Naar Nederlands

Of all the 50 model engine engines I've made over the years I know most of the features and "secrets", but not at what speed they are exactly running.
Some time ago I made a tachometer using a little DC motor from a VCR, see the picture below.

When the rubber wheel is pushed against the flywheel of a model engine, the DC motor rotates and then acts like a voltage generator. The DC voltage generated by this motor is a measure of its rotation speed and thus also the rotation speed of the model engine. By pushing the rubber wheel against the three jaw chuck of my lathe, I was able to determine the (linear) relationship between the known speeds of the lathe and the generated DC voltages. With this calibration procedure and taking the fly wheel diameter of the model engine into account I can determine the speed of the model engine.

The problem however is that the power of most of my small model engines is so that the engine speed is reduced when I push the rubber wheel against the flywheel. Even more so, all of my very small Stirling engines and Flame Eaters are hardly or not at all able to drive this DC motor and they mostly stop running as soon as the rubber wheel makes contact with the flywheel.
So I had to look for a tachometer that does not make any mechanical contact with the flywheel.

A self-made contactless tachometer
I submitted my problem to my son-in-law Maarten Pennings who is very knowledgeable about micro-electronics and the basic software for that. I knew that he regularly makes successful experiments for which he usually purchases the relatively cheap (Chinese) hardware through the Aliexpress website. He accepted the challenge to design a contactless rpm tacho meter for me and immediately started the project with enthusiasm. He ordered a double set of components for this tachometer, namely:
- A microcontroller (ESP8266) that is programmable with Arduino.
- A sensor with infrared emitter & detector.
- A 4-digit 8x8 matrix display.
The price per set was about 6 Euro.
Maarten lives 85 km away from me and therefore he had ordered to deliver the second set directly from China to my home address. With his own set he developed the software for the system.
I have wired these three components according to Maarten's instruction and mounted them to a 8mm thick transparent plastic plate, see the pictures below:

This isn't going to win any beauty contests, but in doing so the micro controller is easily accessible and the assembly is solid and robust which I like.
The picture below shows an example of a measuring arrangement on a Stirling engine on the bases of which I will explain the operation of this system globally:

A reflective strip has been glued on the flywheel. In this case the strip is simply a piece of 8 mm wide sticky white paper tape. The sensor that is placed directly opposite to the flywheel transmits the infrared radiation and also receives it via the reflection on the flywheel material and the white piece of paper. The reflected radiation is shortly higher when the white paper passes the sensor and the program made by Maarten (which is running in the micro controller) measures the time in milliseconds that passes between two passages of the white paper piece and converts it to revolutions per minute. Maarten has made the software for the micro controller such that fly wheel turnarounds that differ significantly from the moving average are ignored, resulting in a stable output on the matrix display. For example such a deviating turnaround time may occur by signal spikes (much like contact bounce) at the transitions.
Maarten has made several software updates to optimize the operation of the system.
Through several Skype sessions Maarten guided me through the software installation on the micro controller.

This is what I understand and how I look at it as a simple user. For more detailed and more responsible information about the hardware and software, I can refer to the Maarten Pennings GitHub page: click here.

The results
1. When I had wired the components according to Maarten's instructions, the output on the display was corrupt and totally unrecognizable. My son Marc who is also an adept at using this kind of micro-electronics and regularly discusses with Maarten about it discovered that the four elements of the matrix display were placed wrongly on the board by the manufacturer (rotated 180 degrees). When he corrected that, the output on the display was as it should be. Maarten could not have seen this error because my display had not been home with him.

2. It turned out that the system does not work well with a white paper strip on a high polished aluminum flywheel like some of my model engines have. Presumably, the reflection difference between the white paper and the blinking aluminum is insufficiently large to generate reliable signal transitions. When I replaced the white paper tape on such a shiny aluminum flywheel by a black tape the system worked fine. In addition, according to Maarten, this is also "by design" because the software measures the time between black and white transitions and it does not matter if the majority of the wheel is white or black.

3. One of my larger 4-stroke engines has enough power to measure the speed with the contact method as described in the introduction. I was able to apply both methods at the same time and it appeared that measurement results were equal. Therefore, I can conclude that the contactless speed counter according to Maarten's design, shows the correct speeds.

4. I made videos of speed measurements on four of my engines: two Stirling engines, a Flame Eater and a 4-stroke internal combustion engine. The YouTube movie bvelow is a compilation of these video's:


The first video fragment was made by Maarten which demonstrates the operation of the rpm tachometer with his (Lego) test setup. He also shows that when the sensor is moved away from the fly wheel, the last measured speed is kept on the display. By pressing the "flash" button on the micro controller the flywheel runtime can also be displayed in milliseconds as well as the number of turns per second (Hz).

Thanks to Maarten for all his efforts and advise and my son Marc for discovering the incorrect assembly of the four display elements.