Post by thread:[PIC]: Precision 0.5V reference for AD

On Tue, 2 Jul 2002 19:11:52 +0200 Alvaro Deibe Diaz <spam_OUTadeibeTakeThisOuTCDF.UDC.ES> writes: > Hello. > > I am planning to read some 0.5V to 4.5V sensor > voltages, and need good 0.5V and 4.5V precission > references. I plan to use Analog Devices REF194 > to get the 4.5V, but don't know how to get the > 0.5V without using resistors, op-amps, etc. > > Any ideas? > How about just subtracting it out in software? If you have enough A/D resolution to give up the bottom end, this approaches the ideal zero parts solution. Harold FCC Rules Online at http://hallikainen.com/FccRules Lighting control for theatre and television at http://www.dovesystems.com Reach broadcasters, engineers, manufacturers, compliance labs, and attorneys. Advertise at http://www.hallikainen.com/FccRules/ . ________________________________________________________________ GET INTERNET ACCESS FROM JUNO! Juno offers FREE or PREMIUM Internet access for less! Join Juno today! For your FREE software, visit: dl.http://www.juno.com/get/web/. -- 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

> I am planning to read some 0.5V to 4.5V sensor > voltages, and need good 0.5V and 4.5V precission > references. I plan to use Analog Devices REF194 > to get the 4.5V, but don't know how to get the > 0.5V without using resistors, op-amps, etc. Use ground as the lower reference. Do you really need the extra precision from a 4V range instead of a 4.5V range? ***************************************************************** Embed Inc, embedded system specialists in Littleton Massachusetts (978) 742-9014, 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

How about a heavily on diode through a resistor divider? +5V | | > < > < > < | +-----+ 0.65V | | | > | < --- > \ / < <----- Vref --- > | | < --- small C to reduce dynamic | | --- source resistance. | | | +-----+-----+ | ----- / / / Shown here Potentiometer, instead of two resistors because I got lazy. It would'nt be all that accurate, due to temperature drift, but it might be good enough for what you want... And besides, the on voltage of a signal diode doesn't drift all *that* much... well, best of luck. --Brendan {Original Message removed}

>I am planning to read some 0.5V to 4.5V sensor >voltages, and need good 0.5V and 4.5V precission >references. I plan to use Analog Devices REF194 >to get the 4.5V, but don't know how to get the >0.5V without using resistors, op-amps, etc. I've used pressure sensors with this same output range. I used a 5 V reference (if you use a 4.5 V, that's even better) and just subtracted off the zero offset (0.5 V nominal) to get an A/D range of 0-920 decimal (nominal). If you need more steps than this, then maybe you can't use this method. But I suspect it should be just fine for your application. --Andrew _________________________________________________________________ Send and receive Hotmail on your mobile device: http://mobile.msn.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

First of all, thanks to all of you. I need those 0.5V extra precission. Using 4.5V instead of 4V range means losing about 12.5% resolution, and this is too much for my application (we intend to validate a new vehicle dynamics mathematic model with the data we collect). I was looking for a good and stable 0.5V referece, because any error in the references will get into the data. The 4.5V reference I plan to use has under 10ppmV drift with temperature and load. But there is no 0.5V references this good. An obvious solution would be using a precission op amp (rail to rail input and output, single supply) and a pot to divide the good 4.5V and get the 0.5V. But I don't like the pots very much (the final pcb will go into a 1:5 scale car, and there are too much vibrations, noise and temperature variations). Other possible solution is to use a good adjustable reference but the adjust is, again, with a pot... Perhaps any of you have a solution for this problem? Thanks again, Alvaro Deibe. {Original Message removed}

>I was looking for a good and stable 0.5V referece, >because any error in the references will get into >the data. The 4.5V reference I plan to use has under >10ppmV drift with temperature and load. But there >is no 0.5V references this good. Check out the Analog Devices AD584 Precision Voltage Reference. I believe you should be able to get what you want from this. It starts off as a 5ppm reference, and you will need to divide from this to get your 0.5V, but your reference output can be 2.5V, so you need only some close tolerance resistors. By using 5 identical resistors in series, and tapping at one resistor from ground you should eliminate any temp co-eff problems, even if the resistors do not have a tight temp co-eff themselves, and with 5 resistors from the same production batch you should get a close tolerance divider. -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.

>The only question is must I "isolate" the reference >point from the (tree) Vref- of the (tree) PICs that >do the AD whith a good op amp? Do I understand you correctly? When you say "tree" do you mean "three" (3) ? I think you are asking can you use the same reference for 3 PIC chips simultaneously, is this correct? If so you may need to have buffer amplifiers between the resistors and the reference inputs, but I have no experience with this. However I would have thought any modern FET input op-amp used in voltage follower mode would be satisfactory. -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.

Sorry for my english: I have three (3) PICs, not "tree"... :-) ----- Original Message ----- From: "Alvaro Deibe Diaz" <.....adeibeKILLspam@spam@cdf.udc.es> To: "pic microcontroller discussion list" <PICLISTKILLspamMITVMA.MIT.EDU> Cc: "Jesús Cardenal Carro" <.....jcardeKILLspam.....cdf.udc.es> Sent: Wednesday, July 03, 2002 12:40 PM Subject: Re: Re: [PIC]: Precision 0.5V reference for AD > Yes, this is a good solution. Simple and easy. No calibration problems (once {Quote hidden}> selected the appropriate resistors), no temperature problems, vibration, > etc. > > The only question is must I "isolate" the reference point from the (tree) > Vref- of the (tree) PICs that do the AD whith a good op amp? > > Thanks again, > > Alvaro Deibe. > > ----- Original Message ----- > From: "Alan B. Pearce" <EraseMEA.B.Pearcespam_OUTTakeThisOuTRL.AC.UK> > To: <PICLISTspam_OUTMITVMA.MIT.EDU> > Sent: Wednesday, July 03, 2002 11:38 AM > Subject: Re: [PIC]: Precision 0.5V reference for AD > > > > >I was looking for a good and stable 0.5V referece, > > >because any error in the references will get into > > >the data. The 4.5V reference I plan to use has under > > >10ppmV drift with temperature and load. But there > > >is no 0.5V references this good. > > > > Check out the Analog Devices AD584 Precision Voltage Reference. I believe {Quote hidden}> > you should be able to get what you want from this. > > > > It starts off as a 5ppm reference, and you will need to divide from this > to > > get your 0.5V, but your reference output can be 2.5V, so you need only > some > > close tolerance resistors. By using 5 identical resistors in series, and > > tapping at one resistor from ground you should eliminate any temp co-eff > > problems, even if the resistors do not have a tight temp co-eff > themselves, > > and with 5 resistors from the same production batch you should get a close > > tolerance divider. > > > > -- > > http://www.piclist.com hint: The PICList is archived three different > > ways. See http://www.piclist.com/#archives for details. > > > > > > > -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.

> I need those 0.5V extra precission. Using 4.5V instead of 4V range means > losing about 12.5% resolution, and this is too much for my application (we > intend to validate a new vehicle dynamics mathematic model with the data we > collect). If you're that close to the edge, then you should be using an A/D with more bits anyway. If you use an external 12 bit A/D, then you can use the 4.5V reference and ground, and still have plenty of precision left. ***************************************************************** Embed Inc, embedded system specialists in Littleton Massachusetts (978) 742-9014, http://www.embedinc.com -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.

Hi Alvaro, you didn't wrote nothing about your AD precision and sampling rates. This two things give you the answer to your question and are the major criteria by choice of the final solution. Andy ------------------------------------------------------------------------ Hello. I am planning to read some 0.5V to 4.5V sensor voltages, and need good 0.5V and 4.5V precission references. I plan to use Analog Devices REF194 to get the 4.5V, but don't know how to get the 0.5V without using resistors, op-amps, etc. Any ideas? Thanks in advance, Alvaro Deibe. _________________________________________________________________ Join the world s largest e-mail service with MSN Hotmail. http://www.hotmail.com -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.

The project is almost finished. There are three PICs (16F877) talking each other with I2C. Only one 16MHz clock. The master PIC talks (RF) to a remote computer, sending the A/D data collected, 100 packets/second (Manchester-like encoding). The remote computer sends (RF) commands (Manchester encoding) to the master PIC, to control 6 RC servos (1us resolution, 10ms period). This master PIC does other things at the same time (like reading channels for steer and throttle/break coming from a standard RC receptor). The data collected comes from 1 triaxial linear accelerometer (Crossbow CXL02LF3), 3 rate gyros (TOKIN 16D0), 1 triaxial magnetometer (Honeywell HMC2003), 4 suspension and 1 steer position sensors (specialiced pots), temperature sensors (LM335) and 4 wheel angular speed sensors (inductive sensors). The data is collected 100 times per second. Not all of the sensors are relevant. Some are for reference purposes (temperature), and others need all the precision we can get out of the PICs (accelerations, angular speeds and magnetic fiels). The software is running (about 95% finished), and the hardware needs only defining the analog A/D part. One of the goals was to keep the parts count as low as possible, to get little PCB size. I hope this defines the problem. Thanks again, Alvaro Deibe. {Original Message removed}

I still think the solution to the 0.5V reference need is to just use the 4.5V reference and a higher resolution A/D. Any attempt at adjusting or dividing down another reference will introduce errors due to the temperature coefficient matching of the resistors in the divider. Harold FCC Rules Online at http://hallikainen.com/FccRules Lighting control for theatre and television at http://www.dovesystems.com Reach broadcasters, engineers, manufacturers, compliance labs, and attorneys. Advertise at http://www.hallikainen.com/FccRules/ . ________________________________________________________________ GET INTERNET ACCESS FROM JUNO! Juno offers FREE or PREMIUM Internet access for less! Join Juno today! For your FREE software, visit: dl.http://www.juno.com/get/web/. -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.

On Wed, 3 Jul 2002 10:38:47 +0100 "Alan B. Pearce" <@spam@A.B.PearceKILLspamRL.AC.UK> writes: > Check out the Analog Devices AD584 Precision Voltage Reference. I > believe > you should be able to get what you want from this. > > It starts off as a 5ppm reference, and you will need to divide from > this to > get your 0.5V, but your reference output can be 2.5V, so you need > only some > close tolerance resistors. By using 5 identical resistors in series, > and > tapping at one resistor from ground you should eliminate any temp > co-eff > problems, even if the resistors do not have a tight temp co-eff > themselves, > and with 5 resistors from the same production batch you should get a > close > tolerance divider. I've used standard resistor networks (with isolated resistors) for this purpose, hoping the tracking would be better than discrete resistors and that they'd all be at close to the same temperature. I've also used $5 resistors to get under 10 ppm/degree C tempco... Harold FCC Rules Online at http://hallikainen.com/FccRules Lighting control for theatre and television at http://www.dovesystems.com Reach broadcasters, engineers, manufacturers, compliance labs, and attorneys. Advertise at http://www.hallikainen.com/FccRules/ . ________________________________________________________________ GET INTERNET ACCESS FROM JUNO! Juno offers FREE or PREMIUM Internet access for less! Join Juno today! For your FREE software, visit: dl.http://www.juno.com/get/web/. -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.

Why not just use 2 precision resistors to make a voltage divider off of your 4.5 volts. If that's not close enough, just trim it in with your voltmeter and 1 more small resistor in series with the appropriate resitor (or a large value in parallel) Or do you want to manufacture these circuits? > >Hello. I am planning to read some 0.5V to 4.5V sensor voltages, and need good 0.5V and 4.5V precission references. I plan to use Analog Devices REF194 to get the 4.5V, but don't know how to get the 0.5V without using resistors, op-amps, etc. >Any ideas? >Thanks in advance, >Alvaro Deibe. > -- Looking forward: Alan Shinn Experience the beginnings of microscopy. Make your own replica of one of Antony van Leeuwenhoek's microscopes. visit http://www.mindspring.com/~alshinn/ -- http://www.piclist.com#nomail Going offline? Don't AutoReply us! email KILLspamlistservKILLspammitvma.mit.edu with SET PICList DIGEST in the body

> -----Original Message----- > From: Alan Shinn [SMTP:RemoveMEalshinnTakeThisOuTMINDSPRING.COM] > Sent: Thursday, July 04, 2002 9:41 AM > To: spamBeGonePICLISTspamBeGoneMITVMA.MIT.EDU > Subject: Re: [PIC]: Precision 0.5V reference for AD > > Why not just use 2 precision resistors to make a voltage divider off of > your 4.5 volts. If that's not close enough, just trim it in with your > voltmeter and 1 more small resistor in series with the appropriate > resitor (or a large value in parallel) > Or do you want to manufacture these circuits? > > The reference inputs on the PIC A/D can take up to 1mA during the aquisition phase, so you'd want as low source impedance as possible if accuracy is important. Regards Mike -- http://www.piclist.com#nomail Going offline? Don't AutoReply us! email TakeThisOuTlistservEraseMEspam_OUTmitvma.mit.edu with SET PICList DIGEST in the body

