Author Topic: Maintain a current carring wire at a chosen voltage relative to another wire  (Read 4853 times)

0 Members and 1 Guest are viewing this topic.

Offline gus789Topic starter

  • Contributor
  • Posts: 47
  • Country: us
Hi - I am trying to drive a small adjustable current through a thin wire while maintaining said wire at a fixed but adjustable voltage relative to a second wire. The voltage on the current carrying wire must be adjustable from positive to negative (say +5V to -5V) relative to the other wire. The second wire can be at ground.  I am wondering if a LM134 will do this, just not sure whether I will be able to adjust the wire voltage from positive to negative. Any suggestions greatly appreciated. Thanks!
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 21609
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
To what end?
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 

Offline gus789Topic starter

  • Contributor
  • Posts: 47
  • Country: us
the adjustable current is for changing the wire's position in a permanent magnet's field. As it is brought closer and closer to the wire at ground a tunnel current develops. The idea is to get the i-v characteristic of this tunnel current. This is used in some studies on conduction through molecule monolayers on the surface of the wires. None of the papers I've seen actually show a circuit for how to accomplish this.
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 21609
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Ah, very interesting!

The tunnel voltage will be very critical then, no?  You'll want to arrange a three or wire probe to do that, if not use a separate wire entirely.

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 

Offline gus789Topic starter

  • Contributor
  • Posts: 47
  • Country: us
Yes, I found it quite intriguing too and wanted to try it out. Apparently this method has pretty good stability against vibration because the thin wires are positioned with some slack which dampens the vibrations. In the reports I've seen they all used the same wire to carry current and to apply the voltage. The ground wire was at virtual ground of a I-V converter. Typically the positioning current was less than 1mA, and wire diameter was around 10 microns with total length of about 1.5 inches. For gold this gives less than 5Ohm resistance, and at 1mA voltage drop would be only ~ 5mV. So I guess voltage uncertainty along the wire is not that large anyway.

The part I am grappling with is how to make it all work stably at different voltages. I was thinking it'd have to be a  floating constant current source/sink, but I have no experience with this sort of circuitry. Looking around for clues led me to the LM134 as well as to Fig 28 of this app note http://www.ti.com/lit/an/snoa624b/snoa624b.pdf.

Anyway, any ideas are appreciated! Thanks -  :)
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 21609
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Since this is lab stuff, there's no problem with just having another freaking power supply on the bench, right?  Something with isolated output, ranging from perhaps a well filtered transformer, rectifier and cap, on up to a full blown programmable bench supply.  (With a current limiting resistor of course.)  Or you could do worse than have a AA cell and rheostat.

I recall reading about the early days of NMR (ca. 1950): they'd power the magnets (1.2T or so -- iron armature, lots of wire, lots of weight, and not much active volume!) with lead acid batteries, because that was the only thing available with the right voltage and current, and good enough stability and low enough noise to actually resolve the phenomenon (NMR is a parts-per-million game, and that goes for all aspects: stability of the magnetic field, its uniformity, frequency resolution of the detector, and the intensity of the absorption peak!).  Of course, both the battery and the electromagnet have large temperature coefficients, so grad students were tasked with opening and closing the laboratory's windows to regulate temperature...

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 

Offline Rufus

  • Super Contributor
  • ***
  • Posts: 2095
Typically the positioning current was less than 1mA

If you are only talking about a few mA you can do it with op-amps. Drive one end of the wire with a voltage, sink constant current to a -ve enough rail at the other end of the wire.

A standard op-amp + N channel MOSFET current sink circuit will do the job, the drive side is just an op-amp voltage amplifier or buffer.

You probably want a voltage reference for stability but the question is how do you propose/intend to set and adjust the voltage and current? 20 turn pots?
 

Offline gus789Topic starter

  • Contributor
  • Posts: 47
  • Country: us
Well, I was thinking of a pot to adjust the current manually and a triangle wave from a function gen to provide the voltage. Data collection on a digital oscilloscope (so I can have the scope on x-y and get the i-v curves at different pot settings.

