Author Topic: LM399 based 10 V reference  (Read 316876 times)

Ismo and 1 Guest are viewing this topic.

Offline plesa

  • Frequent Contributor
  • **
  • Posts: 965
  • Country: se
Re: LM399 based 10 V reference
« Reply #325 on: September 27, 2015, 08:15:24 am »
Sorry, but it is nonsense to buy references on ebay if you can buy them from manufacturer.
You do not know history and reason why it is on ebay and references on pictures does not look to be in best shape.
Buy several (10x) LM399AH from Linear for 9 USD/pc and operate them at 125°C for one month and select pair (2-6) of lowest tempco references in your circuit.
Or Linear is offering free samples for both versions of LM399, so you can have 4 pieces for free.

I have to investigate one fake LTZ1000 in different thread.
« Last Edit: September 27, 2015, 08:19:18 am by plesa »
 

Online Andreas

  • Super Contributor
  • ***
  • Posts: 2413
  • Country: de
Re: LM399 based 10 V reference
« Reply #326 on: September 27, 2015, 10:18:25 am »

what do you think about this offer:

Your opinions?

So it may be better to buy this, slightly more expensive but pre-aged ones:

Price is much too high. (Compared to new LT parts from DigiKey).

Datecode 1209 on a NS-Reference?
I bought my last references from NS in 2008.
That was about the time when they were discontinued.
So in 2012 genuine NS parts are very strange.

If you solder the parts a new ageing cycle may begin.
So it is useless to buy pre aged parts which were already soldered many times.

With best regards

Andreas


 

Offline Edwin G. Pettis

  • Frequent Contributor
  • **
  • Posts: 413
  • Country: us
  • The plural of anecdote is not data.
Re: LM399 based 10 V reference
« Reply #327 on: September 27, 2015, 10:05:46 pm »
Regarding the myth of surplus LTZ1000/A parts, according to Linear Tech, there are no surplus parts or overstocks, every one of them are accounted for, including the ones sold through their online shop.  LT keeps close track of the LTZs that they sell to the OEMs, they trade data back and forth, the OEMS do not, I repeat, do not have surplus or overstock to sell to anybody, they buy just what they need and nothing more.

Unfortunately, the condition of used parts removed from equipment is available before purchase, they may be in good shape but odds are they are not, de-soldering can cause changes or damage to the chip, therefore, if you are trying to achieve spec in an Vref, buy the LTZ1000/A from a reputable source and forget about the sellers on eBay or elsewhere, they are selling you a pig-in-a-poke.
 

Offline ez24

  • Super Contributor
  • ***
  • Posts: 3092
  • Country: us
  • L.D.A.
Re: LM399 based 10 V reference
« Reply #328 on: September 28, 2015, 03:17:59 am »
fmaimon

This post is very long and old.  It is hard for me to understand what is going on.  Can you give a summary of what has happened over the years or start a new post on what you have learned.  (for new comers)

thanks
YouTube and Website Electronic Resources ------>  https://www.eevblog.com/forum/other-blog-specific/a/msg1341166/#msg1341166
 

Offline Mickle T.

  • Frequent Contributor
  • **
  • Posts: 342
  • Country: ru
Re: LM399 based 10 V reference
« Reply #329 on: October 09, 2015, 09:03:20 am »
DC-Cal Lite v1.0 - 10V/10mA/1mA calibrator, based on the AD5791B DAC.
0.1-10 Hz noise: 0V - 0.4 uV p-p, 10V - 4...5 uV p-p.
0-10 V INL <1 ppm (measured with Solatron 7081 DMMs).
Output resistance 0.003 Ohm (DC voltage mode).
TempCo ~ 0.1 ppm/C.
 

Online lukier

  • Supporter
  • ****
  • Posts: 615
  • Country: gb
    • Homepage
Re: LM399 based 10 V reference
« Reply #330 on: October 09, 2015, 09:57:16 am »
Wow, that is a very nice piece of kit and seriously expensive DAC. Wouldn't it be cheaper to do PWM-DAC, like Fluke does in their calibrators?

What is the stuff in the front end in the 8th picture, in the white/yellow box. Some kind of electrometer sensitive circuit with teflon insulated wires? It's not some weird battery with balancing outputs, is it?
 

Offline Mickle T.

  • Frequent Contributor
  • **
  • Posts: 342
  • Country: ru
Re: LM399 based 10 V reference
« Reply #331 on: October 09, 2015, 10:17:22 am »
White/yellow box is a 7 decades PWW voltage divider / resistor network from 1 Ohm to 9 MOhm.
For example: 0...1 uV rande on the N1A nanovoltmeter.
 

Offline bingo600

  • Super Contributor
  • ***
  • Posts: 1367
  • Country: dk
Re: LM399 based 10 V reference
« Reply #332 on: October 09, 2015, 12:04:42 pm »
Mickle

Why do you have two trafos on the DC ref ?

One for analog, and one for digital ?

/Bingo

 

Offline Mickle T.

  • Frequent Contributor
  • **
  • Posts: 342
  • Country: ru
Re: LM399 based 10 V reference
« Reply #333 on: October 09, 2015, 12:07:19 pm »
Yes, it is. One for analog, and one for digital.
 

Online Vgkid

  • Super Contributor
  • ***
  • Posts: 2511
  • Country: us
Re: LM399 based 10 V reference
« Reply #334 on: October 09, 2015, 02:27:00 pm »
Very cool, it is interesting to see what Mickle shows.
If you own any North Hills Electronics gear, message me. L&N Fan
 

Offline plesa

  • Frequent Contributor
  • **
  • Posts: 965
  • Country: se
Re: LM399 based 10 V reference
« Reply #335 on: October 09, 2015, 04:54:50 pm »
Wauu, interesting piece of equipment. MickleT are you going to share schematic?
 

Online Andreas

  • Super Contributor
  • ***
  • Posts: 2413
  • Country: de
Re: LM399 based 10 V reference
« Reply #336 on: October 09, 2015, 07:18:27 pm »
Wauu, interesting piece of equipment. MickleT are you going to share schematic?

Not only the schematic. Also the PCB layout is highly interesting.
Mickle: do you have a RS274 Gerber file for that.
(and of course the software.)
Which processor are you using?

With best regards

Andreas
 

Offline Mickle T.

  • Frequent Contributor
  • **
  • Posts: 342
  • Country: ru
Re: LM399 based 10 V reference
« Reply #337 on: October 09, 2015, 07:49:09 pm »
Schematic is a simplest. ATMega8, some opamps, resistors and that's all. PCB files is not yet ready, because of some critical mistakes in it. May be some later.
 
The following users thanked this post: denimdragon, umbro

Online Andreas

  • Super Contributor
  • ***
  • Posts: 2413
  • Country: de
Re: LM399 based 10 V reference
« Reply #338 on: October 09, 2015, 09:16:51 pm »
Thanks for sharing.

With best regards

Andreas
 

Offline quarks

  • Frequent Contributor
  • **
  • Posts: 821
  • Country: de
Re: LM399 based 10 V reference
« Reply #339 on: October 11, 2015, 01:30:54 pm »
DC-Cal Lite v1.0 - 10V/10mA/1mA calibrator, based on the AD5791B DAC.
0.1-10 Hz noise: 0V - 0.4 uV p-p, 10V - 4...5 uV p-p.
0-10 V INL <1 ppm (measured with Solatron 7081 DMMs).
Output resistance 0.003 Ohm (DC voltage mode).
TempCo ~ 0.1 ppm/C.

Very nice, thanks a lot for sharing
 

Offline bingo600

  • Super Contributor
  • ***
  • Posts: 1367
  • Country: dk
Re: LM399 based 10 V reference
« Reply #340 on: October 17, 2015, 03:07:20 pm »
Guyzz

Would any of the "Volt Gurus" care to have a look at my 731B graphs , and explain why it's moving the voltage "up", even w. same temperature.

https://www.eevblog.com/forum/testgear/is-my-'new'-fluke-731b-stabilizing-or-going-crazy/msg779636/#msg779636

TIA
/Bingo
 

Offline branadic

  • Super Contributor
  • ***
  • Posts: 1487
  • Country: de
Re: LM399 based 10 V reference
« Reply #341 on: November 09, 2015, 12:30:02 pm »
I had a phonecall with Prema today. They currently have our Prema 5017 for cal in their labs. After 2 years without calibration the gear is still within the 90 days spec. Pretty good.  :-+
I asked if they used a "special" treatment before they solder the reference into the boards. I was told that each reference is running in a batch with several other LMx99s for 1 year in a box in normal condition, taking data each month to calculate the drift for each of it. After 1 year the references were selected for drift for the different DVMs, the bad ones for low accuracy devices, the good ones for high accuracy devices.
They also did similar for the input voltage devider/resistor network and selected for T.C., low T.C. for high accuracy devices, worse ones for low accuracy devices.
I found that kind of interesting to hear and thought to share this info with you. Sad that they don't build DVMs anymore.
« Last Edit: November 09, 2015, 06:01:11 pm by branadic »
Fluke 8050A | Prema 5000 | Prema 5017 SC | Advantest R6581D | GenRad 1434-G | Datron 4000A | Tek 2465A | VNWA2.x with TCXO upgrade and access to: Keysight 3458A, Keithley 2002, Prema 5017 SC, 34401A, 34410A, Keithley 2182A, HDO6054, Keysight 53230A and other goodies at work
 

Offline TiN

  • Super Contributor
  • ***
  • Posts: 4053
  • Country: us
  • xDevs.com/live - 24/7 lab feed
    • xDevs.com
Re: LM399 based 10 V reference
« Reply #342 on: November 09, 2015, 03:30:04 pm »
Thanks for info, good to know!  :-+
YouTube | Chat room | Live-cam | Have documentation to share? Upload here! No size limit, firmware dumps, photos.
 

Online Andreas

  • Super Contributor
  • ***
  • Posts: 2413
  • Country: de
Re: LM399 based 10 V reference
« Reply #343 on: November 09, 2015, 08:30:35 pm »
taking data each month to calculate the drift for each of it.

Mhmm,

monthly measurements leave many open points.

attached 2 LM399 references monitored now about 650 days.

The red LM399 (CH6) has lower drift over time.
But obviously has a problem with "popcorn noise".
Jumping around 1-2 ppm within few days.

The green LM399 (CH7) had a large drift in the begin.
But now has calmed down to less than 1 ppm.
So I would trust more the green one.

The blue LTZ1000 is only for reference
(to show the instability of the measurement system)

So how exactly did they select the voltage references:
Including the measurement of the noise?.

With best regards

Andreas

 

Offline branadic

  • Super Contributor
  • ***
  • Posts: 1487
  • Country: de
Re: LM399 based 10 V reference
« Reply #344 on: November 09, 2015, 09:34:09 pm »
Quote
So how exactly did they select the voltage references:
Including the measurement of the noise?.

The references were measured against a Fluke calibrator and drift against the calibrator were calculated. That's all the info I got.
Fluke 8050A | Prema 5000 | Prema 5017 SC | Advantest R6581D | GenRad 1434-G | Datron 4000A | Tek 2465A | VNWA2.x with TCXO upgrade and access to: Keysight 3458A, Keithley 2002, Prema 5017 SC, 34401A, 34410A, Keithley 2182A, HDO6054, Keysight 53230A and other goodies at work
 

Offline Theboel

  • Regular Contributor
  • *
  • Posts: 223
  • Country: id
Re: LM399 based 10 V reference
« Reply #345 on: November 25, 2015, 12:57:48 am »
Hello All,
As far as I understand for LTZ1000 lowering working temperature can increase the stability is this also work for LM399 ? if this lowering temperature also work with LM399 could somebody share their experience and how to do it (current limiter maybe ?)
Thank You
 

Offline Macbeth

  • Super Contributor
  • ***
  • Posts: 2550
  • Country: gb
Re: LM399 based 10 V reference
« Reply #346 on: November 25, 2015, 01:27:32 am »
I believe all the thermal regulation is built in to the LM399, unlike the much simpler LTZ which requires complicated external support circuitry that can be tweaked?
 

Online Andreas

  • Super Contributor
  • ***
  • Posts: 2413
  • Country: de
Re: LM399 based 10 V reference
« Reply #347 on: November 25, 2015, 06:56:18 am »
(current limiter maybe ?)

Hello,

this is no good idea and will lead to instabilities.
(how do you get the inner temperature of the LM399?)

The only practical way is to leave the heater unconnected and put the LM399 in a own oven.
You could also use the LM329 (= LM399 without heater) which is cheaper but unfortunately no longer sold in metal can case.

With best regards

Andreas


 

Online Gyro

  • Super Contributor
  • ***
  • Posts: 4809
  • Country: gb
Re: LM399 based 10 V reference
« Reply #348 on: November 25, 2015, 10:48:22 am »
Quote
if this lowering temperature also work with LM399 could somebody share their experience and how to do it (current limiter maybe ?)

The LM399 datasheet actually has (had) data on the permissible value range of a resistor in series with the heater to limit its peak warm-up load on the power supply. Unfortunately this only increases the warm-up time, the internally regulated die temperature remains the same.

Edit: I suppose you could feed the heater circuit via a series pass element controlled by a temperature sensor attached to the case, but I suspect that this would be extremely unstable due to the small thermal mass and coupling to the sensor. You'd be a lot better not using the heater and building an external oven instead.
« Last Edit: November 25, 2015, 10:56:20 am by Gyro »
Chris

"Victor Meldrew, the Crimson Avenger!"
 

Offline Alex Nikitin

  • Super Contributor
  • ***
  • Posts: 1013
  • Country: gb
  • Femtoampnut and Tapehead.
    • A.N.T. Audio
Re: LM399 based 10 V reference
« Reply #349 on: November 25, 2015, 11:50:00 am »
(current limiter maybe ?)

Hello,

this is no good idea and will lead to instabilities.
(how do you get the inner temperature of the LM399?)

The only practical way is to leave the heater unconnected and put the LM399 in a own oven.
You could also use the LM329 (= LM399 without heater) which is cheaper but unfortunately no longer sold in metal can case.

With best regards

Andreas

Hi Andreas,

I've looked at the LM329 information and just as a side note - the Linear datasheet has exactly the same picture of LF noise there as in the LM399 datasheet (from Linear)  :palm: , and the TI datasheet has a different (from the old National datasheet) picture with much lower noise level. I suspect that the noise level on an unheated LM329 should be considerably lower than on the LM399 at 90C.

Cheers

Alex

 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf