I am having a difficult problem on a current project so I was wondering if anyone could help. This is a project where I have to measure current gener ated by a sample at high temperature inside a furnace. The current is in t he picoamp-nanoamp range depending on temperature and I have been using a K eithley picoammeter. The furnace is controlled by a PID temp controller co nnected to a solid state relay. When the relay is off I can get data with no issues but when the relay is on there is lots of noise and I cannot get good data. Unfortunately, the high temperature environment makes it diffic ult to shield the sample without going to exotic materials like platinum fo il. Is it possible to measure picoamps-nanoamps in the presence of lots of noise by using some kind of instrumentation amplifier or something with ve ry good common mode rejection ratio?
can you overshoot the temp, and turn off the controller during the test phase and take a reading as you cross back through the desired test temp point?
On Saturday, October 24, 2015 at 9:06:57 AM UTC-4, snipped-for-privacy@gmail.com wro te:
f anyone could help. This is a project where I have to measure current gen erated by a sample at high temperature inside a furnace. The current is in the picoamp-nanoamp range depending on temperature and I have been using a Keithley picoammeter. The furnace is controlled by a PID temp controller connected to a solid state relay. When the relay is off I can get data wit h no issues but when the relay is on there is lots of noise and I cannot ge t good data. Unfortunately, the high temperature environment makes it diff icult to shield the sample without going to exotic materials like platinum foil. Is it possible to measure picoamps-nanoamps in the presence of lots of noise by using some kind of instrumentation amplifier or something with very good common mode rejection ratio?
Hmm, So I would ask how the noise is getting into your measurement. A guess is perhaps noise on the heater. I had issues with capacitive pickup like that on a low temp probe. I stuck a cap multiplier on the heater power and life was good. But maybe some simpler filtering would be enough.
How much power is in the heater circuit. (Current and voltage)
So, interference is from coupling (probably capacitive) with an electric heater's wiring?
There are two things you can do; you can disable your measurement while the current is on. It's even possible, sometimes, to measure during a small part of the AC phase where the interference is, briefly, minimized.
Or, you can try to float the meter (or even the furnace) with transformer isolation, from both AC power and ground. This makes any attempts at shielding (guarding) very effective. Instead of removing the current leak path in the furnace, this means removing the ground return path, everywhere.
On Saturday, October 24, 2015 at 9:06:57 AM UTC-4, snipped-for-privacy@gmail.com wro te:
tate relay.
Check your PID temp controler. I think some have a output for a SSR and al so an output for a regular relay. If so you could switch to a regular rela y and DC for the heating power. Add an one shot and a and circuit so the t emp reading can only occur with the heater either on or off , but not when the relay is changing state.
The wiring outside the furnace should be shielded as well as you can. Maybe some of the noise is coupled in the external wiring / measurement circuit. It takes a lot of trial and error.
If the frequency of the noise is high compared to your sample rate, you might be able to add a low pass filter.
relay is on there is lots of noise and I cannot get good data. Unfortunately, the high temperature environment makes it difficult to shield the sample without going to exotic materials like platinum foil. Is it possible to measure picoamps-nanoamps in the presence of lots of noise by using some kind of instrumentation amplifier or something with very good common mode rejection ratio?
You can't use a J type couple that can be embedded? Those normally are incased in a 3/16 inch rod to be connected to ground.
You've received some excellent suggestions re reducing the source of the noise. Assuming you can do nothing about that aspect, the answer to your final question is yes. Back in the late '70's I had to design the production test equipment for Burr Browns then new OPA111 (1pAmax Ib) The production environment was very noisey and little could be done about reducing it. The solution was to force the Ib current onto a capacitor and then measure dV/dT. The result was repeatability/accuracy down in the fA's. The tradeoff was that dT sometimes had to be in the 100mS range. Art
If one is not in a hurry, you can connect a capacitor across a weak current source for a while, and then measure the voltage. A relay can do the switching. Common-mode rejection can be almost perfect, and you can measure femtoamps. Sort of a flying-capacitor integrator.
On Saturday, October 24, 2015 at 7:06:57 AM UTC-6, snipped-for-privacy@gmail.com wro te:
f anyone could help. This is a project where I have to measure current gen erated by a sample at high temperature inside a furnace. The current is in the picoamp-nanoamp range depending on temperature and I have been using a Keithley picoammeter. The furnace is controlled by a PID temp controller connected to a solid state relay. When the relay is off I can get data wit h no issues but when the relay is on there is lots of noise and I cannot ge t good data. Unfortunately, the high temperature environment makes it diff icult to shield the sample without going to exotic materials like platinum foil. Is it possible to measure picoamps-nanoamps in the presence of lots of noise by using some kind of instrumentation amplifier or something with very good common mode rejection ratio?
Thanks everyone for your responses! To answer some of your questions:
I am going to about 1050 C in an oxidizing atmosphere so that rules out man y traditional materials. Basically I have to use platinum wires to connect to the sample. Nobody that I know makes a high temp coax that goes to 105
0 C which is why I have virtually no shielding.
I don't have to make fast measurements, in fact 1 sample measurement per se cond would be fine.
I have tried ramping to a certain temperature and then unplugging the furna ce. All the noise goes away and I can get continuous data with no problem. But that is not a good long term solution since I will need to record dat a over multiple ramp/dwell/cool cycles which is why I am looking for a bett er solution.
Someone suggested an isolation transformer but that might be cost prohibiti ve and same goes for a DC power supply. My furnace runs off 120 V at 14 am ps. However, I did find a line filter at the local surplus electronics sto re but I haven't tried it yet.
I like the idea of integrating a capacitor. Texas Instruments makes the IV C102 which does this but I am sure there are other solutions as well. Has anyone used this IC? Are there other IC's or solutions that someone could recommend? I will be honest, my electronics skills are amateur but I have breadboarded a few things in my time.
Thanks in advance! I really appreciate all your responses.
One can make one's own coax, from platinum tube, platinum wire, and furnace cement (or powdered aluminum oxide, or a few beads on the wire to keep wire away from tube). Need to use a mineral insulation that is still insulating enough at 1050 C.
What is the impedance of the sample, at temperature? What is the physical mechanism of current generation?
Having a measurement gate exactly one second long allows profound suppression of prime power frequencies (50 Hz or 60 Hz) and all harmonics.
It is not hard to arrange 14 amps at 110 volts at DC. All it takes is a 25-amp fullwave rectifier bridge and some filter caps. This will not be expensive, and it is well within the reach of an amateur to build such a thing.
What are the heating elements made of? At high temperatures, some solid materials are subject to electrolysis if powered with DC. The furnace maker will know.
It requires considerable skill to build good femtoamp electronics, so it may be better to figure out how to use the Keithley picoammeter.
Hmm. What is the model number? If it is digital, it should be able to implement measurement windows that are an exact multiple of the period of the power line frequency. This is called NPLS (Number of Power Line Cycles) in the DMM world.
This furnace controller - is it zero-voltage switching? That will sharply reduce noise. One can buy such controllers off the shelf. Try Omega.com. Although one would expect the SSR to be zero-voltage switching. The input filter may be the answer. And ensure that the shall of the furnace is well grounded. If one replaces the SSR with a mechanical switch, do you get noise when the switch is closed?
On Saturday, October 24, 2015 at 7:06:57 AM UTC-6, snipped-for-privacy@gmail.com wrote: I am going to about 1050 C in an oxidizing atmosphere so that rules out many traditional materials. Basically I have to use platinum wires to connect to the sample. Nobody that I know makes a high temp coax that goes to 1050 C which is why I have virtually no shielding. ===========================================================================
This might not be useful here, but you could make your own shielded coax with some platinum tubing, platinum wire, and ceramic beads that barely fit into the tube loosely strung along the wire. It would only have to be long enough to get through the oven wall and then for the temperature from conduction to get down to what something like 316L stainless steel could handle so you could switch to SS wire and tubing to then get far enough away to switch to regular coax, or save the extra connections and just make the platinum section long enough so that the cold end is cold enough to connect to regular coax. Once you get outside the oven you will get plenty of cooling by convection, so my guess is that you would only need a few inches of the platinum coax outside the oven so the cost would not be that outrageous. Since you are working at such low frequency it doesn't matter what the characteristic impedance is, you just have to find tubing and beads that fit together off the shelf.
You could make a test version from all 316SS that would probably survive for a little while at maybe 700C, especially with an inert atmosphere like nitrogen or argon, so you could see how much good it will do under pretty realistic heater power conditions before committing to the platinum.
You could use a small DPDT relay and a capacitor as a flying-cap integrator. Connect the cap to the input leads for one second; then switch it to a DVM and measure the voltage for 100 ms or whatever; switch back to integrate some more. Eventually the cap will need to be discharged, which can be every measurement or just occasionally.
Nuclear particle detectors sometimes work this way. Low currents are integrated into a cap, which avoids the Johnson noise of resistors. Various things, like photofets, are used to occasionally discharge the cap.
relay is on there is lots of noise and I cannot get good data. Unfortunately, the high temperature environment makes it difficult to shield the sample without going to exotic materials like platinum foil. Is it possible to measure picoamps-nanoamps in the presence of lots of noise by using some kind of instrumentation amplifier or something with very good common mode rejection ratio?
Type K couple works in oxidizing environments and is good up to 1350C
Why not run the heater from a commercial PID temperature controller which uses pulse-width modulation to turn the heater on and off (e.g. once a second), and program the maximum duty cycle to (e.g.) 90%, meaning there is guaranteed to be a 100ms off-time every second. You can do the measurement during the off-time.
It might be challenging to disconnect the heater thoroughly enough to get a clean reading, without using a mechanical relay (which will eventually wear out and may get stuck-on when that happens).
Either you would need to find two solid state relays (one for each end of the heater) that can turn off well enough to give good measurements, or you might find a hybrid solution that works ok:
Put one solid-state relay that turns on and off the power to the heater, and then in series with that, right before the heater, put an additional double-pole mechanical changeover relay that only ever switches when the solid-state relay is already off. This double pole mechanical relay would disconnect both ends of the heater from the mains and optionally would also connect the heater to ground (perhaps via a 100k resistor, just in case one side of the relay fails to switch). The relay should last a very long time because it would not be switching any current as the SSR would be off when it switches. You could achieve the correct timing using for example some logic gates and monostables or 555-timers etc. that are processing the output PWM signal from the commercial PID temperature controller. This would also go to a monostable that drives the trigger input of the meter used for measurements during the off-time.
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.