Very brave plan! However, one of the reasons (besides pure nostalgia) that I'm currently reviving my tape equipment is my interest in analog audio reproduction. Digital signal processing would entirely take away my interest in this.
Wow! I have always dreamt of a dd9 mech walkman controlled with an FPGA which could potentially shrink all the sound chips and controller chips into one allowing the walkman to be incredibly compact (and without some of those faulty caps). Did this idea ever become a reality or close to one?
I do designs for FPGAs as my day job. They are digital devices so not suitable for what you suggest. If you are doing something like emulating a Spectrum computer in an FPGA the most difficult part would probably be doing the analogue tape interface (which is why they are more likely to load games from an SD card or similar). What might be possible would be find suitable ICs to do a completely direct drive design using three brushless motors. I believe that has been done on cassette decks before and my SL-F1 Betamax VCRs definitely use that technique. Even washing machines are going direct drive these days eliminating all belts.
Oh I see. Digital reproductions I assume are considerably easier than analogue with FPGA (if even possible). I know very little of how FPGAs work but I thought they were able to replicate analogue components but I guess that isn't very realistic. That's a really cool job. I was wondering if I should learn about FPGA development but I wasn't sure how widespread their application was. I am still better with software design than electrical but I enjoy learning this sort of thing. Are you actually replicating a Spectrum computer or retro game console at your job? Developing a direct drive mech would definitely allow superior quality for companies making new decks.
To layman FPGA code (VHDL is what I use) looks very similar to software. In fact it reminds me of BASIC as you are often using IF THEN ELSE statements and CASE statements. The big differences are that you have much more control over the hardware and it can operate simultaneously. Need a 20 bit register rather than 16 bits? That is a one number change in the code. Need to write into ten registers simultaneously? Again that is easy in an FPGA but in a processor it would take many clock cycles. No. Doing stuff with specialist radio equipment. People might not know that even low end radios now use what they call Software Defined Radio. https://radiojayallen.com/panasonic-rf-2400-amfm-radio/ The big advantages are the elimination of lots of tuning coils etc, and that because everything is digital performance will be completely consistent from one radio to another. In fact if a radio uses one of the common Silicon Labs ICs things like filter bandwidth will be identical even between one manufacturer and another. Some of the Silicon Labs ICs feature allow a choice of about six bandwidths, which would have been very expensive to implement twenty years ago but now just costs an extra button on the control panel. he advantage of an FPGA over a fixed IC is that it can be programmed to give exactly what you want e.g. a filter bandwidth well away from the fixed options offered by any existing IC. FPGAs are more expensive though with the cheapest being around $20 and the most expensive top of the range ones from Intel or Altera being comparable to a top end CPU. From a hobbyist point of view FPGAs are being used to replicate classic computer systems like the Amiga. The advantage is that they can be designed to be clock cycle accurate and can emulate the CPU, Graphics, and other ICs simultaneously, something very difficult to do with emulation using just a CPU. A randomly picked article on why FPGAs are now the big thing for emulating vintage games consoles. https://hothardware.com/reviews/mister-diy-console-fpga When it comes to processors like the 68000 series, by far the fastest (which can actually be plugged into an Amiga) is actually an FPGA. Back to audio, the key point is that once things are being processed digitally the results should be 100% consistent. Whatever you think of MP3 players, if you pick two of the same model there is a very good chance they will sound exactly the same, unlike the Walkman I fitted new belts into this morning, which seems to be running slow and have one channel lower than the other. A final interesting point. FPGAs aren't new. The original Macintosh used programmable ICs called PALs which were the FPGAs distant relative. Thirty years ago I was using four of those to replace about twenty 74 series logic ICs on a board so we could free up enough space to add a DMA controller. Nowadays those four PALs and the DMA controller would easily fit in an FPGA.
Very good point Longman ! Indeed FPGAs are not suitable for this kind of application. If one wants to build logic control to a walkman, it can be done with cheap microcontrollers, which are both much cheaper and much easier to program. But even that has a problem, as you would need to design a custom mechanism (including custom motors, custom gears, etc.), since the only one that is available nowadays is a mechanical one. Doing the sound processing digitally is something that would only make sense if one wants to implement complex noise reduction systems (Dolby C, Dolby S, dbx). But I think that's not very useful, since even Dolby C is not that common. It is nice to have it, but let me give my personal example: I only own 2 units that can record in Dolby C: a Technics deck and the WM-D6C. And have another 2 walkmans that can just play in Dolby C. Given this, I do not record many tapes in Dolby C. Not to mention prerecorded ones, most of them being recorded without any NR and some with Dolby B. I would much rather focus on building a very low noise amplifier chain (from head preamp to heaphone amp), since in most walkmans that I have dealt with, much of the noise is coming from the amplifiers themselves, not from the tape (assuming a decent tape is used). Very few of them are what I would call low-noise. Dolby B can be built with discrete components, especially if it's only a decoder. The WM-7, for example, uses 4 transistors for Dolby decoding (for each channel, in total 8). So this would not add a significant cost to the unit. What I would want a new walkman to have are the following: - EQ switch for playing both type I and typeII/typeIV tapes - quality construction - low noise - low wow and flutter (as low as it can be using belts, because I do consider a disc drive or direct drive unit would not be feasable) - Dolby B NR (with the mention that it would be prefferable, but not essential) - recording with AC bias (would be prefferable to record on chrome tapes, but that would require a dedicated erase head, as a simple magnet would not work). Should have a LINE IN, not just MIC. The recording feature is again, prefferable, but not essential. I would prefer Dolby B decoding instead of a MIC only DC-bias recording. What a new walkman should not have: - built-in speaker - built-in microphone I say this because I want it to be able to reproduce music in high quality, not to be just a dictation machine. - poor quality circuits with lots of noise and lots of distorsion - cheap plasticky construction
