Author Topic: Nanoamp current source troubleshooting  (Read 4691 times)

0 Members and 1 Guest are viewing this topic.

Offline ZeroResistanceTopic starter

  • Frequent Contributor
  • **
  • Posts: 585
  • Country: gb
Nanoamp current source troubleshooting
« on: April 11, 2019, 06:41:20 am »
I built a simple nanoamp current source.
Op amp is the MCP602. The circuit runs on 3VDC.
I have attached the circuit diagram, R7 is the load through which I measure the current.
Instead of V2 I have used LM4041 For the reference, its a 1.225V fixed reference. The series resistor for the reference is 22K (not shown in schematic).

The input at the voltage divider node R8/R9 shows around 12mV in both LTspice and the practical circuit. However when I check the output voltage across the load the real world circuit shows around 130uV while LTspice shows 13uV?

What would be the cause for this anomaly?
« Last Edit: April 11, 2019, 01:51:08 pm by ZeroResistance »
 

Offline ZeroResistanceTopic starter

  • Frequent Contributor
  • **
  • Posts: 585
  • Country: gb
Re: Nanoamp current source troubleshooting
« Reply #1 on: April 11, 2019, 08:05:28 am »
So here are the different voltages LTSpice vs Practical

MCP602
Vout    6.95mV     693uV
Vin-      136mV     659uV
Vin+     1.12mV    556uV

Vref      1.224V
Vref div 11.96mV    12.117mV
VL          133uV        13.3uV

The first column is the practical circuit and the 2nd column is LTSpice. For some reason the Vin- and Vin+ is showing a huge variation. Probably due to the multimeter leads.
« Last Edit: April 11, 2019, 08:08:34 am by ZeroResistance »
 

Offline exe

  • Supporter
  • ****
  • Posts: 2647
  • Country: nl
  • self-educated hobbyist
Re: Nanoamp current source troubleshooting
« Reply #2 on: April 11, 2019, 01:25:59 pm »
I'd try a bipolar power supply for your opamp. The output may not be able to swing that wide. Or, you can increase R4 or add a diode to increase required voltage drop.
 
The following users thanked this post: ZeroResistance

Offline ZeroResistanceTopic starter

  • Frequent Contributor
  • **
  • Posts: 585
  • Country: gb
Re: Nanoamp current source troubleshooting
« Reply #3 on: April 11, 2019, 01:46:31 pm »
I'd try a bipolar power supply for your opamp. The output may not be able to swing that wide. Or, you can increase R4 or add a diode to increase required voltage drop.

Yes I thought so too, but I guess the MCP602 is a single supply opamp. Also when I check the datasheet of the Op-amp the Common mode input range is
Min: Vss- 0.3
Max: Vdd-1.2
So I think the operating range should fall within the spec of the Op-amp.
I  am powering the op-amp using a 3V supply.

I  didn't get the part of increasing R4 or adding a diode?
« Last Edit: April 11, 2019, 01:49:27 pm by ZeroResistance »
 

Offline exe

  • Supporter
  • ****
  • Posts: 2647
  • Country: nl
  • self-educated hobbyist
Re: Nanoamp current source troubleshooting
« Reply #4 on: April 11, 2019, 02:08:56 pm »
I'd try a bipolar power supply for your opamp. The output may not be able to swing that wide. Or, you can increase R4 or add a diode to increase required voltage drop.

Yes I thought so too, but I guess the MCP602 is a single supply opamp. Also when I check the datasheet of the Op-amp the Common mode input range is
Min: Vss- 0.3
Max: Vdd-1.2
So I think the operating range should fall within the spec of the Op-amp.
I  am powering the op-amp using a 3V supply.

I  didn't get the part of increasing R4 or adding a diode?

The input may be fine, but output is not. You are asking your opamp to swing down to less than 1mv to ground. I never seen an opamp that could do that. So, my theory is, you want impossible from opamp. Solutions: 1) bipolar supply 2) move operating point so that output is within the range of opamp capability. So, if you increase R4 then for the same current output should be higher, right? If not, then forget this suggestion. Same with diode, if you place diode on output, then opamp will have to raise the output further away from ground to "overcome" additional dropout. Dave has a video about these technique, in this video:
 
The following users thanked this post: ZeroResistance

Offline ZeroResistanceTopic starter

  • Frequent Contributor
  • **
  • Posts: 585
  • Country: gb
Re: Nanoamp current source troubleshooting
« Reply #5 on: April 11, 2019, 03:45:46 pm »
The input may be fine, but output is not.

