I heard back from him, he said he will be sending the needed software along shortly! Looks like a success to me.

As soon as I get them I will upload them to the Vogons driver library thread, and also put a copy on my Google Drive, so even if my local machine fails before backup, a copy will continue to exist. Considering the Vogons driver library thread will literally be the only place on the internet they will exist to the public, its always good to have a backup, since I doubt we will every be able to get a copy of them again.

Success!

The ZIP provided by the creator of the hardware appears to be essentially a file dump of what they had left file-wise for the SoundJr, but all the pieces seem to be there.

I uploaded it in a post to the Vogons Driver Library thread if anyone wants to download it.

Great news! Thank you very much! 😁

Has anyone popped open a Covox Speech Thing included amplified speaker to see what is is in there? Or is that impossible without destroying it?

I'm curious to know if they put a low-pass filtering cap in there, etc... so we know what their intended cut-off for it is meant to be.

Also, any benefit to making a 9-bit DAC over the standard 8-bit DAC that the Covox had? The .1% tolerance resistors that are being used are nearly twice as tolerant as what is needed to make a 9-bit DAC (and nearly 4 times as tolerant as what is needed for an 8-bit DAC, which is why 1% and even .5% would not be enough to ensure tonal accuracy).

Now it makes sense why the 1% parts weren't good enough, after doing the math, they were 2.5x less tolerant than was needed for an 8-bit DAC. It was essentially increasing the error rate to the point that it was reducing the resolution of the DAC to more than 6-bits, but less than 7-bits.

I also did some math for the low pass filters, anything less than right around 500 picofarads (which should produce a cut-off of approx. 30 kHz, well beyond even the best human hearing) will produce no change that can be heard by a human being, so, why bother including it?

As far as I can tell, the SoundJr. uses a 1 nanofarad low pass filtering capacitor, making its cut off about 15 kHz, which isn't too bad.

That's more than doubled the covox-speech-thing which cut's of a 7khz afaik.

dr.zeissler wrote:

That's more than doubled the covox-speech-thing which cut's of a 7khz afaik.

Yeah, from what I have read, the best guess, without anyone having opened up one of the Speech Things speakers to look for a low-pass cap, is that, by ear, its cut-off is ~7-8 kHz, so the SoundJr. has about double the frequency range, covering most of the human hearing range, ~75%, compared to ~38% for a cut-off of 7-8 kHz. (7-8 kHz is an interesting choice, maybe it was because of the limitations of research at that time. It covers the so called "critical speech domain" or 1 kHz - 6 Khz. But its now well known that speech is most easily interpreted if the range is expanded to ~300 Hz - 10 kHz)

The CVX4 still has them all beat. With the ~500 pF low-pass fllter its cuf-off should be ~30 kHz+, well into the cat/dog hearing range, and exceeding human hearing of even the most acute human ears by at least ~5 kHz+ (~10 kHz+ for an average young human, its almost a guarantee that most of us here, [Assuming you are older than 25] are well below the 20 kHz typical human hearing cut-off). So, if you are using your CVX4 with a low-pass of ~500 pF or less, and your dog or cat starts acting odd, chances are you have some kind of harmonic that is resonating in their hearing range, but not yours 🤣

Cloudschatze wrote:

Wayne Foletta (Silicon Shack / SiliconSoft) produced a neat little LPT DAC known as the "SoundJr" back in the late 1980s. It's n […]
Show full quote

Wayne Foletta (Silicon Shack / SiliconSoft) produced a neat little LPT DAC known as the "SoundJr" back in the late 1980s. It's not as advanced as some of the designs being discussed, but has a few interesting features. Instead of discrete resistors, this device employs a pairing of either 767163103 and 767163203, or 8A103 and 8A203 resistor array ICs. In addition, the SoundJr features software-controllable, 8-step volume attenuation. These (or similar) elements might be worth considering as part of a new design.

