Electronics > Beginners
DMM - Does PTC degrade the accuracy of the voltage divider?
Youpi:
Hello everyone,
For some time now, I study the input protections of the multimeters and how they can influence the global precision of the measure (Precision: I talk of multimeters >= 4,5 digits).
Many topics talks about the input protections but I still have a question and don't find the an answer.
I saw several teardowns and schematics of multimeter where an PTC and an massive resistor are in serie on the voltage input to reduce the peak of current in MOV (with/without GDT).
This stage is classicaly followed by a high precision resistor network (as a caddock type 1776) to divide the voltage from 600/1000V to ~10V for the JFET input buffer.
But, why is the purpose to choose a such high precision network (with very low TCR ~30ppm) if components with strong temperature dependency (PTC, massive resistor) are put in serie?
What is the trick? What I missed?
My neuronnes will become nuts! :wtf:
Thank you very much for your time :-+
Ps: I drawed a simplified schematic on Kicad to be more concrete of an input voltage multimeter.
Kleinstein:
In a circuit as shown the PTC would effect the divider TC, though not that much 3.5 K / 10 M is only some 0.03%. So a 3000 ppm/K TC for the PTC would would result in some 1ppm/K shift for the divider. The PTCs used for protection tend to have a quite nonlinear curve: below some 40 or 50 C the TC is still relatively low and only really goes up at higher temperature. So the PTC has an effect, but it is still small. One would still avoid this in a 6 digit+ meter.
David Hess:
Like Kleinstein says, the effect is small and a good design takes it into account. Shunt leakage through the various protection devices also has an effect. In a higher accuracy design, it is more difficult to implement input protection without compromising accuracy.
iMo:
Off Topic - you can see the temperature effects in LTspice easily - below (an example only) how to step through temperature with resistor's TC..
Youpi:
Thank you for your answers, you brought me something to think about ;)
--- Quote from: Kleinstein on December 15, 2019, 04:05:43 pm ---In a circuit as shown the PTC would effect the divider TC, though not that much 3.5 K / 10 M is only some 0.03%. So a 3000 ppm/K TC for the PTC would would result in some 1ppm/K shift for the divider. The PTCs used for protection tend to have a quite nonlinear curve: below some 40 or 50 C the TC is still relatively low and only really goes up at higher temperature. So the PTC has an effect, but it is still small. One would still avoid this in a 6 digit+ meter.
--- End quote ---
I didn't realize the curve around 40 or 50 K is relatively flat and will be good for 4,5 digits (and confirmed by LTspice). It's right that I don't see high-end multimeter with PTC in input. The PTC was ever on handle multimeter. I need to dig deeper to find how you got 1ppm/K.
--- Quote from: David Hess on December 15, 2019, 06:49:10 pm ---Like Kleinstein says, the effect is small and a good design takes it into account. Shunt leakage through the various protection devices also has an effect. In a higher accuracy design, it is more difficult to implement input protection without compromising accuracy.
--- End quote ---
Yes, thank you for the warning. I remember the first time where I calculs the error introduces by the leakage for high impedance... :'( And it's common to use transistor monted as diode for ultra low leakage protection to avoid leakage.
I see others protections as the datron 1281, where no ptc is used (or i missed it!) and use a low impedance input to decrease noise with a kind of "power watchdog". But I am more septic about high surge voltage rugness.
--- Quote from: imo on December 15, 2019, 07:00:59 pm ---Off Topic - you can see the temperature effects in LTspice easily - below (an example only) how to step through temperature with resistor's TC..
--- End quote ---
It's the first time I read about this parameter "tc" : tested and approved! Before, I just added manually like R(1+tc).
ps : I hope you understand my writings
Navigation
[0] Message Index
[#] Next page
Go to full version