grrrrr

There are active clamps in HEX packages, clamp at X+/-200mV but that's at 20mA, they seem slow but your inputs are thermocoupes(?),maybe you ( not me, rule is less than 60 seconds on any search) can find something better:

Yeah, verily! As I found to my great pleasure, when an OpAmp specified to have less than 4pA of bias current turned out to actually only take an amazing 150aA!

Jeroen Belleman

There's only one person shoveling it around here and I was simply doing the readers a service by pointing that out! At least those that didn't already know, about you!

Jamie

s own output ...

That is what you think you doing. In reality, while you aren't the only so urce of misinformation around here, you do seem to be the one with highest ratio of noise-to-signal. Much of what you post is wrong, and none of it is worth the bandwidth it uses up.

te:

ote:

h his own output ...

ing

y source of misinformation around here, you do seem to be the one with high est ratio of noise-to-signal. Much of what you post is wrong, and none of i t is worth the bandwidth it uses up.

here's some data:

;)

-Lasse

Pure speculation on your part.

Hard to make good judgments on your end when your sources are a random collection of google snips, basically a data junky!

Jamie

te:

h > > > > > his own output ...

y source of misinformation around here, you do seem to be the one with high est ratio of noise-to-signal. Much of what you post is wrong, and none of i t is worth the bandwidth it uses up.

What makes you think that (apart from your self-promoting ignorance)? I cit e wikipedia (after I've checked the content) because the level is about rig ht for here. Mostly over your head, of course, which makes you comment abou t "pure speculation" unintentionally hilarious.

It's certainly not my primary information source. Every now and then I do p ost references to the books I read. I've had a subscription to New Scientist fo r some thirty years, and I've been reading my wife's copies of Physics Toda y for a few years now. For a while I had a subscription to the UK equivalent (Physics World) but that was via my membership of the UK Chemical Institute , which I ended when I left the UK in 1993. I'm still a member of the Royal Australian Chemical Institute, but their popular science stuff is a bit do wn-market from even the New Scientist.

What I actually know - which is a whole lot more than you could imagine - i s actually a fairly coherent body of data. I may like collecting data, but I do fit it together rather than just accumulating random snippets.

That's a really interesting part. It won't quite work for me here, but I'll keep it in mind for other things.

I guess I'll add some rrio opamp followers ahead of the ADC pins. AD8629 is explicitly OK with up to 5 mA into its ESD diodes. It's unusual for a data sheet to be so upfront about that.

Hello John !

May i suggest you to consider page 2 Absolute Maximum Ratings Input Current 10mA continuous ; 100mA momentary

You may add a resistor at analog inputs to prevent excessive current.

BR, Habib.

I would think that if it had an unusual tendency to latch up and caught fire that would be covered by the abs max ratings.rather than the blah-blah of the datasheet.

By my reading, it won't latch up even with 0.1A.

Best regards, Spehro Pefhany

Hi, Habib,

The original (simplified) concept was about like this...

but the ADC specs have me worried. So I'll probably do something like...

which is wandering a bit away from the concept of elegant simplicity.

This ADC at least lets me run the analog common AVSS below ground, which is unusual in a delta-sigma ADC. If they could do that, why couldn't they get the ESD diodes right?

Umm, I smell smoke. ;-)

Very unusual. Neat part.

Delta-sigma ADCs are great for this sort of measurement, and they tend to come in 24 bits. In real life, they can deliver 10 PPM accuracy, so the RTD measurement accuracy will be dominated by my analog parts, specifically the stability of the current source.

Since an RTD changes about 3800 PPM per degree C, the measurement accuracy will be a lot better than the RTD accuracy, so it all works.

RTD_Input.JPG

I largely prefer this concept more than the one described below. Assuming TDR sensor is remotely wired, i suppose the two 1G resistors make the analog inputs to be immunise against noise, surge ...etc. why not do the same at each analog inputs ? BTW don't know if internal ESD diodes are in conduction state in the schematic in reference with resitors as large as 1G.

If i were you i would not be bothered about ESD diodes as long as Absolute maximum values are not reached, furthermore there is a PGA in between Delta-Sigma modulator and analog switches, current sources ... so the first schematic remains the best idea IMHO.

Habib.

RTD_Input_2.JPG

The 1G resistors (or maybe 10G resistors) are for open-transducer sensing. Without them, an open input might hang at its previous value for hours or maybe years.

why

The 0.1 volt limit bothers me, since they don't say what happens if it's exceeded. So far, they won't tell me. If the digitized values got a little goofy, I wouldn't mind. If it latched up or something, I would.

_Input.JPG

_Input_2.JPG

yeh a lot of stuff just too keep that adc happy

why not something like ad7194

or if you really wanna go crazy and can live with 18 bits, AD7609

-Lasse

Nah, V_DD, V_SS, what's the difference? ;)

You EEs are such nitpickers.

Cheers

Phil Hobbs

(Who's a bit punchy from three straight days checking layouts and ordering boards and parts for something that really has to be right first time.)

John,

Did the datasheet say nothing about that ?

I have a question on this

What happens if the RTD is disconnected ? if so are you sure the diode to

Why not insert the RTD in a traditionnal bridge with all resistors with

first react on measuring temperature with a 100Ohm RTD sensor.

Best regards, Habib.