[volt-nuts] A new LTZ1000A transfer standard is born

[Andreas Jahn]

Hello Hendrik,

the first what I would do is some qualification measurements:
since the pricing within the bay is lower as on original manufacturer
they might either be recycled parts or the remaining parts after selection 
for instruments.

On my references I did the following tests:
- noise (does it have the 1.2uVpp?)
- temperature test: how large is the temperature drift
  if you have a temperature chamber this will ease the test.
  I only have a thermoelectric cooling / heating box for around 10-40° 
centigrade.
- Temperature setpoint test: how large is the tempco of the unstabilized 
reference if you do a 10 degree setpoint offset.
   (is usually around 50ppm/K)
- Power supply voltage rejection test
- Influence on EMC (sensibility on capacitive influence by your hand)
- Tilt test (described here from Frank)
I hope that I didnt miss any...

Did you already clean the pcb?
Did you pre-age the resistors before mounting?
Is the LT1013 within a hermetically housing?

After qualification tests and if I had a thermal chamber
I would first do some pre-ageing (thermal cycling) for at least 1-2 weeks
with daily measurement of the voltage.
After this at least further 1000-4000 hours continuously powered to 
stabilize.

> measurements for a few days, although I think that the drift+noise from
> the HP reference and my LTZ are quite close

But: the HP reference has already stabilized and is not freshly soldered.

I would do a final calibration after around 6 months continuously operation.

And if you populate the 2nd PCB you can make difference measurements
in a voltage range with higher resolution on your HP3456 (perhaps 0.1uV in 
200mV range).

With best regards

Andreas

----- Original Message ----- 
From: "Hendrik Dietrich" <don_hendi at gmx.de>
To: <volt-nuts at febo.com>
Sent: Saturday, March 02, 2013 10:14 AM
Subject: [volt-nuts] A new LTZ1000A transfer standard is born


> Good morning fellow nuts,
>
> after a short planning and long wait times to get LTZ from china[1], PCB
> from the Ruhr valley[2] and resistors from the UK[3], I was no longer
> able to wait and slap everything together.
> It is a clone of the Williams design in Linear Technology AN86 with
> 2x25K in parallel as R4, as suggested by Dr. Frank Stellmach from this
> maillist. The interesting resistors are General Resistance Econistors,
> the zener plus is buffered with a further LT1013 before breakout with
> 4mm copper-tellur receptacles[4]. The LTZ1000A has a hat made from a
> little plastic cap and a o-ring to fix it to the LTZ case. The positive
> and negative out and also a hole for a guard are on a slotted part of
> the PCB and around the LTZ I spent some holes to give it a bit of
> stress-relief.
>
> It worked instantly, I didn't even try step-by-step startup, current
> limiting was the only safety measure. No stability measurements yet, but
> the HP34401A at work (where I soldered it during lunch break) and the
> HP3456A at home have shown the same voltage (the 3456A is calibrated
> against my work DMM).
>
> After finishing the circuit board, I have put it into a tinned sheet
> iron case[5] and made a makeshift thermoinsulation from bubble wrap and
> a cardboard box. Result:
>
> http://primeintrag.org/LTZ1000A.JPG
>
> that is the newborn and one of the spare PCBs.
>
> Planned next steps, any hints?
>
> 0. Keep it on the HP3456A today and watch infrequently and hope that
> only the last digit might change
> 1. Put it in the thermal chamber at work (hot-cold-hot cycle, tmax=80°C)
> 2. keep it powered for a few weeks
> 3. Hook it to my 3456A and use EZGPIB to get some voltage-over-time
> measurements for a few days, although I think that the drift+noise from
> the HP reference and my LTZ are quite close
> 4. Beg a cal lab to do a un-accredited HP3458 measurement with results
> written on a napkin to get a number! (Repeat after a year or so)
>
> Regards
>
> Hendrik
>
> [1] ebay seller polida 2088
> [2] http://www.fischer-leiterplatten.de
> [3] http://www.rhopointgermany.com/
> [4] http://www.multi-contact.de/
>
> _______________________________________________
> volt-nuts mailing list -- volt-nuts at febo.com
> To unsubscribe, go to 
> https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts
> and follow the instructions there.