Author Topic: LM399 based 10 V reference  (Read 330584 times)

0 Members and 1 Guest are viewing this topic.

Offline Andreas

  • Super Contributor
  • ***
  • Posts: 2447
  • Country: de
Re: LM399 based 10 V reference
« Reply #125 on: April 09, 2014, 07:05:05 pm »
Hello Branadic,

The Vishay parts are usually with around 12 weeks delivery time. (from Israel).
They are manufactured on demand only. (with except of a handfull of stock parts).
If you ask Vishay Germany for parts in sample quantities you are usually redirected to Powertron / Teltow.
They have also own precision resistors.
Of course you can have quicker turnaround times but at a extra cost.

At your prices I guess that HBE-Shop (a FARNELL reseller) has prices including VAT. (and the others are adding VAT).

SMD-Parts (VSMP) are available at DigiKey. Trimming to the desired resistance is included.
http://www.digikey.de/product-search/de/resistors/precision-trimmed-resistors/66806?k=vishay%20vsmp

with best regards

Andreas

 

Offline branadic

  • Super Contributor
  • ***
  • Posts: 1509
  • Country: de
Re: LM399 based 10 V reference
« Reply #126 on: April 09, 2014, 09:02:53 pm »
What is ignored are the longterm / load life stabilities:

Vishay VSMP: to ± 0.03 % at 70 °C, 2000 h at rated power
Vishay Y1453: to ± 0.005 % at 70 °C, 2000 h at rated power
Vishay Y0062: ± 0.005 % at 70 °C, 2000 h at rated power
Vishay Y0785: to ± 0.005 % at 70 °C, 2000 h at rated power
Rohpoint 8G16: ±35ppm/10,000hours ±50ppm/26,000 hours
Burster 1142: 0.01% over years (whatever that finally means)

That shows that VSMP is a comparable bad choice and 8G16 seem to be best, beside the "higher" TC.
« Last Edit: April 09, 2014, 09:07:28 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 Andreas

  • Super Contributor
  • ***
  • Posts: 2447
  • Country: de
Re: LM399 based 10 V reference
« Reply #127 on: April 10, 2014, 09:15:44 pm »
Hello branadic,

in my opinion the shelf life stability is the more interesting spec.
(rarely found in the data sheet).
Usually you would not use a precision resistor near the rated power.

with best regards

Andreas
 

Offline CaptnYellowShirt

  • Frequent Contributor
  • **
  • Posts: 447
  • Country: us
  • Scooty Puff Jr.
Re: LM399 based 10 V reference
« Reply #128 on: April 11, 2014, 12:15:30 am »
...just sticking my head in to get updates on the thread....
 

Offline branadic

  • Super Contributor
  • ***
  • Posts: 1509
  • Country: de
Re: LM399 based 10 V reference
« Reply #129 on: April 13, 2014, 06:32:49 am »
Quote
in my opinion the shelf life stability is the more interesting spec.
(rarely found in the data sheet).
Usually you would not use a precision resistor near the rated power.

I agree in both points, but unless such values are printed in datasheet you need to compare the one that are. And the load life stability is a good parameter to compare them and better than nothing, right?
I would assume that a resistor with ±35ppm/10,000hours ±50ppm/26,000 hours is more stable compared to the one with ± 0.005 % at 70 °C, 2000 h at rated power.
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 Andreas

  • Super Contributor
  • ***
  • Posts: 2447
  • Country: de
Re: LM399 based 10 V reference
« Reply #130 on: April 22, 2014, 08:55:26 am »

I have done the correction calculation for a linear ageing (linear regression coefficient).

The correction is applied on the averaged measurement values and not for the original 1 minute values.

The spread values with the foam do not really change:
CH6 foam corrected:   4.4 uV = 1.3 ppm (unchanged)
CH7 foam corrected: 10.4 uV = 3.0 ppm (was 10.1 uV)

The spread values for the "cotton" isolation seems even to be better for the slotted PCB:
CH6 cotton corrected:  4.6uV = 1.4 ppm (was 4.1 uV)
CH7 cotton corrected:  2.4uV = 0.7 ppm (was 3.6 uV)

remember that standard deviation for my setup is around 0.25 ppm or 1 uV.


I now have repeated the tilt measurements on the well isolated (cotton) LM399s:

Again the slotted PCB (CH7) seems to be better.
CH6 cotton (aged):  4.1uV = 1.2 ppm   (zener voltage = 2 * 3428 = 6856 mV)
CH7 cotton (aged):  1.9uV = 0.5 ppm   (zener voltage = 2 * 3452 = 6904 mV)

with best regards

Andreas


 

Offline CaptnYellowShirt

  • Frequent Contributor
  • **
  • Posts: 447
  • Country: us
  • Scooty Puff Jr.
Re: LM399 based 10 V reference
« Reply #131 on: April 22, 2014, 02:23:56 pm »
What are the LM399's driving currents?
 

Offline Andreas

  • Super Contributor
  • ***
  • Posts: 2447
  • Country: de
Re: LM399 based 10 V reference
« Reply #132 on: April 22, 2014, 03:17:10 pm »
Hello,

simple 7K5 pull up resistors (metal film) on battery supplied, stabilized supply of 14.08V for CH6 + CH7
so around 1mA.

with best regards

Andreas
 

Offline turbo!

  • Regular Contributor
  • *
  • Posts: 89
Re: LM399 based 10 V reference
« Reply #133 on: April 22, 2014, 05:52:07 pm »
This voltage reference is used in a precision balance I tore down just the other day. The resistor that attaches to it is a 15.00k Ohm, 0.5%, 50ppm/K

Pay close attention to the ground plane "tongue" dips under the resistors. The ordinary 10 ohm resistor right next to it feeds the heating element for the can.

Perhaps the  10 ohm resistor is meant to be a heating element as well to warm up the trace beneath it so the entire cluster remains at the same temperature? The best way to avoid EMF voltage generation aside from using exotic matching is to have the affected components all be at the same temperature.

The six banded resistor then connects to ground through the shunt resistors on stand-offs which are 0.1%. 

The permissible voltage variation span is fairly wide. 6.8 to 7.1v. For this scale, it doesn't really matter what the exact voltage is, because the offset adjustment is made with a calibration. What matters is that it STAYS where it is, whatever that value may be.

« Last Edit: April 22, 2014, 05:54:16 pm by turbo! »
 

Offline CaptnYellowShirt

  • Frequent Contributor
  • **
  • Posts: 447
  • Country: us
  • Scooty Puff Jr.
Re: LM399 based 10 V reference
« Reply #134 on: April 22, 2014, 06:06:21 pm »
@turbo

R43 and R44?
 

Offline turbo!

  • Regular Contributor
  • *
  • Posts: 89
Re: LM399 based 10 V reference
« Reply #135 on: April 22, 2014, 06:18:39 pm »
@turbo

R43 and R44?

43 feeds the heating element. I don't see why its needed, but I think the more important function is to act as a heater itself to heat the copper trace. 
44 is 15kohm 0.5% 50ppm, then it connects to the other shunt. The big diode looking thing on stand offs is a Vishay 0.1% resistor. So, I think they're connected as a divider. The precision scale is functioning, so I'm not going to do too much autopsying. This is a 1980s design by the way.

Interestingly, you'll see the stand-offs for the big shunt uses intentionally thinned out pattern to connect to ground. I think this is to minimize thermal transfer from the trace.
« Last Edit: April 22, 2014, 06:22:21 pm by turbo! »
 

Offline Andreas

  • Super Contributor
  • ***
  • Posts: 2447
  • Country: de
Re: LM399 based 10 V reference
« Reply #136 on: April 22, 2014, 06:58:31 pm »
but I think the more important function is to act as a heater itself to heat the copper trace. 

Interestingly, you'll see the stand-offs for the big shunt uses intentionally thinned out pattern to connect to ground. I think this is to minimize thermal transfer from the trace.

Hello,

perhaps only to limit the inrush current at start up (see datasheet).
In steady state the 20mA * 10 Ohms = 0.2V  -> 4mW will not produce significant heat. (perhaps 2-4 degrees C)

Interestingly the 10 ohms will not significantly dampen down inrush current.
But the output voltage of LM399 has a relative large dependancy on heater voltage/current.
So perhaps the 10 Ohms might be a compensation for this.

Do you know how large the heater supply voltage is? is it +15/0V or +15/-15V?

The thinned out pattern is mainly due to wave soldering.

with best regards

Andreas
 

Offline CaptnYellowShirt

  • Frequent Contributor
  • **
  • Posts: 447
  • Country: us
  • Scooty Puff Jr.
Re: LM399 based 10 V reference
« Reply #137 on: April 22, 2014, 07:07:00 pm »

But the output voltage of LM399 has a relative large dependancy on heater voltage/current.



I've heard this before. Is there a mechanism that causes this?
 

Offline Andreas

  • Super Contributor
  • ***
  • Posts: 2447
  • Country: de
Re: LM399 based 10 V reference
« Reply #138 on: April 22, 2014, 07:18:02 pm »
I've heard this before.

Perhaps in one of my threads.
After finding out that the LM399 is sensitive to tilting, I decided to measure all effects that might have an influence.
The PSRR of the heater voltage was one of the measurements not mentioned in the datasheet.

For the reason I can only guess: perhaps imperfections in the temperature governor to increase stability?
Since there is no capacitor on the chip it can only be a p-governor.

With best regards

Andreas
 

Offline Mickle T.

  • Frequent Contributor
  • **
  • Posts: 342
  • Country: ru
Re: LM399 based 10 V reference
« Reply #139 on: April 22, 2014, 07:52:37 pm »
Some words about tilting.
One of my simple statistical voltage references with the Dewar flask isolation has no orientation effects at all, i.e. 0.0 ppm in any direction. LM399 heater current < 5 mA.
 

Offline Andreas

  • Super Contributor
  • ***
  • Posts: 2447
  • Country: de
Re: LM399 based 10 V reference
« Reply #140 on: April 22, 2014, 08:32:59 pm »
Hello Mickle,

you are always good for a surprise.  :-+

Your idea with the dewar is thermal isolation in perfection.
And 5 mA is really a very low value. Is it for 15 or 30V heater supply?

Is it possible to get more detail photos from the isolation construction?

What is the metal block on the photo? is it containing the voltage divider resistors? thermally stabilized?

with best regards

Andreas
« Last Edit: April 22, 2014, 08:35:59 pm by Andreas »
 

Offline casinada

  • Frequent Contributor
  • **
  • Posts: 596
  • Country: us
Re: LM399 based 10 V reference
« Reply #141 on: April 22, 2014, 09:27:46 pm »
Mickle,
Are the banana connectors stolen from old HP equipment? :)
 

Offline turbo!

  • Regular Contributor
  • *
  • Posts: 89
Re: LM399 based 10 V reference
« Reply #142 on: April 23, 2014, 12:41:08 am »


For those who asked about the circuit above used in an electronic balance.
It takes about 4-5 seconds to settle down to 6.863v.

It runs off of 24v DC from a linear power supply. The 10 ohm resistor feeds the heater and the 15k ohm resistor carries the zener current.

The heater current starts around 12mA and stabilizes to 10mA, so the total resistance is around 2.4k ohm. It's dissipating close to 1/4W which is more than enough to become noticeably warm.
The 15k ohm resistor had about 17v, and with 7v regulation, it adds up right. This comes out to only a tad over one mA of reverse current through zener. The surface temperature runs around 45°C after about half an hour with the cover on.
 

Offline CaptnYellowShirt

  • Frequent Contributor
  • **
  • Posts: 447
  • Country: us
  • Scooty Puff Jr.
Re: LM399 based 10 V reference
« Reply #143 on: April 23, 2014, 03:44:24 am »
I've heard this before.

Perhaps in one of my threads.



Probably, I've read it several places. Also I think DiligentMinds said Bob Dobkins (the designer) said it too during his interview.
 

Offline Mickle T.

  • Frequent Contributor
  • **
  • Posts: 342
  • Country: ru
Re: LM399 based 10 V reference
« Reply #144 on: April 23, 2014, 06:05:38 am »
Hello Andreas,

Stabilized 15V heater supply.
Aluminum block is filled with paraffin and contain a PTFE board with 14 TC-selected and aged hermetic foil resistors for stat. divider. Dividers tempco is ~ 0.04 ppm/C, so there is no thermal stabilization.
Isolation construction is very simple: only Dewar and huge foamed plug in it  :)

Mickle T.
 

Offline acbern

  • Frequent Contributor
  • **
  • Posts: 316
  • Country: de
Re: LM399 based 10 V reference
« Reply #145 on: April 23, 2014, 07:21:56 am »
can you comment on the paraffin. is this for moisture? your resistors are already hermetic. or any other (mechanical/thermal) reason?
 

Offline Mickle T.

  • Frequent Contributor
  • **
  • Posts: 342
  • Country: ru
Re: LM399 based 10 V reference
« Reply #146 on: April 23, 2014, 08:05:16 am »
Paraffin is only for temperature equalization. S5-61 foil resistors have a significant TC difference.
 

Offline Andreas

  • Super Contributor
  • ***
  • Posts: 2447
  • Country: de
Re: LM399 based 10 V reference
« Reply #147 on: April 23, 2014, 08:31:59 pm »
Hello Mickle,

on your cirquit diagram on volt nuts I have seen that you also used 10 Ohms resistors
for your 3*LM399 cirquit in series with the heater (on a 14.5V supply).
Now we have the weight scale also with the 10 Ohms (but with a 24V supply).

Is there a special reason for the 10 Ohms resistor in the heater line of the LM399?

With best regards

Andreas


 

Offline Mickle T.

  • Frequent Contributor
  • **
  • Posts: 342
  • Country: ru
Re: LM399 based 10 V reference
« Reply #148 on: April 24, 2014, 05:02:55 am »
Hello Andreas,

I think this is an accidental coincidence. There is no special reason for the 10 Ohms resistor.

Mickle T.
 

Offline macfly

  • Contributor
  • Posts: 30
  • Country: de
Re: LM399 based 10 V reference
« Reply #149 on: April 24, 2014, 10:44:03 am »
Paraffin is only for temperature equalization. S5-61 foil resistors have a significant TC difference.

Hello Mickle,

very smart, your design!
What are these 'S5-61' resistors?
Where is the manufacturer and what specs they have (TC, long-time stability)?

Regards,

Macfly
« Last Edit: April 24, 2014, 03:40:53 pm by macfly »
Genius is one percent inspiration, ninety-nine percent perspiration (Thomas Alva Edison 1903)
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf