Author Topic: BJT/mosfet/jfet minimal usable current?  (Read 2186 times)

0 Members and 1 Guest are viewing this topic.

Offline exeTopic starter

  • Supporter
  • ****
  • Posts: 2647
  • Country: nl
  • self-educated hobbyist
BJT/mosfet/jfet minimal usable current?
« on: July 19, 2019, 10:52:20 am »
Hello!

I'm designing a low current source (from 100nA to 10mA, may be in multiple ranges). I'm trying to choose best pass element for the circuit. So, candidates are:
1) bjt
2) mosfet
2.5) jfet
3) no pass element at all, use opamp output directly
4) use resistor as "pass" element and pull voltage down with an opamp

My question is: are bjts and mosfets suitable for low currents? (assuming minimum required current is above leakage current). Like, any noise, bandwidth? What Hfe/transconductance to expect? I use low-power devices, like 2n3904 and 2n2700. I also have some small signal jfets like j112.
 

Offline Berni

  • Super Contributor
  • ***
  • Posts: 5050
  • Country: si
Re: BJT/mosfet/jfet minimal usable current?
« Reply #1 on: July 19, 2019, 11:32:11 am »
Nothing wrong with simply using the output of an opamp at these low currents.

Tho the compliance voltage also plays a big part in it. Its a big difference being able to source current at 1mA 5V and another sourcing 1mA 1000V

With the <1uA currents you will need careful design of your circuit to be able to accurately measure and regulate such a low current. And what kind of current source it is matters too. You can have current sources that only sink to ground, ones that can supply only positive voltages into a load or even 4 quadrant sources where its possible to sink or source current at a positive or negative output voltage. Also high or low side current measurement can have a big impact on design.
 

Offline exeTopic starter

  • Supporter
  • ****
  • Posts: 2647
  • Country: nl
  • self-educated hobbyist
Re: BJT/mosfet/jfet minimal usable current?
« Reply #2 on: July 19, 2019, 12:23:34 pm »
Thank you very much for the answer. Compliance voltage is 12V because I'm afraid of working with high voltages. Originally I wanted a high-side current sensing, but now I'm more into low-side sensing. This will make voltage regulation a bit harder, but I don't need much precision. I'd like to control voltage and current from DAC, so reference voltages are referenced to ground. The whole project is just for fun.

PS I wonder how good tl072 for such applications (I don't have a better opamp at hand with high-impedance inputs).
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 22436
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: BJT/mosfet/jfet minimal usable current?
« Reply #3 on: July 19, 2019, 12:44:47 pm »
Amp will be limited by noise in a single range; with switched ranges (i.e., for a Howland current pump*, a selectable series output resistor), a TL072 wouldn't be bad.  I'd think that full range should be reasonable with it.

*The traditional form of this is with symmetrical resistors, but that fully halves the compliance range.  Fortunately, the "top" resistors can be reduced proportionally (or something like that, I forget?) to give the same result with compliance range arbitrarily close to the amp's actual output range.  Note that this exaggerates errors (increased noise gain), making the output noisier, and making trimming attractive (for amp offset and resistor tolerance).

Otherwise, there are small JFETs, like PN4116 series, which would actually have trouble reaching 10mA so you might pick something not quite as tiny, but that are guaranteed for very low leakage.  Small BJTs have good performance, but aren't often rated for it; for example a typical 2N4401 leaks nanoamps at room temperature, but you'll never find a datasheet that agrees.  Apparently some RF types are quite good (which makes sense from the small junction size), but be careful about voltage rating (it's often low; that's part of how they get the speed up) and parasitic oscillation (probably use a ferrite bead or resistor in series with the base/gate).

MOSFETs are similar, usually pretty good but not always guaranteed as such, and smaller is better.  2N7000 is pretty much your introductory jellybean MOSFET, and typically leaks ~nA when fully cut off.  (Note that subthreshold current flow applies -- there isn't a hard pinch-off for MOSFETs, the drain current just keeps falling exponentially below the quadratic current flow region.  Vgs(th) is defined at some modest current in this range; expect to need Vgs(off) much lower than this.  For enhancement mode MOSFETs, drain current is usually low enough to be considered leakage current, for Vgs near zero -- no negative bias needed.)

And not that you would, but just for completeness: vacuum tubes leak quite a lot, I think due to their hot operation (weak emission from everywhere, not just the intended cathode surface).  The dynamic range (cutoff to full conduction) isn't very good, certainly fewer orders of magnitude than you are asking for here.  Though it would be interesting, I guess, to think of ways to trick it into working anyway.  Like, the heater voltage could be reduced on the low ranges, to reduce current capacity and gain, and hopefully leakage.  Or maybe a high voltage or RF type gives lower leakage, again in normal or reduced operation.  If very low bandwidth is acceptable, a high voltage diode could certainly be used, varying the heater voltage to set output current (the thermal time constant of such a heater is fractional seconds).

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 
The following users thanked this post: exe

Offline MagicSmoker

  • Super Contributor
  • ***
  • Posts: 1408
  • Country: us
Re: BJT/mosfet/jfet minimal usable current?
« Reply #4 on: July 19, 2019, 12:55:50 pm »
I almost dashed off a reply that a Howland current source/sink using a CMOS (or JFET) input op-amp would be much better than a discrete transistor based current source, but then I remembered that the effective output impedance depends on the matching of the resistor ratios as well as their absolute values, and to achieve the effective Zout of 120M (12V / 100nA) would require impractically precise matching and/or high resistance values (defined by me as higher than 10M).

