Author Topic: Can an LM399 10v reference run too hot in an insulated case? [ANSWERED]  (Read 7305 times)

0 Members and 1 Guest are viewing this topic.

Offline enut11Topic starter

  • Frequent Contributor
  • **
  • Posts: 933
  • Country: au
  • Love building/modifying/restoring test equipment
Hi
I have just completed two LM399 10v refs on PCBs and they have been running non stop since before Christmas to condition the components. I now want to mount them in boxes with proper power-in and output terminals.

I have a couple of small boxes that I want to house the references in:

1) Plastic box 100x60x25mm
2) Metal box 110x60x30mm

Question is, will the LM399 overheat in a small fully insulated box?
enut11
« Last Edit: January 24, 2017, 02:14:10 am by enut11 »
an electronics nut from wayback...
 

Offline RandallMcRee

  • Frequent Contributor
  • **
  • Posts: 540
  • Country: us
Re: Can an LM399 10v reference run too hot in an insulated case?
« Reply #1 on: January 22, 2017, 05:12:02 am »
No, because its always at 70-80 degrees centigrade.

The LM399's heater has a temperature controller, that, umm, controls the temperature.

So, even in a vacuum flask, that temperature should be maintained. It might decrease the heater current consumption, though.

Randy
 
The following users thanked this post: enut11

Offline enut11Topic starter

  • Frequent Contributor
  • **
  • Posts: 933
  • Country: au
  • Love building/modifying/restoring test equipment
Re: Can an LM399 10v reference run too hot in an insulated case?
« Reply #2 on: January 22, 2017, 05:32:31 am »
Thanks Randy. This is handy to know because I wanted to use a small box suitable for shipping.

Another question, would a sealed box be an advantage for this reference?
enut11
an electronics nut from wayback...
 

Offline enut11Topic starter

  • Frequent Contributor
  • **
  • Posts: 933
  • Country: au
  • Love building/modifying/restoring test equipment
Re: Can an LM399 10v reference run too hot in an insulated case?
« Reply #3 on: January 22, 2017, 09:48:13 am »
LM399 10v ref mounted in a small plastic case with insulation to fill space. Current dropped from 32mA to 22mA compared to unboxed PCB, . The components have been powered up for about 50 days.

Output buffer is an LT1001. Input is reverse voltage protected. Will now let it run overnight with a temperature sensor to see if there is any drift in output.
enut11
an electronics nut from wayback...
 
The following users thanked this post: cellularmitosis

Online Kleinstein

  • Super Contributor
  • ***
  • Posts: 14070
  • Country: de
Re: Can an LM399 10v reference run too hot in an insulated case?
« Reply #4 on: January 22, 2017, 09:59:55 am »
If extremely well insulated it can run too hot, but this is very hard to achieve. If hot the current consumption will go down to the around 2 mA range.

Depending one the thermal design, the resistors and OP will also get hot. To a certain degree (e.g. 40-50 C) this can be good, as it also stabilizes that temperature a little and keeps relative humidity low, but it also accelerates aging of the resistors.

A closed case is good, as this prevents variable air flow and thus variable temperature profiles. A temperature sensor for the inside is a good idea, even if just for curiosity.
 
The following users thanked this post: enut11

Offline Andreas

  • Super Contributor
  • ***
  • Posts: 3221
  • Country: de
Re: Can an LM399 10v reference run too hot in an insulated case?
« Reply #5 on: January 22, 2017, 10:28:31 am »
Current dropped from 32mA to 22mA compared to unboxed PCB, .
Hello,

that is ok. (Current consumption will further drop down by 1-2 mA if you put the housing upside down).
But take care: output voltage will also vary in the low ppm range.

By the way: from where (+part number) did you get the housing and the dual connector?

with best regards

Andreas

Edit: I think the LM399 is already up side down within the housing.
So it has already lowest power consumption.
« Last Edit: January 22, 2017, 10:31:59 am by Andreas »
 

Offline HighVoltage

  • Super Contributor
  • ***
  • Posts: 5453
  • Country: de
Re: Can an LM399 10v reference run too hot in an insulated case?
« Reply #6 on: January 22, 2017, 11:15:59 am »

By the way: from where (+part number) did you get the housing and the dual connector?
with best regards
Andreas

I have used the same or similar before and think it is Keystone Electronics, Part Number 7019
But the metal parts are brass.

There are 3 kinds of people in this world, those who can count and those who can not.
 

