Automatic gain control

Something like the top schematic here:

Note, some ADCs have this function built-in. Usually for a strain gauge you will want very low Vos and TCVos (eg. with 10mV FS and a low-resolution 10-bit ADC, the LSB is about 1uV), so you'll probably want to use an amplifier with very good DC characteristics and take care as to thermoelectric effects and so on.

Best regards, Spehro Pefhany

Yes, that's a practical solution.

Graham

Graham, What's going on with you this past week? Run out of Lithium again ?:-)

...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 | |

"Eeyore" wrote in message news: snipped-for-privacy@REMOVETHIS.hotmail.com...

Try to guess why the feedback signal in some closed loop systems is called the "error signal".

Particularly useful when the forward gain is a known quantity ;-)

...Jim Thompson

Load cells don't usually come calibrated, and typically have, maybe,

Nearly all weighing or strain systems have a processor after the a/d converter. One removes the mechanical load and notes the zero ("tare") load cell output, then applies a test weight or force and again notes output ("span" or "gain" cal) then does appropriate math (in software!) on all future measurements. One could save tables of cal factors for a number of different transducers, provided you have a mechanism to identify each one to the software.

It would be unusual nowadays to do these calibrations with hardware circuits.

John

Or just use a 24-bit delta-sigma whose input range is a bit bigger than you'll ever expect. There are lots of cheap serial d-s adc's that would connect directly to the strain gauge output.

John

;~)

I know, I wrote a software decoder for the WWVB signal about 20 years ago.

be

Just a wild guess:

You'll have an opamp that takes your strain gauge output as its input and amplifies it for input to your DAC. For the gain-setting resistance, use one of those digital potentiometers that remembers its resistance setting (or else use some memory IC, etc, and a "dumb" digital pot). You'll have a toggle switch to select either calibrate mode or normal mode. In calibrate mode, your amplifier's output will be fed into another opamp, which subtracts it from 5 vdc and also integrates that difference. Somebody else here will be better-able to tell you how to use the integrator's output to run a binary counter up or down (depending on the sign of the integrated error signal) to set the digital pot's resistance, which, due to the subtraction from 5v and the integrator, will soon settle to the resistance needed to give you

Tom Gootee

OK, back to electronics. If I understand the whole issue, you want full scale output constant at any full scale output of your bridge, disregarding the sensitivity of your strain gauges. So you have to calibrate your bridge to give you certain, part scale output, and adjust your gain accordingly, either by precisson potentiometer which you lock in position for this measurment or by AGC which should be still somehow locked constant during measurment. The known unbalance you create by applying parralel resistor to one side of your balanced bridge to create a simulation of stress. You have to calculate what stress it represents, depending on resistance of the particular batch of gauges you use. (If you use semiconductor gauges I cannot help you as the resistances have very wide resistance range and its change with temperature may be out of usefull precission.) So for the manual gain control you can use DC excitment of the bridge and get DC output. For AGC try AC, then you go to AC amplifier and life is easy. The detector may be a little more complicated as it has to inform you if the bridge is under tension or compression so your AC acts as carrier to ballanced detector.

Been there, used it!

Have fun

Stanislaw Slack user from Ulladulla.

All very logical, but extremely stupid.

The OP clearly didn't want or need a direct connection between the output and +5V, so the challenge is to work out what he might actually have wanted.

The subject line "automatic gain control" might have given you a clue, if you weren't so dead set on being unhelpful.

See posting 59 in this thread Sun, Jul 9 2006 1:21 am.

You should be able to find it on Google Groups

even if it has fallen off your news server, along with a lot of other information that you seem to have missed.

First, strain gauges are linear only over a limited range. Ill considered bridge configurations do not help and often hinder linearity. I am not sure what you are doing but a set of load cells seems to be called for; perhaps even calibrated ones.

Unless your excitation is on the order of 40 V (probably damaging to destructive for the strain gauges) you can not get 4 to 5 V out of the bridge. The normal signal is millivolt, sometimes even microvolt. Instrumentation amplifiers are required. These are fixed gain critters except for some little fine adjustment for calibration.

Over a decade of having to tackle all the left handed and screwball tests in a test lab teaches a few things.

Boom !!! ;~)

Ah, so it is trans.