Why are Instrumentation amps so expensive?

do a search for Guy Macon

martin

Should I be looking at the insult file?

Well, yes, your post was rather verbose, I sort of fell asleep reading it. Can you repost in sort of 10 lines or less?

martin

Sorry if I offended anyone, but your right, it is wordy. Ver II: (7 lines)

Why are true instrumentation amps so expensive? I understand from AOE that one of the advantages of the instrumentation amp over a difference amp is that resistor matching is less critical, and yet if I look at e.g. the 'INA' range of amps, the differential amps cost less than the instrumentation amps.

I would have expected technology to have produced some sort of chopper amp auto zeroing instr. amp by now, or perhaps there is now a 'better' way to capture a floating differential voltage and present it to an ADC?

Dont know. We will have to wait for Win Hiill or Jim T or other guru's to reply Sorry , just making a point

martin

--
Do you have an actual application in mind, with spec\'s that you have
to meet, or are you just kvetching?

Even from the "verbose" version it's not clear what the OP is trying to accomplish ;-)

Maybe a "flying cap" ??

What input voltage range?

...Jim Thompson

--
|  James E.Thompson, P.E.                           |    mens     |
|  Analog Innovations, Inc.                         |     et      |
|  Analog/Mixed-Signal ASIC\'s and Discrete Systems  |    manus    |
|  Phoenix, Arizona            Voice:(480)460-2350  |             |
|  E-mail Address at Website     Fax:(480)460-2142  |  Brass Rat  |
|       http://www.analog-innovations.com           |    1962     |
             
I love to cook with wine.      Sometimes I even put it in the food.

Wow, up to Version 6.12 ALREADY?

"You are not ANSI compliant and your markup doesn't validate. You have a couple of address lines shorted together. You should be promoted to Engineering Manager."

That's just wrong. ;)

Hello Roger,

It's probably because of the low market volume. They are used in measurement setups and university projects and that's not a high volume market.

For high volume products that require a large CM range engineers usually spend quite some time to make it happen with jelly bean parts. Often the first stage is discrete.

--
Regards, Joerg

http://www.analogconsultants.com

Yes, they exist, for less than a couple dollars in 1K. They're not cheap because there are really no high volume applications for them. I've been able to avoid them in most instrumentation applications, for example.

If it's not moving too fast you can just measure the two voltages and subtract digitally.

Best regards, Spehro Pefhany

--
"it\'s the network..."                          "The Journey is the reward"
speff@interlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com

Roger wrote: >. So the

Ok, lets say that you are hunting for components to acomplish certain function in known enviroment. You are as free as a bird to use what fits. I am test department tech. on hunt for instrumentation amplifiers to amplify, for computer recording, your ADC, physical parameters like heat (thermocouples from copper/constantan to platinum/platinum-iridium and tungsten; strain gauge reading of forces from static to mechanical shock and so on and on) So I end (early '80s) with instruments of Vishay at $400/each (50 qty) and do my testing for more than 8 years without failures. And NOT in lab but in enviroments where the Tektronix scope displayed a spiral trace with the trace moving in oposite direction to the flow of time. (Real SCI-FI). And the technical design department want accuracy

If true then it's a little bit strange.......

Guy Macon sort of promotes himself as an 'Engineering Manager'.

Oh..... I'm being slow :-)

DNA

Don't get confused by AoE. Instrumentation Amps have some indispensable advantages over diffamps or discrete designs. The input impedance is very high, *equal* for both inputs and for *common mode* and *differential mode* signals. This maintains the CMRR for any source unbalancing. The internal resistors are trimmed to very high precision. A single resistor can vary the gain. The AD8225 has a gain error of +/-0.1% and 5ppm/K drift max. Offset error is not so much of an issue today, it can be corrected for in the digital domain. This amp also has a CMRR of >80dB at 10kHz, very difficult to achieve with discrete parts, C-trimmers are needed for that. There are quite a bit of applications where IAs are used: RTI, thermocouple etc. bridges, professional audio, 4-20mA receivers, EKG front-end....

--
ciao Ban
Apricale, Italy

Very well said. I use the AD620 by Analog Devices, and it works very well to measure currents as low as 5 amperes on a 1000A 100 mV shunt. This corresponds to 5 mV, and my 1% accuracy specification means I must have no more than 50 uV of noise or common mode error, where common mode voltage could be as high as one or two volts. I pay about $5 to $8 for these (in small quantities), but it is quite reasonable in a $3000 instrument that is used for calibration purposes. I have also used the INA118 or similar devices, which were somewhat cheaper, but had more noise and drift.

Paul

Yes, me neither. I could build one of these out of an LM324 and use the spare one to drive an LED.

Bastards!

DNA

Nothing. In the verbosr version I even mention that I have a solution for my app.

Mine is purely a curiosity question: "Why are instrumentation amps so expensive?".

I could narrow down the field and ask "Why do the instrument amps tend to cost more than the diff amps in a range of precision amplifer products such as BB's INA range".

Answers here and elsewhere do not convince me:

1) Laser trimming. Well, good diff amps also need laser trimming and in some respects (according to AoE) the instrument amp cct makes some resistors matching less critical.

2) Circuit complexity. Scince when has this been a major factor in IC prices.

3) Quantity. Instrumentation amps have a potentially much larger market, specificaaly they could be used to improve most designs where diff amps are used.

Clearly there is something that makes them so much more expensive to produce than similar precision analog devices.

I also entered into the question of why there are not improved "chopper" versions of instrumentation amps or cicuits which can do the same. I specifically noted that the ICL1706 device shown in AoE appears to be no longer made, nor is there anything like it.

Let's answer this in one. Most ADC's require an input range within the

5/3v3 supply range and only offer psuedo differential inputs which have zero CMRR characteristics, yet many sources cannot be readily adapted to this: Uninsulated thermocouples and bridge sensors optimised for eg. 10V supplies are common examples. Even ground referenced sources (which in some respects includes uninsulated thermocouples) can benefit from differential inputs as sensors may be remote and/or in power cct's and be subject to large CM noise. In low impedence situations it is possible to use diff amps, but I would only bother to make the distinction if the diff amp saved me money, one good low priced instrument amps could eliminate the requirement for a similar diff amp in the same product range.

But I also mused on the idea that there could be ADC's with a floating capacitor input, i.e. FETS which attach a cap to a source, and then switch it into the convertor, so the ADC can measure voltages outside the supply range which the core converter is running at and, by flipping the cap, could null out offset errors at the same time.

Well, I would love an ADC like this, but perhaps that is because I do not know some technique that gets the same results at low cost ;-)

Hello Roger,

...

Like where? I have used diff amps in designs that went into production but I have never used instrumentation amps on any released product. There was no need to.

Ban has brought up good examples but even there cost often rules. Many good ECG (EKG) units go sans instrumentation amp.

Offsets and extended input voltage ranges are often handled by clamping circuitry. In the same way that TV sets restore the DC bias for the video path from a source that isn't DC coupled.

--
Regards, Joerg

http://www.analogconsultants.com

Roger wrote: (snip)

If you need really good common mode range, this is the way to go. One pair of reed relays connect the cap to the field voltage, and after they are opened, a second pair connect the cap to the A/D and signal ground. You need definite break before make contacts.

I have seen thermocouples, with more than a hundred common mode volts, read this way.

Still missing the point :-(

If a precision instrumentation amp cost pretty much what a precision diff amp does then there would seem little point having both in a product range.

So that rules out "quantity" as a factor.