Society of Robots - Robot Forum

Electronics => Electronics => Topic started by: TrickyNekro on June 29, 2011, 12:28:08 AM

Title: Low distortion, high slew rate op amp for Ultrasonic sensor
Post by: TrickyNekro on June 29, 2011, 12:28:08 AM
Hello guys, greetings from Greece


I'm building an Ultrasonic sensor as a summer - vacation project.

What I'm thinking? I'll be using the op amp to amplify the ultrasonic traducer signal from the sine wave I have at the input to a square wave to be able to read it with a microcontroller. This is needed cause I'll be using Frequency modulation to pass a signal with the carrier frequency so that the device is not prone to echoes.

For the FM to be accurate I need a rail-rail, high slew rate op amp. But being in need to amplify very small signals I also need the op amp to be
of very low distortion so that I don't get many garbage in the signal. (which some of them can be processed out by the microcontroller).

To make a list of the features needed I'll post this list:

* 5V supply minimum (3.3V isn't bad at all but can also use 5V)
* high slew rate ~ 4 - 5 V/μS and higher (preferably higher)
* low distortion at 40kHz
* extremely low bias current at inputs ~ 1 - 20 fA (preferably FET input devices)

Right now I'm using LMC662CN but it's not fast enough. It actually doesn't exit the high rail saturation fast enough.
And in such a circuit, it's very difficult to avoid saturation (basically saturation is wanted but I have also some other ideas that it may not be necessary so let's drop that discussion for now)

That's all that I can think right now.
I'm looking forward to your recommendations. I usually use National cause they can supply samples to work with.

I have TI also in mind, I'm currently browsing around their sites right now, but still recommendations come in handy.

Best Regards, Lefteris
GReece
Title: Re: Low distortion, high slew rate op amp for Ultrasonic sensor
Post by: newInRobotics on June 29, 2011, 01:33:35 AM
Hi TrickyNekro,

You might be interested in Maxim (http://www.maxim-ic.com/products/amp_comp/) web page. Just have a look at it and make Your choice (You might even be able to get free samples to prototype Your application).
Title: Re: Low distortion, high slew rate op amp for Ultrasonic sensor
Post by: TrickyNekro on June 29, 2011, 01:59:51 AM
Yup, I always have them in mind for samples but thanks anyways!

I just sampled some OPA4140 for Texas Instruments. They seem to be up to the task!


To tell the truth, I'm more in need of experience here. Someone that has worked before with high gain applications
and has something in mind to recommend, because he actually used it before.

But I'm not in place to turn away any help I can find, so thanks again ;-)


Best Regards, Lefteris
Greece


BTW, if I do manage to build that thingy I'll make it open source. Haven't seen many nice sensors schematics on the internet
and a lot of people just buy that standard products, crippling their versatility.
Title: Re: Low distortion, high slew rate op amp for Ultrasonic sensor
Post by: waltr on June 29, 2011, 09:30:07 AM
Op-amps can be tricky things to spec. Almost all op-amps that their output slammed into a rail will take some time to recover. It is best to avoid that situation by clamping the inputs. If you need high gain then it is usually better to use more than one op-amp is series so that the gain is distributed. When a lot of gain is in one stage then oscillations can become a real issue. I tend to keep gain to 20X or less in one stage and careful to compensation to prevent an peaking.

A good op-amp is from the OP27 family, there are several newer and improved versions of this op-amp offered by several makers.
Some other op-amps that may fit your needs are: ADA4898, LT6230, AD8021, OP2177, OPA827.
Title: Re: Low distortion, high slew rate op amp for Ultrasonic sensor
Post by: Billy on June 29, 2011, 01:12:55 PM
Not a direct answer to your question but here is a thought:

Increase the analog PS voltage to keep signal away from the rails.
If at the point of going digital you need it fast at the digital rail, consider a comparitor (which are designed to be fast at extremes).

And as said before, don't put too much gain into a high frequency amp, it'll oscillate unless very carefully designed (well above my level of ability, not sure about yours).

Good luck, and please let us know how it goes.
Title: Re: Low distortion, high slew rate op amp for Ultrasonic sensor
Post by: TrickyNekro on June 29, 2011, 03:47:22 PM
Op-amps can be tricky things to spec. Almost all op-amps that their output slammed into a rail will take some time to recover. It is best to avoid that situation by clamping the inputs.

That's more than true, still I'm aware of it. Every transistor device will take time to exit saturation. Still the inputs aren't clamp able. The input signal is amplified by two units. The output from the transducer is a 40kHz sine wave of only some mV amplitude. The first stage amplifies the signal to some hundred of mV and the second stage to ~4V volts.

The inputs can't be clamped for two reasons. Amplitude varies with the strength of the signals received and the signal falls drastically as the distance between the sensor and the wall increases.
The second reason is that these voltages are extremely low to be clamped effectively. How to clamp it? With schottky? No go....  :P

If you need high gain then it is usually better to use more than one op-amp is series so that the gain is distributed. When a lot of gain is in one stage then oscillations can become a real issue.

That's what I do  ;D !!! I'll upload the the circuit to give it a look.


A good op-amp is from the OP27 family, there are several newer and improved versions of this op-amp offered by several makers.
Some other op-amps that may fit your needs are: ADA4898, LT6230, AD8021, OP2177, OPA827.
Thanks for the suggestions mate. If the currently used don't come up to the task I will certainly start the shoot out.


Increase the analog PS voltage to keep signal away from the rails.
If at the point of going digital you need it fast at the digital rail, consider a comparitor (which are designed to be fast at extremes).
Hmmm, in a lab set up yes that can be done... But on a robot, it's an over kill to have both 12V and 5V...
I'm building this sensor for the "masses". Hobbyists, who are creative and don't want to pay the the big bucks companies.
Most of them usually use the 5V rail to power electronics so... The simple version will bet on them.

Good luck, and please let us know how it goes.

I sure will mate! Not here, I'll try to have some results first. But I have good news.

I tested some single frequency burst, and it gave a good enough reading about the distance.
I didn't do accurate testings, but the circuit acted good enough. I'll keep building the analog part for sometime
then post my results!!!
Title: Re: Low distortion, high slew rate op amp for Ultrasonic sensor
Post by: Soeren on June 30, 2011, 04:48:56 PM
Hi,

I'm building this sensor for the "masses". Hobbyists, who are creative and don't want to pay the the big bucks companies.
That doesn't ring too well with superfast hard-to-find op-amps (and I'd think that the creative hobbyists make them themselves, or else I don't see where the creativity lies?), but if you need fast, you could use something like LTC6409 (http://www.linear.com/product/LTC6409), which is a 10GHz GBW, 1.1nV/√Hz (this is very low noise, which is far more important than distortion in your app.) Differential Amplifier/ADC Driver, 3300V/μs Differential Slew Rate, 52mA Supply Current, 2.7V to 5.25V Supply Voltage Range.
You can get better, faster etc. amps if you look hard, but OTOH I've seen US sensors made with very lowly op-amps (and even with transistors alone), so the LTC6409 is vast overkill.

You might wanna take a look at the LMV712 from National (which claims: 5MHz GBP, Slew rate 5V/μs, Supply current 1.22mA/channel, VOS< 3mV max, Guaranteed 2.7V and 5V specifications, Rail-to-Rail inputs and outputs, 1.5μA shutdown ICC, 2.2μs turn on, which should do very fine (still a bit overkill for a 40kHz US app. and not as widespread among hobbyists).