Can anyone help with a microprocessor controlled circuit to produce a very stable and settable -10 to +20 Microvolts supply which is temperature stable. The voltages should be selectable and be able to step through a series of selected mivrovolts.
How many discrete voltage do you want?
The easiest approach would probably be to use the microprocessor to drive a DAC offering an output swing of +5V to -5V (or +2.5V to -2.5V)
- hook up a 143k (E96 0.1%) resistor to your DAC output, and a 432k (E96 0.1%) resistor to your negative reference, and 2R (4R) resistor to ground. Farnell stocks suitable parts.
The difficulty is avoiding microvolt thermocouple voltages between the
2 ohm resistor and the point where you want to see the 20uV - the only easy option is to put the 2 ohm resistor at that point. You may then have to be carefull to keep the resistance of your ground return path well below 2 ohms, so the (up to) 10uA return current doesn't add to the 20uV across the 2R resistor, but careful star grounding can get around that.Generating microvolts reliably does take a certain amount of thoughful care, but it can be done.
The acid test of a microvolt generator is always to turn off or detach your biasing source and see how many microvolts your measuring system then sees - it should be zero, and hardly ever is .... a simple mechanical switch to do the job can be handy, and something electronic controlled by the microprocessor might let you do auto-offset corection. A relay might not be a good idea - the coils get hot and can generate thermocople voltages.
-- Bill Sloman, Nijmegen
How many discrete voltage do you want?
The easiest approach would probably be to use the microprocessor to drive a DAC offering an output swing of +5V to -5V (or +2.5V to -2.5V)
- hook up a 143k (E96 0.1%) resistor to your DAC output, and a 432k (E96 0.1%) resistor to your negative reference, and 2R (4R) resistor to ground. Farnell stocks suitable parts.
The difficulty is avoiding microvolt thermocouple voltages between the
2 ohm resistor and the point where you want to see the 20uV - the only easy option is to put the 2 ohm resistor at that point. You may then have to be carefull to keep the resistance of your ground return path well below 2 ohms, so the (up to) 10uA return current doesn't add to the 20uV across the 2R resistor, but careful star grounding can get around that.Generating microvolts reliably does take a certain amount of thoughful care, but it can be done.
The acid test of a microvolt generator is always to turn off or detach your biasing source and see how many microvolts your measuring system then sees - it should be zero, and hardly ever is .... a simple mechanical switch to do the job can be handy, and something electronic controlled by the microprocessor might let you do auto-offset corection. A relay might not be a good idea - the coils get hot and can generate thermocople voltages.
-- Bill Sloman, Nijmegen
Just find a DAC with itty-bitty output transistors. Regards,
Mike Monett
Antiviral, Antibacterial Silver Solution:
A DAC and a voltage divider comes to mind.
Best regards, Spehro Pefhany
-- "it\'s the network..." "The Journey is the reward" speff@interlog.com Info for manufacturers: http://www.trexon.com Embedded software/hardware/analog Info for designers: http://www.speff.com
The problem with two resistors of identical tempco in a divider is...?
-- Many thanks, Don Lancaster voice phone: (928)428-4073 Synergetics 3860 West First Street Box 809 Thatcher, AZ 85552 rss: http://www.tinaja.com/whtnu.xml email: don@tinaja.com Please visit my GURU\'s LAIR web site at http://www.tinaja.com
If their materials are identical, so that there are no thermocouple offsets, and there are no significant temperature gradients, there's no problem. In reality, there are several thermocouples in series in any such circuit, and the OP won't get fractional-microvolt stability without paying close attention to that issue.
Cheers,
Phil Hobbs
If their materials are identical, so that there are no thermocouple offsets, and there are no significant temperature gradients, there's no problem. In reality, there are several thermocouples in series in any such circuit, and the OP won't get fractional-microvolt stability without paying close attention to that issue.
Cheers,
Phil Hobbs
I only require 5 to10 voltages which I need to set to certain values within the required range. I have experimented with a simple bridge circuit using various values of standard resistors (at this stage Not low temp coeff) The output volage obtained is not very stable, although the entire circuit was kept at the same temperature. The problem is probably compounded by the need to have two pots in the circuit to adjust the output volage.
I have never used a DAC before. I do not understand where the resistors are connected. Can you suggest a DAC to use.
Yes, the tempco matching of the resistors is of little importance in this application (1000ppm is 30nV), however typically precision resistors are made with better materials and it may accidentially be of some benefit. In particular, 5% 'carbon film' resistors are usually
*not* carbon film at low values (nickel alloy or something like that). Similarly, the reference stability for the DAC and DC offset if of little importance. You want to avoid thermal gradients and thus keep anything that generates heat away. It's nto that hard to get to a large fraction of a uV with some care and experience, even with inexpensive components.Best regards, Spehro Pefhany
-- "it\'s the network..." "The Journey is the reward" speff@interlog.com Info for manufacturers: http://www.trexon.com Embedded software/hardware/analog Info for designers: http://www.speff.com
On sci.electronics.design it is normal practice to "bottom post" - that is to add your response to the bottom of the message, rather than sticking it on the front, as you have done here. I've copied your response to the bottom of the post, and folowed with my reply.
For this sort of applcation, you'd probably use one of the timing counters inside the microprocessor to set up a pulse-width-modulated (PWM) waveform. If you only require 5 to 10 specific values, an 8-bit counter could do the job - you can usually get a 16-bit counter without trying to hard.
You'd them feed the PWM waveform into the control pin of a multiplexer
- the 74HCT4052 would be nice
because the control voltage can be 0V to +5V while the switched outputs can swing from +5V down to -5V.
One of the inputs to the multiplexer is your positive reference - the LT1004-1.2 could be nice
and another would go to your negative reference, which you'd derive from the +1.200V reference with an op amp and a pair of precision resistors (or a precision divider). I'd go for -0.600V.
That would leave you two free inputs on a 4052, which I'd take to good nearby ground connections, for the microprocessor to pick up when you want to check if 0V reads 0V on your measuring device.
The output from the 4052 would then be fed into a a Sallen-Keys unity gain low-pass filter to get rid of the PWM ripple and the output of the filter would swing from -0.600V to +1.200V.
You'd get your -10uV to +20uV by dividing down this output much as before - 120k into 2 ohms would do the job nicely. You've still got to make sure that the 10uA running back to the power suplies doesn't shift the "ground" at your measuring point by few uV - as little a an ohm resistance in the return path could produce a 10uV drop along the connection.
It can be important - one one occasion we managed to reverse the polarity of an adjustable voltage reference because a dumb fault in the printed circuit layout was running 100mA of digital supply current through a sensitive analog ground connection. Easily fixed when we worked out what was going on, but very embarassing in the meantime.
-- Bill Sloman, Nijmegen (but in Sydney at the moment).
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.