Author Topic: Data on various voltage reference ICs over time.  (Read 4087 times)

0 Members and 1 Guest are viewing this topic.

Offline InsatmanTopic starter

  • Frequent Contributor
  • **
  • Posts: 278
  • Country: ph
Data on various voltage reference ICs over time.
« on: November 27, 2017, 07:07:40 am »
Greetings,  Over the past several months I have been testing a variety of voltage reference ICs.  4.096V, 5V and 10V references ICs are tested from a variety of manufactures.  All ICs were purchased new from Digikey.  Temperature/Humidity data is overlayed where available.  Note; you can use the humidity data for all three charts but only the 10V data has it overlayed.   All measurements are made using an HP34401A.  The unit isn't in official cal, but has been checked against a calibrated reference from VoltageStandards.com and is well within spec. 

Enjoy
Insatman
Retired Pulsed Power Engineer/Physicist...now I just dabble in electronics
 
The following users thanked this post: Andreas, chuckb, zhtoor

Offline Andreas

  • Super Contributor
  • ***
  • Posts: 3221
  • Country: de
Re: Data on various voltage reference ICs over time.
« Reply #1 on: November 27, 2017, 07:44:42 am »
Hello,

Interesting work.

some questions arise:

- how is the setup made? Devices soldered directly on board or socketed or dead bug style mounted?

- VRE305 returned to APEX (why?)
With 2 samples of VRE3050A I had the experience that they have large noise and large hysteresis (T.C. was within or near spec)

- do you know the T.C. of your DMM?

- do you have the values as .CSV or .TXT
  the different scales makes it difficult for me to see details on the more stable devices.

with best regards

Andreas


 

Offline InsatmanTopic starter

  • Frequent Contributor
  • **
  • Posts: 278
  • Country: ph
Re: Data on various voltage reference ICs over time.
« Reply #2 on: November 28, 2017, 09:05:28 am »
@Andreas,
I will try and answer your questions.
 
The setup used.  Most devices are socketed in the fixture pictured in the attachment.   I built this fixture to test and age devices for use later in various projects, so I used sockets.  Note the VRE305 socket is a high quality mil-max socket, the others are standard 8 pin sockets from digikey. 

A few devices are in various reference boxes such as the one marked on the 10V chart as "PTC AD587UQ"  This device is thermally stabilized with a PTC thermistor as an experiment.
 
A list of the devices not in the fixture follows.
10V REF01EP-1 is socketed and housed in a project box.
10V PTC AS587UQ is socketed and thermally stabilized using a PTC thermistor.
10V AD587UQ-1-HP is socketed and housed in a aluminum "hotbox" which is maintained at 52C via a PWM heater servo.
10V MV-106...this is my EDC voltage standard and really doesn't belong on this chart but appears for reference
10V LM399 is soldered into a voltage reference circuit in a project box.
5V LT1460GCZ5-1 is soldered and housed in a project box (same box as REF01EP-1)  I suspect that some of the noise on the other LT1460 devices is from the sockets used.


VRE305.  I returned the first VRE305 because it was clearly out of spec for initial voltage output and it looked like it was also drifting much more than spec.  After submitting my data to APEX they agreed to ship me a new one if I returned the old one.   We each paid for shipping...me to return and APEX to send me a new one (FedEx no less).   The new unit is working much better than the original.  A friend of mine who has extensive experience with the VRE305 says that the units made by APEX are not so good anymore.  The original units (can't remember original manufacturer) were much better.   These chips are pricey...more than LTZ1000.  I don't recommend them.

My DMM is an HP34401A which for this voltage and range setting is specified at 6uV/degC maximum.  I suspect the real value is somewhat lower than that.

The data is in an Excel file which can be exported to CSV ore TXT.   PM me and I'll send you whatever format you perfer via private email. 

Insatman

Retired Pulsed Power Engineer/Physicist...now I just dabble in electronics
 
The following users thanked this post: zhtoor

Offline beanflying

  • Super Contributor
  • ***
  • Posts: 7358
  • Country: au
  • Toys so very many Toys.
Re: Data on various voltage reference ICs over time.
« Reply #3 on: November 28, 2017, 10:23:46 pm »
Interesting information.

On the 5V data any idea what happened to the measured data between about 7-800 hours? There seems to have been some strange positive and negative drifts happening?

Coffee, Food, R/C and electronics nerd in no particular order. Also CNC wannabe, 3D printer and Laser Cutter Junkie and just don't mention my TEA addiction....
 

Offline InsatmanTopic starter

  • Frequent Contributor
  • **
  • Posts: 278
  • Country: ph
Re: Data on various voltage reference ICs over time.
« Reply #4 on: November 29, 2017, 07:12:02 am »
@beanflying,

i'm not sure actually.  Early on I didn't have automatic temp/humidity logging.  So I don't know if that was a factor.   

Insatman.


Interesting information.

On the 5V data any idea what happened to the measured data between about 7-800 hours? There seems to have been some strange positive and negative drifts happening?
Retired Pulsed Power Engineer/Physicist...now I just dabble in electronics
 

Offline Andreas

  • Super Contributor
  • ***
  • Posts: 3221
  • Country: de
Re: Data on various voltage reference ICs over time.
« Reply #5 on: November 30, 2017, 02:04:02 pm »
Hello Insatman,

thanks for the offer. But I fear that the devices of interest will have near equal T.C. to a 34401A if you do not have narrowed down the data sheet value by a measurement.

With best regards

Andreas
 

Offline InsatmanTopic starter

  • Frequent Contributor
  • **
  • Posts: 278
  • Country: ph
Re: Data on various voltage reference ICs over time.
« Reply #6 on: December 01, 2017, 02:44:05 am »
@ Andreas,
I see your point.  At present I lack a method of doing that measurement with sufficient accuracy.   One of the motivations for building/testing various references was to have a stable enough source for such a measurement.  At that point I would have to build some sort of environmental chamber for the device under test (34401A).   I'm not at that point however.   What is useful is looking at the group in total and see which are most stable and of those stable ones to look for patterns of like movement which could infer the meters tempco.  Especially if the sources moving together are unlike devices.   

I've also acquired additional 6.5 digit meters and am comparing them to one another over time.  This may also at least let me bracket the tempco/drift.  At present a 7.5 or 8.5 meter is just too expensive for me, unless a real bargain pops up.   

Insatman

Hello Insatman,

thanks for the offer. But I fear that the devices of interest will have near equal T.C. to a 34401A if you do not have narrowed down the data sheet value by a measurement.

With best regards

Andreas
Retired Pulsed Power Engineer/Physicist...now I just dabble in electronics
 
The following users thanked this post: denimdragon

Offline InsatmanTopic starter

  • Frequent Contributor
  • **
  • Posts: 278
  • Country: ph
Re: Data on various voltage reference ICs over time. Update
« Reply #7 on: January 18, 2018, 08:38:55 am »
Attached in the next two posts (because of the 2M limit) are updated data on my measuring of various voltage references.   New are some 2.048 and 2.5V references.
Hope this is useful to some of you.   All data is measured on a 6.5digit HP34401A except for the LM399 which is measured on an HP3457A.
Retired Pulsed Power Engineer/Physicist...now I just dabble in electronics
 
The following users thanked this post: denimdragon

Offline InsatmanTopic starter

  • Frequent Contributor
  • **
  • Posts: 278
  • Country: ph
Re: Data on various voltage reference ICs over time.
« Reply #8 on: January 18, 2018, 08:40:01 am »
Remaining PDFs for previous post
Retired Pulsed Power Engineer/Physicist...now I just dabble in electronics
 
The following users thanked this post: denimdragon

Offline InsatmanTopic starter

  • Frequent Contributor
  • **
  • Posts: 278
  • Country: ph
Re: Data on various voltage reference ICs over time.
« Reply #9 on: April 16, 2018, 05:46:33 am »
Attached is data collected on various voltage reference ICs over time.  Hopefully some of you will find this useful in deciding which reference IC is best for your application.

Some notes:
1) The temperature drop at ~5800 hours is due to an air conditioner replacement in my test lab.  The new unit has double the capacity, allowing me to keep the lab at a lower temperature.
2)  The VRE305 data on "Best 5V" clearly shows out of specification drift.  This is an APEX branded IC, which is a replacement from the original one I purchased from Digikey.  The original VRE305 was returned to  APEX for being out of spec on initial voltage.  A "new old stock"  Thaler branded VRE305 will be tested in the same chassis after it has accumulated enough burn-in hours.
3) The MAX6250ACPA+ is a cost effective alternative to the VRE305.
4) Relative Humidity data is only presented on the 10V devices, but you can apply the data to all devices.
 
Retired Pulsed Power Engineer/Physicist...now I just dabble in electronics
 
The following users thanked this post: Andreas

Offline 3roomlab

  • Frequent Contributor
  • **
  • Posts: 824
  • Country: 00
Re: Data on various voltage reference ICs over time.
« Reply #10 on: April 16, 2018, 06:26:11 am »
would you mind telling a brief story of your LM399 premod and postmod?
 

Offline Andreas

  • Super Contributor
  • ***
  • Posts: 3221
  • Country: de
Re: Data on various voltage reference ICs over time.
« Reply #11 on: April 16, 2018, 06:48:18 pm »
3) The MAX6250ACPA+ is a cost effective alternative to the VRE305.

obviously: 20 ppm/kHr for the VRE is something I have not seen on other buried zeners after the first kHr(s).
(For me the large noise of my VRE3050s was the reason why I never did some ageing measurements on these).

The MAX6x50 that I have measured have "only" about 10 ppm/year.
The AD587LQ have even better values (2 ppm/year typ) after some kHrs run in time.

with best regards

Andreas
 

Offline InsatmanTopic starter

  • Frequent Contributor
  • **
  • Posts: 278
  • Country: ph
Re: Data on various voltage reference ICs over time.
« Reply #12 on: April 17, 2018, 01:19:47 am »
would you mind telling a brief story of your LM399 premod and postmod?

This particular LM399 reference was my prototype built on perf board.  The original version (pre-mod) used the LT1001 op-amp recommended by LT on the LM399 datasheet tp scale the output voltage to 10V.  This voltage was adjusted using a 5K trimmer in series with 6.8K.   The unit was housed in a cast aluminum enclosure and the LM399 was further enclosed by a small styrofoam cube.  15V was supplied using an LT3080ET precision regulator.   The LT3080 was supplied from an 18V regulated wall-wart type supply.   Schematic Attached.   

The modifications consisted of replacing the op-amp with an OPA177FP (pin compatible), replacing the 5K pot with a 100ohm pot and 1.5K resistor in series and lining the aluminum enclosure in thin foam insulation.  Of those mods I think the op-amp replacement is the most important, followed by replacing the pot with a much lower value.   The fixed LTC resistors are all 15-20ppm parts, while the pot is rated at 100ppm/C.   Better resistors would help, but the cost are also higher.   Pots better than 100ppm are not easy to find or cheap. 

I now have a PCB design for these references that works well.  I have a couple of devices under test currently with this PCB.  Schematic of latest version also attached.
« Last Edit: April 17, 2018, 01:42:26 am by Insatman »
Retired Pulsed Power Engineer/Physicist...now I just dabble in electronics
 
The following users thanked this post: 3roomlab

Offline InsatmanTopic starter

  • Frequent Contributor
  • **
  • Posts: 278
  • Country: ph
Re: Data on various voltage reference ICs over time.
« Reply #13 on: April 17, 2018, 01:49:24 am »
3) The MAX6250ACPA+ is a cost effective alternative to the VRE305.

obviously: 20 ppm/kHr for the VRE is something I have not seen on other buried zeners after the first kHr(s).
(For me the large noise of my VRE3050s was the reason why I never did some ageing measurements on these).

The MAX6x50 that I have measured have "only" about 10 ppm/year.
The AD587LQ have even better values (2 ppm/year typ) after some kHrs run in time.


with best regards

Andreas

I agree about the AD587LQ.  Looking back at my data I tested one early on.   It performed very well.  I'm not sure why I pulled it from the text fixture...probably needed the space at that time for something else.  I still have the IC and will put it under test again when I have an open slot available.
Retired Pulsed Power Engineer/Physicist...now I just dabble in electronics
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf