EEVBlog #72....

[tomvleeuwen]

(...)
There is an interesting quirk on the MAX4329 data sheet, the common mode input voltage range (over the full temperature range) is quoted as  GND-0.05v to Vcc-1.4V; so using a mid point ground at the minimum supply voltage of 2.7v means that the common mode range is exceeded (by 50mV).
Possibly an asymmetric split supply (say +1.6v, -1.4v) would be better.
(...)

I think you had the wrong data sheet, the one from the Maxim site quotes Vee to Vcc. Even the input offset voltage is specified @CM=Vee and Vcc
Yes I looked at the full temperature range.

[jimmc]

My fault, I mistyped the part number it should  be MAX4239 as in Dave's blog.
The figures are correct for the MAX4239.

Jim

[joelby]

If anyone's interested, this month's Circuit Cellar has an article on picoammeter design, which is a bit different to Dave's design.

[Zad]

A guard ring would be a good idea for any future evolution if an order for 10,000 came in (cheap but a minor effect on the performance) but it certainly isn't worth re-designing the PCB for another run of 100. Chances are that EMI noise from the room is going to swamp that current anyway. Equally, if you were after ultimate performance then an offset power supply might be appropriate but I tend to be of the opinion "if it ain't broke, don't fix it". It might seem a simple matter of changing the resistive divider ratios, but seemingly simple things like that have a horrible habit of coming back to bite you in the form of obscure bugs.

This raises an interesting point about engineers. We seem to be wired differently (ho ho) to other people. Whilst a road mender, a doctor, or a lawyer will tend to stand back at some point and say "that'll do", an engineer will tend to want to design the most perfect piece of equipment possible. The trouble is that it takes exponentially more time to produce. Then, having got a brilliant design, we want to make it better... and better... and better, long after our changes would ever be noticed by the customer.

What separates a student or run-of-the-mill engineer from a professional is knowing when to stop, and that is not easy to learn. I wonder if female engineers have this problem too or if it is just a testosterone thing?

Mike

[allanw]

I'm fairly new to PCB layout, but was there any reason why the top side (where the labels are) couldn't have a ground plane? And also, having a ground polygon on the component side?

Thanks.

[Simon]

What separates a student or run-of-the-mill engineer from a professional is knowing when to stop, and that is not easy to learn. I wonder if female engineers have this problem too or if it is just a testosterone thing?

Mike

Yea i'm having that problem with a project I'm on, I'm at version 5 and that the stable base version, I'm still dreaming of extra features

[hans]

A guard ring would be a good idea for any future evolution if an order for 10,000 came in (cheap but a minor effect on the performance) but it certainly isn't worth re-designing the PCB for another run of 100. Chances are that EMI noise from the room is going to swamp that current anyway. Equally, if you were after ultimate performance then an offset power supply might be appropriate but I tend to be of the opinion "if it ain't broke, don't fix it". It might seem a simple matter of changing the resistive divider ratios, but seemingly simple things like that have a horrible habit of coming back to bite you in the form of obscure bugs.

This raises an interesting point about engineers. We seem to be wired differently (ho ho) to other people. Whilst a road mender, a doctor, or a lawyer will tend to stand back at some point and say "that'll do", an engineer will tend to want to design the most perfect piece of equipment possible. The trouble is that it takes exponentially more time to produce. Then, having got a brilliant design, we want to make it better... and better... and better, long after our changes would ever be noticed by the customer.

What separates a student or run-of-the-mill engineer from a professional is knowing when to stop, and that is not easy to learn. I wonder if female engineers have this problem too or if it is just a testosterone thing?

Mike

I think it's just a bit of geekness. If people would feel offended by that, I'm sorry but they probably also shame for it. I don't, if I want my programs or projects to do something into an exact detail which others wouldn't care about, I could put myself on to that.

Otherwise I usually stick to 'it meets the given requirements, so it's good enough'. I have taken this a bit from the projectmanagent we got on college, which means you set a list of requirements you want to meet before a project, then start to work onto meeting those requirements. Of course, if you find out it's impossible to achieve, you can adjust it, but adjusting them upwards usually means you will be busy, and then a bit more, and more which doesn't work out at all.

As David also mentioned; electronics always works in magnitudes of x times. 10x is great, 100x is excellent, but anything above that probably hard to achieve (also because you need nV opamps, 0.01% resistors and that kind of stuff to meet the right specs.

[EEVblog]

I'm fairly new to PCB layout, but was there any reason why the top side (where the labels are) couldn't have a ground plane? And also, having a ground polygon on the component side?

The top side does have a ground plane.
And the ground plane works for visual appearance too, as it increases the contrast of the solder mask enormously. The mask looks wimpy and pale when there is no copper underneath.
Forgot to mention this in the blog.

Dave.

[allanw]

I'm fairly new to PCB layout, but was there any reason why the top side (where the labels are) couldn't have a ground plane? And also, having a ground polygon on the component side?

The top side does have a ground plane.
And the ground plane works for visual appearance too, as it increases the contrast of the solder mask enormously. The mask looks wimpy and pale when there is no copper underneath.
Forgot to mention this in the blog.

Dave.

Ah, I see. Thanks.

I plan on using the pcb-as-front-panel idea for my projects in the future. That's a great tip.

[ToddFun]

I think this new opamp from National Semiconductor would be great for Dave's
µCurrent project.

http://www.national.com/pf/LM/LMP8358.html#Overview

LMP8358 Instrumentation Amplifier 
Max offset voltage of 10 µV
Max offset voltage drift of 50 nV/°C
Includes fault detection circuitry to detect open and shorted inputs
SPI programmable gain or just use two resistors for up to 1000 gain.
low input voltage noise and high gain-bandwidth

I see you can get free samples in 6 weeks for some packages, but WOW these are not cheap! $2.71 each at 1,000+ pcs.
I tried a lot of distributor and nobody has this chip yet.