LEDs and photodetectors have to be a certain size and this would limit the resolution of a small encoder. They got around this using the mask plates (phase plates). When the lines on the wheel and the lines on the mask plates are aligned, then half of the LED light gets detected. When the lines on the wheel is between the lines on the mask plates, then the mask plates block half of the light and the wheel blocks the rest so that the detector is dark. To get quadrature encoding, one set of lines on the mask plate is offset 1/4 of the wheel line period from the other set of lines. The spacing between the line sets can be used for other detectors if differential signals are desired.
I may have mentioned this before. My apologies if I did. I'm old and forgetful. It sounds like you were working at HP when I was at Lovejoy. We had a photographer take a promo photo of the control card in a 25 hp motor speed controller we were doing a burn-in load-test on. So it was spinning the dyno at full rated power when the photographer took the shot from 4 feet away using a big Honeywell Strobonar. It appears the light flash went right through the HP gray fiber-optic receivers and triggered all 3 phases of the GTO gates simultaneously, and 2 of the 3 GTO's grenaded. Miraculously, the photographer wasn't hurt, except for falling backwards onto the floor. The president of Lovejoy called HP and told them about the problem. They said that wasn't possible. But the next day he got a call back saying HP was sending us a box of black photo receivers. It was clearly a very uncommon failure mode, but one that did happen, and could happen again.
What a 'cool' part of modern technic! First time I can see, a rotary decoder 'decoding' when turning the shaft and understanding what happens, ... ! Brilliant device ! Absolut brilliant ! ! ! 73 de Markus - db9pz (JN39fq ; 5km/3miles east of LX)
Interesting stuff. Reflective demo-box: Looks like the knob shaft slides in and out a bit. Am curious if it's designed to work accurately through that type of motion... e.g. maybe it can have a pushbutton under. Also, the index-pulse! Wow, I get it now, it's aligning a very specific marking, as opposed to just having a single slit. With a tiny bit of offset in scaling between the mask and the disk, it could probably be far more precise than a single slit alone.
Speculation: One-sided 'Reflective' encoder could avoid using transparent glass, and thus could (conceptually) be made far more indestructible at lower cost than Thru models. Just in case an operator needs to ride along in a 30,000g smart motor shell.
I'm old enough that I recall seeing some encoders that were not optical, but purely mechanical. They had a rotary platter with conductive intermittent arcs, and wipers (one per bit) brushing along, either making contact or not. They used Gray Code * of course, because the mechanical alignment of the arcs was not very good at the required tiny scale. They were the size of a baseball or even a softball. Number of bits was never 16 (IIRC), but somewhere around perhaps 10 or 12 (so thousands of features per 360-degrees). An advantage was that they were power-on absolute, i.e. they didn't need a full spin to find the zero reference (optical encoders can do that too, if designed for it). * FYI for those following along at home: Gray code is a mixed-up binary where only one bit changes at a time.
Didn't IMASI Guy does those 'mechanical' (with wipers) in vid #770? I mean, you don't have to be to old: they still use the mechanical ones a lot... and China makes a lot of them... for cheap... they are not very robust, or reliable...like the ones you are referring to... but they are cheap. 3 rotary encoder technologies... beside the wiper and contact type... and also way way more expensive are inductive, optical, and magnetic. A high end rotary encoder is ex... pensive!
Also, just to add the mechanical encoder you described, using Grey code, was an 'absolute' encoder. (There are incremental, and absolute, types.) The absolute type used to be much more expensive than the incremental type. There are also optical encoders that are absolute encoders.
Do you want to make a tutorial how to run an alc883 HDA codec on an esp32 or stm32, maybe with usb as input? Some stm32(nucleo F413ZH) have SAI with selectable ac'97 option. I did not find any example, tutorial on the internet how to connect hda(or ac'97) codec to microcontroller, and use it.
Remember those HP-encoders from stuff I worked on around 1990!!! That was when the microprocontrollers were 8052’s, lol, so we used some decoder-chips to process the QPM output to a binary output. Great was my surprise that those things showed up in cheap computer mouses just a few years later. In the 80’s when I started my electronics career, those encodersbwere all analogue synchro-transmitters.
has lots of potential, but contains a dull attitude! says: I just chase my tail ! U spin me round, round! I'm so dizzy, my head is spinning! love, love me. doodoo! space 4 rent! hi:)
That box reminds me of the quadrature encoder in old Quantum ProDrive LPS hard drives. Really cool idea, but this technique to locate tracks was probably never used much later. I think it was in the 52MB, 120MB or 240MB size. Mine all have a (c)1988 DisCache IC. These make for a really fun teardown!
LEDs and photodetectors have to be a certain size and this would limit the resolution of a small encoder. They got around this using the mask plates (phase plates). When the lines on the wheel and the lines on the mask plates are aligned, then half of the LED light gets detected. When the lines on the wheel is between the lines on the mask plates, then the mask plates block half of the light and the wheel blocks the rest so that the detector is dark. To get quadrature encoding, one set of lines on the mask plate is offset 1/4 of the wheel line period from the other set of lines. The spacing between the line sets can be used for other detectors if differential signals are desired.
I may have mentioned this before. My apologies if I did. I'm old and forgetful. It sounds like you were working at HP when I was at Lovejoy. We had a photographer take a promo photo of the control card in a 25 hp motor speed controller we were doing a burn-in load-test on. So it was spinning the dyno at full rated power when the photographer took the shot from 4 feet away using a big Honeywell Strobonar. It appears the light flash went right through the HP gray fiber-optic receivers and triggered all 3 phases of the GTO gates simultaneously, and 2 of the 3 GTO's grenaded. Miraculously, the photographer wasn't hurt, except for falling backwards onto the floor. The president of Lovejoy called HP and told them about the problem. They said that wasn't possible. But the next day he got a call back saying HP was sending us a box of black photo receivers.
It was clearly a very uncommon failure mode, but one that did happen, and could happen again.
What a 'cool' part of modern technic! First time I can see, a rotary decoder 'decoding' when turning the shaft and understanding what happens, ... !
Brilliant device !
Absolut brilliant ! ! !
73 de Markus - db9pz (JN39fq ; 5km/3miles east of LX)
Interesting stuff.
Reflective demo-box: Looks like the knob shaft slides in and out a bit. Am curious if it's designed to work accurately through that type of motion... e.g. maybe it can have a pushbutton under.
Also, the index-pulse! Wow, I get it now, it's aligning a very specific marking, as opposed to just having a single slit. With a tiny bit of offset in scaling between the mask and the disk, it could probably be far more precise than a single slit alone.
Speculation: One-sided 'Reflective' encoder could avoid using transparent glass, and thus could (conceptually) be made far more indestructible at lower cost than Thru models. Just in case an operator needs to ride along in a 30,000g smart motor shell.
Absolutely fascinating!! 🤓
I'm old enough that I recall seeing some encoders that were not optical, but purely mechanical. They had a rotary platter with conductive intermittent arcs, and wipers (one per bit) brushing along, either making contact or not. They used Gray Code * of course, because the mechanical alignment of the arcs was not very good at the required tiny scale. They were the size of a baseball or even a softball. Number of bits was never 16 (IIRC), but somewhere around perhaps 10 or 12 (so thousands of features per 360-degrees). An advantage was that they were power-on absolute, i.e. they didn't need a full spin to find the zero reference (optical encoders can do that too, if designed for it). * FYI for those following along at home: Gray code is a mixed-up binary where only one bit changes at a time.
Didn't IMASI Guy does those 'mechanical' (with wipers) in vid #770? I mean, you don't have to be to old: they still use the mechanical ones a lot... and China makes a lot of them... for cheap... they are not very robust, or reliable...like the ones you are referring to... but they are cheap. 3 rotary encoder technologies... beside the wiper and contact type... and also way way more expensive are inductive, optical, and magnetic. A high end rotary encoder is ex... pensive!
Also, just to add the mechanical encoder you described, using Grey code, was an 'absolute' encoder. (There are incremental, and absolute, types.) The absolute type used to be much more expensive than the incremental type. There are also optical encoders that are absolute encoders.
Do you want to make a tutorial how to run an alc883 HDA codec on an esp32 or stm32, maybe with usb as input?
Some stm32(nucleo F413ZH) have SAI with selectable ac'97 option.
I did not find any example, tutorial on the internet how to connect hda(or ac'97) codec to microcontroller, and use it.
Remember those HP-encoders from stuff I worked on around 1990!!! That was when the microprocontrollers were 8052’s, lol, so we used some decoder-chips to process the QPM output to a binary output. Great was my surprise that those things showed up in cheap computer mouses just a few years later. In the 80’s when I started my electronics career, those encodersbwere all analogue synchro-transmitters.
has lots of potential, but contains a dull attitude! says: I just chase my tail ! U spin me round, round! I'm so dizzy, my head is spinning! love, love me. doodoo!
space 4 rent! hi:)
Capsule corp
That box reminds me of the quadrature encoder in old Quantum ProDrive LPS hard drives. Really cool idea, but this technique to locate tracks was probably never used much later. I think it was in the 52MB, 120MB or 240MB size. Mine all have a (c)1988 DisCache IC. These make for a really fun teardown!
Most interesting!
Brilliant. :)
those ecoders are so smooth to use just like a VFO