1088XLD - Custom A8 Computer in 1050 Drive Case

[+mytek]

MIDI DIN Interface Test

 

https://www.youtube.com/watch?v=KGPNeWckvfw

 

 

 

Revised Schematic (removed polarity inversion following the opto-isolator).

 

 

EDIT: I figured out what caused my mistake. The H11L1 component I was using in my original schematic showed it having an inverted output, which it doesn't. The above schematic shows the corrected symbol.

[ivop]

Quick update: the CS9236 board is working! At first, I didn't get any sound, so I started measuring everything, like clocks and power supply while a MIDI tune was playing. At times, I restarted the MIDI player. Just when I was about to give up, I restarted the player for, like, the fourth time, and I heard a crack on my speakers, turned up the volume and there it was! After some more tests, I have concluded this is because it takes too long to start (longer than the S2) when connected to the motor line. This won't happen in the 1088XLD when it is always powered on.

 

Another thing I noticed is that the output is considerably softer. I'll see if it gets louder if I remove the AC Coupling, which might not be needed after all. ?

 

 

 

[+mytek]

Quick update: the CS9236 board is working! At first, I didn't get any sound, so I started measuring everything, like clocks and power supply while a MIDI tune was playing. At times, I restarted the MIDI player. Just when I was about to give up, I restarted the player for, like, the fourth time, and I heard a crack on my speakers, turned up the volume and there it was! After some more tests, I have concluded this is because it takes too long to start (longer than the S2) when connected to the motor line. This won't happen in the 1088XLD when it is always powered on.

 

Another thing I noticed is that the output is considerably softer. I'll see if it gets louder if I remove the AC Coupling, which might not be needed after all. ?

 

cswbboard.jpg

 

 

 

Congratulations

 

The AC coupling shouldn't affect the volume unless the capacitor values are too low. I would retain it, since my audio mixer circuit works best if everything is capacitively coupled. Maybe use an LM358 to boost the volume?

 

--------------

 

I had a thought last night regarding the SYNTH Source Select Switch on my DIN interface board in regards to using MIDI record/playback programs. After watching a couple of videos of people using basic MIDI trackers, it became apparent to me that it would be nice to have both the MIDI-IN and the MIDI-OUT automatically route to the on-board synthesizer without conflict. This would be especially good if you are using a keyboard with no built-in synthesizer. So for instance you could record a melody and then play it back, all without having to flip a switch back and forth.

 

So with a few added diodes and a resistor, here's what I had in mind...

 

 

For those cases where you wished to only use an external synth, the physical switch would let you turn 'OFF' the built-in synth.

 

I'll give this a try today and see what happens. Any suggestions on the best record/playback app to test this with?

[ivop]

OK, I'll leave the AC coupling in place (removing it did indeed not affect the volume). I read up on the LM358 and it seems most "professionals" think it's a poor choice for audio and it has cross-over issues. They advise the Tl062. No idea if this is overkill or not. Also, it might be that the chip misses the early volume events because it is not ready yet to receive them.

 

About routing IN to OUT, this is mostly done in software and is called software MIDI-THRU. I think even Midi Track III supports it, although I still haven't really dug into that program as its interface is terrible. But that should be the record/replay app to test with. Or a simple BASIC program with the MIDI Mate handler.

[ivop]

Edit time ran out re output levels, I cannot find the webpage again, but I remember reading that the S2 is particularly "hot" and period-cards had a lower output level. BTW I'm also looking to acquire one of the old-school WB boards (http://members.home.nl/c.kersten/) or one of the chips used, but the prices are insane:

 

https://www.ebay.com/itm/Yamaha-DB50XG-extended-General-Midi-XG-4MB-ROM-wavetable-sound-daughterboard/183442902617?epid=7017752222&hash=item2ab60c8e59:g:gh8AAOSw~WtbomyA:rk:5:pf:0

https://www.ebay.com/itm/Waveblaster-Module-MIDI-Interface-Board-Sound-Card-Wavetable-DB50XG-NEC-XR385/282619543082?hash=item41cd6ffe2a:m:mscZ-Tt5Wat-bIAhLKBslDw:rk:2:pf:0

 

[+mytek]

About routing IN to OUT, this is mostly done in software and is called software MIDI-THRU. I think even Midi Track III supports it, although I still haven't really dug into that program as its interface is terrible. But that should be the record/replay app to test with. Or a simple BASIC program with the MIDI Mate handler.

 

I just tested the proposed 'MIDI-OR' circuit and it works great! So now MIDI IN and OUT events are automatically fed to the S2 board when the switch is engaged. Didn't need to run an app to test this, since the power-up intro is a MIDI-OUT event, and of course the Rock Band 3 keyboard is connected to MIDI-IN.

 

The software solution you mentioned wouldn't work with the original circuit, since there would be no way to send anything from the Atari side to the S2 if it's source select was set to a MIDI-IN connection.

 

 

Edit time ran out re output levels, I cannot find the webpage again, but I remember reading that the S2 is particularly "hot" and period-cards had a lower output level. BTW I'm also looking to acquire one of the old-school WB boards (http://members.home.nl/c.kersten/) or one of the chips used, but the prices are insane:

 

https://www.ebay.com/itm/Yamaha-DB50XG-extended-General-Midi-XG-4MB-ROM-wavetable-sound-daughterboard/183442902617?epid=7017752222&hash=item2ab60c8e59:g:gh8AAOSw~WtbomyA:rk:5:pf:0

https://www.ebay.com/itm/Waveblaster-Module-MIDI-Interface-Board-Sound-Card-Wavetable-DB50XG-NEC-XR385/282619543082?hash=item41cd6ffe2a:m:mscZ-Tt5Wat-bIAhLKBslDw:rk:2:pf:0

 

 

Although my set-up sounds very good, I doubt it would pass a true audiophile's idea of what things should sound like. If you really crank up the volume you can hear a little bit of background noise. This is present with or without the synth, so it's just a 1088XEL thing. I also suspect the same would occur on any Atari. So for the absolute highest audio quality, you would want to go with an external synth, or directly route the audio outputs from an internal synth to an external amplifier and speakers. So yes since your board would be universal and could be used on something other than an Atari, you probably would want to keep the sound as clean as possible. But if your main target is intended to have it's audio fed through an Atari (i.e, SIO Audio-In), then the LM358 is probably more than adequate.

 

BTW, those Wave Blaster boards you linked to are enormous .

 

----------

 

Question on MidiMon program. I see a flashing cursor next to Active Sense and nothing after System Clock (see image below). Is this normal?

 

 

[+mytek]

ivop please reference my last post which I think you already read before I added a question about your MidiMon program.

 

BTW, very nice and useful test program .

[ivop]

[snip]I have concluded this is because it takes too long to start (longer than the S2) when connected to the motor line. This won't happen in the 1088XLD when it is always powered on.

 

I have confirmed this by powering the MIDI Muse and WB board separately from the joystick port. I don't boot over SIO anyway, so this was no problem. After that, it worked marvelously!

 

And it has definitely a different character than the S2 board! I'll look into uploading movies to youtube and try to make a small video in the coming days.

 

 

See the challenge to shrink the new CS9236 board to S2 size

 

[ivop]

BTW, those Wave Blaster boards you linked to are enormous .

LOL, yeah, if you compare the size of the original ISA Sound Blaster 16 cards compared to current PCI-e sound cards, it's baffling.

 

Question on MidiMon program. I see a flashing cursor next to Active Sense and nothing after System Clock (see image below). Is this normal?

Yes, that is normal. Any MIDI device that hasn't had any activity on IN or OUT is required to send an Active Sense event at least every 300ms. That's what the blinking cursor is for. After connecting a device, it should start blinking, even before you press any key.

 

The system clock events are used to properly time and sequence a song. For example, if I start the demo song on my 80's Yamaha keyboard, I see System Clock events.

Edited October 14, 2018 by ivop

[+mytek]

Cool .

 

So far all looks very solid, I'm running an 8 meter cable from the Rock Band 3 to the MIDI-IN and don't see any missed events or glitching. Also discovered that through the use of the 1,2, A, and B buttons I can shift up and down the scale, and up and down through the voices, which show up as patch changes in MidiMon.

 

 

 

[ivop]

Cool .

 

So far all looks very solid, I'm running an 8 meter cable from the Rock Band 3 to the MIDI-IN and don't see any missed events or glitching. Also discovered that through the use of the 1,2, A, and B buttons I can shift up and down the scale, and up and down through the voices, which show up as patch changes in MidiMon.

 

RBK_Buttons.jpg

 

P1220082.JPG

 

Oh, patch 17 isn't Capitalized! Fixed at github

 

Funny, this is the second time today I noticed something wrong on an image posted here. The first time (see post #277) I noticed I forgot to turn of my multi-meter

[+mytek]

 

Oh, patch 17 isn't Capitalized! Fixed at github

 

Funny, this is the second time today I noticed something wrong on an image posted here. The first time (see post #277) I noticed I forgot to turn of my multi-meter

 

That was just a fraction of the available voices. Did you check all of them?

 

------------

 

New version of MIDI DIN Interface PCB layout completed.

 

Almost ran out of room, so I guess I'll call this done .

 

I also added some extra holes to the DIN connector footprints to make it more universal as to the choice of manufacturers that can be used. Jameco has 5-pin DINs with larger spaced front mounting pins for only $1.05 each. So with this latest change the Jameco part may now be used as well.

 

Manufacturer: GOLDSUN ELECTRONICS

Part Number: B-1001-R

Jameco: 29399

1,600 to be in stock end of the month

 

 

This is the one I originally specified that will also work.

 

Manufacturer: Switchcraft

Part Number: 57PC5F

Newark: 44F7434

Mouser: 502-57PC5F

 

I have some of the Jameco ones, and they look quite good and sturdy. So for less than half the price of the other ones, it's quite a cost savings.

[+MacRorie]

 

Um, no. If it ends up as an essentially-reconfigured 1088XEL, there will be PIC microcontroller chips for the keyboard, mouse and VGATE functions, there will be an FTDI serial-to-USB interface board, an Ultimate 1MB board and likely some variant of the logic setup for an XEL-CF PBI hard disk interface.

 

Plus the possibility of assembly errors, the risk of a bad solder joint or - yes - a bad Atari chip. They *do* fail occasionally, believe it or not.

 

 

They fail more often than most people think. I have had builds go south 3-4 times because of a bad or marginal Atari LSI. All were straight from a vendor and not pulls.

[_The Doctor__]

most of the folks who end up with bad chips are of the I don't believe in ESD or static affecting chips group. A vendor doesn't necessarily follow those procedure either, I've got chips sent to me without ESD bag or pads, and some of them have been pulls from whatever they had laying around... of course chips do go bad.. but it's been a very very very rare thing. What lightning storm or power failure flicker? It's gonna be fine... zzzap.. Watcha mean assembling stuff while not grounding myself and wearing house slippers in low humidity while wearing natural and synthetic fibers is a bad idea.. weird I didn't see the spark jump and kill it... but the wife says there was a little flash... hmmm it's dead JIM... must have been a defect.. it was old... You mean doing this stuff on a carpet of static and debri is bad? no way... how comes this stupid thing does random letters on the screen sometimes... lol

nothing like working on stuff when you got awesome static cling going on...

your hair charged to the point you look like a mad scientist just means your ready to operate on them there circuit boards..

Edited October 15, 2018 by _The Doctor__

[+mytek]

Thus far I've only come across one bad chip, and that was a Sally I got from a beta tester's 1088XEL that was being flakey. I even put a Sally in backwards by accident, and left the board powered for several minutes before i discovered my mistake, and it survived (still in use on my 1088XEL to this day). The Atari chips are pretty tough, but just like anything there will be failures.

[Level42]

I agree, compared to some other brands......you know what.....the A8 VLSIs are very robust.

 

I only once received a bad POKEY which sings out of tune. Came from a former Atari Benelux Employee. Kept it for fun....

[+mytek]

So it looks like we now have a complete working MIDI solution for the 1088XLD project, which has also benefited XEL owners as well. Implementation of the built-in S2 Synth board will be easy and similar to how the U1MB installs, resulting in a simple plug in installation. Same goes for the optional DIN connection board, however this aspect will require some modification to the rear section of the upper 1050 case to allow access to the DIN connections. I will have to play around with this idea to see how best to pull it off, but it's looking like a combination of vent removal and a bezel plate might be the way to go. Of course if all you want is the internal Synth playback capability, no case mods will be necessary.

 

Anyway I'll get more into that aspect down the road. Be interesting to see what comes of it.

 

For now it's time to get back to the XLD motherboard PCB trace routing .

[ivop]

And it has definitely a different character than the S2 board! I'll look into uploading a movie to youtube and try to make a small video in the coming days.

 

 

 

Songs (partially) played:

 

1. Grabbag (Duke Nukem 3D)

2. Lufia Fortress Of Doom

3. Chrono Trigger Battle

4. Chrono Trigger Theme

5. Dr. Who Theme

6. Doom E1M1

7. Relax (FGTH)

 

Re 7 : We need more cowbell!

[+mytek]

Sounds good .

 

Looking forward to doing a comparison myself. So any idea of the overall parts cost?

 

Big oscillator, any reason not to go with 1/2 size can for size reduction later on?

 

Nice job Ivo .

[ivop]

Thanks!

Looking forward to doing a comparison myself. So any idea of the overall parts cost?

The most expensive parts is the CS9236 ($12-$15). After that comes the CS4333 DAC (~$3), 74HC04, oscillator, LM1117-3.3, PLCC28 socket, DIP14 socket, 12 capacitors, 8 resistors. 2x13 pin header.

Overall, I think about $25,= should get you all the parts. PCB was $2,50 a piece.

Big oscillator, any reason not to go with 1/2 size can for size reduction later on?

 

No reason. I already received a smaller (square) oscillator

[+mytek]

Making Progress...

 

 

Schematic: 1088XLD_Preview3_10-19-2018_schema.pdf

 

Having to move a few things here and there as I go to make for a more efficient layout. At this point I am probably 75% complete, with a few challenging long distant runs to figure out (gets tougher as more stuff starts getting in the way). Once all that has been completed, then I'll get to move on to the Stereo Pokey, IDE, Power, and Video sections. At the end of this process, I'll need to move some of the component ID labeling to avoid being obliterated by all the vias.

 

I'm saving the easiest routing for last .

[+Stephen]

Man would it be awesome if the actual PCB was coloured exactly like that!!!

[Mr Robot]

So it looks like we now have a complete working MIDI solution for the 1088XLD project, which has also benefited XEL owners as well. Implementation of the built-in S2 Synth board will be easy and similar to how the U1MB installs, resulting in a simple plug in installation. Same goes for the optional DIN connection board, however this aspect will require some modification to the rear section of the upper 1050 case to allow access to the DIN connections. I will have to play around with this idea to see how best to pull it off, but it's looking like a combination of vent removal and a bezel plate might be the way to go. Of course if all you want is the internal Synth playback capability, no case mods will be necessary.

 

Anyway I'll get more into that aspect down the road. Be interesting to see what comes of it.

 

For now it's time to get back to the XLD motherboard PCB trace routing .

 

 

Before you move back to the XLD, just one quick question...

 

Will this MIDI solution work for playing MIDI Maze?

[_The Doctor__]

it's midi, it should one way or another...

[+mytek]

Yes it will .

 

However it will require the complete solution, meaning the DIN connection board needs to be plugged in, which becomes a daughter board in the 1088XLD. The same DIN connection board will work in an XEL via a cable connection to the XEL-MIDI board. The user just needs to figure out where to locate it, and drill some holes in the case (easier said then done, and not recommended for some one without the skills to do so).