Seems like you are on the dot here.
Don't know how I missed that  :palm:

 

Offline ZeroResistanceTopic starter

  • Frequent Contributor
  • **
  • Posts: 585
  • Country: gb
Re: Nanoamp current source troubleshooting
« Reply #6 on: May 03, 2019, 02:24:29 pm »
I'd try a bipolar power supply for your opamp. The output may not be able to swing that wide. Or, you can increase R4 or add a diode to increase required voltage drop.

Would it be possible to use the existing opamp MCP602 which is not bipolar and shift the operating point to VDD/2 so that the opamp operates at the center of the supply rails ?
 

Offline exe

  • Supporter
  • ****
  • Posts: 2647
  • Country: nl
  • self-educated hobbyist
Re: Nanoamp current source troubleshooting
« Reply #7 on: May 03, 2019, 03:44:49 pm »
Would it be possible to use the existing opamp MCP602 which is not bipolar and shift the operating point to VDD/2 so that the opamp operates at the center of the supply rails ?

Short answer is yes :). The thing is, division between bipolar and non-bipolar opamps is... sort of artificial. An opamp doesn't care about power supply at all as long as input/output within the range of what the opamp can do. So, bipolar  power supply just offsets the input signal to make it within the range (from opamps point of view) and extends output range. So, when adding, say, -3V supply to opamp, for opamp the input signal becomes shifted by +3V. May be I'm not the best to explain this, try watching some youtube videos, they should be better at it.
 
The following users thanked this post: ZeroResistance

Offline ZeroResistanceTopic starter

  • Frequent Contributor
  • **
  • Posts: 585
  • Country: gb
Re: Nanoamp current source troubleshooting
« Reply #8 on: May 04, 2019, 06:06:02 am »
Ok so I changed the circuit to use dual supply's but after this the circuit does not respond well to when I change the input voltage? It does respond well to changes in output load, it keeps the current constant there.
But when I change the Input voltage from the reference the current does not linearly increase?
 

Offline mvs

  • Frequent Contributor
  • **
  • Posts: 370
  • Country: de
Re: Nanoamp current source troubleshooting
« Reply #9 on: May 04, 2019, 07:48:26 am »
But when I change the Input voltage from the reference the current does not linearly increase?
It should be linear.  But linear means in general case f(x)=a*x+b and your b is not zero.
MCP602 has +/-2mV input offset voltage. It is around +/- 45nA of output current in this schematic.
 
The following users thanked this post: ZeroResistance

Offline ZeroResistanceTopic starter

  • Frequent Contributor
  • **
  • Posts: 585
  • Country: gb
Re: Nanoamp current source troubleshooting
« Reply #10 on: May 04, 2019, 08:31:35 am »
But when I change the Input voltage from the reference the current does not linearly increase?
It should be linear.  But linear means in general case f(x)=a*x+b and your b is not zero.
MCP602 has +/-2mV input offset voltage. It is around +/- 45nA of output current in this schematic.

Woah! +/-45nA is huge! I am practically trying it out with an LMC662 and that has a offset voltage like 6mV!!
Also caught an error in the schematic the lead R7 had to be 1Meg i have entered 1M and Ltspice considers that as milli :-/O
 

Offline mvs

  • Frequent Contributor
  • **
  • Posts: 370
  • Country: de
Re: Nanoamp current source troubleshooting
« Reply #11 on: May 04, 2019, 10:34:36 am »
Input offset voltage is more or less constant, so you can try to add correction and calibrate it out.
Or you can use chopper stabilized opamp (Zero-Drift, Auto-Zero, etc) like MCP6V* series.
 
The following users thanked this post: ZeroResistance

Offline Kalvin

  • Super Contributor
  • ***
  • Posts: 2145
  • Country: fi
  • Embedded SW/HW.
Re: Nanoamp current source troubleshooting
« Reply #12 on: May 04, 2019, 01:41:38 pm »
In this particular case where using a battery, it is possible to introduce approx. 200 mV offset voltage between the signal ground and the negative power supply using a simple [schottky] diode.

Without the offset voltage created by the diode, the rail-to-rail op amp will not be able to swing its output all the way down to the signal ground. However, introducing the diode, the negative power supply is now about 200 mV below the signal ground and the op amp is now able to swing its output down to signal ground potential without difficulty.

Although the signal reference ground will have some ripple (60mV in this particular case) relative to the op amp's negative power supply even with the 10uF bypass capacitor, the typical power supply rejection ratio (PSRR) of an op amp will be able to take care of this "problem". Similarly, any temperature effects introduced by the diode will be attenuated by the op amp's PSRR.

