > reference and a 1V unamplified input signal, one ADC step equals (roughly) > 1.95mV, and you won't see an ADC reading over 512. Amplify the input > signal x2 and you will get ADC readings up to 1023, but now each ADC step > represents 3.9mV. No, this is backwards. Each A/D step will still be 2 / 1023 = 1.95mV at the PIC pin because nothing has changed there (still using 2V reference). However, that A/D step size gets *divided* by the opamp gain when projected back to the original signal. Each step now corresponds to a 978uV change in the input signal. Another way to see this is that the 0 to 1V input signal maps to the full range of the A/D, regardless of how we got there. 1V / 1023 = 978uV. Unless I needed very high accuracy, I would multiply the input signal by a little less than 5 and dispense with Vref altogether. In otherwords, use the 5V power regulator as the reference. ******************************************************************** Olin Lathrop, embedded systems consultant in Littleton Massachusetts (978) 742-9014, RemoveMEolinTakeThisOuTembedinc.com, http://www.embedinc.com -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.

> Sorry, you're right -- the ADC step itself doesn't double, its > relationship to the original signal, which is half of Vref, does. Now > you're taking 1023 steps to measure twice the original signal, rather than > 512 steps to measure the original signal. It's the same thing, isn't it? If you multiply the original signal by 2 and use a Vref of 2V, you are taking 1023 steps to measure the whole original signal. ******************************************************************** Olin Lathrop, embedded systems consultant in Littleton Massachusetts (978) 742-9014, spamBeGoneolinspamBeGoneembedinc.com, http://www.embedinc.com -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.

There, you see, finally Olin made sense of my addled attempt at answering the OP's question. Thanks, Olin. Yes, if you amplify the original signal you get *higher* resolution relative to the original signal. In a project just recently finished we had to divide a higher (0-20V) input signal by 4, which gives you *lower* resolution relative to the original signal. OK, did I finally get it right? 8-P Damn, it's a good thing I'm not actually working on a PIC project right now, I could blow the house up. Two weeks off work to move into a new house is counter-productive - you get no "vacation" from work, and there's two weeks' worth of crap waiting when you return to work as well. That and being away from my bench and the projects for even longer and the rust rally sets in, doesn't it? Dale -- Hallo, this is Linus Torvalds and I pronounce Linux as Leennuks. Hallo, this is Bill Gates and I pronounce 'crap' as 'Windows'. On Thu, 11 Oct 2001, Olin Lathrop wrote: > > Sorry, you're right -- the ADC step itself doesn't double, its > > relationship to the original signal, which is half of Vref, does. Now > > you're taking 1023 steps to measure twice the original signal, rather than > > 512 steps to measure the original signal. It's the same thing, isn't it? > > If you multiply the original signal by 2 and use a Vref of 2V, you are > taking 1023 steps to measure the whole original signal. -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.

Dale wrote: >There, you see, finally Olin made sense of my addled attempt at answering >the OP's question. Thanks, Olin. Yes, if you amplify the original signal >you get *higher* resolution relative to the original signal. In a project >just recently finished we had to divide a higher (0-20V) input signal by >4, which gives you *lower* resolution relative to the original signal. >OK, did I finally get it right? 8-P The resolution does not change because of the analog scaling applied. It's determined by how well the max input is matched to the Vref. The smaller the input max is compared with Vref, the worse the resolution. Forget what the original input range was - it's not relevant after scaling. >Damn, it's a good thing I'm not actually working on a PIC project right >now, I could blow the house up. Two weeks off work to move into a new >house is counter-productive - you get no "vacation" from work, and there's >two weeks' worth of crap waiting when you return to work as well. That >and being away from my bench and the projects for even longer and the rust >rally sets in, doesn't it? Dale plugs in his new PIC controlled toaster into new house... WOOF ! SIZZLE ! POP ! :-) Regards... -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.

