VOGONS


FMonster, the monster of sound synth

Topic actions

Reply 180 of 253, by eeguru

User metadata
Rank Newbie
Rank
Newbie

I recently took detailed measurements from a number of sounds cards to get the WB footprint correct. It seemed like everything was snapped to an imperial grid. I arrived at the same 5050 x 3100 mil hole differences as Tiido with the same 250 mil offset to pin 1. Those translate to the same mm dimensions as above. It was refreshing to convert my part footprint to millimeters and see the same numbers on his drawing. Thanks for that.

It really doesn't matter where you center the WB card itself as long as you have room for the standoffs to land free of traces and the board edges to clear tall components, etc.

I do plan on adding a 10 pin additional header to my footprint to allow for ID of newer WB cards (I2C) and passing of digital audio (I2S) back from the Ext.WB cards to my current sound card project. Of course any older daughter card would still use the legacy 26-pin header. But I am designing a WB daughterboard that will accept a Raspberry Pi Compute Module 3 (CM3) to provide MUNT emulation true'r to MT-32 and the like. I figured having a digital audio output option in 2019 would be ideal. Maybe we can all adopt it as a standard? I've also added an intermediate mounting hole so shorter cards (like the CM3 Pi) can have more than one support boss. I thought about adding a support for Dreamblaster X2 proprietary hole, but my card already has one-integrated on-board and I'm running short on PCB space.

Attachments

  • wb_dimensions.png
    Filename
    wb_dimensions.png
    File size
    18.31 KiB
    Views
    479 views
    File license
    Fair use/fair dealing exception

Reply 181 of 253, by Tiido

User metadata
Rank Oldbie
Rank
Oldbie

It is nice to know that someone else landed on same measurements ~

As far as expansions go, I have added SPDIF output to the NC pin between the two GNDs. The other NC could in theory be SPDIF input for the cards that do ADC business like DB60XG. I do like the I2S sound idea, but I'm not super fond of the placement of the header for it though, it is right in the area where I'd put all analog related things and that area is quite premium. There's usually plenty of capacitors etc. in that general area and it will probably collide with them on many cards. Something in line with existing connector is less likely to collide with anything on the host card I imagine.

T-04YBSC, a new YMF71x based sound card & Official VOGONS thread about it
Newly made 4MB 60ns 30pin SIMMs ~
mida sa loed ? nagunii aru ei saa 😜

Reply 182 of 253, by eeguru

User metadata
Rank Newbie
Rank
Newbie

Perhaps we should start a separate thread?

I can handle S/P-DIF output from the WB as well. That might be sufficient. Extending the 26-pins by 4-6 would work too I suppose. I don't have my final board fully layed out yet. But tentatively I was thinking I2C SDA/SCL for an optional eeprom on the daughter board that could carry identification info. My sound card project is just a big FPGA+DSP. So I added I2S and a few dynamically assignable pins based on the eeprom ID. For example, the CM3 Pi WB board could run the console to those pins and the FPGA could enable a UART at a COM[1-4] address to connect to it.

My use-cases are a bit one-off. But having some more options from a connector designed 25 years ago would be nice.

Reply 183 of 253, by Tiido

User metadata
Rank Oldbie
Rank
Oldbie

As far as identification goes, you could just use MIDI output functionality and issue a special sysex that the WB card will respond to. Less signals needed, perhaps bypassing need for any extra interface at all if SPDIF is sufficient (I don't actually like the format but many things have it...).

Separate thread is best I guess, and these posts could perhaps be moved to the new thread too...

T-04YBSC, a new YMF71x based sound card & Official VOGONS thread about it
Newly made 4MB 60ns 30pin SIMMs ~
mida sa loed ? nagunii aru ei saa 😜

Reply 184 of 253, by Fagear

User metadata
Rank Member
Rank
Member
Tiido wrote:

I have been using these measurements for my WB cards...

eeguru wrote:

I recently took detailed measurements from a number of sounds cards to get the WB footprint correct. It seemed like everything was snapped to an imperial grid. I arrived at the same 5050 x 3100 mil hole differences as Tiido with the same 250 mil offset to pin 1. Those translate to the same mm dimensions as above.