Br,
Kalvin
« Last Edit: May 04, 2019, 02:22:16 pm by Kalvin »
 
The following users thanked this post: ZeroResistance

Offline ZeroResistanceTopic starter

  • Frequent Contributor
  • **
  • Posts: 585
  • Country: gb
Re: Nanoamp current source troubleshooting
« Reply #13 on: May 04, 2019, 04:02:39 pm »
In this particular case where using a battery, it is possible to introduce approx. 200 mV offset voltage between the signal ground and the negative power supply using a simple [schottky] diode.
I had done something similar and tried it out practically, with the LMC662, based on what exe had suggested above. Basically he said to insert a diode in series with the load so that the op-amp works extra and is forced to raise the output above 0.7V, above its output swing limit of 0.1V, however I found something interesting in the datasheet.

The output swing seems to be load dependant and seems to go as high as 0.7V when the load is 600ohm or so. Pls have a look at the attached image.

I inserted a 1N4148 diode, and then 2 and then went upto 4 in series to mitigate this issue. However I found that at loads above 100K to 1Meg, the circuit wasn't performing well, I sort of thought that well the diode wasn't able to conduct at these low currents, and gave up on the idea. But I may well be wrong.
 

Offline exe

  • Supporter
  • ****
  • Posts: 2647
  • Country: nl
  • self-educated hobbyist
Re: Nanoamp current source troubleshooting
« Reply #14 on: May 05, 2019, 08:04:25 pm »
The output swing seems to be load dependant and seems to go as high as 0.7V when the load is 600ohm or so. Pls have a look at the attached image.

To me it looks like Vmin and Vmax are swapped in the datasheet. Anyway, that's expected, as there is a current limiting resistor. The more current you draw, the more dropout there. But it shouldn't be a problem when there are nanoaps.

I inserted a 1N4148 diode, and then 2 and then went upto 4 in series to mitigate this issue. However I found that at loads above 100K to 1Meg, the circuit wasn't performing well, I sort of thought that well the diode wasn't able to conduct at these low currents, and gave up on the idea. But I may well be wrong.

Did you check on a real circuit, or in a simulator?
 

Offline ZeroResistanceTopic starter

  • Frequent Contributor
  • **
  • Posts: 585
  • Country: gb
Re: Nanoamp current source troubleshooting
« Reply #15 on: May 06, 2019, 08:34:40 pm »
Did you check on a real circuit, or in a simulator?
I checked in a real circuit.
My next step is to connect a bipolar supply to it. Hopefully I won't run into any issues.
 

Offline exe

  • Supporter
  • ****
  • Posts: 2647
  • Country: nl
  • self-educated hobbyist
Re: Nanoamp current source troubleshooting
« Reply #16 on: May 06, 2019, 09:33:00 pm »
If it's a real circuit you can measure current and voltage drop on diodes, as well as opamp output voltage swing.
 

Offline ZeroResistanceTopic starter

  • Frequent Contributor
  • **
  • Posts: 585
  • Country: gb
Re: Nanoamp current source troubleshooting
« Reply #17 on: May 09, 2019, 01:17:23 pm »
In this particular case where using a battery, it is possible to introduce approx. 200 mV offset voltage between the signal ground and the negative power supply using a simple [schottky] diode.

Without the offset voltage created by the diode, the rail-to-rail op amp will not be able to swing its output all the way down to the signal ground. However, introducing the diode, the negative power supply is now about 200 mV below the signal ground and the op amp is now able to swing its output down to signal ground potential without difficulty.

Although the signal reference ground will have some ripple (60mV in this particular case) relative to the op amp's negative power supply even with the 10uF bypass capacitor, the typical power supply rejection ratio (PSRR) of an op amp will be able to take care of this "problem". Similarly, any temperature effects introduced by the diode will be attenuated by the op amp's PSRR.

