Author Topic: Ultra Precision Reference LTZ1000 (Read 1418874 times)

MK · « **Reply #925 on:** March 29, 2015, 08:34:54 am »

With a 7 volt source and the 70K resistor for R2, what is the room temperature superdiode that any of you get? i.e. gnd to pin 7, pins 4 and 5 connected to 70K and other end of 70 K connected to 7 or 10 volts? that way we can get the scaling of the transistor and get a better handle on the vbe vs temp. dont connect up the zener or heater for this and you can be certain it is at room temp.

janaf · « **Reply #926 on:** March 29, 2015, 10:48:33 am »

Quote from: MK on March 29, 2015, 08:34:54 am

With a 7 volt source and the 70K resistor for R2, what is the room temperature superdiode that any of you get? i.e. gnd to pin 7, pins 4 and 5 connected to 70K and other end of 70 K connected to 7 or 10 volts? that way we can get the scaling of the transistor and get a better handle on the vbe vs temp. dont connect up the zener or heater for this and you can be certain it is at room temp.

Please re-formulate a bit clearer. Do you mean the voltage over the diode at room temperature? It's current dependent. If you search my previous posts in this thread, you will find some plots. Anyway, i'ts around 6.6V for the zener only at typical 5mA.

MK · « **Reply #927 on:** March 29, 2015, 03:23:16 pm »

Quote from: janaf on March 29, 2015, 10:48:33 am

Quote from: MK on March 29, 2015, 08:34:54 am
With a 7 volt source and the 70K resistor for R2, what is the room temperature superdiode that any of you get? i.e. gnd to pin 7, pins 4 and 5 connected to 70K and other end of 70 K connected to 7 or 10 volts? that way we can get the scaling of the transistor and get a better handle on the vbe vs temp. dont connect up the zener or heater for this and you can be certain it is at room temp.
Please re-formulate a bit clearer. Do you mean the voltage over the diode at room temperature? It's current dependent. If you search my previous posts in this thread, you will find some plots. Anyway, i'ts around 6.6V for the zener only at typical 5mA.

I mean, just the vbe of the pin4 pin5 pin7 transistor at about 10 uA and at room temperarure. Pin 4 and pin 5 connevcted together and dont connect up the zener or heater, I thought that was quite clear.

http://www.ti.com/ww/en/bobpease/assets/www-national-com_rap.pdf
go to page 115 and you will then understand why I am asking this.

janaf · « **Reply #928 on:** March 29, 2015, 04:48:23 pm »

OK, get it! And when Bob Pease is involved, it always get interesting!

I have just gone by the datasheet "very predictable 2mV/K" and experimentally in the full circuit; put a trimmer on the temperature adjustment divider, turn down until the output voltage gets unstable, then turn it up again by suitable margin.

The datasheet states that the CE voltage is uncertain to within 10C.
But I will surely do the measurement, maybe later tonight.

MK · « **Reply #929 on:** March 29, 2015, 05:28:08 pm »

with that single measurement and Bob Pease's graph you can predict the temperature to about +-1 degree at a current of your choice. That will also help you fine tune the current through the zener with any particular R1 value and allowing for the temperature rise that imposes on you.

janaf · « **Reply #930 on:** March 29, 2015, 06:11:55 pm »

Measurements attached. Pick your current. Note the slight temperature drift.

(MOD updated attached file)

Jan

janaf · « **Reply #931 on:** March 29, 2015, 06:27:32 pm »

Looking at the data, the difference between the two measurement sets is in the order of 0.01 volt which would be equivalent of 5C

Can't explain that other than "something" was moving, maybe meter, actual devise temp (handled and moved it before measuring...). With the very low current (mod): and 1mS measurement time (/mod), it could hardly be self heating? I also waited something like half a minute between each data point.

I can leave it for a while to stabilize and try again...

Andreas · « **Reply #932 on:** March 29, 2015, 06:41:24 pm »

Hello Jan,

I think that the x-axis in the diagram (current) is scaled by a factor 10 too low.

And a big

for doing all the measurements.

With best regards

Andreas

janaf · « **Reply #933 on:** March 29, 2015, 07:08:59 pm »

Hi, Just realized. The scale is right, the measurements wrong. Should be in the order of 0.1mA, not 0.01mA. New results coming....