Thanks! Those measurements almost match mine (within +/- 0.6 mm error). 😎
I thought that DB50XG and SCB-55 were a bit wider according to some photos... but I guess those are not.

eeguru wrote:

I do plan on adding a 10 pin additional header to my footprint to allow for ID of newer WB cards (I2C) and passing of digital audio (I2S) back from the Ext.WB cards to my current sound card project.

As Tiido said, it probably reasonable to utilize unused pins on WB header and perform synth detection via MIDI means. 😕
Also, earlier (?) WB boards had components facing soundcard, so those can conflict with additional header.

Creative WaveBlaster (CT1900).jpg
Filename
Creative WaveBlaster (CT1900).jpg
File size
280.02 KiB
Views
428 views
File comment
Creative WaveBlaster (CT1900)
File license
Fair use/fair dealing exception
Creative WaveBlaster II (CT1910).jpg
Filename
Creative WaveBlaster II (CT1910).jpg
File size
214.82 KiB
Views
428 views
File comment
Creative WaveBlaster II (CT1910)
File license
Fair use/fair dealing exception
Orchid Wave Booster 4FX.jpg
Filename
Orchid Wave Booster 4FX.jpg
File size
239.61 KiB
Views
428 views
File comment
Orchid Wave Booster 4FX
File license
Fair use/fair dealing exception
eeguru wrote:

I figured having a digital audio output option in 2019 would be ideal. Maybe we can all adopt it as a standard?

For me there is no reason to do it (for this project). Almost all blocks on FMonster output analog audio, only OPL2 and OPL3 use digital output, but it is not some S/PDIF or I2S, those require proprietary DACs. So my board has no use for I2S data unless I'll put a DAC on it.

"2nd stage" started to get shape...

monster_option_2019-08-15.png
Filename
monster_option_2019-08-15.png
File size
78.84 KiB
Views
428 views
File comment
FMonster "2nd stage" @2019-08-15
File license
Fair use/fair dealing exception

New BIG soundcard: FMonster.
Covox Sound Master replica
Innovation SSI-2001 replica & DuoSID.
My audio/video collection.

Reply 185 of 253, by Fagear

User metadata
Rank Member
Rank
Member

Thinking about usability...
There will be a separate manual supplied for configuring the board. But there will be the moment, when the FMonster board will be without manual somewhere.
So, I think, there should be brief manual on the board itself.

I've updated my project and I need some feedback.

fmonster_2019-08-18.png
Filename
fmonster_2019-08-18.png
File size
226.17 KiB
Views
360 views
File comment
FMonster @2019-08-18
File license
Fair use/fair dealing exception

Is anything understandable? Where confusions are? 😕

New BIG soundcard: FMonster.
Covox Sound Master replica
Innovation SSI-2001 replica & DuoSID.
My audio/video collection.

Reply 186 of 253, by root42

User metadata
Rank Oldbie
Rank
Oldbie

That looks crazy. It will never work. 😁

One question regarding OPL2: Will it always listen on 388h, for AdLib compatibility? Even when jumpered for 220h/222h?

YouTube and Bonus
80386DX@25 MHz, 8 MiB RAM, Tseng ET4000 1 MiB, SnarkBarker & BlasterBoard, PC MIDI Card + SC55 + MT32, XT CF Lite, OSSC

Reply 187 of 253, by Fagear

User metadata
Rank Member
Rank
Member
root42 wrote:

That looks crazy. It will never work. 😁

Maybe. 🤣

root42 wrote:

One question regarding OPL2: Will it always listen on 388h, for AdLib compatibility? Even when jumpered for 220h/222h?

By default you select either OPL2 will listen to 388h or 220h or any other port.
Probably I can add some "schematic OR" for this, so if you'll install jumper both on 220h and 388h it will listen to both addresses and not burn out outputs of 74HC138. 😕

New BIG soundcard: FMonster.
Covox Sound Master replica
Innovation SSI-2001 replica & DuoSID.
My audio/video collection.

Reply 189 of 253, by root42

User metadata
Rank Oldbie
Rank
Oldbie
Fagear wrote:
root42 wrote:

One question regarding OPL2: Will it always listen on 388h, for AdLib compatibility? Even when jumpered for 220h/222h?

By default you select either OPL2 will listen to 388h or 220h or any other port.
Probably I can add some "schematic OR" for this, so if you'll install jumper both on 220h and 388h it will listen to both addresses and not burn out outputs of 74HC138. 😕

Still not following... The original Sound Blaster listens on 388h always, in addition to whatever port you chose. Does the FMonster do the same? Software for Adlib will write to 388h, whereas Blaster software will use whatever BLASTER environment or auto detect tells it to do.

YouTube and Bonus
80386DX@25 MHz, 8 MiB RAM, Tseng ET4000 1 MiB, SnarkBarker & BlasterBoard, PC MIDI Card + SC55 + MT32, XT CF Lite, OSSC

Reply 190 of 253, by Scali

User metadata
Rank l33t
Rank
l33t
root42 wrote:

Still not following... The original Sound Blaster listens on 388h always, in addition to whatever port you chose. Does the FMonster do the same? Software for Adlib will write to 388h, whereas Blaster software will use whatever BLASTER environment or auto detect tells it to do.

I think the idea is to choose whether the OPL2 or the OPL3 should be listening to 388h for AdLib compatibility?

http://scalibq.wordpress.com/just-keeping-it- … ro-programming/

Reply 191 of 253, by Fagear

User metadata
Rank Member
Rank
Member
root42 wrote:

The original Sound Blaster listens on 388h always, in addition to whatever port you chose. Does the FMonster do the same?

At the moment - no, it does not. If you set the OPL2 jumper to 388h - it will respond as AdLib. If you set it to 220h - it will not. But you can, at the same time, leave OPL3 at 388h.
Or you can leave OPL2 set to 388h/388h (as AdLib) or to 388h/38Ah (as PAS) and set OPL3 part for 220h/240h/260h (as SB).

74HC138 doesn't have open collector type output, so setting two jumpers at once (220h AND 388h) will overload its outputs.
But if such behavior is desired, I can try to add some sort of "diode OR" for jumpers, so you will be able to set 220h + 388H for the OPL2 block alone. 😕

Is the same functionality desired for OPL3 block also?

New BIG soundcard: FMonster.
Covox Sound Master replica
Innovation SSI-2001 replica & DuoSID.
My audio/video collection.

Reply 192 of 253, by Benedikt

User metadata
Rank Member
Rank
Member
Fagear wrote:

I've updated my project and I need some feedback.

fmonster_2019-08-18.png

Is anything understandable? Where confusions are? 😕

Would it be possible to consolidate at least some of the 74xx logic into a bunch of GALs?

I'd probably just build a tiny board with an FPGA and a DAC. It's obvious that you want original and period accurate components, so it can't be an FPGA, but GALs might still be an option when it comes to part count and size reduction.

Reply 193 of 253, by root42

User metadata
Rank Oldbie
Rank
Oldbie

Oh, I didn’t think that the OPL3 part could do the 220h. So maybe that feature isn’t really needed. I was merely thinking about how SB and other cards do it.

YouTube and Bonus
80386DX@25 MHz, 8 MiB RAM, Tseng ET4000 1 MiB, SnarkBarker & BlasterBoard, PC MIDI Card + SC55 + MT32, XT CF Lite, OSSC

Reply 194 of 253, by Fagear

User metadata
Rank Member
Rank
Member
Benedikt wrote:

Would it be possible to consolidate at least some of the 74xx logic into a bunch of GALs?

