Voltage reference 2

[hoda0013]

I am building a device for corrosion testing. To do corrosion testing, the part being tested is submerged in an electrolyte and a current is passed through it. The voltage source is controlled and the current is measured. These are plotted against one another and at a certain voltage, a breakdown of the part can be seen (corrosion).

The device consists of a precision power supply capable of outputting a voltage from -0.6V to about 0.8V and then back to -0.6V. The voltage is increased at a rate of 1mV / second and needs to be accurate to within 1mV. The current is measured using a picoammeter and the current ranges from tenths of microamps all the way up to about 10 milliamps. The current measurement problem can be solved by purchasing a picoammeter which I have done. However, the voltage source may pose an issue. I haven't been able to find voltage source that is both programmable, low-noise (0.5mV or less), and capable of positive and negative voltages.

The closest thing I have found is a power supply (BK Precision 9150) that can output from 0V - 5V and is very precisely controlled (0.1mV line regulation , 0.5mV load regulation)as well as programmable. The problem is that it does not output negative voltages which I need. Is there a way to add in a negative voltage to the output of this power supply to bring the voltage down to -0.6V?

I am thinking about using an op amp to level shift the voltage into the negative region, but the output from the op-amp will be as noisy as it's rail voltages right? I am worried that the op-amp stage will potentially ruin my clean signal coming from my power supply.

Any ideas would be greatly appreciated.

Thanks

 

[DRWeig]

Simple answer might be to get a second BK 9150, put it in series (assuming it's isolated) with the one you already have, and use the common point as your zero. If you keep these two outputs isolated, you'll have a ±5 precision source.

I may be missing something, though -- let some of the others pitch in too.

Good on ya,

Goober Dave

 

[hoda0013]

@DRWeig: I hadn't thought of that and I'm sure that would be a good solution. However, at $1,800 per power supply, this isn't economical for me.

I don't have a ton of experience in analog electronics, so I can't be sure this is possible, but I'm thinking if I found a low noise, adjustable negative voltage source, then I could add it into the BK9150 output voltage to make my negative voltage. However, finding something with the precision I need has been tough.

 

[DRWeig]

Yokogawa GS200?

It's the closest I could come with my Google kung fu...

If none of the regular contributors in this forum (of whom I stay in awe) come through for you, you might try re-posting in the measurement and control instrumentation engineering forum...

Good on ya,

Goober Dave

 

[IRstuff]

Whenever I think of voltage sourcing, I think of Keithley:

http://www.keithley.com/products/dcac/voltagesource/broadpurpose/?mn=2400

On its 2V range, its programming resolution is 50 uV, and measures down to 10 pA resolution.

Yokogawa, which used to be the manufacturer of the HP4145 semiconductor parameteric analyzer might also be a possible source of similar equipment.

TTFN

FAQ731-376

 

[IRstuff]

One primary advantage of the Keithley approach is that any measurement is directly correlatable to the voltage supplied, since it's from the same machine. That cannot be said of a discrete solution.

TTFN

FAQ731-376

 

[Compositepro]

I have an Intertech PGS151 Potentiostat/Galvanostat which was designed specifically for electrochemical measurements. I don't know if they are still in business, as I can't find them with Google. There are only a few references to the instrument in some publications.As I recall the instrument was about $5000 many years ago. Their phone number in the manual is 4082637335 (San Jose, CA). The input impedance is 10^10 ohms. I don't know how accurate you need to be but these kinds of measurements can require great sensitivity and there are many details to be aware of. These instruments are so expensive because most of these details have already been addressed by the instrument manufacturer.

 

[btrueblood]

Um, I may be missing something too, but instead of two power supplies, how about an H-bridge relay circuit to essentially reverse the leads from a single power supply?

 

[IRstuff]

Keithley 2400 $3150:

http://www.ecrunch.com/listing/Keithley_2400_Digital_Sourcemeter_200V_1A_20W.html

TTFN

FAQ731-376

 

[Warpspeed]

If you only need to source picoamp current, why do you need a hulking great power supply ?
An opamp should do it easily.

How about a bipolar digital to analog converter with a precision voltage reference, and enough bits to give microvolt resolution.

 

[hoda0013]

Problem solved. Someone was selling a corrosion testing setup on ebay.

@Warpspeed: I tried doing exactly that, but ran into problems with noise. With the resources at hand, I managed to minimize the noise to 20mV, but need more like 0.5mV.

 

[Warpspeed]

Twenty millivolts of noise is huge.
A suitable L/C low pass filter should be able to attenuate that to well within your requirements.

Anyhow, glad to hear you have now found a solution.

 

[roydm]

How about using a higher voltage power supply with a voltage divider circuit to reduce to the level you require.
If you had a ±50 Volt supply with a 10:1 Voltage divider that might work. I haven't thought it through, the idea just jumped out.

 

[eeprom]

Why not use op amps with batteries as a power supply. You could easily control the voltage with a pot. And there would be no noise.