LED reference current source

Poor little sensitive John and Michael find me too rough. What a pity.

-- Bill Sloman, Nijmegen

What I find you is nasty and boring and stuck deep in the past.

I would have offered you some consulting business if you weren't such a toxic toad. I found someone else, and he's flexible and cheerful.

--

John Larkin                  Highland Technology Inc
www.highlandtechnology.com   jlarkin at highlandtechnology dot com   

Precision electronic instrumentation
Picosecond-resolution Digital Delay and Pulse generators
Custom timing and laser controllers
Photonics and fiberoptic TTL data links
VME  analog, thermocouple, LVDT, synchro, tachometer
Multichannel arbitrary waveform generators

Sloman still suffers from delusions of adequacy, and of being human. What a waste of oxygen. :(

But, Bill's point remains. The use of AC excitation buys your way out of several traps. If Johnson noise is higher than the temperature signal, the AC method works. The best chopamp doesn't. If self-heating of the sensor matters, RTD sensors will always give more accurate measurements with small excitation signals. AC doesn't lose accuracy. Chopamp loses accuracy.

Any given temperature control problem can be solved the 'best' way in an engineering sense, by using the least-cost option; but you cannot ignore the AC option just because it didn't win at the last (or last two) design reviews. For elephants, you'll need the elephant gun.

That's not true. An AC measurement has an inherent 3 dB bandwidth penalty, so for the same transient response, the SNR will be 3 dB lower. Johnson noise is white--do the math. AC measurements help with DC drifts and 1/f noise, if you do them well, but are no help whatsoever in the flatband.

For temperature stabilization, you don't care about dissipation, unless it's time-varying or causes turbulent convection. Otherwise you get a constant offset, which you normally don't care about, because it sits still.

Can you identify a single actual noise source that is not removed by putting the chop amp and all the sensing circuitry in the controlled volume? I've put that question several times in this thread, and all I've heard in response is huffing and puffing.

Doing that gets rid of the variation in thermocouple offsets, gradients, wind, and everything else, if you do it right. As JL said, the thermomechanical design is the fiddly part, not the sensing bridge and front end amplifier.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC
Optics, Electro-optics, Photonics, Analog Electronics

160 North State Road #203
Briarcliff Manor NY 10510
845-480-2058

hobbs at electrooptical dot net
http://electrooptical.net

--
And, by casting stones, you intimate that you're not?

on the

he

nd

led

t of

al,

Rejecting one of the side-bands could eliminate that disadvantage if it was - in fact - a real problem. The fact that nobody bothers is a clear enough indication that it isn't. In practice, AC excitation justifies itself by getting rid of 1/f noise.

th

Nothing helps to get rid of Johnson noise, short of sinking the whole measuring system in liquid helium. Basically what you are saying that AC excitation is merely perfect.

e

accuracy.

You do care about it, to the extent that it constrains the stability required from your excitation source.

o)

Gerhard Hoffmann came up with

which makes it clear that you can reduce thermal gradients within a controlled volume, but can't eliminate them. That's rather more than huffing and puffing. You argument is actually not that you can "remove" them but merely that you can reduce them to a tolerable level in the applications that you are currently interested in. Suddenly claiming that you can "remove" them comes perilously close to huffing and puffing.

It reduces them, but doesn't "get rid of them".

This can be true, for some applications. It's not universally true - not every temperature stabilising system has to fit down a bore-hole - and in some application, like Priel and Larsens' (very different) well- stirred water baths, the thermomechanial design is trivial.

-- Bill Sloman, Nijmegen

't

Like all the other people who can't be bothered to flatter you.

I doubt it. Jim Thompson would probably have persuaded you that it was subsidising un-American activities or some similar nonsense.

And flatters you. I hope he's as good as designing circuits as he is at massaging your ego.

-- Bill Sloman, Nijmegen

d.

's

sn't

n.

I don't suffer from such delusions - I enjoy them. And they may not even be delusions. Michael Terrell isn't actually infallible, nor even all that reliable,

-- Bill Sloman, Nijmegen

One of the most productive things you can do in life is find and sustain long-term relationships with talented, generous, fun people that you can trust. People like this help one another and care about one another in the long term. If they can design electronics or write code or so some sort of serious science, all the better. If they ski, yet better.

JT is of course not one of those kind of people. You aren't either.

--

John Larkin                  Highland Technology Inc
www.highlandtechnology.com   jlarkin at highlandtechnology dot com   

Precision electronic instrumentation
Picosecond-resolution Digital Delay and Pulse generators
Custom timing and laser controllers
Photonics and fiberoptic TTL data links
VME  analog, thermocouple, LVDT, synchro, tachometer
Multichannel arbitrary waveform generators

The Johnson noise depends on system bandwidth. Modulation doesn't help.

The best chopamp doesn't.

I don't see why. A chopamp will drift nanovolts longterm, and it can be inside the oven, right next to the RTDs. An AC system has lots more ways to make errors.

AC doesn't lose accuracy.

Why?

--

John Larkin                  Highland Technology Inc
www.highlandtechnology.com   jlarkin at highlandtechnology dot com   

Precision electronic instrumentation
Picosecond-resolution Digital Delay and Pulse generators
Custom timing and laser controllers
Photonics and fiberoptic TTL data links
VME  analog, thermocouple, LVDT, synchro, tachometer
Multichannel arbitrary waveform generators

Amen! Jamie

the

accuracy.

It's really not very difficult to control the dissipation of the bridge to any accuracy you might remotely need. If it dissipates a couple of milliwatts, even a TL431 would get the change in dissipation down to ~

100 nW, without breaking a sweat. Spending another dollar on the reference will get you an LM329B, which would drop the dissipation tempco to a few PPM, i.e. several nanowatts per degree, even with the reference at ambient. That will be way below the level of the residual thermal forcing.

It gets rid of them to the same level as the thermal forcing rejection of the controller, i.e. probably a factor of 10**4 for a reasonable design--Karlquist et al got 10**6-ish. With a multizone two-layer system, which is what I built, the forcing rejection is higher.

The very little that remains is almost entirely a gain error, because there's little or no opportunity for the gradients to change in a way that isn't correlated to the sensor temperature change. So yes, to any reasonable engineering approximation, it gets rid of them as a source of thermal control errors.

This is engineering, Bill. Getting rid of them doesn't mean setting them to 0.0, just that they go away to the accuracy of the rest of the measurement or better.

Trivial? To get microkelvins in the face of serious thermal forcing?

Don't think so.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC
Optics, Electro-optics, Photonics, Analog Electronics

160 North State Road #203
Briarcliff Manor NY 10510
845-480-2058

hobbs at electrooptical dot net
http://electrooptical.net

So, the ability to choose a frequency with ONLY Johnson noise, and severely limit the bandwidth, wins. AC excitation does both, chopper amps allow the second (but don't facilitate it).

accuracy.

It's not about dissipation alone, the heat of the sensor and the amplifier cause thermal currents (flowing away from those sources) and that implies temperature gradients, inside the 'controlled zone'. It probably doesn't matter much for a 2W 65C heater, but in a cryostat at 4 Kelvin, where thermal conductivity of materials is low, that's a killer.

Sure; thermocouple effects due to heat generated in the sensor and the amplifier will cause a DC error- an offset. You can hope that it will be a constant, but (for instance) varying the output voltage of an op amp changes the bias currents inside the chip. Gradients will also respond to atmospheric pressure or gas composition. That DC error is NOT a reliable constant.

Nice pictures here if anybody is curious:

Good story here:

--
These are my opinions.  I hate spam.

HP (and probably others) often had 2 versions of gear.

One had a high quality crystal. The other had a low cost crystal with the expectation that you would be using an external clock input from your local fancy lab standard.

--
These are my opinions.  I hate spam.

ld.

e's

esn't

.

I've got a few friends like that.

That would be James Arthur, whose ideas about economics are roughly equivalent to flat earth geology or Young Earth Creationist biology.

It does seem unlikely, but your opinions on the subject aren't really all that reliable.

Probably not, but your opinions in this area are distorted by your need for continual approbation. And your capacity to assess serious science - as revealed by your blind acceptance of anti-climate change denialist propaganda - isn't impressive.

-- Bill Sloman, Nijmegen

Lark>

High time you learned to snip. Phil Hobbs parting shot was actually a non sequitur, so the rest of the post was entirely irrelevant.

-- Bill Sloman, Nijmegen

rely on the

in the

s, and

trolled

s

out of

signal,

p with

ver

more

lose accuracy.

ss

t a

Another advantage of DC excitation. Of course, at that level, you have to start being careful about the resistances of your tracks and solder joints, and you wouldn't want any stray magnetic fields driving AC around that particular loop.

nnot

t two)

ll

ts,

We are communicating in English, not some new language you've just invented for your own convenience.

AC excitation gets rid of thermocouple voltages and other sources of 1/ f noise. Compact design merely reduces them to acceptable levels in some situations.

Read the literature.

-- Bill Sloman, Nijmegen