MK · « **Reply #934 on:** March 29, 2015, 07:49:05 pm »

Quote from: janaf on March 29, 2015, 07:08:59 pm

Hi, Just realized. The scale is right, the measurements wrong. Should be in the order of 0.1mA, not 0.01mA. New results coming....

I believe that 6.8 Volts and 70K is approx 10uA, and that Bob Deakin has recommended running those transistors at less than 50uA. They are not designed to cope with the full zener current ever.

Also for 10uA I would have thought you might want a longer measurement to allow it to settle. 50 micro watts is not going to warm the device, to any degree that you could measure.

janaf · « **Reply #935 on:** March 29, 2015, 07:59:23 pm »

Fooled me too but ; 7V/70K = 1/10000A = 0.1mA = 100uA

I'll upload a new file now. The measurements are now very much more consistent. Just remember that the datasheet specs are "within 10C" for this.

Dr. Frank · « **Reply #936 on:** March 29, 2015, 08:16:09 pm »

Quote from: janaf on March 29, 2015, 07:59:23 pm

Fooled me too but ; 7V/70K = 1/10000A = 0.1mA = 100uA

I'll upload a new file now. The measurements are now very much more consistent. Just remember that the datasheet specs are "within 10C" for this.

Jan, that's great!

Would you mind uploading the original xls file, instead of that crappy pdf assembly?

THX

Frank

janaf · « **Reply #937 on:** March 29, 2015, 08:47:58 pm »

Sorry Frank.

I uploaded the correct PDF now, the messy one should never be there

I'm keeping the spreadsheet for now, it's too messy.

Sometime I'll summarize it all in a neat format, graphs and data.

MK · « **Reply #938 on:** March 29, 2015, 09:30:49 pm »

Quote from: janaf on March 29, 2015, 07:59:23 pm

Fooled me too but ; 7V/70K = 1/10000A = 0.1mA = 100uA

I'll upload a new file now. The measurements are now very much more consistent. Just remember that the datasheet specs are "within 10C" for this.

Ooops, should have looked closely at my spreadsheet, I had an extra 0 for the resistance....
Also cannot see the pdf now?

janaf · « **Reply #939 on:** March 30, 2015, 08:47:09 am »

Hi,

The data is attached to this post: https://www.eevblog.com/forum/projects/ultra-precision-reference-ltz1000/msg640218/#msg640218

TiN · « **Reply #940 on:** March 30, 2015, 05:32:10 pm »

Hi all.

Sorry to interrupt, but can anyone highlight usual tempco of LTZ reference? I am testing my K2001 and comparing LTZ direct ref to EDC MV106 DC box, and getting little misleading data, to which it seem that LTZ have ~1ppm/°C sensetivity, which is kinda high. I suspect it's more of a drift of 2001 itself, rather than reference, but
will have time only next week to check with thermostat box. Resistors are VPG foils, which specced below 2ppm/C.

Andreas · « **Reply #941 on:** March 30, 2015, 06:55:54 pm »

Quote from: TiN on March 30, 2015, 05:32:10 pm

Hi all.

Sorry to interrupt, but can anyone highlight usual tempco of LTZ reference? I am testing my K2001 and comparing LTZ direct ref to EDC MV106 DC box, and getting little misleading data, to which it seem that LTZ have ~1ppm/°C sensetivity, which is kinda high. I suspect it's more of a drift of 2001 itself, rather than reference, but
will have time only next week to check with thermostat box. Resistors are VPG foils, which specced below 2ppm/C.

Hello,

thats what I have on my 2 references:
With respect to 25 deg change in PCB temperature (or 33 deg C environment).

LTZ1000A #1 around 4.5uV (* 2) = -9uV change / 7V = 1.3 ppm / 25K = -0.05 ppm/K
LTZ1000A #2 around 5uV (* 2) = 10 uV change / 7V = 1.4 ppm / 25K = 0.06 ppm/K

all with UPW50 resistors specced with 3 ppm/K (not tested)

Edit:
Y-Axis Output voltage in mV (after 2:1 precision divider LTC1043 based)
X-Axis temperature in deg C

red: ltz output voltage
pink: averaged value (noise reduced)
blue: temperature sensor on PCB near LTZ1000
green: (right Y-axis) temperature sensor outside the aluminium box (environment).

With best regards

Andreas

janaf · « **Reply #942 on:** March 30, 2015, 09:59:29 pm »

Nice!

So both boards increased output with temperature? Can't quite interpret the scales of the plot.

janaf · « **Reply #943 on:** March 30, 2015, 10:11:47 pm »

Today's measurement: Temperature versus power (dissipation) in the LTZ1000ACH.

On an IC hooked up with mini-hooks
Only Q1 with 0.1mA (0.05mW) and the heater where connected.
The temperature was calculated as 0.002V/C, starting from room temperature.
No insulation or soldering, or PCB. Just a sheet of paper over, to stop wind

RESULT: Surprisingly linear, around 0.36C/mW (360C/W) and quite near the 400C/W stated in the datasheet.
Note that the zener itself dissipates about 6.5V*5mA = 33mW, which could cause a temperature rise to around 35C without any heater. In reality, the self-heating would be lower, when soldered to a PCB.

In the attached doc there is one plot for a narrow range around 30-40C and another over a wider range.

Andreas · « **Reply #944 on:** March 31, 2015, 05:05:39 am »

Quote from: janaf on March 30, 2015, 09:59:29 pm

Nice!

So both boards increased output with temperature? Can't quite interpret the scales of the plot.

Sorry
as common with my measurements:
Y-Axis Output voltage in mV (after 2:1 precision divider LTC1043 based)
X-Axis temperature in deg C

red: ltz output voltage
pink: averaged value (noise reduced)
blue: temperature sensor on PCB near LTZ1000
green: (right Y-axis) temperature sensor outside the aluminium box (environment).

so LTZ #1 has a negative T.C.
and LTZ #2 has a positive T.C.

With best regards

Andreas

macfly · « **Reply #945 on:** March 31, 2015, 07:15:27 am »

Quote from: janaf on March 30, 2015, 10:11:47 pm

....
RESULT: Surprisingly linear, around 0.36C/mW (360C/W) and quite near the 400C/W stated in the datasheet.
Note that the zener itself dissipates about 6.5V*5mA = 33mW, which could cause a temperature rise to around 35C without any heater. In reality, the self-heating would be lower, when soldered to a PCB.
...

Hi janaf,

35 degree's ? I can't retrace this:

dT = Pv x Rth = 0.033W * 400K/W = 13.2 K

Regards,

macfly

janaf · « **Reply #946 on:** March 31, 2015, 10:19:18 am »

macfly: 22.7C surrounding temp plus dT 13.2K temperature rise = 35C (approx) Slightly less with my 360C/W

EDIT: This is an indication on why temperature setting below R4/R5 of 1:12 (nominally near 40C) gets difficult.

macfly · « **Reply #947 on:** March 31, 2015, 04:27:33 pm »

Quote from: janaf on March 31, 2015, 10:19:18 am

macfly: 22.7C surrounding temp plus dT 13.2K temperature rise = 35C (approx) Slightly less with my 360C/W

EDIT: This is an indication on why temperature setting below R4/R5 of 1:12 (nominally near 40C) gets difficult.

Sorry, I overlooked the '... to around'

Regards,

macfly

janaf · « **Reply #948 on:** April 02, 2015, 07:52:09 am »

Edit: (SOLVED)
Does someone have a spare LTZ1000CH ie the "non-A" version to sell me? It can be used, if there are no new ones around, as long as it's fully OK. I'd like to do some measurements on it. Just buying a single one from LT will cost me around $100 incl shipping, a bit steep....

TiN · « **Reply #949 on:** April 04, 2015, 08:09:43 pm »

Hi janaf,

I might order some stuff from LT next month (may), including few LTZs, i could add one more for you to save shipping cost for you, if you are not in rush.


EEVblog Main Site	EEVblog on Youtube	EEVblog on Twitter	EEVblog on Facebook	EEVblog on Odysee

Author Topic: Ultra Precision Reference LTZ1000 (Read 1418874 times)

Share me