With such a mosfet current sink wouldn't I be limited to work with only positive voltages though?

@Tim - I thought about the two power supply idea, but it seemed hard to control the two power supplies simultaneously so as to maintain a fixed voltage difference. Only if I get one grad student to control each supply - but that becomes expensive  :) 
 

Offline Rufus

  • Super Contributor
  • ***
  • Posts: 2095
Well, I was thinking of a pot to adjust the current manually and a triangle wave from a function gen to provide the voltage. Data collection on a digital oscilloscope (so I can have the scope on x-y and get the i-v curves at different pot settings.

With such a mosfet current sink wouldn't I be limited to work with only positive voltages though?

You could drive the wire directly from a half decent function generator then. If the current sink is sinking to a -ve enough rail then all voltages on the wire will be positive. If the generator can do say +/-10v you need maybe a -15v supply for the current sink. A 1k sense resistor in the current sink would give you 1mA/volt and a pot across a cheap 2.5v shunt reference like a TL431 would give you 2.5mA full scale.
 

Offline AG6QR

  • Frequent Contributor
  • **
  • Posts: 857
  • Country: us
    • AG6QR Blog
Are both wires supposed to be carrying current?  The same amount of current in each wire?

If so, I would connect the wires using a variable resistor at one end (or perhaps an adjustable voltage regulator, which is sort of a "electronically controlled variable resistor) and put a lab power supply in constant current mode on the other end of the pair of wires.


But if my hypothesis about the wires is wrong, a general way to do it, maintaining control of everything independently, is to use three lab power supplies, each with isolated outputs.

Short one wire across one supply, and let the supply go into constant current mode, using the current knob to regulate current through that wire.

Short the other wire across another supply, again letting the supply go into constant current mode, using the current knob to regulate current through the wire.

Finally, hook the final supply between the two wires.  With no other current path between the wires, this supply will be in constant voltage mode, and you can adjust the voltage between the wires using this.

You can connect a ground reference to any one place in the circuit, if desired.

Repeating for emphasis, this strategy only works if the supplies have isolated outputs (which is common among lab power supplies).
 

Offline gus789Topic starter

  • Contributor
  • Posts: 47
  • Country: us
Quote
Are both wires supposed to be carrying current?

No, only one of the wires carries the current. The other is stationary.

Rufus - is the attached image more or less what you meant in terms of using a current sink? Also looked at the TL431 current sink but I'd probably go with the op-amp one.

Thanks!
 

Offline Rufus

  • Super Contributor
  • ***
  • Posts: 2095
Rufus - is the attached image more or less what you meant in terms of using a current sink?

Yes but the TL431 is supposed to be used as a 2.5v reference (above -15v) which the pot goes across to give better stability of the sunk current.

The op-amp could use 0v for its +ve supply you don't really need +15v.

You need to decide on what wire current range you want to design this properly.

With that circuit you have 1v/mA across the sense resistor and need to allow maybe 1v minimum across the MOSFET so -10v on the wire and 4mA through it is about the limit.

 

Offline gus789Topic starter

  • Contributor
  • Posts: 47
  • Country: us
Quote
Yes but the TL431 is supposed to be used as a 2.5v reference (above -15v) which the pot goes across to give better stability of the sunk current.
Got it! Will have to order some of these.

Quote
The op-amp could use 0v for its +ve supply you don't really need +15v.
I thought about that initially, but was worried that the positive voltage on the wire could cause some strange behavior.

Quote
You need to decide on what wire current range you want to design this properly.
Indeed - that will depend on how strong my magnet is, which I hope test tomorrow by slapping this circuit together and testing it.  But I suspect 4mA should be more than enough for the purpose. I don't want the wire to heat up too much.

Thanks a lot for all the comments and suggestions!
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf