EEVblog Electronics Community Forum
Products => Test Equipment => Topic started by: quantumvolt on June 17, 2013, 07:39:23 am
-
Another Voltage Reference thread? Yes. This one is tested out and is for almost everyone.
The AD588 gives 10.000 Volt +- 1 mV (0.01% or 100 ppm) with tempco 1.5 ppm/degC without any external components (except decoupling capasitors for the batteries/power supply). The chip is trimmed/calibrated/sorted (?) from the manufacturer so you do not need to adjust the output voltage - just hook it up and you are guarantied an output between 9.999 and 10.001. You can trim it if you have access to a suitable meter.
It is not the best reference in the world - sure. But it is probably the best of the ones that need no supporting circuits or devices around them. When doing a search on this forum, I amazingly get no results for AD588 (in spite of all the threads about voltage references). This chip is a real simple and cheap way to get a reference without any measuring/trimming. And you DO get your 10 V ref. calibrated - the chip via the datasheet is of course in effect an implicit calibration traceable to Analog Devices (unless there are melons).
You can buy it on eBay for around USD 10 and up including shipping (when I searched now, I got 26 sellers). Note that you probably want the AD588BQ because of higher initial precision. Here is the datasheet and schematic/wiring (for 10 V, 5 V and more):
http://www.analog.com/static/imported-files/data_sheets/AD588.pdf (http://www.analog.com/static/imported-files/data_sheets/AD588.pdf)
My test-chip has been beaten for weeks - one pin broken, soldered in/out multiple times, shorted pins, half the supply voltage gone for hours and so on. It is still very precise and stable. I have ordered 3 more (from different sellers). When paralleling 4 units through 100 Ohm resistors I will statistically - given that the chips' errors are independent - improve the performance with a factor Sqrt(4)=2 i.e. 0.005%. Hence max. deviation from 10 Volt will statistically be 0.5 mV.
(https://www.eevblog.com/forum/index.php?action=dlattach;topic=17905.0;attach=51865;image)
The chip has been handled too much and too badly by me. The pin for the negative supply broke. I soldered a piece of wire to it. No problem :-DMM
Typical application :-DD
(https://www.eevblog.com/forum/index.php?action=dlattach;topic=17905.0;attach=51867;image)
1x AD588BQ, 4x 9V batteries, 1x LM7815, 1X LM7915, 2x 220 microF, 2x 0.1 microF, optional 1x 0.068 or 0.1 microf noise red. cap. (see datasheet). That's all ... :-+
-
Very interesting. 28USD from DGK seems legit. O0
-
I wouldn't know as I live in Thailand. But for those of you who do not want to buy from Hong Kong / mainland China, I just checked the lowest price on eBay from US sellers (shipping is for Thailand):
Price:
US $16.95
Shipping:
$6.16 (approx. THB188.31) USPS First Class Mail Intl / First Class Package Intl Service
Since I am talking about China - some of the Chinese have fun with this chip also :o :
Link: http://www.microcontrol.cn/bbs/MINI/Default.asp?101-34077-0-0-0-0-0-a-.htm (http://www.microcontrol.cn/bbs/MINI/Default.asp?101-34077-0-0-0-0-0-a-.htm)
(http://www.microcontrol.cn/HomeImages/DIY/MRB588/MRB588_01.jpg)
-
IIRC Silicon Chip did a reference using this chip a while ago. I was looking at getting one, but RS did not stock them, so got an AD581KH instead.
-
If you hookup the AD588 to a 6.5 digit multimeter, and if you blow air on the chip, does the digits drift?
-
I can't see the pictures in your post... :(
-
I am sorry if there is something wrong with the small pictures. I will look into it.
I do not have a AD581KH, but I read somewhere about the AD587 used in a reference for sale to hobbyists. That is how I found the AD588 - I simply wanted 0.01% without depending on other people, instruments and a trim pot. While searching I suddenly saw:
Tolerance: ? AD588 0.01% ? ADR43x, AD780, REF195 0.04% • Drift (TC): ? AD586, AD588 1?2ppm/°C ? AD780, ADR42x, ADR43x, 3 ppm/°C ADR01, ADR02, ...
and other similar tables.
About blowing on the chip - haven't tried it. But I have heated it with the soldering iron and cooled it with ice. I will make a video - I just got a cheap thermometer with probe from eBay.
Now that I have decided to do something with this chip - I will use some time to try to do some measurements/testing. But I only have a DMM Agilent 34401A - I have no AC millivoltmeter for noise, no scope, no data logger etc. But I will do what I can ...
For now I only have this video - meant for one of my friends who insisted I needed a shielded box. I have ordered some aluminium project boxes, but the board in the video is just a copper clad ground plane (cheap PCB from eBay). All the components are fixed with tape before soldering and wireing .... :-//
https://www.youtube.com/watch?v=twbxo6ALq04 (https://www.youtube.com/watch?v=twbxo6ALq04)
-
Found a Russian 8) version of "JIM ROWE Everyday Practical Electronics, June 2011". Looks good if you are going to do a decent build ...
Link: http://radio-hobby.org/modules/news/article.php?storyid=1226 (http://radio-hobby.org/modules/news/article.php?storyid=1226)
(http://radio-hobby.org/uploads/schemes2/1226/r-3.png)
The trimming is optional - you need access to a calibrated precision Voltmeter with minimum resolution 0.1 mV. If not - just do as in the datasheet pg. 8 and you get the 0.01% (max. error 1 mV) by default.
-
This seems to be the new version of the old AD2712LN. The AD2712 has better line regulation, identical load regulation, better TempCo but worse long term stability and noise. The AD2712 is obsoleted so its pretty pricey from any reputable distributors. As an aside, I am going to be getting a batch of AD2710L's from a reputable US distributor. The 2710L's have the same 1mV at 10V specs as the AD2712, but are only +10V. I plan on putting the ones I don't need up for sale on here. Depending on the shipping cost to me, they will likely be around $10 each + shipping. When I have more finalized info I'll post in the B/S/W forum if anyone is interested.