fear-Resistors in Microcurrent (before and after a OP-amp)

[Jannehrl]

Hello Community,
I am an apprentice and currently working on the microCurrent project from Dave. I do understand the whole circuit except
these 270? Resistors. I did ask my instructor, but neither did he understand why they are there. He said, it may be for a voltage drop, but he wasnt shure.
I highly appreciate any conclusion to solve this mystery.

In the attachement is the original schematic from Dave.

[Gyro]

No, I don't understand the reason for the value used for R12 (270) either.

It's purpose in an op-amp circuit is to minimise the offset due to input bias currents on the two op-amp inputs. It should be equal to the parallel combination of the feedback resistors (R3,6k8+R11,2k2) and the feedback resistor to ground (R5,1k), so around 900R would be appropriate.

Edit: I see there are several other 270R resistors (not sure which one you are asking about now), R4 is to limit the LED operating current and R8 and 10 are there to prevent the op-amps from oscillating if asked to drive a capacitive load. As there are other 270R resistors in the circuit then the choice of R12 is probably just arbitrary for cost reduction. 1k would have been a better choice, but the only other one in the circuit is a more expensive close tolerance one.

[zapta]

My guess is that R8 is to protect the op amp in case of a short of the output. It doesn't affect the gain.

[pmbrunelle]

No, I don't understand the reason for the value used for R12 (270) either.

It's purpose in an op-amp circuit is to minimise the offset due to input bias currents on the two op-amp inputs. It should be equal to the parallel combination of the feedback resistors (R3,6k8+R11,2k2) and the feedback resistor to ground (R5,1k), so around 900R would be appropriate.

This opamp draws almost no current from its inputs, so there's no need to make sure both inputs see the same DC source resistance.

Therefore, the resistor serves more of an ESD protection purpose as suggested by blueskull.

Standardizing parts values is a good way to avoid mistakes, reduce costs, etc.

[c4757p]

That resistor isn't going to offer much ESD protection. If you applied e.g. 1kV ESD, it would still permit 3.7A to flow, and that's assuming it's ideal. Remember, it's the discharge current that matters. It could permit even more, if for example the ESD arcs across it as ESD is prone to doing.

It does, however, protect the opamp against a large (but not ESD-large) voltage being applied to the input. Imagine applying 10V directly to the inputs in the lowest range (10k sense resistor), a fairly easy mistake to make. It could even happen just due to the DUT attempting to draw too much current and "saturating" the current shunt, which happens at a very low current with such a large sense resistor. You'd have 10V applied to the opamp input, 8.5V above V+, without that resistor. With the resistor, the current through the opamp's diodes will be limited to around 29mA - still not great, but definitely better than a direct connection.

[c4757p]

In any case, the other purpose seems more important to me - high voltage applied to the inputs is going to happen very often with this device. Consider connecting a power supply to anything with decoupling capacitors, the uCurrent is going to "saturate" as the caps fill up. With such high sense resistances, this is going to be a very common event, something much more essential to protect against (as it happens in the normal course of using the device) than just ESD.

[Gyro]

R12 is way undersized for any sort of sensible overload protection. On the nA range (10k shunt) the op-amp is very exposed. The esd protection diodes are probably rated for only 5mA continuous, so this limit will get exceeded at about 2V above supply rail.  At anything much above that you can at best expect the input characteristics to start degrading.

I can't understand why protection diodes weren't included in this design, you might cite leakage, but at 1nA they would only have 10uV across them and at higher currents their effect would soon become insignificant.

Trying not to be too critical, I would also have used a 2 pole 6 way rotary switch rather than the slide switch and put in most of the missing 4 decades of current shunts. That would minimize the voltage burden across a much wider range of currents, whilst reducing the demands on op-amp and protection diode specs.

[Jannehrl]

We do have tought about more Resistors too. To make Currents, which are at sizes between the current ranges more precise. We did also think about a polarity protection & voltage protection. But we didnt make it, because we were worried, that it could make troubles in relation with the Op-amp. 

Neverless, in my opinion the main question about the 270 Ohm before and after the OP-amps haven`t been answered.
I`m thankful for all the present answers, but most answers were contradicting each other. I would really like an answer which clarrify not only why they are there, but also
why it was taken the size of 270 Ohm. 

No, I don't understand the reason for the value used for R12 (270) either.

It's purpose in an op-amp circuit is to minimise the offset due to input bias currents on the two op-amp inputs. It should be equal to the parallel combination of the feedback resistors (R3,6k8+R11,2k2) and the feedback resistor to ground (R5,1k), so around 900R would be appropriate.

If you arent bothered, could you explain that further Gyro?

I dont understand this, because the input resistance of the Op-amp is already very big, so how is the 270Ohm resistance making a difference?

[dom0]

The input resistance of the op is not "physical", it does not exist: it is the result of a regulation process. If the device is powered off, or the limits are exceeded, input resistance is very non linear ; basically determined by the ESD diodes to either supply rail.

[Kleinstein]

R12 makes a little sense in protecting the input from excessive currents through internal diodes of the OP. However the value of 270 Ohms is rather low, so for a real protection they should have used something like 10 K. My best guess for the value is, that they already have it in the circuit, so shorter BOM.

For the output of the OPs the 270 Ohms are about right to isolate the outputs from capacitive loads. Though this would have needed to have C4 from the OPs output and not behind the resistors - this way the resistor makes things even worse.

[Jay_Diddy_B]

Hi,

I think that you might be 'over thinking' these resistors.

I understand the concept of making the resistance on the +ve and -ve inputs equal to minimize the effect of input bias currents. But on the MAX4329 the input bias current is around 1 pA. With the Thevenin equivalent of the divider R5, R3 and R11, 0.9k Ohm 1 pA = 900pV of offset on the input, (1+ gain) x offset on the output 10uV. This is further amplified by U4 (x10) to 100uV.

I think that the 270R resistors are being used as jumpers. The 270R resistor was chosen for the R4, to set the LED current. The lower value resistors R1 and R9 are precision (0.05%) so not economic for jumpers.

So 270R were used for 'jumpers'

Why the jumpers? Because the PCB is also the front panel, to keep the layout free from via or tracks on the front panel side.

Regards,

Jay_Diddy_B

[Gyro]

If you arent bothered, could you explain that further Gyro?

I dont understand this, because the input resistance of the Op-amp is already very big, so how is the 270Ohm resistance making a difference?

It probably doesn't make any difference in this case because the opamp has very low input bias current. I suspect it is just there as a not very successful protection resistor (value chosen just because there were others on the board).... However in a standard opamp circuit it is good practice that both inputs see the same input resistance. There is always some bias current on opamp inputs (check datasheet specs), if the resistances seen by the inputs are different then this current will become an additional offset voltage (different voltage drop across the resistances). I hope this is more clear.

As I say this is general good practice, but much less important in low input current opamps (though still relevant in very high precision situations), Edit: or where the input resistance values are particularly high, eg. even with 1pA input current, a 1M input resistance would result in 1uV offset.

[Kleinstein]

The OPs in this circuits are chopper-stabilized OPs, and for these OPs the bias current at the inputs is usually not the same, but even tends to have opposite sign, especially for low nose ones with higher bias current.  So same size resistors generally do not help with AZ OP, unless at high temperatures.

[Gyro]

Good, point - in that case reserve my advice for non-AZ opamps, particularly bipolar ones or when working with high impedances. 

[Gyro]

I think that the 270R resistors are being used as jumpers. The 270R resistor was chosen for the R4, to set the LED current. The lower value resistors R1 and R9 are precision (0.05%) so not economic for jumpers.

So 270R were used for 'jumpers'

Why the jumpers? Because the PCB is also the front panel, to keep the layout free from via or tracks on the front panel side.

Yes, they could be serving that function (although personally I still think the ones on the opamp outputs are providing stability and short protection) however there still should be a protection resistor of sensible value (not 270R) at R12 as it's the only protection that the opamp has against any sort of overload.

[dom0]

Good, point - in that case reserve my advice for non-AZ opamps, particularly bipolar ones or when working with high impedances. 

Another kind of op amp where (artificially) created equal impedances can hurt performance are the ones with bias compensation, because, much like AZ op amps, they often will have essentially random, uncorrelated input currents. => no gain in offset, but more noise.

[Gyro]

 

[Mr Smiley]

Why not just ask the man himself 


Dave 

 

[EEVblog]

Others have said it.
There is nothing magical, they aren't strictly necessary, just a common "best practice" thing.
Having the output of an opamp directly connected to anything the user interfaces with is not the best idea.
Having the input directly connected is not a good idea either, so good practice to whack a resistor in series.
Can potentially help with opamp stability with capacitive loads (coax etc), same for the virtual ground.
Also I knew they'd be handy for jumpers (I think on the old design it was necessary.
I did not investigate if they were strictly required in the final build design, so did not optimise them out.
Why 270? BOM consolidation, it was already the LED dropper.

[EEVblog]

I did ask my instructor, but neither did he understand why they are there.

If your instructor couldn't at least guess at some potential reasons, then I'd say he's a not a very good instructor 

[Gyro]

Why 270? BOM consolidation, it was already the LED dropper.

If just for BOM consolidation then it would probably be worth retrofitting R12 with 10k to provide some degree of input overload protection.

[dadler]

Why 270? BOM consolidation, it was already the LED dropper.

If just for BOM consolidation then it would probably be worth retrofitting R12 with 10k to provide some degree of input overload protection.

The only other 10k is a precision resistor.

[Gyro]

Yes, I pointed that out in my original reply (reply #1). I was talking about retrofitting a 10k resistor to improve protection, not BOM consolidating the close tolerance one.

Edit: Actually, BOM consolidating with the existing 10k one would probably cost very little in the scheme of things.