Vložit
- čas přidán 19. 04. 2021
- Episode 770
I've wondered how the cheap rotary encoders work. Let's take a look. I also show how a fancy (expensive) optical encoder works.
www.banggood.com/custlink/K3K...
www.banggood.com/custlink/mDG...
www.banggood.com/custlink/KDv...
Be a Patron:
/ imsaiguy - Věda a technologie
Thank you very much! This is about the only video really opening up one of these devices. Concerning the LED, I suspect the smallest die is a current mirror to supply a constant current to the LED die.
Excellent content! Very informative! Thank you.
Rich.
So the 90 degree phase shift is what indicates which direction the shaft is being turned. Really interesting. I had no idea that HP even designed and manufactured their own optical, rotary encoders. What a powerhouse of a company they are! Thank you IMSAI Guy for continuing my informal education fine sir! Fred
The IC in series with the IR LED is probably a constant current source.
Great recon. Thanks for the look! Enjoyed.
All the information you say I know, but your approach to giving information is great and fun 🌟🌟🌟🌟🌟🌹🌹🌹🌹🌹
Thanks, have never seen the insides of a rotary encoder before!
Thanks for the teardown and making of the video.👍👍❤️❤️
Thank you for that. I'm looking into how to best handle switch bounce on these cheap encoders, and it was _hugely_ helpful to see inside.
Nice teardowns. I have never had the guts to dig into a unit on my HP gear.
The progress in mech contact types since the earliest units is significant for a number of reasons. The first mass-market I worked with were from the early 1980's microsoft mouse, which used a PCB disc with the quadrature pattern and a common ring on it in copper and three stationary contact (common, channel A and channel B). Really unreliable, but they worked well enough. Several generations later, we have a four-point rotating contact, five contacts for each channel, offset by roughly 90deg. for 20 cycles per rotation. Much better for reliability, having the spring contact moving and being a single, self-referencing part. The phasing in either case is +/- maybe 30deg, but fine for a manual encoder.
The encoder can be treated as a mechanical implementation of a Mealy state machine, with the state being the quad position and the input defining transition being rotation direction. This makes designing hardware around them a LOT easier in some cases.
First optical encoder I worked with was in a Boeckeler 'electronic dial indicator" used as a position pickup on an optical comparator. No LEDs. Just a grain-of-wheat incandescent bulb. It was periodically necessary to open the unit and adjust trimmer resistors for each channel phototransistor. Circa late 1970's. It was a bodge on a conventional dial indicator with the interrupt disk replacingthe needle, and a light-tight case added with the electronics. By the mid-1980's off the shelf opto-interrupt units with LED and phototransistor were available.The unit I first dug into is still in service, on 24/7, many decades later. The negative sixteenth (thirteenth? Ok.
The "slotted disk + light emitter + light receiver" is a classic! It would be nice to see that encoder biased under the microscope (most cameras can see infrared)
Not sure if anyone cares but if you guys are bored like me atm then you can watch all of the latest series on InstaFlixxer. Been binge watching with my brother for the last couple of weeks =)
@Ryland Trenton definitely, been watching on InstaFlixxer for months myself :)
Thank You...
I've been trying to learn SDR stuff and my reaction to "its high for two counts" was - QUADRATURE SAMPLING!
Thanks!
A truth table like that is called a Gray Code (or reflected binary code). The main feature is that only 1 bit changes at once. With regularly binary counting, more than 1 bit can change at a time. This is a really simple way to ensure there are no race conditions. For example, if the switch goes from 11 to 00, you may read it after only the first wiper has changed and get 01 instead of the real value. That can't happen with Gray code.
Camera Tip: Please try to use the highest F-stop for the available light, this will give you a greater depth-of-field and have more of your object in-focus. In some case you have to add light but with most sensors, you can usually stop them down all the way with bright-light.
Thanks for the optics information :). you are correct except it gets complicated by video. You are forced to use shutter speeds that are a multiple of the light flicker. so 1/60 1/30 ... So 1/30 is the slowest for motion blur. So the f/# is matched to 1/30 shutter. Then the only other variable is ISO. I have the ISO set to 400 and don't like to run it higher. That limits my range of f/# choices. It is also complicated by the fact I'm using a macro lens that has a shallow depth of field. I could take still photos with small apertures. Reminds me of the f/64 club from the old photography days. I took some nice photos with my 4x5 view camera at f/64. You mention introducing bright studio lights, and that would help, but is way more work than I care to do.
@@IMSAIGuy OK, I understand about the light synchronization that would definitely be a limiting factor. Long ago a friend of mine lent me a hand built 4 x 5 press camera with an extraordinarily great lens. It had F/64 and you could mount it on a tripod and take just extraordinarily great photos. We used to buy bulk 8 x 10 film and cut it down using a film cutter; also we did all of our own dark room processing--I really miss those days, they were ones where you were very careful before you squeeze the trigger! Anyway love your channel and videos. 73...
Presumably, if the radial spacing of the detecting photodiodes is different from that of the slots in the stainless disk you can make a moire effect and get massively better resolution?
Otherwise, why have more than 2 photodiodes (for direction sensing)?
Please ask your mate!
This is great! thanks for making this. Those last 2 bangood links are for the optical encoders?
yes, various styles, not optical
Thanks for the video. One question: if my mechanic rotary encoder works only every second step (both directions), is it the encoder fault or other component on the board?
Thanks for your help in advance.
could be either, but mechanical things usually break before electronics.
@@IMSAIGuy Unfortunately it is a new Aixun T3a soldering station (I made a YT video showing the problem). Do you have any idea how can I measure what is the problem without oscilloscope?
@@norbithehun there are two wires that give the pulses. so you could measure the voltage of one and turn the wheel and see if you get pulses, then look at the other and see if that one is giving pulses too. if only one works you may have a solder bridge shorting out things (seen that before). if it counts up and down turning left or right then you don't have a short. the encoders are very inexpensive do you might just change it out and see if that fixes it. Of course you can't use your new soldering iron 😀
@@IMSAIGuy Thanks for your help. It is working both directions, but only every second "klick". Anyway, I will measure both signal wire with my multimeter in max/min measuring and I will see. I will change it if it's faulty, but I have to find the same rotary encoder.