How to choose the best ultrasonic sensor

Hi, I'm preparing for my bachelor's project and I'm not really sure how to choose the best ultrasonic sensor. I need something reliable, with very small angle of detection something which wouldn't report crazy made up distances like HC-SR04. The maximum distance for the sensor to measure would be 2 meters and I'm going to use two of these sensors side by side. Has anybody got experiences with ultrasonic sensors?

For instance, can anybody compare these three sensors - HC-SR04, US-100 and Maxbotix HRLV-EZ1?

luckyx182:
...something which wouldn't report crazy made up distances like HC-SR04...

Define crazy distances.

Pick a sensor that has the most comprehensive data sheet. In particular, look for the behaviour when no object is detected. The first cheap HC-SR04 I used had no automatic timeout so if nothing was detected it would just wait for ever and ignore attempts to send another pulse. The only way to get it working again was to recycle the power.

The general problem with all cheap sensors is that they don't give you an option to ignore objects that you know are too close to be the thing that you are interested in. They only give the distance to the closest object, not a list of distances and signal strengths as you would get with a more expensive sonar system.

mikb55:
The general problem with all cheap sensors is that they don't give you an option to ignore objects that you know are too close to be the thing that you are interested in.

I don't need that. I will monitor the distance to the closest object and when the distance shortens (somebody going by) I need to trigger an event. But the HC-SR04 reacts differently on various surface types, the sound gets reflected here and there and I get wrong distances to the closest object.

luckyx182:
I don't need that. I will monitor the distance to the closest object and when the distance shortens (somebody going by) I need to trigger an event. But the HC-SR04 reacts differently on various surface types, the sound gets reflected here and there and I get wrong distances to the closest object.

The basic of Ultrasonic rangefinder is Ping - Pong, measure time and convert to distance, so every U.S sensor have the same problem with material and object surface.
In my oppinion HC-SR04 is fine but its better using SR05. Its not much more expensive but reliable, Im able to measure from 2cm up to 600cm (under 450cm is more accurate)

If you need narrow beam, look for Maxbotix sensors, ofcourse more expensive.
Or you can add some other type of sensor like IR distance sensor from Sharp to make sure you get the right result.

Minh.

I think the Parallax Ping ultrasound sensor is better than the HC-SR04 sensors.

I've also used a Maxbotix which I think works well.

I haven't used the Maxbotix HRLV-EZ1 but my guess from reading the datasheet is it's better than most other sensors.

As osnoneo points out, all ultrasound sensors will have similar weaknesses when attempting to measure objects which absorb ultrasound.

There are some nice hobby level laser rangefinders. If you need to reliably measure distances were ultrasound sensors fail, a laser rangefinder may be the better sensor for this application.

SparkFun sells LidarLite and Parallax sells a variety of laser rangefinders. I tend to think the ones sold by Parallax are better quality but I'm biased. I occasionally do freelance work for Parallax in exchange for parts.

I have also had issues with the HC-SR04. Particularly, the sensor has a significant limitation with how many samples per second one can collect as a ping out has to be initialized before the receiver can "listen".
In terms of accuracy/resolution, since the HC-SR04 is 40kHz, the maximum possible resolution is +/-0.3cm since that is the wavelength of sound at that frequency. To get +/-0.03cm accuracy, it is possible but it would require an ultrasonic sensor with a 400kHz transducer. There are of course tradeoffs when increasing the frequency, the biggest being the loss in signal strength. This would significantly reduce the maximum distance that can be read reliably.
The HC-SR04 also has several manufacturers that have slightly different variations in their circuit design. I have brought over 2 dozen of these sensors from different sources and had them perform very differently.
Right now I'm working to design a 400kHz ultrasonic sensor similar to the size of the HC-SR04 but without the limitations. The transducers themselves can be purchased for around $15 from China but the electrical circuit to use them must be designed using a variety of processors, resistors, capacitors, step up transformers, filters, etc...