On Wed, 10 Oct 2001, Olin Lathrop wrote: > > But then, is there a method to do an A/D on an input > > signal of 0V to 1V on the 16F873 using Vrefs? > > Not if you want full range of the A/D without external parts. To get the > full A/D range from a 0 to 1V signal requires an external amplifier. > > There is a method Olin, without any external operational amplifier. You need only to sum the input signal with a 2V dc. This is quite simple if you have an AC input signal. However method depends on the source impedance of the input signal. Another tested method is to generate a zero impedance floated point somewhere at 2V. ( but this one required an OA ) Thus you can measure both positive and negative signals in an unorthodoxe way. Of course diferential amplifiers is doing the same thing. Vasile -- http://www.piclist.com#nomail Going offline? Don't AutoReply us! email TakeThisOuTlistservEraseMEspam_OUTmitvma.mit.edu with SET PICList DIGEST in the body

>Olin Lathrop wrote: > > > > But then, is there a method to do an A/D on an input > > > signal of 0V to 1V on the 16F873 using Vrefs? > > > > Not if you want full range of the A/D without external parts. To get the > > full A/D range from a 0 to 1V signal requires an external amplifier. Vasile wrote: > There is a method Olin, without any external operational amplifier. >You need only to sum the input signal with a 2V dc. This is quite simple >if you have an AC input signal. However method depends on the source >impedance of the input signal. By this do you mean shift the original up closer to Vref? How does this increase resolution? The shifted signal is still 1V in amplitude. Maybe I've misunderstood what you mean. I can't see how shifting it makes a difference. Only a signal with same range as Vref gives full resolution. The 1V signal is half of the min Vref, so there's no other way to get the full A/D resolution without amplifying it. Am I missing something here? Regards... ___________________________________________ David Duffy Audio Visual Devices P/L U8, 9-11 Trade St, Cleveland 4163 Australia Ph: +61 7 38210362 Fax: +61 7 38210281 New Web: http://www.audiovisualdevices.com.au ___________________________________________ -- http://www.piclist.com#nomail Going offline? Don't AutoReply us! email RemoveMElistservTakeThisOuTmitvma.mit.edu with SET PICList DIGEST in the body

Of course you have right David, I was thinking just to the first question, resolution can't be improved only if the signal have the same amplitude like reference. For 87x, unfortunately the +vref can't be +1V. Vasile On Fri, 12 Oct 2001, David Duffy wrote: {Quote hidden}> >Olin Lathrop wrote: > > > > > > But then, is there a method to do an A/D on an input > > > > signal of 0V to 1V on the 16F873 using Vrefs? > > > > > > Not if you want full range of the A/D without external parts. To get the > > > full A/D range from a 0 to 1V signal requires an external amplifier. > > Vasile wrote: > > There is a method Olin, without any external operational amplifier. > >You need only to sum the input signal with a 2V dc. This is quite simple > >if you have an AC input signal. However method depends on the source > >impedance of the input signal. > > By this do you mean shift the original up closer to Vref? How does this > increase resolution? The shifted signal is still 1V in amplitude. Maybe > I've misunderstood what you mean. I can't see how shifting it makes a > difference. Only a signal with same range as Vref gives full resolution. > The 1V signal is half of the min Vref, so there's no other way to get the > full A/D resolution without amplifying it. Am I missing something here? > Regards... > > ___________________________________________ > David Duffy Audio Visual Devices P/L > U8, 9-11 Trade St, Cleveland 4163 Australia > Ph: +61 7 38210362 Fax: +61 7 38210281 > New Web: http://www.audiovisualdevices.com.au > ___________________________________________ > > -- > http://www.piclist.com#nomail Going offline? Don't AutoReply us! > email listservEraseME.....mitvma.mit.edu with SET PICList DIGEST in the body > > -- http://www.piclist.com#nomail Going offline? Don't AutoReply us! email EraseMElistservmitvma.mit.edu with SET PICList DIGEST in the body

On Fri, 12 Oct 2001, Vasile Surducan wrote: > Of course you have right David, I was thinking just to the first question, > resolution can't be improved only if the signal have the same amplitude > like reference. For 87x, unfortunately the +vref can't be +1V. But no one stop you to set -vref = +2.0 V and +vref = +4V. Summing to variable input ( which is 0 to 1V ) a constant +2V will alow to improve the resolution better than measuring 0...1V with -vref = gnd and +vref = vcc I forgot to tell you that in my last mail. ( also small negative signals can be measured see: http://www.geocities.com/vsurducan/electro/PIC/f877.html ) Vasile {Quote hidden}> > On Fri, 12 Oct 2001, David Duffy wrote: > > > >Olin Lathrop wrote: > > > > > > > > But then, is there a method to do an A/D on an input > > > > > signal of 0V to 1V on the 16F873 using Vrefs? > > > > > > > > Not if you want full range of the A/D without external parts. To get the > > > > full A/D range from a 0 to 1V signal requires an external amplifier. > > > > Vasile wrote: > > > There is a method Olin, without any external operational amplifier. > > >You need only to sum the input signal with a 2V dc. This is quite simple > > >if you have an AC input signal. However method depends on the source > > >impedance of the input signal. > > > > By this do you mean shift the original up closer to Vref? How does this > > increase resolution? The shifted signal is still 1V in amplitude. Maybe > > I've misunderstood what you mean. I can't see how shifting it makes a > > difference. Only a signal with same range as Vref gives full resolution. > > The 1V signal is half of the min Vref, so there's no other way to get the > > full A/D resolution without amplifying it. Am I missing something here? > > Regards... > > > > ___________________________________________ > > David Duffy Audio Visual Devices P/L > > U8, 9-11 Trade St, Cleveland 4163 Australia > > Ph: +61 7 38210362 Fax: +61 7 38210281 > > New Web: http://www.audiovisualdevices.com.au > > ___________________________________________ > > > > -- > > http://www.piclist.com#nomail Going offline? Don't AutoReply us! > > email RemoveMElistservEraseMEEraseMEmitvma.mit.edu with SET PICList DIGEST in the body > > > > > > -- > http://www.piclist.com#nomail Going offline? Don't AutoReply us! > email RemoveMElistservspam_OUTKILLspammitvma.mit.edu with SET PICList DIGEST in the body > > -- http://www.piclist.com#nomail Going offline? Don't AutoReply us! email RemoveMElistservTakeThisOuTspammitvma.mit.edu with SET PICList DIGEST in the body

> There is a method Olin, without any external operational amplifier. > You need only to sum the input signal with a 2V dc. This is quite simple > if you have an AC input signal. However method depends on the source > impedance of the input signal. > Another tested method is to generate a zero impedance floated point > somewhere at 2V. ( but this one required an OA ) Thus you can measure both > positive and negative signals in an unorthodoxe way. Of course diferential > amplifiers is doing the same thing. You can't sum voltages with passive components, only average (or weighted average) them. This will always attenuate the input signals at least a little bit. With a resistor or two you could level shift the 0-1V signal to a different range (and loose some range in the process), but that doesn't solve the problem. A 0-1V range can't be converted to the full A/D range by a PIC, no matter where in the 0-5V range this input voltage is presented. The restriction comes from the spec that Vref+ - Vref- must be at least 2V to ensure the full accuracy. ******************************************************************** Olin Lathrop, embedded systems consultant in Littleton Massachusetts (978) 742-9014, EraseMEolinspamspamBeGoneembedinc.com, http://www.embedinc.com -- http://www.piclist.com#nomail Going offline? Don't AutoReply us! email RemoveMElistservKILLspammitvma.mit.edu with SET PICList DIGEST in the body

> But no one stop you to set -vref = +2.0 V and +vref = +4V. > Summing to variable input ( which is 0 to 1V ) a constant +2V will alow > to improve the resolution better than measuring 0...1V with -vref = gnd > and +vref = vcc True, but why not just Vref+ = 2V, Vref- = GND, and use the orginal 0-1V signal unmodified. That will give you 9 bits of accuracy on a 16F87x, same as your method above. ******************************************************************** Olin Lathrop, embedded systems consultant in Littleton Massachusetts (978) 742-9014, olinSTOPspamspam_OUTembedinc.com, http://www.embedinc.com -- http://www.piclist.com#nomail Going offline? Don't AutoReply us! email spamBeGonelistservSTOPspamEraseMEmitvma.mit.edu with SET PICList DIGEST in the body

On Fri, 12 Oct 2001, Olin Lathrop wrote: > > But no one stop you to set -vref = +2.0 V and +vref = +4V. > > Summing to variable input ( which is 0 to 1V ) a constant +2V will alow > > to improve the resolution better than measuring 0...1V with -vref = gnd > > and +vref = vcc > > True, but why not just Vref+ = 2V, Vref- = GND, and use the orginal 0-1V > signal unmodified. That will give you 9 bits of accuracy on a 16F87x, same > as your method above. > Because you can't set Vref+ below +2.5 V. I haven't read the last data sheet for 877 because I can't see the Microchip site but my tests were done on two Microchip engineering sample and one purchased chip ( when there was no info in datasheet about the range for Vref+ and Vref- ) I've quess nothing was change from then. Also I found that SMP bit from SSPSTAT register is only READABLE ( tested with simulator, and debugged with LED ) even datasheet say different. Best regards, Vasile -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics