Discovering the mysteries of the datasette

[eslapion]

A possible solution...

Groepaz said:

the filtering happens at a frequency that is way beyond what matters for a tape loader

The figures I noted above proved that to be completely false. The datasette is really designed for a specific bit rate of operation.

If you want a faster datasette transfer speed, the 1st problem of the standard Commodore protocol is that it just saves/load the same information twice then compares the 2. It is much more efficient to have saves working a bit like information on a floppy/CD, sectors of a specific size followed by a checksum of some sort.

Then you have then frequency of the filter which you can probably double or triple and increase the bit rate accordingly.

Then finally, you can easily double the tape speed with the alteration of the pulley size or the adjustment of the motor speed but that's really the last step.

If you're a crackpot for performance, you could probably find yourself an open-reel tape system and remove it's analog electronics and replace it with a modded datasette electronics board on which the filter's frequency response was increased.

[Forbidden64]

If you're a crackpot for performance, you could probably find yourself an open-reel tape system and remove it's analog electronics and replace it with a modded datasette electronics board on which the filter's frequency response was increased.

*Lights crack pipe.*

That's exactly what I want to do! I want to get an old Akai 8mm reel to reel and turn it into a datasette. Anyway, like I said before I don't pretend to be an expert at all. I am a total novice. I remember reading your stuff about the PLA replacement Eprom/why that specific chip would work where other's had issues and thinking geez it will be years before I can be that good(I also bought a whole bunch of them in case I have a PLA issue later).

There is one obvious clue though...which is that he elected to go with that particular speed, and not a faster one. From what I have read it seems to be quite reliable though. If the author is anything to go by. Rabbit stores 300k on 30 mins of tape, or 10k/min or ~166 bytes per second. Not bad for an old datasette!

I am not exactly impressed by the figures you're providing here. I think a normal Commodore save/load is around 120 bytes per second. I think Turbotape (Compute's Gazette) provides about 300 Bytes per second.

If we go by the author's estimation, it is 4x faster(same article cited). That was the figure he gave for data/time "300KB on a C30 tape" within the article as well which seems. I have not tested it 'shrugs'. I use 'fast tape' from 'The Cassette Book for the Commodore 64' by Abacus software. Which is easily 10x faster in practice and I stand by that figure. I was reading it and thought...meh why not...and type it all in. The commodore rom loader writes at 150 baud from what I understand, and since it writes the data twice, that is effectively 75 baud. The speed of rabbit, again according to the author, was about 4x faster than the original. Not the fastest out there by far, but for 1983 it's pretty good especially since it was a cartridge that you didn't have to load separately and make SYS calls, like most turbo tapes including mine. I managed to lower the load time down considerably by saving the memory segment that it was stored in, instead of using a basic loader. Then waiting a few seconds before holding down record in order to shorten the leader tone.

Now here is marko makela to set the record straight on rabbit with some hard numbers!

The tape format of Vic Rabbit
reverse engineered and documented by Marko Makela, September 22, 2002

The Vic Rabbit format appears to use three different pulse widths.
The pulse widths vary surprisingly much, and there doesn't seem to be
too much error tolerance. The difference of the widest short pulse
and the narrowest medium pulse I observed is only 24 microseconds.
The difference between the widest medium pulse and the narrowest long
pulse is 32 microseconds. I wonder if this format tolerates any tape
speed variations.

The initial sync is CDBBBDBDDBBBBBBBBBBBBBBBBB repeated 1200
times. (B=short pulse, C=medium, D=long). CD appears to mark the
start of a byte. The tape header is 26 times CDxxxxxxxx where each x
is either B=0 or D=1. This header is followed by a C (medium) pulse
and 80 times CDBBBDBDDBBBBBBBBBBBBBBBBB again, followed by one copy of
the program, again followed by a C pulse. Unlike in the normal
format, the data bytes are recorded most significant bit first.

Each block seems to start with a $0f byte. The last two bytes of a
block are a 16-bit sum of the data bytes, excluding the $0f byte and
the checksum bytes themselves.

In the tape header, the file name (16 bytes) comes first. Then are
two or three unknown bytes followed by the start and end address of
the program block. The program block contains just the raw program
data between the $0f byte and the checksum.

This would tend to agree more with the article in that it would do a long pulse then short pulse for instance..and interpret it by the space between. In so doing it uses more than one pulse per bit. At least that's what I get out of there along with all the other treasure.

[Forbidden64]

On a side note, and in defense of what eslapion is doing to those nay sayers and contrarians, it's not enough to say that someone else did something that 'solves' the problem a long time ago. There is always a better mouse trap. I see that quite frequently on this forum and it is really frustrating to read; muddy's up the threads with useless filler in between relevant posts with pointless arguments. The voyage of discovery is 75% of the fun of using/modifying/interfacing/programming/designing things for vintage computers. Some of us are just starting our journey, like myself. Others are quite advanced in their knowledge like eslapion. Working within limitations is a game unto itself. To say that someone made something a while ago so don't bother is not constructive and can even discourage people from trying new things that turn into ideas/products we love and use within the VC community as a whole. Even if they don't end that way, everyone is better for them having tried it. We all learn something. With eslapion's track record of making really neat stuff and finding solutions in a variety of systems, I am very excited that he is taking an interest in the datasette. I _love_ tape drive systems! I will of course help in any way I can although with my inexperience I'm not sure how to help. I am a fast learner though, so if any 'dogs body' stuff I can help with that and probably learn a lot while I am doing it.

Also thanks for clarifying the schematic eslapion. One follow up question I have about the schematic is this. If the 5V with the 3.3 ohm line is there to pull up the signal to give a constant high when there is nothing on the line which makes perfect sense as it is tied in with the serial line, how does it ever go low? I don't understand how that happens without any circuitry to select whether or not that 5v is fed into the line or not. I read about how this is also done at start up with the cpu and rom select etc at address 0 on the c64 , and had the same question. Perhaps you can answer that? I'd really appreciate it.

[eslapion]

Anyway, like I said before I don't pretend to be an expert at all. I am a total novice.

It doesn't matter. I started my engineering degree at the age of 33 so for the first 20 years I owned a VIC-20, I was a novice too. I really took the course that would allow me to understand the most at the age of 35 and it took 3 months.

https://www.etsmtl.ca/Programmes-Etudes ... gle=ELE200

Which is essentially the content of that book:
https://www.amazon.com/Electronic-Princ ... 0071261915

I think everyone can learn that by themselves given enough time.

I remember reading your stuff about the PLA replacement Eprom/why that specific chip would work where other's had issues and thinking geez it will be years before I can be that good(I also bought a whole bunch of them in case I have a PLA issue later).

A simple notion called "slew rate"... I suppose this means you have a small stock os M27C512-90B6

There is one obvious clue though...which is that he elected to go with that particular speed, and not a faster one. From what I have read it seems to be quite reliable though. If the author is anything to go by. Rabbit stores 300k on 30 mins of tape, or 10k/min or ~166 bytes per second. Not bad for an old datasette!

I am not exactly impressed by the figures you're providing here. I think a normal Commodore save/load is around 120 bytes per second. I think Turbotape (Compute's Gazette) provides about 300 Bytes per second.

If we go by the author's estimation, it is 4x faster(same article cited).

In order to get a clearer answer, I would have to do something I haven't done in more than 30 years... save a 16k file to a datasette with Commodore's normal tape protocol! (and clock it...)

If the 5V with the 3.3 KILO ohm line is there to pull up the signal to give a constant high when there is nothing on the line which makes perfect sense as it is tied in with the serial line, how does it ever go low? I don't understand how that happens without any circuitry to select whether or not that 5v is fed into the line or not.

On the IEC side, this line is tied to 7406 circuits (in 1541 and other IEC drives) which are "open collector output". These chips can only pull their output low and they are used either to convert a signal from low to a higher voltage (5V to 12V, usually) or to signal on a line which is shared by multiple devices - the case of the IEC.

The computer (C64, VIC-20 or C128) is not equipped with the means to pull this line low itself since both CASS-RD and !SRQ IN always receive data from an external device, wether it be a datasette or disk drive in these 2 inputs.

If you connect a bunch of chips which can only pull low on a piece of wire, "something" has to be responsible for pulling high - that's the resistor's job.

[Forbidden64]

Took me a second, but I totally get it I think lol! So the 7406 inverters, being open collectors pretty much suck most the current away so that the flag pin on the 6526 is then low, otherwise is it high all the time. I think Ill do a mock circuit like this and noodle around with it later today.

Yes I bought a small stock of M27C512-90B6's on aliexpress.com. It's too bad you can't use any 27512, as I have 100 of some other kind of that chip which I bought for multi-carts(A lifetime supply lol since they were 1$ each). They are all in storage at the moment along with most of my stuff :/
Want to hear something funny? I am 33 now! lol. Maybe that is my queue to learn EE instead of lolligagging around.

[eslapion]

So the 7406 inverters, being open collectors pretty much suck most the current away so that the flag pin on the 6526 is then low, otherwise is it high all the time.

Correct!

Yes I bought a small stock of M27C512-90B6's on aliexpress.com.

If you didn't buy them from Digikey or Mouser then there is a great deal of chance they are fake.

These fake chips destroyed the reputation of the M27C512-90B6 as a good PLA sub. In reality, they are re-branded Winbond ordinary 27C512 and the pin 1 notch is clearly different.

Fake:

The package and the 2 notches are identical on Winbond chips

Genuine:

The contrast of the ink is much lower on real chips
Only ST PROMs have notches like that - mine were ordered from Mouser which is an official authorized distributor for ST.

Want to hear something funny? I am 33 now! lol. Maybe that is my queue to learn EE instead of lolligagging around.

Excellent idea!

[eslapion]

A very large datasette with 20MBytes capacity...

https://www.youtube.com/watch?v=7Lh4CMz_Z6M

That's gotta be the most crackpot datasette I have ever seen.

[Forbidden64]

Really fascinating design...typical of IBM then and now. Spare no expense, not even for pity sake. Did you see those solid gold connectors? what on earth were all those wires doing?! It looked like a AMD processor glued to a rope. I like their solution to apparently not having stepper motors. That mechanical clutch style was widely used in cars at the time. My first car was a 1973 merc, and it had an entirely mechanical clutch fan! Now they just have cheap plastic electric fans. That thing was made out of machined aluminum, much like that IBM fan assembly in that video. Very elaborate and very cool. I want to hook that up to my commodore lol.

DC motors, like all others before and after. Commodore's datasette is definitely the result of a tape deck fornicating with a box window fan. Being the only tape drive system I know of that has a 9V AC motor. How awesome was that punched card machine too!? If you were offered free accommodations for a year, or that system, which would you choose?? I'd go with sleeping under a bridge with my IBM vacuum column without thinking. Not like anyone can steal it! It's too darn heavy!

I heard that the US army is actually phasing out their IBM system 1 machines that are running our defense networks :<. Now they are going to replace something that ran reliably for 60 years with something that will barely crawl past 6, and is made by a country that willfully makes fake chips/failure prone ones. Unless of course its another IBM system heh. They haven't let their standards slack at all.

Now being a guy who has a DEC PDP 11/03 clone, in the form of a Heathkit H-11, I can attest to the strength of those 8" floppy drives/construction of commercial grade drives. The dual drive weighs 90 lb(45kg) and is solid machined metal ($$$2600 clams in 1978 money _just_ for the dual drive!). You couldn't wipe the data on those disks in the drive if you had a defribulator handy. I really want to make a large scale reel to reel tape drive with RS232 I/O...that way I can use it with an RS232 adapter on the user port of the commodore and use the built in RS232 routines I read about for passing bytes out, or in etc in the kernal, or with TU58 format on the H-11 which has native rs232 for, well pretty much everything. Even for the terminal. I'm not sure if it is considered heresy to mention Heathkits here though lol...I think that is more VCfed territory, but hey since we are talking vintage mainframe tape drives . But its why I mention RS232 being a goal if I eventually made one. Then set the baud rate within the standard.

Another tape system in the mainframe world to look at is the DC300, and subsequent DC100 drive HP scaled down for the HP85. which was way more complicated, but much faster. I don't even think we've touched on the major elephant in the room though. Which is semi-modern tape systems such as LTO 1,2,3,4,5etc.., which are quite inexpensive now. There is also the AIT drive Sony tape drives etc. One drive I saw and I wished I had memorized the make and model, was a tape backup that used regular cassettes but held a whopping 600 mbs, or at least that was the claim! LTO-1 can easily hold 100gb, and AIT I believe holds 20 megs uncompressed. I have one of those backup drives on my 386 portable. I did open it up to see where the cpu battery was/why it wasn't working, but beyond getting it working, I didn't check the tape drive model. Might be an avenue to explore if one isn't looking for vintage eye candy like a RTR tape machine.

The other cool eye candy cassette tech is the "Open cassette" which is crazy cool looking. has metal reels and looks like a miniature RTR spool with metal guards on the sides guarding the actual tape. you spool it into the tape, and then install it into the regular deck. Another idea I had was to make a micro retro RTR deck using this type of setup, As a spool of cassette tape is quite small, as would be the mini-RTR home-brew deck. Which I would paint light blue and make it to look like a good ol' mainframe floor model tape system of course...

[eslapion]

Back on the VIC-20 storage topic, IMHO, the best magnetic Commodore equipment based storage for the VIC-20 is the 1581 disk drive with JiffyDOS.

BTW, the Commodore datasette motor runs at 6V, not 9V - there is a zener diode to control the output voltage from the power transistor in the VIC-20/C64/C128.

Sidenote: Did you check if your ST PROMs are fake ?

[Forbidden64]

Back on the VIC-20 storage topic, IMHO, the best magnetic Commodore equipment based storage for the VIC-20 is the 1581 disk drive with JiffyDOS.

Ya, but it's not clunky and giant, and there aren't any arrays of wildly blinking lights! I Have to agree though. It's the best plug in solution to magnetic storage for the vic, and disks are still very very easy to find.

BTW, the Commodore datasette motor runs at 6V, not 9V - there is a zener diode to control the output voltage from the power transistor in the VIC-20/C64/C128.

Good to know! funny enough I have never actually looked at the datasette's schematic. I have been more focused on logic circuits, and less on analog stuff. Now, I am thinking I have that kind of backward. Although I am glad I learned all that, I wouldn't even be able to build a linear power supply to power it haha! I do have a few books on that stuff, as well as a kit (one of those 500 in 1 circuit lab products). So that should carry me through the general theory etc as that book you had linked and its small enough to fit in my cramped quarters. I am more and more of a mind to formally study this as a professional, I just need to get things stabilized so I am not possibly dumpster diving in a few weeks, and then focus on EE. One of my former co-workers moved into a trailer in the desert in Arizona to afford schooling to become an EE. Now that's commitment!

Sidenote: Did you check if your ST PROMs are fake ?

I plan to check next time I visit my storage unit, which will likely be soon. I know exactly where they are. I do recall them having a yellowish dull label, but I could be mistaken, and I'd like to be sure. The notches you pointed out will be a dead giveaway if they are fakes. At the moment I am more concerned with making sure I keep the unit lol! *sigh*

[eslapion]

One of my former co-workers moved into a trailer in the desert in Arizona to afford schooling to become an EE. Now that's commitment!

Well, I didn't have to go that far but I had to make serious sacrifices to get there.

It was worth it.

[joshuadenmark]

A very large datasette with 20MBytes capacity...

https://www.youtube.com/watch?v=7Lh4CMz_Z6M

That's gotta be the most crackpot datasette I have ever seen.

I want one of these

[eslapion]

I want one of these

Oh My! I sure don't ! There is no room in my small apartment for a tape drive the size of a refrigerator that can only store 20 MBytes...

However, I would love to have color adhesives I can put on my fridge that makes it look like it's a vacuum column tape drive. I can only imagine the looks on the face of people coming to my place for the 1st time.

Now that would be real cool!

[joshuadenmark]

I live in a house by myself, got lots of space for a couple of these beauties.
Nice idea with the fridge sticker though

[Forbidden64]

However, I would love to have color adhesives I can put on my fridge that makes it look like it's a vacuum column tape drive. I can only imagine the looks on the face of people coming to my place for the 1st time.

Now that would be real cool!

That is such a fun idea! Paint it sky blue and then go to a business vehicle decal place and get some magnetic signs to stick on there.

Code: Select all

\\\\\
 -D-D
   _\
 \___