Author Topic: Analog devices ADA4805 opamp offset voltage  (Read 1871 times)

0 Members and 1 Guest are viewing this topic.

Offline brumbarchrisTopic starter

  • Regular Contributor
  • *
  • Posts: 223
  • Country: ro
Analog devices ADA4805 opamp offset voltage
« on: March 24, 2020, 12:50:28 pm »
Hello everyone,
Please consider the schematic and layout portion I represent in the pictures below. It is essentially a non-inverting amplifier, with a gain of some 55, built around the precision Analog Devices ADA4805 opamp . "AMP_OUT" output is disconnected from load. "VC_OUT" input is pulled to ground by making a short circuit across D312 (nothing drives it, anyway). Yet, for some reason, the "AMP_OUT" output measures -50mV, while I would have expected something in the -6.8...+6.8mV (max opamp offset voltage is 125uV at room temperature multiplied by the circuit amplification factor, which is 55). I have checked that pins 3 and 4 of the opamp measure really 0.0000V. I measured with several instruments. I replaced the opamp with a new one. The power supplies are clean, with no unwanted artifacts on them. I suppose layout does matter, but to what extent? There is a solid GND plane (not illustrated in the attachment) on the very next layer under the surface layer represented in the picture.

So why this large offset outside of the expected range?
« Last Edit: March 24, 2020, 02:49:59 pm by brumbarchris »
 

Offline Vovk_Z

  • Super Contributor
  • ***
  • Posts: 1478
  • Country: ua
Re: Analog devices ADA4805 opamp offset voltage
« Reply #1 on: March 24, 2020, 01:15:01 pm »
Does R340 really have 1.8k (have you checked it's value, and R341)?
 

Online Wolfram

  • Frequent Contributor
  • **
  • Posts: 414
  • Country: no
Re: Analog devices ADA4805 opamp offset voltage
« Reply #2 on: March 24, 2020, 01:29:32 pm »
Beware that this amplifier has a maximum input bias current of 800 nA, which would lead to an offset of 1.4 mV across 1k8||100k. If this was the cause of your problem, it would also be observable when measuring the voltage on the input pins.
 

Offline Vovk_Z

  • Super Contributor
  • ***
  • Posts: 1478
  • Country: ua
Re: Analog devices ADA4805 opamp offset voltage
« Reply #3 on: March 24, 2020, 01:51:20 pm »
+1. I just wanted to say the same.
 Both input pins have to have some voltage. For example, pin 3 = 0,0 mV (because it's grouded), but pin 4 = 0,9 mV.
 

Offline schmitt trigger

  • Super Contributor
  • ***
  • Posts: 2431
  • Country: mx
Re: Analog devices ADA4805 opamp offset voltage
« Reply #4 on: March 24, 2020, 02:01:34 pm »
Could you momentarily short R341 and measure again?

Sorry, did not read the whole datasheet. This only works if the  ADA4805 is unity gain stable
 

Offline Vovk_Z

  • Super Contributor
  • ***
  • Posts: 1478
  • Country: ua
Re: Analog devices ADA4805 opamp offset voltage
« Reply #5 on: March 24, 2020, 02:09:00 pm »
It is unity-gain stable. So TS has to have about 0.9 mV on pin1 (out) with shorted R341.

It seems that we have a pretty close to typical current here.
470 nA is "Low" dependably on with what opamp we compare. There are a lot of opamps with pico- and fetto-amps bias current.
« Last Edit: March 24, 2020, 02:18:02 pm by Vovk_Z »
 

Offline StillTrying

  • Super Contributor
  • ***
  • Posts: 2850
  • Country: se
  • Country: Broken Britain
Re: Analog devices ADA4805 opamp offset voltage
« Reply #6 on: March 24, 2020, 02:25:25 pm »
R340 being open circuit would just about explain it, -0.50uA input bias current through the 100k = -50mV.
.  That took much longer than I thought it would.
 

Online SiliconWizard

  • Super Contributor
  • ***
  • Posts: 15800
  • Country: fr
Re: Analog devices ADA4805 opamp offset voltage
« Reply #7 on: March 24, 2020, 02:51:17 pm »
R340 being open circuit would just about explain it, -0.50uA input bias current through the 100k = -50mV.

Yep.
 

Offline brumbarchrisTopic starter

  • Regular Contributor
  • *
  • Posts: 223
  • Country: ro
Re: Analog devices ADA4805 opamp offset voltage
« Reply #8 on: March 24, 2020, 03:18:42 pm »
R340 is not open, I just triple-checked it.

I took a multimeter with more digits and better precision and I measured the input pins again (in the original configuration):
Pin 3 (non-inverting): +0.0073mV
Pin 4 (inverting):        +0.0473mV
Output:                -50.2mV
Having reconsidered the effect of the bias current, as some of you guys pointed out, I actually believe the limits for the output should now be +/-84mV. So the output is indeed within limits, but I would have expected to see the effect of the input bias current at the input pins themselves. At such a large output voltage, I would have expected to see some 0.77mV actual difference between the two input pins, which I do not see.

Changed R340 to 180R and R341 to 10k, so as to keep their ratio constant, and measured again:
Pin 3 (non-inverting): +0.0076mV
Pin 4 (inverting):       +0.045mV
Output:                -2.7mV
In this case the limits are +/-14mV. So well within limits in this case.


If I just reconfigure it to a voltage repeater (R340 removed and R341 short-circuit), I get:
Pin 3 (non-inverting): +0.0085mV
Pin 4 (inverting):       +0.05mV
Output:                +0.05mV
In this case the limits are +/-0.125mV. So well within limits in this case.


Why is the effect of the input bias current not visible on the input pins, in the case of the original configuration?

Best regards,
Cristian
« Last Edit: March 24, 2020, 03:38:50 pm by brumbarchris »
 

Offline voltsandjolts

  • Supporter
  • ****
  • Posts: 2549
  • Country: gb
Re: Analog devices ADA4805 opamp offset voltage
« Reply #9 on: March 24, 2020, 03:20:23 pm »
Also might be worth checking the op-amp output is indeed steady-state DC.
i.e. measuring with a DMM might show some offset that is actually an AC signal caused by e.g. coupled noise from DCDC converter.
 

Offline brumbarchrisTopic starter

  • Regular Contributor
  • *
  • Posts: 223
  • Country: ro
Re: Analog devices ADA4805 opamp offset voltage
« Reply #10 on: March 24, 2020, 03:39:26 pm »
I've checked with an oscilloscope, it is really DC.
 

Offline StillTrying

  • Super Contributor
  • ***
  • Posts: 2850
  • Country: se
  • Country: Broken Britain
Re: Analog devices ADA4805 opamp offset voltage
« Reply #11 on: March 24, 2020, 03:47:18 pm »
"R340 is not open, I just triple-checked it."

Have you checked its 0V end is connected to a good 0V (on resistance).
You do have VC_OUT and AMP_OUT following a long parallel path, but if you've checked the -50mV is not oscillation there's not much left. The +/- 5V decoupling doesn't look tooooo bad.

Both input pins have to have some voltage. For example, pin 3 = 0,0 mV (because it's grouded), but pin 4 = 0,9 mV.

I think this could be it, 0.55uA of bias current flowing though the grounded +IN is 0V, but 0.55uA flowing through the 1.8k|100k = 0.98mV, times 55 = 54mV.
So the +IN needs 1.8k in series with it to cancel the 2 input bias current voltage offsets.

« Last Edit: March 24, 2020, 04:29:19 pm by StillTrying »
.  That took much longer than I thought it would.
 

Online SiliconWizard

  • Super Contributor
  • ***
  • Posts: 15800
  • Country: fr
Re: Analog devices ADA4805 opamp offset voltage
« Reply #12 on: March 24, 2020, 05:16:03 pm »
So the +IN needs 1.8k in series with it to cancel the 2 input bias current voltage offsets.

I would also try that. (Unless I'm mistaken, I think that would be more like 1.8k // 100k, but for all practical purposes, it won't make a big difference.)
 

Offline Vovk_Z

  • Super Contributor
  • ***
  • Posts: 1478
  • Country: ua
Re: Analog devices ADA4805 opamp offset voltage
« Reply #13 on: March 24, 2020, 06:31:55 pm »
That opamp definitely doesn't like large resistor nominals, smaller R have to be about 20-200 Ohms (look at the datasheet typical circuits). It can be even better to use 18 R at that place, I mean 1000R/18R (and 18 R at another input).
« Last Edit: March 24, 2020, 06:35:21 pm by Vovk_Z »
 

Offline brumbarchrisTopic starter

  • Regular Contributor
  • *
  • Posts: 223
  • Country: ro
Re: Analog devices ADA4805 opamp offset voltage
« Reply #14 on: March 26, 2020, 08:49:47 am »
Thank you all for your great help. Indeed I can mitigate the effect of the input bias current by decreasing the resistor values. However, what still puzzles me, is that I cannot see the effect of this offset current when measuring the voltage at the inputs of the opamp. Just to reiterate, the measurement results with various resistor configuration are as follows:


Results for R340=1800R and R341=100k:
Pin 3 (non-inverting): +0.0073mV
Pin 4 (inverting):        +0.0473mV
Output:                -50.2mV

Results for R340=180R and R341=10k:
Pin 3 (non-inverting): +0.0076mV
Pin 4 (inverting):       +0.045mV
Output:                -2.7mV

Results for R340=open and R341=short:
Pin 3 (non-inverting): +0.0085mV
Pin 4 (inverting):       +0.05mV
Output:                +0.05mV

I would have expected to see a significantly larger voltage difference between the inverting and non-inverting inputs of the opmap in the first case (R340=1800R and R341=100k), something around 0.77mV. However, that is not the case. Any idea why?

Best regards,
Cristian
 

Offline Vovk_Z

  • Super Contributor
  • ***
  • Posts: 1478
  • Country: ua
Re: Analog devices ADA4805 opamp offset voltage
« Reply #15 on: March 26, 2020, 09:26:16 am »
It can depends on your ground point.
 

Offline brumbarchrisTopic starter

  • Regular Contributor
  • *
  • Posts: 223
  • Country: ro
Re: Analog devices ADA4805 opamp offset voltage
« Reply #16 on: March 26, 2020, 09:48:32 am »
I have tried several: GND point of D312, GND point of R340, GND point of the decoupling caps, they all show similar readings, no notable difference.
 

Offline Jay_Diddy_B

  • Super Contributor
  • ***
  • Posts: 2766
  • Country: ca
Re: Analog devices ADA4805 opamp offset voltage
« Reply #17 on: March 26, 2020, 10:52:28 am »
Hi,

The issue here is bias currents. These are included in the LTspice model for AD4805.

Model



I have two models, one is the OP original circuit. It gives an output voltage of 47.5mV, V(out)

In the second model I have included R3 so the impedance at both of the input terminals is approximately equal. The output, V(out_2) is 430uV

Results





Try replacing your input short with a 1.8k \$\Omega\$ resistor?

Regards,
Jay+Diddy_B

 

Offline StillTrying

  • Super Contributor
  • ***
  • Posts: 2850
  • Country: se
  • Country: Broken Britain
Re: Analog devices ADA4805 opamp offset voltage
« Reply #18 on: March 26, 2020, 11:46:33 am »
<snipped junk idea> :)

It just needs the inputs' DC resistances matching.

« Last Edit: March 26, 2020, 12:19:08 pm by StillTrying »
.  That took much longer than I thought it would.
 

Offline Kevin.D

  • Frequent Contributor
  • **
  • Posts: 290
  • Country: england
Re: Analog devices ADA4805 opamp offset voltage
« Reply #19 on: March 26, 2020, 12:16:10 pm »

I would have expected to see a significantly larger voltage difference between the inverting and non-inverting inputs of the opmap in the first case (R340=1800R and R341=100k), something around 0.77mV. However, that is not the case. Any idea why?

Best regards,
Cristian

When neg feedback is applied  the differential input voltage measured at the input feedback node  will be reduced approx by factor of the loop gain (Vin=Vin/(loop gain+1). Whats left is what I call the 'loop gain error voltage' referred to the input which  for an ideal amplifier with infinite voltage gain would be zero.
The 'Loop gain' is the difference between amplifier open loop gain and the signal gain,  e.g At DC Opamp   Aol= 100dB   signal voltage gain = 35dB  so loop gain= 100-35 =65dB your diff input voltage will be reduced by factor of ~ 65dB.
« Last Edit: March 26, 2020, 12:17:57 pm by Kevin.D »
 

Offline brumbarchrisTopic starter

  • Regular Contributor
  • *
  • Posts: 223
  • Country: ro
Re: Analog devices ADA4805 opamp offset voltage
« Reply #20 on: March 26, 2020, 01:55:20 pm »
Everything is now crystal clear regarding the quations I had. Thank you very much to everybody who contributed.

Best regards,
Cristian
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf