exact current and potential source

One can purchase pre-trimmed reference voltage components, like AD581, with various precisions ($5 to $40). Standard cells (with a calibration certificate) are higher.

Nothing will get you to .001 mV unless your laboratory setup includes VERY tight temperature controls; wiring thermocouple effects make anything below .05 mV into a difficult exercise.

A calibrated voltage source, an attenuator (some of the old potentiometers, available on the internet at bargain prices, will hold four-digits of accuracy easily) and a few precision resistors makes it possible to generate lots of voltages and currents with simple operational amplifier circuits. Alas, you DO have to know and treat all the important error sources to do this kind of precision work.

Voltmeters in parallel have the same applied V; ammeters in serial have the same applied I. Everything else, you gotta keep track of errors and error propogation.

Sorry I made a mistake. I meant .001A and .001V not mA/mV so....

I will google for the AD581.

I need it for calibrating the oscilloscope. I don't have another source for calibration.

I am writing an oscilloscope program to Velleman PCSGU250. It is at it's early stages - ALPHA! But can be had from

If you don't have the hardware connected it starts in a mode where you can play with sin waves and select X-Y Plot from menu. Basic principle to me is that I will make possible to simulate scope without hardware and therefore develop the program.

Velleman is at

The frequency analyzer and bode plotter has proven to be very difficult to me.

NOTE !

makes and Open Source version of oscilloscope, which works with Windows, Linux and MacIntosh.

I am now going to spend 1-2 weeks to add support for Velleman PCSGU250 to it.

Best Regards Kari

--
PIC - ARM - DISPLAYS - RELAYS - MODULES - CONVERTERS - I2C - SPI -
KEYPADS - ACCESSORIES
http://www.byvac.com   (I am just a satisfied customer)

--- What you're looking for is called a "current shunt" and is essentially a low resistance precision resistor which is wired in series with the load and the voltage dropped across it [the shunt] measured in order to determine the current through it.

In order to comply with your spec, what you need is a shunt with a sensitivity of 1 millivolt per ampere, which is readily available and is called a "100 millivolt shunt" in the trade.

That 100 millivolts refers to the voltage which will be dropped across it with 100 amperes through it, and an example would be at:

Unfortunately, there's an error in their literature which specifies that an ammeter is used in conjunction with the shunt in order to measure current, when what they mean is 'voltmeter'.

JF

--
Equally unfortunately, there's also an error in my description, since a
100 millivolt shunt is a shunt that's made to drop 100mV with a
_specified_ current through it, not just 100A.

For example, the Delteco MKA-100-100 is a 100A 100mV shunt, while the
Delteco MKA-10-100 is a 10A 100mV shunt. 

JF

--
I misread your requirements, please disregard my earlier posts.

JF

Not an error at all; the shunt is one half of a pair of conductors, the second of which is a meter with its own internal resistance (for a typical d'Arsonval meter, a few dozen ohms). The shunt is taking most of the current, being lower resistance, but the meter IS a low impedance (ammeter). Current metering shunts are intended to have the same temperature coefficient as the meter, so are NOT simply accurate resistors. It's a current divider function with two elements, in the classic meter scheme.

The difference between a 'voltmeter' and 'ammeter' in this case is just the difference between 'drops-lots-of-voltage' and 'drops-negligible-voltage'.

--
In the sense that all d'Arsonval movements require the flow of charge to
work, they are ammeters.

However, when one is used to measure voltage, it's called a 'voltmeter'
and its resistance is quite high,(ideally infinite) while when used to
measure current it's called an 'ammeter' and its resistance is quite
low. (ideally zero ohms)

In the situation you describe where a (say) 1 ampere meter is used to
measure 100 amperes, a shunt is placed in parallel with the meter so
that with 100A through the assembly, 99A is passed through the shunt and
1A through the meter, with the meter dial indicating "100" at the
full-scale 1 ampere point.

The type of shunt I was referring to, shown at:

http://www.deltecco.com/MKA-DC.html

_is_ a temperature-compensated precision resistor, with Kelvin
connections, and is used to drop a voltage which is precisely related to
the charge flowing through it, that voltage being read by a high
impedance voltmeter for essentially zero loading on the shunt.

JF

That seems to be correct. I have never heard of a shunt that matches the tempco of the meter, except that both are constructed to be as stable as possible with a zero tempco. However, some analog meters are supplied with special leads which are calibrated to be used with a particular shunt. The meter load is only significant for shunts under 500 mA where the typical 5 mA analog meter affect the reading by about 1%.

I'm not sure what the OP means by an accuracy of 0.001mV/mA. Since he asked for a current and voltage source, it may just mean 1 uV or 1 uA depending on what is needed. And usually absolute accuracy is not required, especially for calibrating an oscilloscope.

Paul