The below circuit at Analog Devices looks more practical, though cleanliness of the pcb and guard rings will almost certainly be needed if any kind of accuracy at 100nA is required:

https://www.analog.com/en/design-center/reference-designs/circuit-collections/precision-nanoamp-bidirectional-current-source.html#cc-overview
 
The following users thanked this post: exe

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 15156
  • Country: de
Re: BJT/mosfet/jfet minimal usable current?
« Reply #5 on: July 19, 2019, 01:03:35 pm »
The Tl072 is relatively high drift and noisy at low frequency. So it's not a really good choice.
 
The following users thanked this post: exe

Offline exeTopic starter

  • Supporter
  • ****
  • Posts: 2647
  • Country: nl
  • self-educated hobbyist
Re: BJT/mosfet/jfet minimal usable current?
« Reply #6 on: July 19, 2019, 02:01:05 pm »
Thank you guys very much for detailed answers.

The reasons I don't want current pump are: 1) my load is floating, so low-side sensing is fine 2) it doesn't let setting compliance voltage. As for output impedance I don't really need precise output because I don't have precise equipment to measure it. I'm also not sure if Howland circuit is easy to modify the gain (= multiple ranges). Otherwise I like it.

> The below circuit at Analog Devices looks more practical

Saw it, thank you. Doesn't let setting compliance voltage either. The performance they claim is very impressive to me. It is also bipolar, which is cool, but not required for me :). My understanding is it's not that different from a typical current sink, apart from the differential amplifier and buffer to provide ground-level reference.

I probably don't really compliance voltage to be adjustable, but for fun why not. I'll try to build a circuit with output directly tied to opamp for the sake of simplicity. May be not that great because of self-heating, but at least it will give some taste of what it feels like working with nanoamps :).
 

Offline exeTopic starter

  • Supporter
  • ****
  • Posts: 2647
  • Country: nl
  • self-educated hobbyist
Re: BJT/mosfet/jfet minimal usable current?
« Reply #7 on: July 19, 2019, 07:55:14 pm »
Wow, I just hooked LMC662AIM opamp, ~1M resistor, made a typical current sink and it worked. At least it shows 0.09-0.11uA when measured by my DMMs. I didn't even clean the rosin flux.

I also measured current by connecting DMM with ~10M impedance in voltage mode, shows 0.98V for ~0.099 Vset (tried to dial-in 0.1V on my analog power supply). So, there is some error, but I'm more than happy, considering there is no decoupling, long leads, and thin ground wire is shared between Vsupply and Vset. Power supply was unipolar 12V.

Opamp seems to be noisy in specs, but I don't know if I should worry about 22nV*Hz^(1/2). I mostly play with bjts, leds, etc. So I don't think noise is a problem. Do you guys have a suggestion for a better opamp that wouldn't break the bank?

Next step gonna be hook it to stm32f373 with integrated DAC and ADC, add a display and it will be a beast.

UP: ah, I know why I have LMC662 on my shelf: I tried to build this circuit, but wasn't able to get 1G resistor: http://www.vk2zay.net/article/251 . So, if it's good for picoamps, then it's good for nano-amps.
« Last Edit: July 19, 2019, 08:03:06 pm by exe »
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 15156
  • Country: de
Re: BJT/mosfet/jfet minimal usable current?
« Reply #8 on: July 19, 2019, 08:43:29 pm »
As so often its a balance between current noise and voltage noise. So the best OP depends on the impedance / current range.
A few good candidates, about with increasing price tag:
TLC272, TLC277,  LMC6482, OPA197, max4238, LTC2050, OPA145

The AZ OPs (max4238, LTC2050) may be interesting for DC / low frequency. They do however have a little more bias and current noise and also higher frequency voltage noise.

For the noise, my feeling is that the normal higher frequency noise is the lesser problem. The more critical part could be 1/f noise and thus higher noise at a low frequency (e.g. 0.1 Hz). Low current experiments tend to be more on the slow side.

For the configuration, I would also consider having a TIA at the low side and just drive the high side with an OP.
 
The following users thanked this post: exe

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 17428
  • Country: us
  • DavidH
Re: BJT/mosfet/jfet minimal usable current?
« Reply #9 on: July 19, 2019, 09:51:34 pm »
Otherwise, there are small JFETs, like PN4116 series, which would actually have trouble reaching 10mA so you might pick something not quite as tiny, but that are guaranteed for very low leakage.

PN4116s could be used in parallel or one could drive a bipolar transistor.

Quote
Small BJTs have good performance, but aren't often rated for it; for example a typical 2N4401 leaks nanoamps at room temperature, but you'll never find a datasheet that agrees.

Typically common bipolar parts like the 2N5087/2N5089 would be used at low currents but you are right, the datasheets will never give guaranteed leakage specifications that low simply because testing at low currents costs too much.  The only common low current tested part I am aware of is the 2N4117 JFET and even it lacks a cutoff current specification.

The BC550/BC560 may also be suitable.  Look for bipolar parts which have specified and high hfe at low collector currents.  Also look for superbeta parts and ones intended for logging applications.  It is actually pretty amazing how low you can go.

For what it is worth, common small signal Darlington transistors can work down to at least 1 microamp.



The Howland current pump is not the only way to use an operational amplifier to make a current source.  It just applies for a bilateral current source.

Oh, and for such a large output range, I would use a bipolar anti-log converter to get rid of the resistor and range switching.  Forget trying to go from 10 milliamps to 10 nanoamps with resistors unless you want a lot of current shunt switching.  Or look up how multimeter ohms converters work if you must use that method.
« Last Edit: July 19, 2019, 09:59:19 pm by David Hess »
 
The following users thanked this post: T3sl4co1l


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf