VOGONS


Crazy Idea - MT32 WaveBlaster Card

Topic actions

Reply 60 of 100, by RJDog

User metadata
Rank Member
Rank
Member
RJDog wrote:

The TI part I'm looking at says its switching time is 4.7ns, which equates to roughly 2MHz, so that should be okay, I think?

I just read a blog post where the guy was able to cleanly pass a 860kHz (430,000 baud?) signal through the 74HCT125, so that sounds pretty okay to me.

Reply 61 of 100, by ab0tj

User metadata
Rank Member
Rank
Member

4.7ns is more like 213MHz. 74 series logic is routinely used in circuits operating at 10+MHz, so I'm sure you'll be fine. If you're only level converting the MIDI signal, even a resistor divider would probably do just fine.

Reply 62 of 100, by gdjacobs

User metadata
Rank l33t++
Rank
l33t++

A resistor divider would work fine for conversion of 5V to 3.3V as long as your input Z is high and output Z is low (compared to your network Z). Buffered logic is insensitive to impedance. Essentially, choose the tool you require.

All hail the Great Capacitor Brand Finder

Reply 63 of 100, by Ariakos

User metadata
Rank Member
Rank
Member
RJDog wrote:

Well.. there do seem to exist PCMCIA MIDI cards....

Yes, they do exist. But they are few in actual numbers and even if you happen to have one it's always in General MIDI, not LA Synth. Quite understandable since LA synth was pretty much old news when PCMCIA cards came available and GMidi was The Thing. You can easily emulate MIDI stuff with modern laptops through DosBox+Munt. Maybe even go with USB MIDI interface and real MIDI modules. But only plausible way to do MIDI with any true hardware solution in genuine retro laptops is through those PCMCIA sound cards. And good ol' LA synth in them is nonexistent.

Ah, but I digress. I just wanted to express my fondness for genuine hardware laptop MIDI solutions. I understand that this daughtercard project is it's own thing. 😎

Reply 64 of 100, by PhilsComputerLab

User metadata
Rank l33t++
Rank
l33t++

I've mucked around with a MeeGoPad Compute Stick with the older Z3736F Atom and it ran DOSBox games as well as Munt without any issues. I can see similar tablets selling for around $80 to $100, so I think this is the direction I'm leaning towards. All we would need is a MT-32 themed tablet stand 🤣

YouTube, Facebook, Website

Reply 65 of 100, by RJDog

User metadata
Rank Member
Rank
Member
gdjacobs wrote:

A resistor divider would work fine for conversion of 5V to 3.3V as long as your input Z is high and output Z is low (compared to your network Z). Buffered logic is insensitive to impedance. Essentially, choose the tool you require.

It's true, for 5v -> 3.3v a resistor network would be fine; but I need to go 3.3v -> 5v as well, so I think I will need at least one active chip to perform that task, and if I've got one chip to go one direction I might as well have an active chip to go the other direction. The buffer chips are pretty cheap anyway, like $0.65... almost cheaper than resistors.

I had to think about this a lot over the weekend, thinking that I would throw out the whole idea of a WaveBlaster card and just go with the standalone unit and do the whole thing in 3.3v logic, but I think I came up with a good design, I think better than the first revision, so I'm happy; just means going back to the drawing board and entering t he new design into KiCAD again...

Reply 66 of 100, by ab0tj

User metadata
Rank Member
Rank
Member
RJDog wrote:

It's true, for 5v -> 3.3v a resistor network would be fine; but I need to go 3.3v -> 5v as well, so I think I will need at least one active chip to perform that task, and if I've got one chip to go one direction I might as well have an active chip to go the other direction. The buffer chips are pretty cheap anyway, like $0.65... almost cheaper than resistors.

Well, I'm just being pedantic at this point, but a 3.3V '1' is still a '1' in a 5V system (well into the '1' range in TTL logic, and just above the cutoff but still within the range of a 1 for a CMOS device)...

That said, yeah a buffer isn't a huge expense anyway and I was mainly just illustrating the point that at these speeds you don't need anything fancy.

EDIT: Ok, I take that back about CMOS logic. The spec apparently says 3.5V minimum, but the datasheet for e.g. a 74HC125 says 3.15V so YMMV

Reply 67 of 100, by gdjacobs

User metadata
Rank l33t++
Rank
l33t++
ab0tj wrote:
Well, I'm just being pedantic at this point, but a 3.3V '1' is still a '1' in a 5V system (well into the '1' range in TTL logic, […]
Show full quote
RJDog wrote:

It's true, for 5v -> 3.3v a resistor network would be fine; but I need to go 3.3v -> 5v as well, so I think I will need at least one active chip to perform that task, and if I've got one chip to go one direction I might as well have an active chip to go the other direction. The buffer chips are pretty cheap anyway, like $0.65... almost cheaper than resistors.

Well, I'm just being pedantic at this point, but a 3.3V '1' is still a '1' in a 5V system (well into the '1' range in TTL logic, and just above the cutoff but still within the range of a 1 for a CMOS device)...

That said, yeah a buffer isn't a huge expense anyway and I was mainly just illustrating the point that at these speeds you don't need anything fancy.

EDIT: Ok, I take that back about CMOS logic. The spec apparently says 3.5V minimum, but the datasheet for e.g. a 74HC125 says 3.15V so YMMV

The low threshold of a 3.3V TTL high signal is in an indeterminate range for 5V TTL. Lots of situations where that could get you in trouble. If you know you're always getting 3.2-3.3V into the gate, you're probably okay, but buffering eliminates those sorts of concerns.

All hail the Great Capacitor Brand Finder

Reply 68 of 100, by RJDog

User metadata
Rank Member
Rank
Member

So, I feel like I'm resurrecting an old thread here, but really I have been quite actively working on this one 😊

So I have what is now Revision 3 of the wavetable board. Admittedly, I am starting to become less convinced that a wavetable card is really the way to go, as it might not be practical for most users as opposed to a standalone-only device. This standalone "host" is obviously designed to accept the wavetable card and turn it into standalone unit, but it is also designed to be able to accept other "generic" wavetable cards and turn them into a standalone unit... yes, very similar to dreamblaster's WaveBlaster Module MIDI Interface Board, but from a "generic" perspective this board also has midi out and thru, and has a USB MIDI interface (auto switches the input to the wavetable card from the MIDI 1.0 ports to the USB MIDI when USB is plugged in). That said, my testing will be limited to ensuring the "MB-32" wavetable card works with it.

I changed to surface mount components for the ICs... mostly because it seemed to be practical to get Atmel microcontrollers with the features and pin count that I want that only come in surface mount packages, so I said screw it, and chose surface mount packages for the other ICs as well.

I'm pretty committed to this design now, and actually have all of the components ordered and received to build it... just waiting on the PCBs now...

Attachments

Reply 69 of 100, by hard1k

User metadata
Rank Oldbie
Rank
Oldbie

Hell yeah! Looking forward to your news!

Fortex, the A3D & XG/OPL3 accelerator (Vortex 2 + YMF744 combo sound card)
AWE64 Legacy
Please have a look at my wishlist (hosted on Amibay)

Reply 72 of 100, by Rackatack

User metadata
Rank Newbie
Rank
Newbie
RJDog wrote:
PhilsComputerLab wrote:

The display is part of the experience too.

Yeah, I thought that too. Maybe I'll make a WaveBlaster card, and then a custom enclosure the card can be inserted into to make it a standalone unit with MIDI interfaces and LCD..?

So why not a display that fits in the tower (like a DVD or floppy drive) with something like this?

Attachments

  • maxresdefault.jpg
    Filename
    maxresdefault.jpg
    File size
    75.64 KiB
    Views
    1426 views
    File license
    Fair use/fair dealing exception

Reply 73 of 100, by gdjacobs

User metadata
Rank l33t++
Rank
l33t++

I've done a proof of concept for that:
Roland MT-32, CM-32L + General MIDI for $50 Building a MIDI box

Let me know if you want the source modifications for mt32d.

All hail the Great Capacitor Brand Finder

Reply 74 of 100, by hard1k

User metadata
Rank Oldbie
Rank
Oldbie

That would be too good to be true.

Fortex, the A3D & XG/OPL3 accelerator (Vortex 2 + YMF744 combo sound card)
AWE64 Legacy
Please have a look at my wishlist (hosted on Amibay)

Reply 75 of 100, by gdjacobs

User metadata
Rank l33t++
Rank
l33t++

Also, any information on the actual output quality of the Orange Pi? The Raspis have a notoriously bad codec driving the headphone jack, so an external audio device (either on an expansion plate or via USB) is virtually a must.

All hail the Great Capacitor Brand Finder

Reply 76 of 100, by RJDog

User metadata
Rank Member
Rank
Member
Rackatack wrote:

So why not a display that fits in the tower (like a DVD or floppy drive) with something like this?

On the MB-32 wavetable card the "host" connector is intentionally designed to be dual purpose; to not only provide the required connections to the MB-33 host (the standalone enclosure with all I/O, LCD and front panel), but it can also provide a connection to a LCD panel. I designed it that way on purpose so that it could be adapted to drive an LCD when used internally as a WaveTable card with the LCD in, say, a 5 1/4" drive bay. I am designing the board to do this, but don't intend to have the PCB made in this prototyping phase.

gdjacobs wrote:

Also, any information on the actual output quality of the Orange Pi? The Raspis have a notoriously bad codec driving the headphone jack, so an external audio device (either on an expansion plate or via USB) is virtually a must.

It is supposed to be 24-bit DAC, and a respectable 96 dB SNR, notably better than the Raspberry Pi which don't really have a DAC, just a reasonably filtered PWM ouput.
This is one of the things that attracted me to the OPi. I (unfortunately) only just received the Orange Pi Zero this past week (yay shipping from China) and havent had a chance to really test it yet.

Reply 77 of 100, by Super_Relay

User metadata
Rank Newbie
Rank
Newbie

there is an alternative audio driver on the pi that sounds a lot better with its default analog out. by default its not enabled because it uses a lot of compute power from the graphics chip.

see this thread
https://www.raspberrypi.org/forums/viewtopic. … p?f=29&t=136445

Reply 78 of 100, by RJDog

User metadata
Rank Member
Rank
Member

Starting to get a little more real...

Attachments

  • 20170618_001908.jpg
    Filename
    20170618_001908.jpg
    File size
    1.67 MiB
    Views
    1260 views
    File license
    Fair use/fair dealing exception

Reply 79 of 100, by hard1k

User metadata
Rank Oldbie
Rank
Oldbie

Looking great! 😀

Fortex, the A3D & XG/OPL3 accelerator (Vortex 2 + YMF744 combo sound card)
AWE64 Legacy
Please have a look at my wishlist (hosted on Amibay)