Author Topic: Project KX : DIY calibrator / reference source/meter (Read 180715 times)

Dr. Frank · « **Reply #50 on:** March 27, 2014, 07:00:46 pm »

Quote from: TiN on March 27, 2014, 05:35:04 pm

I think around +55-60°C, as it can be +35°C here from time to time.

OK, well, and in your - mysterious - country, you don't have basements in your houses, usually? There the temperature would be stable to 21°C, presumably.

Ah well, I thought over your description, that your ref somehow takes a while to stabilize..

in temperature and in voltage output.. that sounds a little bit strange to me, but familiar in another aspect:

The reference should be spot on within a few seconds.
If I switch off my LTZ references, and switch them on again, they reach their original output within a few seconds. There might be a drift over the next few minutes of a few ppm, or a few tenths of a ppm, only.

But there is definitely no longer termed drift over many ppm after switch on...

Please describe in more detail... there might something be fishy.

By your description I was remembered to my own fault in the circuitry.. I used the original datasheet (version a), and reversed the polarity of the heater resistor of the LTZ, due to easier layout.

At that time, the parasitic diodes, which to-day were drawn in the datasheet, had not been published by LT.

The effect was, that the temperature regulation did not work correctly, and I also saw a creepy drift of the output. After asking LT, I corrected the circuit, and since then, it's quickly spot-on, as described.

Frank

bingo600 · « **Reply #51 on:** March 27, 2014, 07:01:36 pm »

What's the price of the boards ?

I'd like to join the LTZ club.

As posted here
https://www.eevblog.com/forum/projects/ultra-precision-reference-ltz1000/msg414172/#msg414172

Where are you located ?

/Bingo

Andreas · « **Reply #52 on:** March 27, 2014, 09:19:42 pm »

Quote from: Dr. Frank on March 27, 2014, 05:32:53 pm

I thought, the LTZ reference had a negative TC of around -50ppm/K. Therefore the voltage should decrease with higher temperature.

Hello Frank,

at least my two LTZ1000A have a TC of around +50 ppm /K. Voltage goes up by 3.7mV as temperature setpoint is increased by around 10 degrees. (from 50.5 to 60.1 deg nominal)

With best regards

Andreas

TiN · « **Reply #53 on:** March 28, 2014, 03:12:21 am »

Quote

What's the price of the boards ?

Bare boards are free, just a shipping cost (20USD will do fine for most locations).
I'm located in Taiwan, shipping by EMS (similar to USPS).

Quote

Please describe in more detail... there might something be fishy.

I think the caveat is due to resistors used with big tempco.
I don't know yet when I'll get proper resistors.

chickenHeadKnob · « **Reply #54 on:** March 28, 2014, 03:53:20 am »

Hello TiN

Your photography is a joy as usual. Noticed the boards appear to be ENIG finish. I was wondering about this, wouldn't HASL have lower tempco from seebeck? I am asking because I don't know. One of the main problems every experimenter faces in this area with an ambitous project like yours is too many options to explore in a proper scientific way.

TiN · « **Reply #55 on:** March 28, 2014, 11:47:00 am »

Got a setup for test, will see when it stabilizes. DAQ measure temperatures every 30 sec.

TiN · « **Reply #56 on:** March 28, 2014, 03:21:22 pm »

Regarding finish - it's <3 microinch thickness for gold and <160 microinch for nickel, while copper thickness is >40umm so I assume it does not matter.

Startup temperatures and heater voltage.

Voltage source for module: +15.00V 55mA Keithley 2400
Voltage output measurement: Keithley 2001 (cal'd by official Tek services in beginning of february 2014)
Power supply for sensors MAX6610: +5.00V 100mA max Agilent E3649A
DAQ system for datalogging: Agilent 34970A with mux card 34901A.

Thermal photos

Just powered on.. I put piece of matte thermal conductive pad to capture temperature of shiny LTZ can surface.

Around 1c variance, not bad. This is fully open-air without any thermal/airflow shielding, so when it's inclosed should be much more better.

Heater BJT and it's control opamp. Probably will add more separation to get rid of those 2C raise on that side of PCB.

Need better thermal coupling between opamp and MAX6610 sensor here.. Will probably have sensor MAX6610 placement really close to opamp, so it could touch it, and have thermal grease for better contact.
Overkill, but it's easy to do.

5 minutes more...

Can see that LED and it's resistor making a hot spot, so that will be removed in next board revision.

And bonus - PCB cross-section macro shot.

Close-up:

Normal macro:

CaptnYellowShirt · « **Reply #57 on:** March 28, 2014, 03:55:20 pm »

Ok... I'll say it... I'm jealous.

TiN · « **Reply #58 on:** March 28, 2014, 03:57:22 pm »

List of samples for boards is still available, lol.

TiN · « **Reply #59 on:** March 31, 2014, 10:32:57 am »

Working on labview app to get data off Keithley's..
Temperatures:

Got ADC/DACs as well.

TiN · « **Reply #60 on:** April 22, 2014, 10:57:28 am »

Btw, if anybody have question regarding Vishay PG resistors, shoot, i'll have VPG rep for tutor/demo here next week.

CaptnYellowShirt · « **Reply #61 on:** April 22, 2014, 03:52:28 pm »

Ask them if they do any type of die-level stress relief after the resistors are laser trimmed. In the LTZ1000 forum, we've been talking about strain effects on chips and its effect on drift.

TiN · « **Reply #62 on:** April 28, 2014, 02:47:03 pm »

Project hit a bump on a road, called long leadtime for precision resistors. But it's not dead yet, while I'm waiting on VPG order, managed to score some custom Fluke wirewounds, VPG VPR5's and VCS101's for current domain of project.

Also some time ago I did got current sensor, almighty zeroflux LEM Ultrastab 600S. It's the only current transducer with specced 1ppm linearity error and 0.2ppm/K tempco, with veeery wide range

As with most magnetic current transducers, performance likely to be be limited to burden resistors and attentuation circuitry on secondary.

Anyway, enough text, here are photos of goodies:

(Everything clickable as usual)

What is that?

Those are wire-wound Fluke resistors. They are huge.

This is 250Kohm, 2W 0.05% 5ppm/°C one.

See that tiny spot in center? That's generic SMD resistor 1206 390ohm.

These are constructed like an inductors, so no doubt they are inductive, but for DC circuitry that can be workarounded.

50K measurement:

Also scored Vishay PG VPR5 1.0000R's (7pcs)

Aren't they beautiful?

Good match to LEM's.. tiny thing in center - LTZ1000.

Quick check, source 10mA from Keithley 2400, measure voltage on sense terminals with 2002.

Also got some constantan (Copper+Nickel 44) wire for finetuning, made in 1988

.

If believe factory label, it's 0.07 mm diameter, 127 ohm per meter. It's coated wire for isolation.

So that's all for today:

Total list of parts: http://dev.xdevs.com/issues/1068

I'll meet VPG rep tomorrow, they promised some demo

Notice - if anyone decided up for assembled LTZ1000's VREFs - this week is last call, as I'm going to order LTZ's from Linear next monday. This is hobby project funded from own pocket, so there will be no excessive stock of expensive parts.

Vgkid · « **Reply #63 on:** April 28, 2014, 04:53:52 pm »

So it was you who snagged the lot of fluke wirewound resistors on ebay. I was watching those...

CaptnYellowShirt · « **Reply #64 on:** April 28, 2014, 05:21:05 pm »

Quote from: Vgkid on April 28, 2014, 04:53:52 pm

So it was you who snagged the lot of fluke wirewound resistors on ebay. I was watching those...

TiN snags ALL the good stuff!

TiN · « **Reply #65 on:** April 29, 2014, 05:06:37 am »

It's a big (maybe even too big) project which needs lots of supply.
If it will get nowhere, then someday all stuff will be for sale again

Also some wirewounds are damaged, so need some surgery with microscope.

TiN · « **Reply #66 on:** April 29, 2014, 04:56:27 pm »

Fixed up few resistors (broken wire due to mechanical damages).
Test of one 250Kohm after fixing:

Ambient temperature, +25C:

Submerge into LN2 (-196°C or around that)

KC0PPH · « **Reply #67 on:** April 29, 2014, 06:07:59 pm »

Reading your thread here is giving me some good entertainment. I see there is another DIY guy out there trying to make something on the ridiculous side of things. Im currently building a SMU. Plans are to have it be a 0-5V with resolution down to 1uA and 1uV. Hopefully it will evolve into something much better than that. I am waiting on a few parts to come in and will start building my reference PCB here in the next few weeks (Im using 4 LM399's). I was thinking about using some Chromel to wind my own resistors. Maybe i can find some other wire to build my resistors with (Not even sure the tempco of chromel).

Anyways Ill be following your thread (I wish I had KeithLey equipment like that -- odd thing is my name is Keith Lee)

Golana · « **Reply #68 on:** April 29, 2014, 06:18:45 pm »

Hi, KC0PPH. That sounds like an interesting project. I would definitely be interested in a DIY SMU if you feel like posting any details.

CaptnYellowShirt · « **Reply #69 on:** April 29, 2014, 06:29:48 pm »

I'd like to see the details too. Especially on the 4 V_ref setup.

KC0PPH · « **Reply #70 on:** April 30, 2014, 12:00:05 am »

Quote from: Golana on April 29, 2014, 06:18:45 pm

Hi, KC0PPH. That sounds like an interesting project. I would definitely be interested in a DIY SMU if you feel like posting any details.

When i start getting serious Ill make a thread for it. Today just perfected the Switching Pre-Regluator for the design. (Im going for a basic CV/CC PSU as the first step of the build +/-24V 3A) Ill write up a more formal definition of the plans...

Anyways inb4 hijacking this thread I will return it to the owner

TiN · « **Reply #71 on:** April 30, 2014, 10:01:37 am »

Okay, time to do some other stuff.

Anyone want to make a guess, what is this will be?

Back

Front

Side

PCB from heatsink side, heatsink hidden for clarity

PCB from back side, board hidden for clarity.

Black thing in center - cap. Black things around - SO8 packages

jeremy · « **Reply #72 on:** April 30, 2014, 11:25:29 am »

TiN: I noticed in your last few pictures that you were measuring resistance on an Agilent "switch unit". What did you have the resistor connected directly to?

TiN · « **Reply #73 on:** April 30, 2014, 01:15:48 pm »

It was just connected to 34901A mux card. All my Keithley gear is at home, at work I have only 34970

CaptnYellowShirt · « **Reply #74 on:** April 30, 2014, 02:25:25 pm »

The new project is a very un-user-friendly hand warmer to deal with all that LN2.

What's the LN2 used for at work?


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

Author Topic: Project KX : DIY calibrator / reference source/meter (Read 180715 times)

Share me