[volt-nuts] J511 current regulator diode (FET) temperature coefficient

[David]

I doubt the temperature coefficient of D717 matters in this circuit.

Feedback is taken off of the output through TR705 or TR706 (x10 mode)
back to the inverting input of IC705.  The current source was used to
prevent any significant shift in the operating point of TR703 for
frequency compensation reasons because the switched cascode has
voltage gain which is inside of the operation amplifier feedback loop.

The drift in the cascode bias from D716 is irrelevant for the same
reason.

Having voltage gain inside of the feedback does suggest another
failure mode that could cause drift or noise in the output.  Could
IC705 or TR703 be oscillating?  Maybe C714 dried out and has low
capacitance.

On Thu, 5 Jul 2012 10:52:59 +0100, "David C. Partridge"
<david.partridge at perdrix.co.uk> wrote:

>Yes the TC of these devices is awful.
>
>I need to check against other meters, but I'm fairly sure this will turn out to be a "feechur" of the circuit design.
>
>So I'm thinking in terms of replacing this with something better, but what?  I'd rather not (being lazy) try to design a good zero TC bipolar 4.7mA current source for this.  
>
>Do you have suggestions - ideally with a low part count - the LT3092 looks attractive, but the distributors don't seem to stock them, while the LM334 needs an extra diode and resistor.  
>
>Thanks
>Dave
>-----Original Message-----
>From: volt-nuts-bounces at febo.com [mailto:volt-nuts-bounces at febo.com] On Behalf Of Charles P. Steinmetz
>Sent: 05 July 2012 03:34
>To: Discussion of precise voltage measurement
>Subject: Re: [volt-nuts] J511 current regulator diode (FET) temperature coefficient
>
>Dave wrote:
>
>>Measuring the voltage drop across R718 which is the 100 ohm collector 
>>resistor for TR703, I see a huge change just taking the lid off the 
>>meter:
>>
>>Lid on: 0.45200V after 8 hours powered on
>>      *     *     *
>>The voltage across this resistor was almost 0.5V at power on from cold.
>>
>>Are these diodes subject to failing with poor temperature regulation?
>
>They don't have very good temperature regulation to begin with.  Vishay specifies the J511 with a tempco of -0.34% per degree C.  (If I recall correctly, this is significantly less than some other manufacturers claim for the J511 -- but I long ago gave up on CRDs, so that's an aged recollection.)
>
>If yours changed from 5 mA cold to 4.25 mA hot, and I've done the maths right, that implies a temperature change of 28C if the tempco is -0.34% per degree C -- which does not sound at all implausible between ambient temperature rise in the meter and self-heating of the J511.  Of course, if your J511 is from another manufacturer and has a higher tempco, it would take less than a 28C change to produce the same shift (I also did not try to guess how much less than 5 mA yours sourced when cold).
>
>Note that (like zener diodes) the J5xx series goes through a zero tempco point between the J507 (nominal 1.8 mA, +0.08% per C) and J508
>(2.4 mA, -0.05% per C), so you might be able to do better with a carefully chosen parallel combination.  Or, if good temperature compensation is needed at this location, you may want to build a temperature-compensated current source with an LED reference and a transistor (that is probably what I'd do), cascoding it with a FET if necessary to lower the output conductance.
>
>However, if this particular meter is performing differently than others of the same model with the same CRDs, the tempco of the current source may not be the source of the problem.  (Or does that meter have a fan that is not working properly, or obstructed internal ventilation paths?)
>
>Best regards,
>
>Charles
>
>
>
>
>
>_______________________________________________
>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.
>
>
>_______________________________________________
>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.