this is exactly what I am talking about! using 7474s or similar (just more in one chip is good rather than several in 'ripple'/series), Im doing this to A) have a RTC, B) a Breadboard clock, and C) because its a long IC chain, I get those 'Cs' for free.
I rather like 1024, 512, 256, 128 and 64, thats 5 notes on a monotonic scale, really, 64/128, 256 or 512 and 1024 or 2048 gives 3 'distinct' high medium and low sounds. Somewhere in there, I think this 'nice side effect of my breadboard clock' got turned into more than it needed be, and Im very interested in this, outside of 6502 context per se.
I do not think I can fit all that on the 54x70mm perfboard for this project.

GARTHWILSON is correct in the use of the 6522 as a single voice frequency generator. As I am unlikely to use a 7474 or 74574 or 74273 or something on the tiny can board, the 6522 use is really very helpful and good advice, as it not only can recreate the 'free Cs' from the RTC (easy to do chimes eh..., I might build that as a edu-solder kit...), the 6522 can make any other frequency too. When I move to the can, I might do that to get back these 'lost tones'


Im doing several smaller experiments over on the EEVBlog forums, and a lot of this sort of thing is definitely on the list of shirt term projects. Again, Im not trying to solve the 'create a human audible tone' problem, which leads to the next bit...


the overall goal is to make a tiny 6502 in an mint can.
the immediate goal is setup a breadboard clock thats 'better' than the analog adjustable one built on 555s from the videos...

I noticed that in the course of doing so, to maximize the use of my RTC and a few frequency dividers, I get 'free C note sounds'


In the course of becoming aware I could get 'free tones' out of my Clock circuit, and in discussing things here, it occurred to make a different project later, with an audio synth made with a PLL and divide-by-n counter.
If I can, Id use a PSG, if I could get one, a SN76489 or more likely a knockoff AY-3-8913 rather than any other solution, though as stated, there are a lot of ways to make an audible tone or 3. I would love, to get a 4 voice-channel PSG or similar. An old SID would be fun, though whatever I use, since I intend later to make this a kit/DIY-tutorial, I want it to be easily found.
it looks like you can get knockoffs of either on the 'bay, and they are under $2 each.

this is hands-down, near as I can tell, exactly what I want: SN76489 or SN76496
It does everything one could need for 8 bit audio, on a small package.
It is basically perfect and if I had to choose in I/O with a 6522 and a serial port, and this, Im not sure I can justify a 6522
the AY-3-819x series is 'a little better', though a bit more complex.


I am looking at that actually, not 100%, the through hole sockets will fit perfboard ok, breadboard ok, and are not bad to solder, a little worried about a bus and space for it, I'll see how it goes in the next step...


I might do that.

____________________________________________________________________________________________


yes Im looking at that for the breadboard, instead of like a potentiometer to select a 2, 4 or 8 Hz, or free running 32768 clock (since breadboards are slow) and since I was doing that anyway, Id use the beeps. I guess I could be clearer that Im doing the C-notes on the breadboard, because Im using my RTC can to double as my human-speed board clock. you are 100% correct that this would require extra hardware, and on my breadboard, because Im doing this anyway, it is convenient to do so.

"In the can", I really want a PSG the previously mentioned Ti tone generator,
its an altoids can, not the jobsite/kayak/hiking computer that takes PC Cards...

_________________________________________________________________________________________________



If I could get THAT kind of function out of a 6522, its worth it to include, unfortunately, the 6522 audio generator trick has but one voice, so suitable for beeps and boops, If Im adding a chip for sound, Im adding Sound Chip...

spoiler alert, Im adding a sound chip.
_________________________________________________________________________________________________

totally get that, and see something like this as a 'user modification'. something at a Chain Store for under $2 would be okay I guess.
Altoids cans are 'traditional' and cheap and ubiquitous.


yes, on my breadboard though. The logic ICs are in Narrow form, and not too big, its a simple circuit and yes, once you add the gate hardware to select notes (its likely to just be 3 tones or so) which is not a lot of read/write to turn tones on or off every 1/8th of a second or so... Mind you, Im just planning to do this on the breadboard, because I dont like a potentiometer based variable clock, I like stepping through a series of discrete choices, that circuit will be fun to build, I want to use TimeKeeper RAM, which has an RTC in MMIO addresses. so to get more used to that, I want to start with an RTC of some kind, there are several. If I can get TK-RAM, I'll use it 'in the can', for the BB, I want to stay simple. One of the first projects I intend to do, is just an RTC decoding to some 7-segment displays or such. Its good basic I/O and easy to see errors along the way.

So, because I am using the RTC and BB clocks off the 32768 can for now, and I need to breakout the frequency in a set of resettable hex or octal flip-flops, so I can get a clean , 1, 2, 4 or 8 Hz breadboard clock, I get these mentioned 'C notes' for free.

still, for just beeps and boops, the 6522 is probably better use of real estate and does a lot more tones. seems like the 6522 can pretty easily double as a single voice PC speaker driver too...


yeah the TK RAM has a built in battery, its pretty cool and you can get a 2k or 8k chip for like $5, maybe $10 for the bigger ones. It gets pretty expensive at a point, users can choose to have more RAM, Im just wanting 'enough', cheap and not in conflict with 6502 vectors (one of the chips date or time values is right over the 6502 hardware vectors in $FFxx, so you cannot use that size.

I "do not like" either SPI or IIC, I am slowly opening up to various serial protocols, it makes sense in certain contexts and as I explore 1- and 2- bit computing. Still, I prefer parallel MCU access. I do not understand SPI RAM. Why is this exists?


I might add one, it does do a lot... its really hard to justify in such a small device, with virtually no intended I/O

I tend to agree. to what shall I interface the altoids can to?
I am reading the 6522, its a really powerful port expander and has a ton of functionality, with some tricks, you can run several devices off each port...
great timers... I think if I understand them, if you wanted you could take the lower bits of T1 (or T2?) and write those to an address with a latch and each 'pin' on that address would have the corresponding frequency, as each pin would change accordingly to output a lot of 'flat C notes', how often you refreshed those would create your playback/sample rate. Combined with the other Timer and maybe some other bits for noise, you can build a rather basic sound chip with several 'Cs' a single voice in any other note and a bit of noise. Mixing this can be done flat and sent to an op-amp. This would actually be 'good' PC speaker and a 'very bad' PSG. Still, I think you could pull off some basic melody and tunes with a 6522...

Id really rather have that 76489

yeah reading the data sheet, it can certainly meet and exceed the pinouts on a binary divider...
I think you can get that out of a 6522 as well, if your source clock is set up to be 32768 (or another good multiplier, such as NTSC colorburst which will grant 432Hz or "some people's" A7 evenly), you take the binary counter especially the lower byte, output that to I think a Latched Register that can be sent to a speaker, the refresh is the sample rate (should be twice highest note) and you AND that with different gates to play any of those 8 notes...

so, not only can I use the 6522 the way you mentioned, I think I can get it to do what I am talking about as well, with about the same amount of extra hardware... wasnt one of those 6522 bugs solved by a flip-flop? if so, the having a multi-gate IC, means if you have any left over, if yo have a counter running at all, you can break out a few beeps and boops, and only require a read/write operation during an update loop, and the instructions to turn notes on or off as needed. you would have your say, octal flip flop (on a board bigger than this) and use a few channels on it for a few extra notes.
so, the VIA can be a really crumby sound chip too, in more ways than one. I'll look further into this later. Beyond the single voice at any note you mentioned, you can get 'reference notes' pretty easy too.


that is exactly what Im thinking of and called it the other thing.
yes, an rs-232 serial port. I can cram one on here. Its in the images I am posting below. I have no idea, what I would connect this to.
USB sounds a lot more 'current user friendly', while the DB-9 is pretty cool and allows a lot of fun stuff, I have no idea what anyone would do with it these days, other than get in trouble...


I was generically referring to both options and any others... 'encode' or 'program' is probably correct, Write to ROM is an oxymoron...
So, as I am not certain what form of Hard Storage, if any will be used, if it will be flash, a ROM chip, a ROM space on an MCU, an SD card, etc
what is the most generic term for 'make bits on chip align correctly'?

I do not know if adding software or changing programs will be possible without diassembly of the device... Im leaning toward an MCU with built in usb, or maybe a USB IC with charging and such.. I'll look into it


I just have no room on this board to do much, even with the smaller packages.
On a full size system, you might have enough stuff for all of this... I am not certain I will have much if any devices to connect to, even ICs here...
Anything I build beyond an absolutely minimal system here, will really be done on another project.



Id have to check the chart again, the PIC24 looks good for the 65816 in terms of similar data/address size, it can do a lot as a companion chip... the PIC16 or 18 is similar enough to the 6502, its not to recreate what a 65xx does, its to take loads off its back and organize whats coming in and do what it can
_____________________________________________________________________



if these were not stupid expensive, this is basically what I was talking about doing, yet in my head, you would select on if the Reference notes and use that to create the octave from its frequency, these are basically the same things...
_______________________________________________________________________________________



so this has 2x PLCC 44 sockets, a couple of placeholders for RAM and ROM, an oscillator can, a power jack, an RS-232 DB-9 Male(?), and a place for the ribbon to go to the screen, an 2 pin header for the speaker, it needs pins in for the buttons and an on-off switch, still its just to see how things fit.

I think Ill redraw tomorrow with a USB and room for the 76489 PSG instead... I simply have no real reason to include the 6522 on this design. (except I will probably use it later). For sound, Id use a PSG... I have real need for I/O on this.
CPU talks to RAM, Talks to Screen, Talks to sound IC, reads ROM, Reads and Writes to RAM, Reads 8 buttons for input...
Thats all it does I think, until you add some kind of storage (might not really change much or need a 6522), or I use USB (a USB chip is better) or a serial/rs-232 port...
anyway, yes you can fit one on here, though I am 'of a mind' that the 6522, and the db-9, are in 'user extension space' and I am more likely to add a headphone jack and a better sound chip personally, others might want a UART, or a better video chip, or more RAM or Storage, maybe an SD card slot instead...


6502 on left, RAM in center, ROM on bottom left, the thing in middle left is the connector for the screen, it doesnt look right call it a place holder. 6522 on right, , Oscillator, power jack leads and a serial port under that. The Caps are for decoupling. Its not really right, it does offer some clarity on how much space there is...

I am very torn on the 6522, and am now very interested in A) putting one on here to gain experience, or B) building a desktop computer with one in it later...

idk, I want to keep scope down, make this as hobby friendly and approachable as possible, and be 'finishable' and 'startable' for young people, old people, people that have shaky hands or are bad eyesight, etc. This should really have a focus on 'what can you do with it' after its built, something missing out there in many places.

6502, RAM, ROM, RTC, Screen, Buttons, Speaker, Battery.

Adding a PSG would be 'real nice'. Adding a 6522 lets it...?

_________________
http://WilsonMinesCo.com/ lots of 6502 resources
The "second front page" is http://wilsonminesco.com/links.html .
What's an additional VIA among friends, anyhow?

Actually, USB FIFOs such as the FTDI FT245 look pretty darn friendly to use, if I'm understanding the datasheet correctly. They appear as a USB serial interface on the host computer (the Linux kernel comes with drivers for these; Windows 10/11 either has them or downloads them automatically) except that it doesn't care about the baud rate and the like. On the device side you've got 8 bidirectional data lines, two pins to indicate ready for read and write, and two pins to strobe read and write. The chip has a 256 byte transmit (to host) buffer, and 128 byte receive (to device) buffer, so you can quickly load it up with data to send and need not worry too much about getting to it immediately when data become available.

On my List Of Things To Do is to add one of these one of my SBCs or microcomputers. The chip appears to handle 5 V TTL levels and the timings I think are fine for my little 8-bit systems (mostly running at 1-4 MHz), so I think it may work hooked directly up to the bus with just a bit of address decoding.

_________________
Curt J. Sampson - github.com/0cjs

I have very positive experience with FT245 module. It enables minimal designs without need of a serial port device, but the downside of FT245 is about 10 times more expensive than the cheap CP2102 USB-serial adapter.
Bill

A while back I picked up three FT245 development boards for under $20, including shipping, from AliExpress. I just checked and prices for that seem still about the same. The cheap USB serial interfaces are about a third the price ($6 for three, including shipping), but then you still have to buy the UARTs as well.

_________________
Curt J. Sampson - github.com/0cjs

Good to know Aliexpress has better price for FT245. I’ll buy a few; it is quite useful for high speed transfer, like video data for animation. Aliexpress’ CP2102 price is around $1.25, similar to eBay price.
Bill

yea i can understand why you would think of USB as a black box compared to older communication standards, because it pretty much is.
but if that's a good or a bad thing depends entirely on the person asked.
and i can tell that you like being able to look deep into what is going on in your data lines and such, and USB is simply not made for that. but that doesn't mean it's completely unusable by hobbiyst.


why does it matter how many pages the spec has? the SD card's simplifed physical spec alone has 363 pages, but it's still a pretty common storage option for hobbiysts.
plus there is rarely a reason to even look at such spec sheets anyways since the most important information has likely already be extracted by someone else online. (though the more obsture the info you need, the more likely it becomes that you actually have to open some spec sheet)

and yes, the lowest layer you have with USB (the packet stuff) is way more complicated than dealing a UART (even the fancy ones), so if someone doesn't want to learn about all of that there are 2 options.
1. stick with a UART because it's simple to understand.
2. use some IC that does all the work for you. like the below mentioned FT245/FT240X
ironically those turn the entire thing around becuase they're actually much easier to deal with than a dedicated UART IC (both in terms of hardware and software).


the FT245 module is quite expensive, but the IC itself is pretty cheap.
of course there is also my favorite, the FT240X... very similar features compared to the FT245R but at only half the price! (~2 USD instead of 4)
either chip are a great replacement for a UART if (like me) you'd never use the RX/TX lines for anything else besides connecting to a PC.

Garth’s comment about vertical space remind me of this 2.1”X2.8” overclocked 6502 SBC, stacked 3 chips high.
https://www.retrobrewcomputers.org/doku ... ome:oc65r0

I don’t have an Altoid, but it may fit minus the CP2102 USB-serial adapter.
Bill

that is both amazing and very cursed.

i wonder if you could take it to another level, instead of an IDE connector sticking out the side, have a CF Connector on the back with the card poking out over the PCB edge juuust enough that you can remove it without tools.

My default standard is to install a six-pin connector (though only four/five are needed) which fits one of those handy FTDI breakout cables; provides power at 5v (note that the 3v3 version still provides 5v on the Vcc line!) plus convenient USB terminal comms with no issues. I've used the FT232 (or equivalent) on an old circuit in which I had to bit-bang the serial at 2400 baud.



I think this would be possible, particularly if you kept the sampling rate to something like 16kHz. Additive synthesis from an existing waveform sample - normally a sine wave but it doesn't have to be, of course - is just a question of picking the next sample from the waveform, and that's just a question of adding modulo the desired output frequency. Not much more effort to do multiple samples rather than one, but it depends how complex you get with attack and decay which you'd want to manage without a multiply or divide... convenient that the ear hears logarithmically. Less convenient that the ear is incredibly good at hearing sampling noise, so the bigger the waveform the better, as a general rule.

I wonder if you could use the internal timers in a minimum pin-count arm or similar to generate a handful of convenient tones? Top of my head there are STM parts with four or more timers with four outputs each that can be assigned to pins...

Neil

_________________
x86?  We ain't got no x86.  We don't NEED no stinking x86!

but again, it's a subjective thing. i think USB is hobbyist friendly enough to be used in projects, as example i use an FT240X for serial and a CH376 for mass storage. both chips give the benefits of USB (ie compatibility with modern devices/systems) but without the drawbacks of the raw interface.


i really doubt all of those pages are actually filled with useful information. also if simple data transfer is the whole goal, why was USB invented in the first place and toke over most other interfaces? it must've done something right... i mean even ethernet is using a packet system and it's just as widespread.
hmm, maybe it was just the raw speed it had compared to everything else at the time?

also, obviously off topic, but how fast could you make a UART run? aassuming you're using an FPGA for the controller+buffers, and diffirential pairs for the RX and TX lines... how much bandwidth could you get out of that?


did you know that you're one of the only people on this forum who occasionally throws a word at me that i have to google because i have no idea what it means?
so today i learned what "drivel" is.
ah, the joys of a non-native english speaker

I'd say this rather misrepresents both the problem USB is solving and the solution they came up with.

The problem is not as simple as it looks. USB is not just about "get data at relatively high speed (for the time) from point A to point B over a serial connection." That had already been solved much earlier via RS-422. But the difference from USB is clear if you try to go out and buy an RS-422 keyboard and RS-422 disk drive, plug them into a computer, and have them just start working. RS-422 doesn't handle power delivery, speed negotiation or automatic device description, nor does it provide a vast set of protocols about that for a wide variety of devices and data types.

If you wanted to do a simple implementation of just the differential serial signalling that RS-422 provides and make it work on any USB system you need to implement only a few pages of the USB 1.1 spec. For simplicity in your implementation you'll probably want to support only low-speed (1.5 Mbps) mode. Actually making it do anything after that you'll need to implement some of the basic discovery protocol (one or two dozen pages) and one of the USB device specifications (one or two dozen more pages). A few years ago I did a rough working-out of how one might implement a USB keyboard using a 6502 and a few other parts, and it wasn't too bad. Not quite as simple as the Atari 800's SIO interface, but not all that much worse.

That said, as a hobbyist you probably wouldn't want to do that unless you had a particular interest in USB, any more than you'd want to build your own CPU from gates. Just as you can simply go out and buy a microprocessor, you can go out and buy a chip like the FT245 that provides a very simple interface and does most of the work for you. And using that chip is significantly simpler than using a UART and setting up an RS-232-style serial connection, even if you run that at TTL levels. In fact, it's hard to think of a communications interface that's simpler, on either end (USB host or hobbyist-built board), than FT245.

_________________
Curt J. Sampson - github.com/0cjs