BASIC audio files for testing

[Sideburn]

It would be a two-step process.

Certainly, you should download and install the WinOSI emulator from Mark's site if you haven't already. It doesn't directly read audio files, but it does let you LOAD text files as simulated "tapes". And you can use the OSIKCS program, from the same site, to convert valid Kansas City Standard audio files into text files you can use for that purpose.

Thanks, yes I have been using OSIKCS to generate wav files to test reading with and comparing my saved audio against. My frequencies when saving seem to be off and LOADS are doing nothing at all.

I have reverted the mods that were done to U62 and U63 and now I am getting a higher pitched carrier signal but it still doesn't seem right. It seems too high frequency.

I need to focus on either getting SAVE working or getting LOAD working.. Tackling them both makes me run around in circles.

Do you know where on the board the carrier signal is generated. I know it comes form the XTAL but having trouble finding where it is getting divided down to the desired frequency.

my most recent SAVE audio attached.

Can you have a listen and let me know what you think?

[Mark]

OSIKCS (based on Ed's CUTS program) generates ~2% faster ~2450/1225 Hz frequency 8bit mono 22050 samples/sec WAV files since the generated wave forms align & and can be compressed greatly. (100x+)
The OSI doesn't care when decoding but you can hear the slightly different frequencies. The native OSI will sound slightly lower in frequency.

The mod done to your OSI as you received it seemed to have increased the clock to the ACIA by 2x (600 baud), but the tones generated were correct (2400/1200hz)

Your latest sample seems to be using 4800/2400hz tones, and appears to be 600 baud maybe? Move everything down one on the clock divider chain!

For more detail you can see the waveform of your sample in Audacity. Count the cycles: 2400hz should have 8 cycles per bit, 1200hz should have 4. Use the spectrum analyzer feature to determine frequencies involved.

[Sideburn]

OSIKCS (based on Ed's CUTS program) generates ~2% faster ~2450/1225 Hz frequency 8bit mono 22050 samples/sec WAV files since the generated wave forms align & and can be compressed greatly. (100x+)
The OSI doesn't care when decoding but you can hear the slightly different frequencies. The native OSI will sound slightly lower in frequency.

The mod done to your OSI as you received it seemed to have increased the clock to the ACIA by 2x (600 baud), but the tones generated were correct (2400/1200hz)

Your latest sample seems to be using 4800/2400hz tones, and appears to be 600 baud maybe? Move everything down one on the clock divider chain!

For more detail you can see the waveform of your sample in Audacity. Count the cycles: 2400hz should have 8 cycles per bit, 1200hz should have 4. Use the spectrum analyzer feature to determine frequencies involved.

Thanks, I will see if I can figure this out in audacity.

I am having trouble figuring out why it is not back to 300 baud. When you say to move everything down on the clock divider chain, could you point me where to look on the board and schematics? Or how to go about doing this? I’m just trying to get the board to match the original schematics.

I pulled the 6850 out and tested it with my chip tester and it is good but that high pitch carrier signal is being derived by a different circuit. Is my carrier signal the correct frequency?

[Sideburn]

To clarify a little more I’ve reverted the board to the stock configuration where pin 3 & 4 of the 6850 are tied together and go to pin 9 of u63. So maybe I have a bad chip, I am trying to figure out how the clock signal dividing is working so I can isolate the issue.

[Sideburn]

And a couple more head scratchers that maybe you can shed some light on?

• There's some more mod wires over on the 6850 (u14), the 7484 (u18) and the 74LS20 (u9) Pin 5 of the 6850 (bRTS) has a jumper going to pin 9 of the 7484 (u18) any idea what all this is for ? should I remove it as well?

IMG_7775.jpeg (2.6 MiB) Viewed 210 times

• My board does not have U67 or U68, is this to be expected?

IMG_7776.jpeg (2.88 MiB) Viewed 210 times

[Sideburn]

On the audio frequency / audacity analysis front, I took a sample of output generated from osikcs to use as a base (should be 300 baud output) and opened it in Audacity and selected Plot Spectrum. It looks like the frequency is peeking at around 1200hz (where my cursor is placed) from what I can tell which should be as expected. Am I doing this correctly?






I am seeing what you are describing (4800/2400) on my OSI output. Now I need to fixure out how to divide the clock in half.

[Sideburn]

Sorry for the trail of posts as I sort this out but I am on the right track now. I disconnected the stock trace that had been previously cut that is going form U57 Pin 2 to U69 pin 14 and changed it to pin 13 and now I believe I am getting the correct frequencies.

1200hz:

1200hz.jpg (336.71 KiB) Viewed 208 times

2400 hz

2400hz.jpg (358.2 KiB) Viewed 208 times

wav file attached. Does this sound right? If so now I just need to fix the RX side! I wonder if those mod wires on dRTS of the 6850 have something to so with it not working on loading...

[Mark]

The frequencies look good.

You can sample the normal idle frequency for a 2400Hz test.

Use this poke to set the ACIA in BREAK mode to hear the 1200Hz tone

Code: Select all

REM SET SERIAL BREAK CONDITION

REM FOR C4P
POKE 64512,241
REM RESTORE WITH POKE 64512,177

REM FOR C1P
POKE 61440,241
REM RESTORE WITH POKE 61440,17

hi-lo-tones.zip

Two frequency test

(729.11 KiB) Downloaded 12 times

Here is a sample OSI BASIC generated audio file for testing/comparison.
If the tone decoder is working it should show up on your screen after a LOAD command

hello_world.zip

Sample KCS WAV and program

(805.89 KiB) Downloaded 10 times

My KCS interface looks very different from yours! When I moved to a C4PMF I needed a way to load my old OSI tapes and quickly bread-boarded the KCS interface along with a baud rate generator & RS-232 interface. Since some of my OSI tapes from an OSI 500 board C2 had the baud rate adjusted too high, I also had to add a variable speed setting. One jumper allows me to send the tone decoded input back through RS-232 so I can decode the old tapes to any serial device!

Custom OSI KCS.jpg (141.21 KiB) Viewed 207 times

[Sideburn]

The frequencies look good.

You can sample the normal idle frequency for a 2400Hz test.

Use this poke to set the ACIA in BREAK mode to hear the 1200Hz tone

Code: Select all

REM SET SERIAL BREAK CONDITION

REM FOR C4P
POKE 64512,241
REM RESTORE WITH POKE 64512,177

REM FOR C1P
POKE 61440,241
REM RESTORE WITH POKE 61440,17

hi-lo-tones.zip



Here is a sample OSI BASIC generated audio file for testing/comparison.
If the tone decoder is working it should show up on your screen after a LOAD command
hello_world.zip


My KCS interface looks very different from yours! When I moved to a C4PMF I needed a way to load my old OSI tapes and quickly bread-boarded the KCS interface along with a baud rate generator & RS-232 interface. Since some of my OSI tapes from an OSI 500 board C2 had the baud rate adjusted too high, I also had to add a variable speed setting. One jumper allows me to send the tone decoded input back through RS-232 so I can decode the old tapes to any serial device!
Custom OSI KCS.jpg

Wow impressive. You know your serial stuff, I sure don't! Thanks for the code and sample files. Yes I just compared my output against a file generated by osikcs and they are close enough.

Now onto getting LOAD to work. I am getting absolutely no activity from a load. should I be getting garbled characters or something if i just produce noise on the input jack?

What is the POT for that is closest to the edge connector? I know the other one is controlling the video out brightness but I haven't figured out what the other one is for yet.

And what do you think these wires are that are jumping over the pin 5 of the 6850?

[Sideburn]

Using your POKE's I wrote a BASIC prog to toggle between the two.
Output attached.

If all looks good then on to sorting out what could be wrong with loading...