REF5010ID 10V standard and resistor

[jwm]

Hi all, on my last mouser order I decided to add a REF5010ID 3ppm/C 10V voltage reference figuring it would be handy to embed in a nice case with bananna jacks for testing multimeters and whatnot. (I have a stack of 3478A's I rescued that I am going to supercap-mod to allow easy changing/testing of the battery backup)

I additionally added a 10ppm/C .1% 10k resistor to my order, so my question is, if I add the resistor in series to the REF output, am I hurting it's usefulness as a voltage reference at all? My intuition says no as the multimeter itself has a megaohm or so resistance in voltage mode in it anyway, adding another 10k even if its imprecise should not affect the voltage measurement at all, is that true?

My reason for wanting to add the resistor is both to protect against an accidental short (not sure if the REF can handle that or not, spec sheet doesn't explicitly mention it but it does list 'short circuit current 25mA' in the parameters) but also so I can use it as a semi-precise 1mA reference. 10ppm .1% resistor and 3ppm 0.05% voltage combine to what accuracy of current reference?

Any other advice on how to set this up other than just wire the chip directly to the output for use as a stable lab reference? Is it normal to add a cap to the output? Thanks all!

    John

[madires]

Ohms law answers all your questions :-) Let's assume we allow a maximum drop of 1mV by the resistor. The current would be I = V / R = 1mV/10kOhm = 0.1µA. Now let's calculate the minimum total resistance for 0.1µA at 10V: R = V / I = 10V/0.1µA = 100MOhm. If your DMM has a input resistance less than 100MOhm you will get a voltage drop larger than 1mV at the 10k resistor. A typical DMM has got 10MOhm, so we have I = 10V / (10M +10k)Ohm = 1µA and get a voltage drop at the 10k resistor of V = I R = 1µA 10kOhms = 10mV.

The idea of using the resistor to create a current reference has one major problem. Your load would have to be 0 Ohms. Any larger resistance would lower the current.

[jwm]

Thanks for the great response. even though I actually said "megaohm" I still had "infinite impedance" in my brain. 

Hitting my head against the wall also made me realize how to protect the chip against overcurrent in a better way, add a lm334  to the _input_ voltage of the reference, not the output. perfect.

I think I will cut a bananna plug in half to make a precision 10k passthrough (bananna plug on one end, jack on the other) so I can just plug it in series if I want it for some reason or to get my almost-sorta 1mA reference.

Actually, what is  the best way to get a precsion current source from a precision voltage (or precision resistance)? Can it be done with easily ottainable parts likely to be in my junk bin? an lm334 with a precision resistor would introduce a big temperature coefficient. ovenize it? Hmm...

    John

[madires]

Please have a look at http://extras.springer.com/2006/978-0-387-25746-4/Chapter_16.pdf for the basic circuits. For a current reference you could build a current regulation, i.e. a sense resistor, an OPamp with driver transistor and a reference voltage. The OPamp will drive the transistor to pass current until the voltage drop across the sense resistor matches the reference voltage.

[jwm]

Thanks. Just what I was looking for!