All I want the load cell to do is trigger the 555 timer circuit (by bringing Pin 2 down below 1/3Vcc momentarily) when enough force is applied, e.g., half of its rated load. I don't need to measure or calibrate the load cell's output, and I don't need to take the analog signal to digital. So I am using a Burr-Brown INA125 and a LM393 comparator before the 555. See schematic.

I'm about to breadboard this and hoping to get advice or tips to improve the design here. J1 goes to a 350Ω full-bridge load cell which runs on 5VDC (max. 8V) and outputs 2.5mA at full load of 10Kg (on a 5V supply).

INA125 instrument amplifier is supplying +5V bridge excitation from its internal supply. R5 is bias current return path. RG and RGa will set Gain to 1000 or 2000 for output at load of 2.5V or 5V into the LM393. R6 and pot R7 are resistor divider to set threshold voltage for trigger. LM393 should be wired to sink current (circuit leading to Pin 2 on 555 chip).

Any critiques, ideas? Should it work?

 

Sorry, forgot to attach datasheets for the INA125 and LM393.

 

Also, pins 3 and 7 on the LM555 are not connected. Error in the drawing.

 

No opinions on this circuit? There has to be something wrong with it, or ways to impove it, right? I'd appreciate any observations. Thanks.

 

Your first PDF comes up blank, but that may be because I use Foxit Reader.

J1 goes to a 350Ω full-bridge load cell which runs on 5VDC (max. 8V) and outputs 2.5mA at full load of 10Kg (on a 5V supply).

Why do you want to use a dual comparator?

Load cell bridges are differential voltage output structures. One side will have an increase in voltage over the unstrained condition; the other will have a lower voltage. Sensitivity is expressed in millivolts per volt of excitation. Where do you come by the 2.5 ma figure? Your IA is responding to the differential voltage. Its CMR lets it respond to the difference voltage, and not the 2.5 volt offset.

One hint - strain gauge bridges have sensitivities in proportion to the excitation voltage. 350 ohm bridges can use 10 VDC to give an output of 3 mV/V. That may not be significant for your application.

 

Beenthere,
I'm using Foxit Reader version 3.3.1.0518 on Windows XP Pro SP3 32bit, and I have no problems opening/viewing Robertwests' schematic. Perhaps your Foxit Reader is outdated or the installation is damaged, or maybe your HDD is getting full? You might try downloading the latest version:
http://www.foxitsoftware.com/products/reader/

Robertwest,
You don't have values specified for the gain resistors.

You don't have values for the LED current limiting resistors. I'll suggest 390 Ohms to 620 Ohms; more if you are planning on using 9v "transistor" batteries. You could also increase R2 and R1 to reduce current consumption if desired.

D2 can be a 1N914/1N4148; a good bit smaller than the 1N400x series.

You don't have a bypass capacitor for the LM393's power rail.
The LM393 can sense down to ground; you really don't need to tie the GND rail to -9v; that would cause the 555 trigger to be pulled below ground momentarily and exceed the 555's input range. Tie it to gnd instead.

You don't show the unused comparator in the LM393; its' inputs should be grounded to prevent unwanted oscillations.

Place a small cap from U2 pin 2 to ground; resistors are noisy, and a cap will help reduce that noise.

You have no hysteresis feedback for the comparator. With no hysteresis, the output may oscillate unpredictably when the inputs are nearly equal. Even a small amount of hysteresis will take care of that.

R6 is 47k, the pot R7 is 100k. You might want to use a single 10k pot and a 20k to 47k resistor on either side of it; otherwise you might find that the adjustment is pretty coarse.

You have pot R4 wired as a fixed 1M Ohm resistor. If you want the 555 timer to be adjustable, you'll need to connect the wiper to one of the ends.

R8 has no value assigned. The output of the 555 timer will be ~1.3v less than your +9v supply, as it is a Darlington follower configuration; you'll only have about 7.7v available for your LED and relay coil; if you are using a relay coil rated for 9v, it may not be reliable. You may need to use an NPN transistor or N-ch MOSFET to sink current from the relay coil instead. If you are planning on running the circuit on 9v PP3 "transistor" batteries, energizing a relay coil will suck them dead in a big hurry.

Energizing a relay coil will put quite a spike on your power rail.
555 timers require a minimum of 1uF and 0.1uF bypass capacitors; the 1uF can be much larger. If you don't use bypass caps on the 555 timer, it will wreak havoc on your power rails, as when the 555 changes output states, Vcc is momentarily shorted to ground.

 

Upgraded to Foxit 5.0, PDF shows up fine. Thank you.

 

Thanks beenthere, I made a mistake in saying mA, meant mV, for the load cell output.

Dual comparator--I did not know a model number for a single comparator chip, only duals and quads, so ordered a dual.

I would run the bridge on 10V, but the load cell is rated 8VDC max. Thank you for the comments and questions.

 

You don't have values specified for the gain resistors.

You don't have values for the LED current limiting resistors. I'll suggest 390 Ohms to 620 Ohms; more if you are planning on using 9v "transistor" batteries. You could also increase R2 and R1 to reduce current consumption if desired.

D2 can be a 1N914/1N4148; a good bit smaller than the 1N400x series.

You don't have a bypass capacitor for the LM393's power rail.
The LM393 can sense down to ground; you really don't need to tie the GND rail to -9v; that would cause the 555 trigger to be pulled below ground momentarily and exceed the 555's input range. Tie it to gnd instead.

You don't show the unused comparator in the LM393; its' inputs should be grounded to prevent unwanted oscillations.

Place a small cap from U2 pin 2 to ground; resistors are noisy, and a cap will help reduce that noise.

SgtWookie,

Thanks very much for all the notes, insights and fixes. I am planning on ~30Ω for the Gain resistor (or maybe 2 30Ω in series with a jumper to select both or one). 30Ω Gain=2000, 60Ω=1000.

I was planning 390Ω for the LED resistors, but this will be running on 2 "transistor" type 9V batteries, so I guess I should plan for maybe twice that to stretch battery life.

Great tip on D2--I have plenty 1N4148 and I can save some space, thanks.

Since I posted this I have added .1uF bypass caps to both sides of the LM393 power pins--sounds like I should tie the LM393 to GND (0V)--do I even need a cap on that pin?

Thanks for the other grounding/decoupling tips--I haven't even gotten to the rest of your post yet!

 

Thanks again SgtWookie. I have made the changes I think I understood, revised schematic is attached, as is datasheet for the Magnecraft 276AXXH-5D relay. It is nominal 5V coil but operating range is 3-24V.

I do have 2 questions: is the hysteresis for U2 correct, and are the inputs to U2 correct for U2 output to go from open to ground when the INA125 output voltage is higher than the reference voltage? Or should these inputs be reversed? (2 for 3 and vice versa)

 

Oh by the way, per the datasheet the relay draws 100mA to energize the coil. Note however that the 555 is set up to one-shot pulse about 30-50ms long, so the relay is not held closed for long. The "duty cycle" when the device is in use averages to about 15-20 pulses every minute, for a session of about 30 minutes, at most an hour. Batteries can be swapped out between races.

I have successfully employed versions of the last stage of this design (the 555, triggered by a microswitch)--although it looks like my design could use improvement (bypass caps etc.). Only the load-cell amp/switching circuit is new.

 

Ok I have the first two stages of the circuit up and running on protoboard and it seems to work fine. I had to delete the 10KΩ resistor from one of the bridge signal wires to ground (Input Bias Current Return Path), the circuit was not working with it in. I assume that the full bridge (load Cell) has its own bias current path (I do not have access to the schematic, but I know there is e.g. temp conpensation circuitry embedded in the load cell).

I had to swap the inputs into the LM393 in order to get the desired behavior from the circuit--i.e., output of LM393 goes to ground and current flows when voltage out from INA125 (pins 10 and 11, now going into Pin 2 of U2) exceeds the reference voltage (now going into Pin 3 of U2).

Next I need to determine if the hysteresis feedback for U2 should simply be one resistor between Pin 1 and Pin 3, rather than the two resistors with one connected to Vcc.

I also want to add a low battery warning (LED), maybe using the spare comparator on the LM393? Any ideas?

 

I do have 2 questions: is the hysteresis for U2 correct,

I'll have to say "no". You have a 100k feedback resistor which is OK, but the instrumentation amplifier will override any hysteresis provided by it. You will need to add a resistor between the output of the INA125 pins 10 and 11, and the noninverting input of U2.

Note that the LM393 has limited output sink current; 6mA is maximum. Instead of a 1k pull-up resistor, you should use 1.5k or higher, as 9v/1.5k=6mA.

...and are the inputs to U2 correct for U2 output to go from open to ground when the INA125 output voltage is higher than the reference voltage? Or should these inputs be reversed? (2 for 3 and vice versa)

I'm sorry, I really don't have time to simulate it at the moment; I'm quite pressed for time. My online time is necessarily limited nowadays.

 

Thanks very much for the help, SgtWookie. I have incorporated various changes into the attached. Also added a low-battery indicator (LED) circuit using the unused comparator in the LM393 chip.