Hi All, I'm working on something that requires a stable voltage reference. Long term stability isn't needed, but good thermal stability is. In other words it would be nice to have sub ppm/C temperature, but a few ppm/month or even per day isn't an issue. Anyway, it got me thinking about building a reference for fun, with some of the newer IC's available, but doing so in an ovenized housing. I know the LM399's and LTZ1000's are out there, but has anyone looked at any of the newer IC's out there? For example this one seems to have pretty good ppm/C drift: http://www.digikey.com/product-detail/en/VRE3025JS/598-1916-ND/2036526 However it's not cheap. And this one looks pretty good for the sub $10 price. http://www.digikey.com/product-detail/en/ADR4550BRZ/ADR4550BRZ-ND/3196874 This is the one I was thinking about parking in an oven. It has a turnover around 70C. Has anyone played with inexpensive reference chips, and tried to make a good reference out of them? Any thoughts or comments welcome! Dan

Hello Dan, If you look for sub-ppm then also noise or hysteresis might be a issue. Nearly all good references have around 0.6ppm low frequency noise with respect to output voltage. (giving 3uVpp for 5V). I have tested 2 VRE3050A for noise and one for TC. In the range of 16-40 deg C average tempco was low (about 8 ppm over 16-40 deg) (nearly zero around 28 deg C) but hysteresis at 25 deg was 4 ppm and 0.1-10 Hz noise around 9uVpp. The second VRE I have only tested for noise which gave also around 9 uVpp. I do not know if both devices are "mondays" devices since I bought them at a time shortly after Thaler was sold to APEX. And current consumption was far below the value of the datasheet. But for me I decided that they are no good value for money for my needs. For the same money you get 5 AD586LQ with <3uVpp noise and nearly no hysteresis in the above temperature range. Ok perhaps you have to select them for tempco. In my case tempco is compensated by a microcontroller with a (NTC) temperature sensor. The ADR4550 suffers like nearly all plastic housings from large hysteresis and perhaps from humidity sensitivity. I measured a nearly linear average tempco of -2ppm/K on one sample. Hysteresis around 16ppm at 25 deg C. Noise was around 3uVpp like good buried zener devices. The LM399 that I have tested are far better in T.C. than the 1ppm/K that are stated in the datasheet. At least this is true around room temperature (10-40 deg C). Only noise is rather different from device to device. Some around the 0.6 ppm others are a factor of 2 or so larger in noise. So perhaps you will have to select for noise. Some Info of other features of LM399 (tilting drift, termal isolation) can be found here: http://www.eevblog.com/forum/projects/lm399-based-10-v-reference/ If you need low T.C. and low noise there is nearly no way not to use the LTZ1000 http://www.eevblog.com/forum/projects/ultra-precision-reference-ltz1000/ With best regards Andreas Am 31.10.2014 um 20:09 schrieb Dan Kemppainen: > Hi All, > > I'm working on something that requires a stable voltage reference. > Long term stability isn't needed, but good thermal stability is. In > other words it would be nice to have sub ppm/C temperature, but a few > ppm/month or even per day isn't an issue. > > Anyway, it got me thinking about building a reference for fun, with some > of the newer IC's available, but doing so in an ovenized housing. > I know the LM399's and LTZ1000's are out there, but has anyone looked at > any of the newer IC's out there? > > For example this one seems to have pretty good ppm/C drift: > http://www.digikey.com/product-detail/en/VRE3025JS/598-1916-ND/2036526 > However it's not cheap. > > And this one looks pretty good for the sub $10 price. > http://www.digikey.com/product-detail/en/ADR4550BRZ/ADR4550BRZ-ND/3196874 > This is the one I was thinking about parking in an oven. It has a > turnover around 70C. > > Has anyone played with inexpensive reference chips, and tried to make a > good reference out of them? Any thoughts or comments welcome! > > Dan > > > > > > > _______________________________________________ > volt-nuts mailing list -- volt-nuts@febo.com > To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts > and follow the instructions there.

-------- In message <54549B2C.9070908@t-online.de>, Andreas Jahn writes: >Ok perhaps you have to select them for tempco. In my case tempco is >compensated by a microcontroller with a (NTC) temperature sensor. Don't underestimate how easy it is to make a oven with a peltier. If you just want to stabilize the temperature around ambient, you need very little power to a peltier to do that, and it can move heat both ways. -- Poul-Henning Kamp | UNIX since Zilog Zeus 3.20 phk@FreeBSD.ORG | TCP/IP since RFC 956 FreeBSD committer | BSD since 4.3-tahoe Never attribute to malice what can adequately be explained by incompetence.

Hello Poul, You are right: Since I use usually battery supply I do not want any heating/cooling. And since the references that I use are only for measuring voltages in a kind of "voltmeter" I can easily compensate the effect of temperature by calculation. On the other side also a reference can be stabilized by a NTC in a small temperature range. See Geller Labs SVR-T. http://www.gellerlabs.com/svr%20series.htm Further Peltiers have one disadvantage that I do not want to have in precision cirquits: The cold side is always collecting humidity. with best regards Andreas Am 01.11.2014 um 12:13 schrieb Poul-Henning Kamp: > -------- > In message <54549B2C.9070908@t-online.de>, Andreas Jahn writes: > >> Ok perhaps you have to select them for tempco. In my case tempco is >> compensated by a microcontroller with a (NTC) temperature sensor. > Don't underestimate how easy it is to make a oven with a peltier. > > If you just want to stabilize the temperature around ambient, you > need very little power to a peltier to do that, and it can move > heat both ways. >

-------- In message <5454C970.40606@t-online.de>, Andreas Jahn writes: >Further Peltiers have one disadvantage that I do not want to have in >precision cirquits: >The cold side is always collecting humidity. Only if you cool below the dewpoint. -- Poul-Henning Kamp | UNIX since Zilog Zeus 3.20 phk@FreeBSD.ORG | TCP/IP since RFC 956 FreeBSD committer | BSD since 4.3-tahoe Never attribute to malice what can adequately be explained by incompetence.