Opamp noise

If I see that circuit with a low impedance transducer and you're telling me you get signals from people moving around, I suspect that some input is floating without you knowing it or it is oscillating ar a high frequency.

You also don't have a filter cap at the output of the regulator. Some regulators oscillate heavily without at least 10uF on the output. I have seen them produce a sawtooth output of 3Vpp with too small caps....

Another thing I am missing is a decoupling cap directly on top of the amplifier. A circuit like this on a breadboard with such long wires is prone to unwanted oscillations.

Meindert

me

one

I want to thank everyone for there input. After doing some modifications with the circuit, I have not got the level drift to just a 2 to 3mV and also so is down to 8mV Peak to Peak. I ended up putting two low pass filters (4.8K resistor with 1nF) and then put a

I am not statisfied with the signal and will do a couple of small changes and then commit to a printed circuit board.

Again, thanks for all the input. My Op amp book didn't cover any of this.

James

I thought you were interested in frequencies from 0-15Hz. Why then use a low-pass filter ar 33kHz? Increase the 1nF to 2.2uF. Then your low pass filter is at 15Hz.

Meindert

Manhattan style is a fancy name for living bug style. Probably for New Yorkers :-)

manufacturers:

Only two ways to find out. Either make a prototype layout or build it dead but or living bug style. Depends on you budget and time schedule. If you need to knows this by tomorrow you can only build it by hand on copper clad.

There is no way to completely shut out the entire spectrum above 20Hz with solderless boards. Sure, you can muffle everything up to a few MHz quite well but above that it'll be wide open.

If you happen to have a metal shed without windows you could try out if the "moving person" effect is much less in there. If it is, then you can be pretty darn sure that RF interference is the problem. The cargo room of a van or truck may also work but only if there are no windows or fiber structure areas, must be metal.

Mostly that doesn't help. RF gets rectified at the first BE junction it sees. Then it becomes DC or in the case of cell phone a low frequency burst pattern. Whether that gets amplified in one fell swoop or in stages usually doesn't make much of a difference anymore. Murphy's law says that RF is like termites, if there is a BE junction it finds it ;-)

Yes, that would make sense if it can be done.

Do yourself a huge favor: Spring for a full ground plane, even if that means you have to go from a 2-layer to a 4-layer. If it has to be

Most books never really cover this stuff. That's why there is usenet :-)

ng is

rail

r the

am

is

This is a very common problem with a very easy solution. You excite your bridge with AC, say 1KHz sine wave and not DC. AC couple your input amp, use a phase sensitive rectifier on its output.

xt -

Would I then use just a basic bridge rectifier to the the DC back at the output? And then of course I loose 1.404 V in the conversion. Or is there another way of doing this?

xt -

Also, what is a phase sensitive rectifier and where can I acquire them? One other question, which out sounding too stupid, they mentioned ferrite beads on the leads, where they referring to ferrite cores in which you wrap you cable through one loop to help shield?

Thanks

Google for 'lock-in amplifier'.

My first question to this is, do you do this before amplifing the orignal differiential signal or after. Next, if using filtering and phase sensitive rectifier, how would I get my negative DC swing from the transducer. I can see have the positive is developed from the rectifier, but not the negative.

comp.arch.embedded.piclist wrote: ...

Hi Jim

My remedy:

(1) I think you have to solder your prototype circuit together - and with as few solder joints as possible.

You can not use a prototype circuit, unless both wires and board springs have non-corroding metal contacts (gold...).

I have measured voltage "steps", in a cheap switch, while conducting current, because of small corrosion island (I assume). The switch is used for voltage division resistor ladder, in a home-made active analog voltage/current meter. It has a gain switch to change full scale between

If either your wires - or springs have contaminations og corrosion island, you have got noise.

Airborne smog/pollution will corrode, what can be corroded (with "help" from humidity and airborne saltcrystal). It can induce voltage, when united with the difference in springs and wire metal (alloys):

(I lowered the contact noise and contact resistance from 10 mOhm to 9 mOhm by greasing the switch contact plates and spring with super lube - silicone oil + very small teflon balls). I have measured the contact resistance on a separatly greased switch after 1, 2 and 3 years - and it still has the low 9 mOhm.

I know that power switches must not be greased, because the sparks burn some of the grease, transforming the grease to other chemicals that may be conducting electricity much better.

(2) Make sure your variable resistors has been gold plated - or maybe use a little to big resistor in parallel with a "big" variable resistor. Then the variable resistor will have the least noise addition (I think).

Some years ago I got a fine looking TV, that worked very erratically. Just by pushing the frame a little (or not at all), the red, green or blue color vanished, it shutdown or more obscure problem (dis-,re-)appeared.

The problem was that all the ICs was in transparent plastic IC-sockets. The sockets internal pin springs could be seen, and was black of corrosion, and that was the cause of the erratic behavior. Very small black flakes could be seen through the plastic.

Nowadays nearly all IC-sockets are non-transparent black - why? ;-)

INA126:

regards,

Glenn

t):

C..."

ed.

What type of caps are the best to use for this. Ceramic, mylar, mica, tantalum ...?

Hi Jim

Some (most?) variable resistors induce bad noise after a shorter or longer time (1/2-2 years). It can be heard in many transistor radios, pre-amplifiers:

DC on analog audio pot causing noise:

Repair Briefs - Audio Amplifers, CDs and Radios:

regards,

Glenn

You feed the sensor with AC obtained from the electronics in the lock-in circuit.

You amplify the output of the sensor with an AC -coupled amplifier, which can contain band- pass filtering on the excitation frequency, to combat interference and noise. You use a copy of the excitation signal to synchronously detect the output of the AC amplifier, and low-pass filter the result.

(alloys):

measurement:

For bypassing and suppressing RF? Ceramic SMT caps in 0603 size. 0805 size is mostly ok, too. 0.1uF directly from each supply pin to the ground plane. Directly, not 1/2" further down the road. For input signal filters you'll have to coarsely calculate the roll-off. Also, I wouldn't push it too low since ceramic caps have somewhat of a memory effect and act as a microphone to some extent. Use X7R ceramic on the supply and smaller C0G caps on the signals (unless you need large values).

Or use a filter and then an "ideal rectifier" circuit built with an opamp and a couple of regular diodes. But pick an excitation and filter frequency that is clean, maybe around 20Hz. Avoid any popular noise areas such as line frequency and all its harmonics. Also 25Hz and 30Hz because those are popular frame rate settings on interlaced monitors. Below 10Hz isn't that good either because then you get into the spectrum of mechanical rumblings.

text -

You need to do some reading. I expect there are some circuits around on the web.