Br,
Kalvin

I built the circuit as suggested. And then connected another 2 stages across, the load 1 a unity gain buffer and then a 100 gain non-inverting amp however at the final output I can see 50Khz noise, this noise seems to be riding on the dc signal, around 200mV p-p. My input is a 20mV dc signal.
I was hoping to get some mains noise instead I end up with 50Khz noise.
Circuit is powered by 6V battery and op-amp is LMC662
« Last Edit: May 09, 2019, 01:38:51 pm by ZeroResistance »
 

Offline exe

  • Supporter
  • ****
  • Posts: 2647
  • Country: nl
  • self-educated hobbyist
Re: Nanoamp current source troubleshooting
« Reply #18 on: May 09, 2019, 01:58:45 pm »
I built the circuit as suggested. And then connected another 2 stages across, the load 1 a unity gain buffer and then a 100 gain non-inverting amp however at the final output I can see 50Khz noise, this noise seems to be riding on the dc signal, around 200mV p-p. My input is a 20mV dc signal.
I was hoping to get some mains noise instead I end up with 50Khz noise.
Circuit is powered by 6V battery and op-amp is LMC662

Did you put compensation? Without compensation network it may not work stable as each element in feedback loop creates a delay. Delay == phase shift :). May be at 50KHz phase shift is 180 degrees.
 

Offline ZeroResistanceTopic starter

  • Frequent Contributor
  • **
  • Posts: 585
  • Country: gb
Re: Nanoamp current source troubleshooting
« Reply #19 on: May 09, 2019, 02:14:56 pm »
No I didn't put any compensation. Do I need to put in compensation for the 100 Gain amplifier?

« Last Edit: May 09, 2019, 02:21:07 pm by ZeroResistance »
 

Offline ZeroResistanceTopic starter

  • Frequent Contributor
  • **
  • Posts: 585
  • Country: gb
Re: Nanoamp current source troubleshooting
« Reply #20 on: May 09, 2019, 02:38:37 pm »
Did you put compensation? Without compensation network it may not work stable as each element in feedback loop creates a delay. Delay == phase shift :). May be at 50KHz phase shift is 180 degrees.

1. Would decoupling caps do the trick? I didn't bother to add decoupling since it was battery powered.
2. I thought that heavy capacitive loads on the op-amp output is one of the  factors that cause unstability. However I don't have any load here except a scope probe.
« Last Edit: May 09, 2019, 02:59:37 pm by ZeroResistance »
 

Offline Kalvin

  • Super Contributor
  • ***
  • Posts: 2145
  • Country: fi
  • Embedded SW/HW.
Re: Nanoamp current source troubleshooting
« Reply #21 on: May 09, 2019, 05:50:12 pm »
Try adding a 1 uF ... 10 uF capacitor in parallel with the BAT54. The capacitor will keep the ground reference stable and reduce noise. While you are at it, add 100 nF ... 1 uF capacitors across the op amps' U1, U2 and U3 power supply pins (although the circuit is battery powered) which will reduce noise in the power supply and reduce noise coupling in the circuit.

You can try to tame the oscillation further by adding a small capacitors (1 nF, for example) across U1's and U3's In- and Out. Just check with simulator that the circuit is still stable and performs as you expect. Tweak the capacitors as necessary in the real circuit.
 
The following users thanked this post: ZeroResistance

Offline ZeroResistanceTopic starter

  • Frequent Contributor
  • **
  • Posts: 585
  • Country: gb
Re: Nanoamp current source troubleshooting
« Reply #22 on: May 09, 2019, 08:21:40 pm »
Try adding a 1 uF ... 10 uF capacitor in parallel with the BAT54. The capacitor will keep the ground reference stable and reduce noise. While you are at it, add 100 nF ... 1 uF capacitors across the op amps' U1, U2 and U3 power supply pins (although the circuit is battery powered) which will reduce noise in the power supply and reduce noise coupling in the circuit.

You can try to tame the oscillation further by adding a small capacitors (1 nF, for example) across U1's and U3's In- and Out. Just check with simulator that the circuit is still stable and performs as you expect. Tweak the capacitors as necessary in the real circuit.

I add a 22uF cap across the schottky but didn't make any difference.
The 0.1uF decoupling made lot of difference and the 50Khz noise seems to have diminished now I get spikes at the output at 10ms intervals.
Please see attached.

 

Offline Kalvin

  • Super Contributor
  • ***
  • Posts: 2145
  • Country: fi
  • Embedded SW/HW.
Re: Nanoamp current source troubleshooting
« Reply #23 on: May 10, 2019, 07:20:21 am »
10 ms equals to 100 Hz, so your circuit is picking up some noise from the rectified 50 Hz mains.
 

Offline ZeroResistanceTopic starter

  • Frequent Contributor
  • **
  • Posts: 585
  • Country: gb
Re: Nanoamp current source troubleshooting
« Reply #24 on: May 10, 2019, 07:32:20 am »
10 ms equals to 100 Hz, so your circuit is picking up some noise from the rectified 50 Hz mains.
I have a laptop next to the circuit but that's again not connected to mains but running on battery. I can see the noise even on the input. Is there a way to reject the power supply noise?
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf