[volt-nuts] Power consumption of 3458A reference board / fittingone in a 3457A

Andreas Jahn Andreas_-_Jahn at t-online.de
Sun Sep 24 10:34:58 EDT 2017


Hello,

> but if the difference was
> 10 mV or so, maybe a 3457A would be good enough.

just mentioned that my last answer did not go to the list.
If you have 2 or more LTZ1000 it is usually possible to record 
differences between them.
I have 6 LTZ1000A and all are within 60 mV.
So I can do measurements either with a K2000 or a 34401A 6.5 digit 
instrument in 100mV range.
Just keep one LTZ under normal conditions and vary the other in 
environment conditions.

I also average the measurement values over 1 minute to reduce noise of 
the DMM.
http://www.eevblog.com/forum/metrology/ultra-precision-reference-ltz1000/msg881955/#msg881955
(see picture "DMMs" for the 6.5 digit measurement instruments in a 
differential TC-measurement).

So there is a good chance that you can also do some measurements on the 
3457A in 30 mV range.
you just need enough samples to get all within 30 mV.

> I don't know anything about that.

120R is the resistor for the current setpoint
of the data sheet of the LTZ1000 with the 13K temperature setpoint.

If you have a orignial 3458A reference board this resistor is reduced
to 111 Ohms because of the higher temperature setpoint.

> Of course that introduces another source of error too, as the buffer would
> have some temperature coefficient.

If you use a chopper amplifier like the LTC2057 the additional
temperature drift is below the claimed 0.05 deg C of the LTZ1000.
Additionally you can adjust the overall tempco of the reference by 
variation of the 400K resistor.
(see EEVBLOG in the long long LTZ1000 thread).

with best regards

Andreas

Am 24.09.2017 um 12:16 schrieb Dr. David Kirkby (Kirkby Microwave Ltd):
> On 23 September 2017 at 22:48, Andreas Jahn
> wrote:
>
>> Hello Dave,
>>
> Hello Andreas
>
>
>> there are several things to consider for a travel standard:
>>
>> - first you might want to have some voltage regulator to stabilize the
>> voltage.
>>    the voltage coefficient that I have measured is around 0.3-0.4 ppm/V
>>    and between full and near empty of 15 cells are 21.8  to 17.2 V.
>>
> Good idea. But fairly easy to implement.
>
> I wonder if I could do any useful tests with the 3457A (6.5 digit), by
> looking at the difference between the various sources. I don't know how
> similar the voltages from these boards will be, but if the difference was
> 10 mV or so, maybe a 3457A would be good enough.
>
> I did buy a cheap (£350 GBP I think) Keithley 2001 7.5 digit meter, but it
> has a few issues, and I deceided to return it.
>
> - the original 3458A board works with a relative high setpoint temperature
>> of the reference.
>>    you might want to reduce the temperature setpoint (save current) if the
>> environment temperature is near room temperature.
>>    The 15K metal foil (2ppm/K) of the HP-board can be reduced to 13K (metal
>> foil) like in the data sheet.
>>    Or if you have a original board you can also parallel 100K (metal film
>> 15ppm/K) to the 15K.
>>
> Good idea.
>
>    Or even further down to 12K5 if the environment of the LTZ is not above
>> 30-35 deg C.
>>
> I'd probably not go that low. It can get quite hot in here sometimes if the
> aircon is off.
>
>
>>    But you might  also want to adapt the 111 R to 120 R to compensate for
>> the higher Vbe of the transistor when the temperature is lower.
>>
> I don't know anything about that.
>
>
>> - If you reduce the setpoint temperature you can also reduce power supply
>> voltage to 15V or even below.
>>     I am using 12 * NiMH and stabilize it down to 14.0 V. (with a low
>> power/low noise LT1763)
>>     The voltage regulator should not be on the same board or near the LTZ
>> to achieve optimum voltage coefficient.
>>
>> http://www.eevblog.com/forum/metrology/ultra-precision-refer
>> ence-ltz1000/msg875274/#msg875274
>
> Cheers
>
>> - you might also want to have some sort of buffer for the output voltage.
>>    otherwise in case of a short of the output (or excess current) the
>> temperature setpoint of the heater goes to infinite ...
>>    the result is a quick ageing of the reference with several ppm shift and
>> around 6-12 months of larger drifts than usual.
>>
>> http://www.eevblog.com/forum/metrology/ultra-precision-refer
>> ence-ltz1000/msg951358/#msg951358
>
> Of course that introduces another source of error too, as the buffer would
> have some temperature coefficient.
>
>> with best regards
>>
>> Andreas
>
> Dave
>



More information about the volt-nuts mailing list