While I was thinking of how I can eliminate the ADC conversion, and the possible problems with the needle and vinyl (needle drop, bad signal quality, signal levels, worn vinyl etc) I had some ideas. My main goal is to convert the rotation of the vinyl somehow, other than capturing the audio.
First I searched for some magnetic encoders, which is could assemble around the shaft and in the center of vinyl, and measure the rotation. I think Numark CDX has something like this. A 10 inch special plate with something in the middle. I probably cannot make it work without modifying my turntable. I need wires for the power and the outgoing signal, so I dismissed the solution for now.
The next idea was to measure the record movement with an optical mouse sensor. I found some very poor quality home made mouse scratch devices on the web. So It could probably work. I tried my wireless mouse and moving the record under the sensor made my cursor also move.
So from now I’m gonna focus on this soulution parallel to the Tascam project.
I disassembled my Microsoft Wirelless Mobile 1000 mouse. I found a PAW3204DB sensor chip inside. It is made by Avago, which was acquired by Pixart. I found the datasheet on the pixart webisite, as well as 20 other optical mouse sensor chips. I studied all of them, to found out which one could work for me easiest. Unfortunatelly it is hard to tell which mouse contains which chip. There are 4 kind of chips:
– very low cost with integrated USB interface. I cannot use these kind of chips, because I can not acces them directly and accessing them through a USB Host interface is too slow and complicated (However i could do it with the DISCOVERY board)
– laser mouse sensors. I don’t want to use them. These are made for gamer mouses with ultra high resolution, but there are a lot of problems with them, and they are expensive, so the mice are expensive also that contains these chips.
– different type of sensors with lower DPI resolution, lower framerate, and direct access to each pixel of the CMOS camera inside (usually 16×16 pixels or similar). There are some project on the net to use them. These are very old I think, and also cannot guess which mouse is assambled with them. The advantage of these is that their interface is a 4-wire normal SPI.
– PAW3204 and 3205 chips, which are designed for wireless mice. You cannot acces the pixels directly, but I don’t want. It is easier to find donor mouse, because I think most of the wireless ones use the same chips inside. And I know at leas one product that has this chip. Microsoft Mobile 1000 which I have. The disadvantage is that they use a 2-wire half duplex SPI. Which sux, but can be managed with a bit of work.
The conclusion is, I have to use the last type of chips. There is a big chance, that it can be rebuilt by anyone, using a commercial cheap wireless mouse anywhere. Don’t have to buy a bounch of devices to found the righ chip inside. I bought 3 random mice. 2 wired and the cheapest 2.4GHz wireless. I haven’t had luck with the wired ones, they contained the integrated chep USB chipsets. But the wireless one had the PAW3205DB chip. BINGO!!
The mouse is: HAMA AM-7300
The difference between PAW3204 and PAW3205 is that the last one does not have the pin #1 connected, and does not have a Quick Burst mode. But I can live with it. Maybe the CPI resolutions are different, I don’t remember. The datasheet for 3205 is not too useful, I found a detailed on for the PAW3204 which is good. Actually, I realized during reverse engineering, that the PAW3205 datasheet is not correct. The register default values should be different from PAW3204 default values regarding the datasheet, but they are actually the same. Never mind, I used the PAW3204 datasheet.
To summerize. The right choice is cheap a 2.4GHz wireless mouse with a PAW3204/5 inside. The default resolution is 1000 count/inch, and 3000 frames/sec. The resolution can be set to 400/800/1000/1300/1600 counts per inch. I don’t know which will work the best for the vinyl scratch movements. Maybe lower ones.
Next post will be about reverse engineering the communication inside the optical mouse. And later, connecting the chip to my STM32F4 DISCOVERY board. I don’t really know now, if the optical sensor will be accurate enough when assambled above the vinyl. I’m going to have a lot of work to connect it to my board and test things out.