It is certainly possible, but I'm not doing it. The theme for the project: "no programmable devices". That includes GALs, FPGAs, MCUs, ROMs.
I'm looking into making next generation of FMonster as USB-connected box with same synth chips, THERE I probable will go with MCU+FPGA for all external logic.

root42 wrote:

Oh, I didn’t think that the OPL3 part could do the 220h. So maybe that feature isn’t really needed. I was merely thinking about how SB and other cards do it.

I'll try to include "address OR" functionality for OPL2 block either way, if I can.

New BIG soundcard: FMonster.
Covox Sound Master replica
Innovation SSI-2001 replica & DuoSID.
My audio/video collection.

Reply 195 of 253, by Scali

User metadata
Rank l33t
Rank
l33t
root42 wrote:

Oh, I didn’t think that the OPL3 part could do the 220h. So maybe that feature isn’t really needed. I was merely thinking about how SB and other cards do it.

SB Pro puts the OPL3 at 220h (or whatever base address is configured).
Port 388h is ONLY for OPL2 on SB cards. The second register bank is not exposed there, as far as I know.

http://scalibq.wordpress.com/just-keeping-it- … ro-programming/

Reply 196 of 253, by Fagear

User metadata
Rank Member
Rank
Member

root42, Scali either way, I've added "diode OR" for all jumper positions on OPL2 block.
Now it is possible to set jumper to ALL locations at once and the board will work just fine. So It can be: SB (L:0x220/R:0x222, LR:0x228), AdLib (LR:0x388), PAS (L:0x388/R:0x38A) and THE SAME TIME if you really want it to. 😲
I've used Schottky diodes to make that "OR" and for YM3812 maximum voltage for "low level" input is 0.4 V, so voltage drop across Schottky diode should be within the limit.

OPL2_2019-08-22.png
Filename
OPL2_2019-08-22.png
File size
80.04 KiB
Views
157 views
File comment
2xOPL2 @2019-08-22
File license
Fair use/fair dealing exception

I've also made some rough calculations on current consumption. Those are primarily based on static current worst-case specs, but here those are.
FMonster.
+5V #1: 99 mA
+12V: 579 mA
-12V: 77 mA
+5V #2: 70 mA

FMonster's "2nd stage".
+5V #1: 49 mA
+12V: 35 mA
-12V: 29 mA
+5V #2: 4 mA

Strangely enough, it seems to fit within ISA bus specs (1.5A per rail) with quite a margin. 😕

New BIG soundcard: FMonster.
Covox Sound Master replica
Innovation SSI-2001 replica & DuoSID.
My audio/video collection.

Reply 197 of 253, by root42

User metadata
Rank Oldbie
Rank
Oldbie

Just put a molex connector on the card. Just like today's GPUs. 😉

YouTube and Bonus
80386DX@25 MHz, 8 MiB RAM, Tseng ET4000 1 MiB, SnarkBarker & BlasterBoard, PC MIDI Card + SC55 + MT32, XT CF Lite, OSSC

Reply 198 of 253, by Fagear

User metadata
Rank Member
Rank
Member
root42 wrote:

Just put a molex connector on the card. Just like today's GPUs. 😉

Molex won't provide -12V rail. 😕 And it will introduce power and ground loops.

New BIG soundcard: FMonster.
Covox Sound Master replica
Innovation SSI-2001 replica & DuoSID.
My audio/video collection.

Reply 199 of 253, by root42

User metadata
Rank Oldbie
Rank
Oldbie

Well, it was just meant as a joke. As the -12V is the biggest load on the board, I think it is fair to assume this will have to come via the ISA bus, and can't be generated via DCDC converters on the board itself. What is that actually used for? SID?

YouTube and Bonus
80386DX@25 MHz, 8 MiB RAM, Tseng ET4000 1 MiB, SnarkBarker & BlasterBoard, PC MIDI Card + SC55 + MT32, XT CF Lite, OSSC