I have an extra unit, if there's any interest in a reverse-engineering effort. Otherwise, there's a passthrough version of the SoundJr on eBay that could be looked at, and isn't too terribly priced. The listing even includes some oddly nice photos of its PCB.

I got Cloudschatze's spare SoundJr. (Thank you very much!) and started poking around inside of it. I am quite impressed with the product, especially for 1992 (which as the first version of his design, he produced a second revision which is the passthrough design mentioned above by Cloudschatze). Wayne Foletta was clearly gifted regarding sound hardware design. He produced a product that had 8 step software volume control, double the frequency range of the Covox Sound Thing, and could directly drive a 32 Ohm pair of headphone due to the integration of a simple amp, and even a sleep-mode so it would stop drawing power if you were running off of a laptop battery, all working off a measly 3 mA provided by the LPT. As far as I can tell, it is the most advanced product of its type ever sold to the public during that time period. The next best was obviously the Covox Speech Thing/Disney Sound Source. All the rest appeared to be sort of 1/2 assed copies of the original Covox (they seem to have run into some of the same issues that were encountered here in the reverse engineering process, but never bothered to work them out and just said "Meh... Good enough" and chucked their hobbled product onto the market for a few bucks).

I plan to thoroughly go through the entire product and create a schematic from it, since I believe the CVX project could actually adapt some of the design decisions that were made here to provide additional features, such as an on-board amp that doesn't need an external power supply, so people could actually directly drive a pair of headphone without an external amp if they want.

I'm going to need some help with the software side though, I'm far more of a theory and hardware person. I'll need someone to assist with pulling the software apart so we can figure out how the 8 step volume control works, since that would also be a nice feature to add directly to the CVX as well. At the moment, the most I can tell is that the volume control signaling is handled through pins 1, 14, 16, 17, and pulling all those pins high drives the on-board amp to full output.

If the best features from this can be incorporated into the CVX design, it will easily be the best "Covox" ever made, quite a bit superior to even the original. Now we just need to convince Dreamblaster to start offering the CVX with a DSS daughterboard, and it will be the ultimate Covox solution, all the retro people that want to produce sound from a system that can't or doesn't have ISA expansion will have the best possible option at a very reasonable price. Sound like a winning solution for all involved, at least to me.

Paralel wrote:

At the moment, the most I can tell is that the volume control signaling is handled through pins 1, 14, 16, 17, and pulling all those pins high drives the on-board amp to full output.

Those 4 pins are the only other 4 output pins on a parallel port, aside from the 8 data lines.
My guess is that one of them is used for signaling that a new volume is to be set.
The other 3 are then probably interpreted as a 3-bit number, yielding 8 discrete values.

If my theory is correct, that would mean that:
- Whenever a volume is set, pins 1, 14, 16 and 17 will at least briefly be set to the 4-bit value to set that particular volume
- For every one of the 8 volumes, one of the 4 pins will be enabled. So we will see the following pattern, in whatever order the pins have been chosen:
Volume 0: 1000
Volume 1: 1001
Volume 2: 1010
Volume 3: 1011
Volume 4: 1100
Volume 5: 1101
Volume 6: 1110
Volume 7: 1111

Perhaps you can check this by just setting all the 8 possible volumes, and monitoring the pins, without having to pull apart the software.

Paralel wrote:

"If the best features from this can be incorporated into the CVX design, it will easily be the best "Covox" ever made, quite a bit superior to even the original. Now we just need to convince Dreamblaster to start offering the CVX with a DSS daughterboard, and it will be the ultimate Covox solution, all the retro people that want to produce sound from a system that can't or doesn't have ISA expansion will have the best possible option at a very reasonable price. Sound like a winning solution for all involved, at least to me."

Great concept, interested in this, when I find the time. (my attention goes to the OPL2 and OPL3 chips on parallel port, right now)

problem is not to hook up a OPL2 to LPT, its the Software.

Are you be able to write a driver to redirect port access to lpt ? (emm386..)

http://www.raphnet.net/electronique/adlib/adlib_en.php

as you can see its pretty easy to hookup the OPL2/3 to LPT.

Take a look at the OPL LPT project forum thread.

Afaik the HW interface should be the same and there is a TSR available on Github.

Scali wrote:

Some things I may add in the near future:
1) Commandline switch to select LPT port
2) Commandline switch to select sample rate

I have updated CvxPlay again, adding these features.
It can be downloaded here:
https://www.dropbox.com/s/d676aihs71fw2bk/cvxplay.zip?dl=0

Usage is simple:
Usage: CvxPlay <filename.ext> <sample rate> <printer port>

Sample rate in Hz, is in decimal, and is optional, default is 44100 Hz. It only supports 16-bit values, so don't go over 65535.
Printer port is in in hex, and also optional (but you need to specify a sample rate as well, to use this), default is 378h

For example:
CvxPlay test.raw 22050 -> Plays the file 'test.raw' at 22050 Hz
CvxPlay test.raw 44100 3B8 -> Plays the file 'test.raw' at 44100 Hz, on port 3b8h

Waw scali, that's great !

Is there interest in another batch of CVX4 ?

I would buy one if another batch is made.

And here is my blog on the Covox/CVX-4:
https://scalibq.wordpress.com/2017/11/28/the-covox-years/

Wow Scali, that's great to read, thanks !
I will post the link on my serdashop cvx page.
Do you see anything that needs to be changed ?
Maybe I do another batch early next year.
Grtz
S

dreamblaster wrote:

Wow Scali, that's great to read, thanks !
I will post the link on my serdashop cvx page.

You're welcome!

dreamblaster wrote:

Do you see anything that needs to be changed ?

You mean to the CVX4? Well no, as a Covox clone, it is everything I hoped for, and more.
I can only repeat some ideas I had earlier:
1) Having some kind of pre-amp to go with the CVX4 would be nice. Perhaps something built around an LM386 or so, which feeds off +5v that you could tap from a USB port or an old PS/2 or XT/AT keyboard port.
I see this for example: https://www.hackerstore.nl/Artikel/310
You wouldn't have to make/supply one yourself, but you could put a link up there as suggestion, and also include some info on how to make cables to get the +5v without needing an external PSU.

2) A 'low-cost' version, that just gives you a bare minimal Covox, but at the same great quality (with default Covox values). So you would drop the DIP switches from the design. I suppose you could also replace the 2x100k resistor pairs by 200k resistors, that shouldn't affect quality much, if at all. That way you need less parts, so it's cheaper and easier to assemble. You could also make the PCB smaller. I don't know how much you would save with that, but if they would be less than 15 euros, it may just convince a whole lot more people to get one (instead of the super-cheap crappy clones out there that really aren't worth it).
I suppose you could make a rough estimate of the cost, and put it on the webshop to see what the interest is. You wouldn't actually have to design the PCB and build the devices unless you know there's actually more interest for it than the CVX4, so it's actually worth the trouble.

I would like to add that once I find my old MDA card, I can put a second printer port in one of my old PCs, and I will see what happens if I create a stereo version of CvxPlay then.

Scali wrote:

You mean to the CVX4? Well no, as a Covox clone, it is everything I hoped for, and more. I can only repeat some ideas I had earl […]
Show full quote

You mean to the CVX4? Well no, as a Covox clone, it is everything I hoped for, and more.
I can only repeat some ideas I had earlier:
1) Having some kind of pre-amp to go with the CVX4 would be nice...
2) A 'low-cost' version, that just gives you a bare minimal Covox...

Yeah good ideas, I'll consider. The batch size is quite determing for the cost as well,
as the assembly factory is quite expensive for small runs. (and in fact all my products are small runs).
A solder kit could be quite cheaper (but through hole 0.1% resistors are even more expensive)
I'll consider the options.
(but now busy with other sound related projects 😈 )

Which one is cheaper, bunch of 0.1% resistors, or a ready-made R2R ladder that should have more matched resistors? I think you can get R2R ladders below 1 euro..