I still like the idea of using a smal DAC to provide the .5V reference. With 4.5V as it's high-end reference, it would produce 0.494V Pretty close, I think. oh, well. --Brendan > The reference inputs on the PIC A/D can take up to 1mA during the aquisition > phase, so you'd want as low source impedance as possible if accuracy is -- http://www.piclist.com#nomail Going offline? Don't AutoReply us! email RemoveMElistservTakeThisOuTmitvma.mit.edu with SET PICList DIGEST in the body

> Extending your idea a little more, I've searched Maxim's web for serial I2C > DACs (I already have I2C, so it makes little sense using a parallel DAC, > isn't it?) and found the MAX5822 dual 12-bit i2c DAC. It has very little I was thinking that you could just hard-wire the parallel DAC, and not worry about any software control of it. --Brendan -- http://www.piclist.com#nomail Going offline? Don't AutoReply us! email listservEraseME.....mitvma.mit.edu with SET PICList DIGEST in the body

> The reference inputs on the PIC A/D can take up to 1mA during the aquisition > phase, so you'd want as low source impedance as possible if accuracy is > important. > > Regards > > Mike Even more, because the acquisition is synchronized in all the PIC's, and takes place at the same time. That's why I want an op-amp to buffer the voltage refferences. Alvaro. -- http://www.piclist.com hint: To leave the PICList RemoveMEpiclist-unsubscribe-requestEraseMEEraseMEmitvma.mit.edu

{Quote hidden}> Alvaro, > > Sounds like you may be using standard 'automotive' sensors, > which are powered from a nominal +5V supply. > > If so, the output from these sensors is usualy ratiometric > (the output sensitivity is proportional to their +5V supply voltage). > > Usual automotive practice is then to use an AD with sufficient > resolution (say 10 or 12 bit) and use the 5V sensor supply as > your AD reference voltage. Your sensor readings will then track > any variation in the 5V supply/refence due to their ratiometric > characteristic. No need for expensive reference sources. > > You don't say what sort of sensors you are using, but most > automotive sensors only have an accuracy of +/-3%, so there > is no merit in cranking up the AD resolution way beyond the > basic accuracy of the sensors. > > regards, Richard Golding Well, not so. Automotive sensors are, frequently, ratiometric with the supply. But the ones I choose are not. The triaxial accelerometer Crossbow CXL02LF3 (it has Analog Devices MEMS accelerometers in it) gives a sensitivity of 1V/g, and it goes from -2 to +2g. Erros like Transverse Sensitivity, Non linearity or Alignment are large, but can be corrected in software. Errors like Noise Density (130ug/Hz^(1/2)) or Noise (1.5mg rms) are in the range of the 10 bits A/D. The triaxial magnetometer Honeywell HMC2003 has, again, its own internal voltage reference, so it's output is not ratiometric with the supply. The figures of this kind of sensor goes well beyond the resolution of the 10bits A/D (using set/reset pulse circuitry to cope with problems like hysteresis). I've tried Murata and Tokin piezo gyros, and will probably use Tokin CG-16D0 (the differences, however, are not that much). With this kind of sensors would probably be enough 8 bits A/D. Thanks for all, Alvaro Deibe. -- http://www.piclist.com hint: To leave the PICList RemoveMEpiclist-unsubscribe-requestspam_OUTKILLspammitvma.mit.edu

LTC1728 has tolerance of 0.05% http://www.linear.com/go/1389 Andre Abelian -----Original Message----- > I am planning to read some 0.5V to 4.5V sensor > voltages, and need good 0.5V and 4.5V precission > references. I plan to use Analog Devices REF194 > to get the 4.5V, but don't know how to get the > 0.5V without using resistors, op-amps, etc. Use ground as the lower reference. Do you really need the extra precision from a 4V range instead of a 4.5V range? ***************************************************************** Embed Inc, embedded system specialists in Littleton Massachusetts (978) 742-9014, 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 PICList is archived three different ways. See http://www.piclist.com/#archives for details.