Offline enut11Topic starter

  • Frequent Contributor
  • **
  • Posts: 933
  • Country: au
  • Love building/modifying/restoring test equipment
Re: Can an LM399 10v reference run too hot in an insulated case?
« Reply #7 on: January 22, 2017, 11:20:54 am »
Thanks Andreas

Dual connector here 19mm spacing:
http://www.ebay.com.au/itm/122168561468?_trksid=p2057872.m2749.l2649&ssPageName=STRK%3AMEBIDX%3AIT
Looks very similar to the one used in Calibratory 105D precision reference

Plastic box here:
http://www.ebay.com.au/itm/291918025308?_trksid=p2057872.m2749.l2649&ssPageName=STRK%3AMEBIDX%3AIT

enut11
an electronics nut from wayback...
 
The following users thanked this post: Andreas

Offline enut11Topic starter

  • Frequent Contributor
  • **
  • Posts: 933
  • Country: au
  • Love building/modifying/restoring test equipment
Re: Can an LM399 10v reference run too hot in an insulated case?
« Reply #8 on: January 22, 2017, 08:43:17 pm »

By the way: from where (+part number) did you get the housing and the dual connector?
with best regards
Andreas

I have used the same or similar before and think it is Keystone Electronics, Part Number 7019
But the metal parts are brass.

Thanks for the connector drawing HighVoltage.
Yes they are made of gold coated brass. BTW, I believe a better way to assemble the dual 4mm connector is to locate the flat washer at the base of the red and black hex heads. This provides a proper metal surface for clamping.
enut11
an electronics nut from wayback...
 

Offline enut11Topic starter

  • Frequent Contributor
  • **
  • Posts: 933
  • Country: au
  • Love building/modifying/restoring test equipment
Re: Can an LM399 10v reference run too hot in an insulated case?
« Reply #9 on: January 23, 2017, 02:22:27 am »
I can only give a brief report on performance until I get my graphing system working.

The 10v ref drifted from a setting of 10.00000v to 10.000009v while it was warming up overnight. In the morning I reset the output to 10.00000v and monitored the output intermittently during the day. From 9am to 1pm the output sat on 9.99999v. Room temp was 27C, RH65%. Temperature probe very close to the 399 chip was reading 46C. Current from the 15v supply was a steady 22mA. Temperature on the surface of the black plastic box was 30C. Voltage measurements were taken with my HP3456A.

So it appears stable now but more will be known when I can graph the output. The next test will be to find out how long it takes to settle from cold.
enut11

Edit: It has now been 12 hours at running temperature since I adjusted the output to 10v and the HP3456A is still reading 9.99999v
« Last Edit: January 23, 2017, 09:37:34 am by enut11 »
an electronics nut from wayback...
 

Offline enut11Topic starter

  • Frequent Contributor
  • **
  • Posts: 933
  • Country: au
  • Love building/modifying/restoring test equipment
Re: Can an LM399 10v reference run too hot in an insulated case?
« Reply #10 on: January 24, 2017, 02:12:57 am »
A forum member asked by PM about the design for this project. It is based on the Portable Calibrator circuit from the NS Application notes, page 9. The layout was done by t2kv who also included a mosfet to protect against reverse voltage input. He included a buffered 10v output and an unbuffered Vz out. We found the 5Kohm pot a bit fiddly to set the output so a revised circuit was done to include two pots (3k + 100 ohms).
In the version posted here I used a fixed resister in series with the 100 ohm pot. As every LM399 has a different Vz, the fixed resistor needs to be selected by trial and error.
enut11
an electronics nut from wayback...
 
The following users thanked this post: mycroft

Offline blackdog

  • Frequent Contributor
  • **
  • Posts: 736
  • Country: nl
  • Please stop pushing bullshit...
Re: Can an LM399 10v reference run too hot in an insulated case? [ANSWERED]
« Reply #11 on: January 26, 2017, 12:58:09 pm »
Hi enut11,

Do not forget that a big part of the drift is dependend on the resistors R3, R4 and your trimpot.
You need minimal resistors of 15PPM/C and mounted close thogetter.
And then you wil also have te lang time drift of these resistors...

Think about the position of the LM399 en de resistors R3, R4 and the trimpot.
You do not want the LM399 to heat resistor R3 more then R4.

My 2 cents  :)

Kind regarts,
Blackdog
Necessity is not an established fact, but an interpretation.
 

Online Kleinstein

  • Super Contributor
  • ***
  • Posts: 14070
  • Country: de
Re: Can an LM399 10v reference run too hot in an insulated case? [ANSWERED]
« Reply #12 on: January 26, 2017, 05:11:57 pm »
The same temperature for R3 and R4 has a limited effect: the TC can be different for these two resistors unless you get a matched pair. At least in part they might follow a similar nonlinear R(T) curve if they are made from the same material.
Having the whole board slightly hot from the LM399 heater already slightly dampens temperature changes, as the temperature of the resistors might be at about half way between ambient and the stabilized LM399 internal.
 

Offline blackdog

  • Frequent Contributor
  • **
  • Posts: 736
  • Country: nl
  • Please stop pushing bullshit...
Re: Can an LM399 10v reference run too hot in an insulated case? [ANSWERED]
« Reply #13 on: January 26, 2017, 05:51:55 pm »
Hi Kleinstein,  :)

I wish the topic starter only to point out to what is importend to obtain good stability.
And even with the same two resistors values, you can have different drift behavior, life of a electronic designer is a bitch  :-DD
There are no resistors being sold, to make a 10V reference of a LM399, we know why.

It's really hard to make a PCB with a LM399 reference on it, to properly characterize it, sometimes you are just lucky  :D

Kind regarts,
Blackdog
Necessity is not an established fact, but an interpretation.
 

Offline MisterDiodes

  • Frequent Contributor
  • **
  • Posts: 457
  • Country: us
Re: Can an LM399 10v reference run too hot in an insulated case? [ANSWERED]
« Reply #14 on: January 26, 2017, 11:46:03 pm »
Hi Kleinstein,  :)

...There are no resistors being sold, to make a 10V reference of a LM399, we know why.


Kind regarts,
Blackdog

Talk to Edwin Pettis here on the forum.  He'll make you some resistors of whatever value you need, to whatever practical tolerance you need - as a good PWW, and good ratio TC.

A lot of high-end production Vrefs are made with custom resistors anyway - sometimes that's the only best & economical way.

Typically you burn in and measure your basic Ref first, then calc your custom resistor values to trim your output right where you need it.  A good WW pot for fine-tune can be used for final polish of output value, if you need to correct for inevitable drift now and then.  Quality WW pots are fine and are rock-solid for decades if you don't have a lot of vibration - and you stay away from the cheaper, noisy junky ones (cermet for instance). 
 

Offline enut11Topic starter

  • Frequent Contributor
  • **
  • Posts: 933
  • Country: au
  • Love building/modifying/restoring test equipment
Re: Can an LM399 10v reference run too hot in an insulated case? [ANSWERED]
« Reply #15 on: January 27, 2017, 11:05:17 pm »
A bit more experimenting. Andreas has pointed out more than once the importance of chip orientation for an LM399 voltage reference.

I ran the reference above with the LM399 upright as well as upside down. The difference was very noticeable.

When the LM399 was upright it settled in to 10v after about 40 min and more or less stayed there.

When the box was oriented so that the LM399 was upside down, it settled closer to 9.99995v, ie 50uV lower, and never got back to 10v.
enut11
an electronics nut from wayback...
 

Offline Theboel

  • Frequent Contributor
  • **
  • Posts: 278
  • Country: id
Re: Can an LM399 10v reference run too hot in an insulated case? [ANSWERED]
« Reply #16 on: January 28, 2017, 12:44:12 am »
Hi Enut11 and Andreas,
just curious about this "orientation effect". are Your LM399 has a gap with the PCB ?
I can not find this difference with my LM399 maybe because I made a thick teflon (5mm) insulator between LM399 and PCB so LM399 leg do not expose to air movement 
 
The following users thanked this post: enut11

Offline Andreas

  • Super Contributor
  • ***
  • Posts: 3221
  • Country: de
Re: Can an LM399 10v reference run too hot in an insulated case? [ANSWERED]
« Reply #17 on: January 28, 2017, 10:35:29 am »
Hello,

a good thermal insulation around the LM399 and the pins (from both sides of the PCB) will reduce the zener voltage change when tilting. (a large amount of heat is dissipated by the pins).
The amount of the effect also varies from device to device.

Perhaps you have one of the LM399 near the sweet spot (6.875 V).

With best regards

Andreas
 
The following users thanked this post: enut11

Offline enut11Topic starter

  • Frequent Contributor
  • **
  • Posts: 933
  • Country: au
  • Love building/modifying/restoring test equipment
Re: Can an LM399 10v reference run too hot in an insulated case? [ANSWERED]
« Reply #18 on: January 28, 2017, 11:08:07 am »
Hi Enut11 and Andreas,
just curious about this "orientation effect". are Your LM399 has a gap with the PCB ?
I can not find this difference with my LM399 maybe because I made a thick teflon (5mm) insulator between LM399 and PCB so LM399 leg do not expose to air movement

Hi Theboel. My 399 was mounted about 5mm above the PCB. As you can see from Reply#3 (photo P1230610.jpg) the PCB was insulated on all sides so any circulating currents would be confined to a very small volume. The box lid was on during all tests.

In the case of the 'upside down test', it took over 10 hours for the voltage to get near 10.00000v. Compared with the 'upright test' where 10.00000v was hit after 40 minutes.
enut11
an electronics nut from wayback...
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf