Project KX : DIY calibrator / reference source/meter

[Vgkid]

What type of coax is that?
Looking forward to the video.

[TiN]

It's the one which was in K7172.

[TiN]

Time for tiny bit of progress...
Got front panel from CNC tonight. Just a window for a screen and 8 holes for binding posts for now. I did not make up my mind about keyboard layout and buttons type, so will mill that later.



Rear view on the panel. Screws are temporary too, they will be hidden under face mylar in final assembly.



I put some soft tape around VFD opening to avoid glass damage and chips.



Comparison to K2000 size. Case looks big, but after installation of pair transformers from Fluke 332 and thermal chamber for references there will be not that much space left.

I need to find screen panel for VFD, you know those dark-blue semi-transparent plastic sheets, to make VFD contrast good. I don't know correct english name for it .

[ManateeMafia]

The panel looks good, plenty of room for buttons... How are you getting the mylar panel overlay made? Print it yourself or have it custom made? I think there are supplies where you can print it yourself and use a clear protective overlay.

The window should probably be made of polycarbonate. I quick search found a 12" sq. sheet 1/8" thick for less than $10. I need to replace the one on my 9520 and I might go with a different color too.

[TiN]

It's long way till panel design. I'll use it as is, with remote control over UART. Reason why I wanted screen first, so I can have realtime logs displayed during prototyping.
As of windows, yea, polycarbonate likely will do but I think they have special colors for VFDs, not just blue? 

[ManateeMafia]

I didn't realize the VFD displays were using special filters like these http://noritake-vfd.com/cu20045-uw5j.aspx

[Astrid]

For interest. My solution reference with  LTZ1000A.
Goals
- only two-layer
- as much as possible to eliminate extremely expensive resistor
- in addition to the primary reference low noise output Vref
- processor for  temperature monitoring



Top right corner. Low noise applifier/atteunator with noise <10nV/Sqrt{Hz, for me set output voltage on Vref +5V
Bottom STM32F0 MCU for serial output temperature of LT1000A and Resisotr array and board.
R1 is resistor array (reralized in oil filled metal DIL14 ( primary uset for crystal oscilator),
four resistors are composed of the resistors 11pcs 0805 metall foil 0.1%, TCC 5ppm.
The resulting accuracy is better than 0.001%, resistors have a very narrow thermal bonding.
What the results would be seen :-).

[TiN]

Nice looking board, interesting oil resistor idea.
Perhaps we can peek on schematics?
U2,U3 are thermal sensors? I'd try to more decouple them thermally from surroundings.
Also that thick connection on the left of LTZ area will suck lot of thermal mass.

Do you plan running whole board enclosed in hermetic box? That was idea behind my size/positioning of elements, as well as use of 4 layers (extra copper on inner layers to add more mass around the board).
Also what are foil resistors on top (R25 and above it)? Divider for 5V output ratio?

U7 is LDO? I'd move it out of the board, and provide already regulated separate voltage input to MCU. This to avoid heatsource nearby.

Be careful with SMAs, not all of them have copper contacts/pin. Hope you use differential output (as outer casing of SMA is usually brass).

[TiN]

Few folks were asking me for LM399 version reference boards, but I have no extras left.
So I made some tweaks for PCB and added updated Gerbers in case one interested.



All data is in article with details.

[TiN]

Pimped up my resistance box a little (added 100 ohm and 20k wirewounds) before calibration. 
Also replaced that 10Meg contraption with USF340 from Caddock.

Final looks:



Side closeups:



Will see how she works to transfer resistance calibration from various meters.

[Berni]

Just came across this project and i must say it looks pretty impressive.

Not the sort of project id expect to see from someone who already has a bunch of nice Keithley gear. But i do love the monster VFD display on there.

[TiN]

Toolkit update, so I don't feel too bad about posting high resolution numbers anymore.



Both 2002's are very happy now, and become sealed boxes for time being. Sorry for toothpics in second meter, I'm sure someone would understand pain of having shrouded banana jacks on 8.5-digit DMM. 

Also got some more upcoming things and bits.



Another version of basically same thing:



Any guesses what these are? 

Oh, in case someone missed my post in LM399 thread, here's updated dual reference 20V low-cost ref:

[Dr. Frank]

Toolkit update, so I don't feel too bad about posting high resolution numbers anymore.

Both 2002's are very happy now, and become sealed boxes for time being. Sorry for toothpics in second meter, I'm sure someone would understand pain of having shrouded banana jacks on 8.5-digit DMM. 

Also got some more upcoming things and bits.

Any guesses what these are? 

So you finally are able to get SI-Volt for your LTZ1000 assemblies, with < 1ppm uncertainty, if both 2002 agree, and your lab is good for that level of uncertainty??
Did you also let one of your LTZs be calibrated?


These tiny PCBs look like either small SMPS in TO220 shape, like LM78xx, or small references.
This 10pin IC indicates, that it's a regulator, though.

Frank

[chickenHeadKnob]

I don't like these quiz type questions usually, too much cognitive work. Because  I am like "just tell me already" before I die. I have reached an age where I don't even risk buying green bananas.

At first I thought those little boards could be some kind of bolt on temp sensor, but with a to-220 profile and the silk screen I am sure Dr. Frank is right. He ist the winner-dinner-chicken-sinner. So what is his reward?

[TiN]

chickenHeadKnob
Sorry to hear, but whole life is a continuous learning process. You welcome to skip on, if it disturbs too much.

Dr. Frank
I'm still processing all data and calculating uncertainties. I'll post more on this once have data refined.
All my available LTZ modules were tested too, including 4 chips in socket board I shown before, as well all fancy resistors I have.
I have good confidence now that I'm within 1ppm abs on DCV, and around 3ppm on resistance.

It's in between of your assumptions for modules.



These are little ultralow noise LDOs based on LT3042, which have lowest noise I'm aware of LDO, specified to 0.8?VRMS (10Hz to 100kHz). They are available only in MSOP and DFN packages, so not quiet easy to prototype for our usual analog tinkering. Hence to "adapter" to TO220. Setting of output voltage is done via single resistor, with range up to 15VDC. Output current is 200mA.

Bigger "TO247" one is same thing, but with NPN pass transistor to raise max output current to 1A. It's based on schematics on page 26 of LT's datasheet.

As for reward, Dr.Frank can have two assembled units of both type, if he desire and send me his shipping information via PM, for his enjoyment

[TiN]

Same thing with pass transistor, assembed for test.

[Macbeth]

What a nice idea with those LDO's. I see you left us guessing what pass transistor you are using? 

Why the respin of the LM399 board? Are you thinking of selling "low cost" references calibrated to the TiN Labs Volt?

[TiN]

Last few days I needed some AC high (200ish) voltages for testing, but I don't have suitable source for required frequency. Of course, easy choice is to buy Fluke 5700 and be done with it, but that's not the jedi way.
All I need from signal is to be stable for 1 minute and to be 30kHz sine.

So I decided to give a try to LT1210 cascode amp with FETs to see how feasible it is.



Used SMT breadboard to assemble the thing, powered with +20/-20V for initial checks.
Resistors - metal film 25/50ppm/K.

Contraption photo:





Now I can see that resistors will dissipate lots of power, so it's not that precision grade source
Sine from HP3245A at 1.365VAC 30kHz gave 10.007V output, and during this power consumption from PSUs was 31mA + 30.7mA
Load for this amp is megaohms, so I don't need amps of output power either.

Ideas?

[2N3055]

How about putting together a pretty much an few watt audio amp, step up transformer, and levelling feedback from the secondary...

I'm pretty much certain i had a schematic for something like this.. I will try to find it...

Since this is sine wave , levelling loop doesn't even have to be RMS, just peek value, and on a single freq, easily calibrated... Of course if load is not a complex impedance that will distort waveform..


Just a thought.

Or you could use a modified piezo driver amplifier... They work in that frequency and voltage range...

Take a look at Apex Microtech app notes.. you might get some ideas..

[teksturi]

I'am missing this project  . Are you going to continue this ever?    

[RikV]

has this project been abandoned? Or has it moved to somewhere else?

[TiN]

It has been abandoned because rebuilding commercial calibrator happen to be easier and quicker job to complete  .

That project was covered here:
https://www.eevblog.com/forum/metrology/(volt-nut)-zombie-5700a/