Author Topic: The LTFLU (aka SZA263) reference zener diode circuit (Read 204209 times)

SilverSolder · « **Reply #400 on:** August 27, 2019, 06:21:49 pm »

The reason for asking is that there are several LTFLU based references here that could be fun to compare and characterize, but I am not sure how much the measurement bias currents would influence the devices or measurement process overall (when comparing via a KVD or other resistive device to null the voltages)?

dietert1 · « **Reply #401 on:** August 30, 2019, 08:27:39 am »

Our AD587 (Geller) reference drops about 10 ppm when connecting a 10 KOhm load. So 1 mA test current is high, but 1 uA should be OK unless one aims for nV.

Meanwhile i got some more LTFLUs and made a little test rig (example). It is about the minimum to put the zener and the transistor into operation. One can check the transistor beta, should be about 150.
One of the new chips came bad, with a short between collector and Gnd. Those two chips with strong negative TCs mentioned before have low beta of about 30. All others exhibit low TCs at about 40 to 45 uA and will probably make good references.

Regards, Dieter

MegaVolt · « **Reply #402 on:** September 05, 2019, 12:02:45 pm »

Quote from: Dr. Frank on February 19, 2015, 10:34:56 pm

Well, and JJ arrays on liq. Nitrogen temperature also have been demonstrated.. about 1e-8 uncertainty..See here:
http://juwel.fz-juelich.de:8080/dspace/bitstream/2128/2069/1/19406.pdf

Unfortunately, this link is not available. Maybe someone saved this file?

mycroft · « **Reply #403 on:** September 05, 2019, 12:43:30 pm »

I think the article is at http://juser.fz-juelich.de/record/26808/files/19406.pdf

Quote from: MegaVolt on September 05, 2019, 12:02:45 pm

Quote from: Dr. Frank on February 19, 2015, 10:34:56 pm
Well, and JJ arrays on liq. Nitrogen temperature also have been demonstrated.. about 1e-8 uncertainty..See here:
http://juwel.fz-juelich.de:8080/dspace/bitstream/2128/2069/1/19406.pdf
Unfortunately, this link is not available. Maybe someone saved this file?

dietert1 · « **Reply #404 on:** September 06, 2019, 05:59:57 pm »

Today i tried to run one of the LTFLU references at higher zener currents. I remember some statements that the LTFLU zener is similar to the LTZ1000 zener, except the LTFLU has four of them. The differential resistance measurements of about 7 to 8 Ohm for the LTFLU and 30 Ohm for the LTZ1000 point to the same direction. So i found it plausible to run the LTFLU at 12 mA instead of 3 mA. Of course the reference needed readjustment for output voltage and TC, but it works well. Have to compare noise yet.
Anybody tried this? What are the long term results?

Regards, Dieter

iMo · « **Reply #405 on:** September 06, 2019, 08:13:34 pm »

I've seen a lot of information on the LTFLU on the bbs.38hot.net site, but it is not accessible today..

Or lymex'es new site:

http://bbs.1ppm.cn

branadic · « **Reply #406 on:** November 11, 2019, 04:20:20 pm »

Received some LTFLU boards today, fabricated on aluminium. Need to solder a first board the next days.

-branadic-

branadic · « **Reply #407 on:** November 13, 2019, 08:35:29 pm »

Almost done with assembly of the first board. All parts are 0805 except the NTC, which is a 100k in 0402. We are getting closer to the first working unit

-branadic-

branadic · « **Reply #408 on:** November 25, 2019, 05:58:30 pm »

First temperature measurement performed on the aluminium core board, looks okay. Now I need to trimm the output voltage to 10,00000V and repeat the measurement, to find the final oven temperature setpoint.

-branadic-

dietert1 · « **Reply #409 on:** November 30, 2019, 09:10:03 am »

What is your experience with handsoldering wires and SMD tuning parts to that board? I guess the LTFLU is mounted as an SMD part.

Here an image of my latest creation: LTFU with higher currents, Nomca replaced by 10x UPF50B100RV, ADA4622-2 => OPA2189, npn transistor as buffer for reference voltage. The divider now has an output impedance of about 200 Ohms and i hope this will help. The transistor base current contributes roughly 1 ppm for every 10 Ohms of divider impedance. The LTFLU was mounted into the board, since i am using its transistor(s) as temperature sensor.

Regards, Dieter

branadic · « **Reply #410 on:** November 30, 2019, 10:18:43 am »

Hi Dieter,

my experience with SMD tuning parts are fine by now. Still I have to tune the output by changing one SMD resistors to what I think is 10,00000V as I want to avoid any pot. However, I'm still about 8ppm off, which in theory is okay, but I want to see the limits I can achieve. Tuning takes some time, as with every change in resistance I have to perform a new temperature profile to verify the result.
Also the wires are no problem by now as I currently run the profile using my temperature chamber and not with a heater resistor attached to the back, thus not the final oven. But I'm sure I have to use smaller wires in later assembly.

-branadic-

iMo · « **Reply #411 on:** November 30, 2019, 11:17:57 am »

Quote from: branadic on November 30, 2019, 10:18:43 am

Still I have to tune the output by changing one SMD resistors to what I think is 10,00000V as I want to avoid any pot.

How the TC of that tuning resistor translates into the total TC?

branadic · « **Reply #412 on:** November 30, 2019, 08:14:38 pm »

Don't know and I currently don't care about it.

-branadic-

Andreas · « **Reply #413 on:** November 30, 2019, 08:39:04 pm »

Hello Branadic,

thanks for the paper.

If you consider the LT1006 as output buffer for the LTFLU then another possibility would be also the OPA187.
This is the low power version of the OPA189 with comparable noise to the LT1006 but lower supply current as the OPA189/ADA4522.
So battery life could be increased.

For my AD587LW references I will check wether the OPA187 is suitable. (The AD587 has 2-3 uVpp of noise so the 0.4uVpp will not be mentioned but the 0.1 mA supply current will extend battery life by 20-30% compared to the LTC2057).

with best regards

Andreas

Kleinstein · « **Reply #414 on:** November 30, 2019, 09:18:58 pm »

The NOMCA resistor array may introduce quite some noise. There may be a slightly less noisy resistor configuration to build a 10 to 7 V divider.
The idea is to have a more even voltage across the resistors and thus better averaging of the 1/f excess noise. I would expect about a 25% reduction in excess noise - so maybe not worth a new board. Similar the matching should also be slightly better, as the weight of the resistors is more even.

In the shown diagram R8 is used for trimming to something a little less than 7 V. It's best use / value depends on the exact ratio needed.

I would be somewhat careful with an AZ OP in the circuit. There is no need for super low drift, as there is some extra gain (e.g. 50-100) from the transistor, that attenuates OP drift. However the extra spikes can be a problem. Different from the LTZ1000 circuit, there is no need for a single supply OP. I would more like consider an OPA196 for low current.

For an extra output buffer an AZ OP is OK.

branadic · « **Reply #415 on:** December 01, 2019, 04:45:11 pm »

Andreas,

already to late, the other boards are already assembled with LT1006. Since LT1006 is the single version of LT1013 and I have only one in the circuit current draw is already reduced.

Quote

Different from the LTZ1000 circuit, there is no need for a single supply OP. I would more like consider an OPA196 for low current.

I wanted a single supply opamp, since it simplifies battery power supply and charge circuitry.

-branadic-

branadic · « **Reply #416 on:** December 11, 2019, 06:26:25 pm »

I measured the LTFLU reference in my thermal chamber at constant temperature (48.2°C). However, I'm somewhat limited by my meter R6581D and its tempco of about 1ppm/K or at least by stability of ambient temperature, which changes internal temperature of the meter.
Still waiting for some resistors to further trimm the output voltage.

-branadic-

Noopy · « **Reply #417 on:** February 05, 2020, 07:25:34 pm »

Thanks to branadic

I was able to take some more pictures of his LTFLU:

https://www.richis-lab.de/REF04.htm

Quite interesting!
(I didn´t read the whole thread. Sorry...

)
Only very few transistors are connected to the circuit (tempco-mathing i assume)...
There seems to be a optional heater…

Nice big fusable links:

iMo · « **Reply #418 on:** February 05, 2020, 07:45:33 pm »

It would be great to find a final result based on your and mine attempts to draw a schematics of the chip - especially the pretty difficult block around the 4 zeners..

Noopy · « **Reply #419 on:** February 05, 2020, 07:49:57 pm »

Interesting…

Q9... Hm... The two transistors in the middle between the four zener?
In my view that transistors connect the "anode" or "z" with the substrate. Probably to give it some negative potential...

iMo · « **Reply #420 on:** February 05, 2020, 08:00:37 pm »

There are 2 resistors at each zener, one from the cathode to the emitters and the second one going somewhere - in my case to the base of the Q9.
The sheet resistance - my values are rather high, it will be much less, but the first step is to understand how the block around the zeners is actually wired

branadic · « **Reply #421 on:** February 05, 2020, 08:14:23 pm »

Richard, SZA263 and LTFLU are two different stories:

-branadic-

Noopy · « **Reply #422 on:** February 05, 2020, 08:15:25 pm »

I interpreted the structures this way:

From each zener there are four "resistor wires".
Every resistor is connected to the cathode.
Two parallel resistors are connected to the bondpad in the upper right corner.
One is connected to the emitter of the transistor batch. I assume it´s some current dividing.
The last one is a possible low impedance path instead of the resistor for dividing the current.

I see nothing more…

Quote from: branadic on February 05, 2020, 08:14:23 pm

Richard, SZA263 and LTFLU are two different stories:

-branadic-

Uuuuh, thanks for the hint... Sorry!

Edit: Corrected the mistake on my site…

branadic · « **Reply #423 on:** February 05, 2020, 08:52:27 pm »

Btw, one unit was marked as LTFLU-1CH while the other was marked as LTFLU-1ACH. I leave it up to you to find out what's the difference between both.

-branadic-

dietert1 · « **Reply #424 on:** February 06, 2020, 09:48:54 am »

Anyone with the capability to sample the zener voltages of an open LTFLU chip? I am wondering whether it makes sense to implement four seperate zeners on one die. This may be more or less useless because the zeners will be extremely similar.

Regards, Dieter


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

Author Topic: The LTFLU (aka SZA263) reference zener diode circuit (Read 204209 times)

Share me