VOGONS


CVX4 : high quality covox adapter

Topic actions

Reply 320 of 484, by gdjacobs

User metadata
Rank l33t++
Rank
l33t++

That's essentially what the opamp does for you (within the current draw limits of the chip of course). Low output impedance from the amp prevents any line in from being attenuated.

All hail the Great Capacitor Brand Finder

Reply 322 of 484, by gdjacobs

User metadata
Rank l33t++
Rank
l33t++

You'd be able to drive anything compatible with the voltage swing and max current supply of the opamp. Generally, that includes line level inputs and some headsets. Any speakers or headphones which demand higher voltage or higher current from their supply would need another amplifier in the reproduction chain.

All hail the Great Capacitor Brand Finder

Reply 323 of 484, by Jepael

User metadata
Rank Oldbie
Rank
Oldbie

While some power can be sourced from parallel ports for some applications, in general, I think there's a lot of incompatibilities when doing so.
For instance, when powering some SPI EEPROM or small microcontroller for parallel port "dongle" for software activation, at least there you have a specific software handling the parallel port in specific way. For instance idling data pins high a lot for power harvesting into storage cap, and only shortly transfer data so the cap voltage does not drop too much. And even then those did not work on all computers..

But for general purpose covox application, there is no standard in what way they set the control pins, and since the data is audio, there is no guarantee how much on average the bits are low or high to provide power. Also parallel ports from different eras have different output stages so the logic levels might be 5V or 3.3V to begin with, and early logic chips had pretty high output impedance, so the voltage will fluctuate a lot depending on how many bits are turned on (when harvesting power from them that is). So for R2R DAC + opamp, I don't see a general solution that would work well on all machines, even if it does work on certain parallel ports with strong output drivers. And because of the high output impedance of the data pins, wouldn't that also change audio amplitude depending on how many bits are on, so that would distort the audio as well (unless it's a DAC chip).

Having said that, the Covox patent did include designs that harvested power from parallel port and had a transistor as emitter follower (buffer amplifier). But did they produce any buffered designs, I have no clue. If not, chances are they found out they don't really work, which is why the plug came with externally powered speakers so they can have the amplifier there. I believe some early designs had two wires from the speaker to the plug, the other carrying power for some electronics in the plug.

Fortunately, much more power for active covox plug can be harvested from elsewhere, like keyboard/mouse/joystick ports, serial port..

Reply 324 of 484, by gdjacobs

User metadata
Rank l33t++
Rank
l33t++

It would definitely be a tricky project. I had considered output impedance on the port, although I wasn't sure if it would be surmountable. Different logic levels would make it even more of a tangle.

All hail the Great Capacitor Brand Finder

Reply 325 of 484, by Scali

User metadata
Rank l33t
Rank
l33t

Is it confirmed though, that there exist parallel ports that output 3.3v rather than 5v?
I mean, the standard has always been +5v, so a number of devices would have problems if someone suddenly changed it to 3.3v.
So I can't really imagine that this actually occurs.

Also, it's not related to the base voltage of the motherboard, CPU or anything. You have the same problem with COM ports for example, and there are standard chips to solve that, like the MAX232: http://www.ti.com/lit/ds/symlink/max232.pdf
I would imagine that a similar chip is used for modern parallel ports, and they're still 5v-level on the outside.
If not, let people with 3.3v ports sort that out for themselves. Let them put MAX248 or equivalent chips in between to bump the voltage back up to 5v.

I think it's a reasonably safe assumption that people will want to use a Covox on a Pentium or lower. I tried the CVX-2 on my Pentium Pro 200, and didn't notice a difference with my 286-20, so I suppose they are both +5v.

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

Reply 326 of 484, by Jepael

User metadata
Rank Oldbie
Rank
Oldbie
Scali wrote:

Is it confirmed though, that there exist parallel ports that output 3.3v rather than 5v?

I mean, the standard has always been +5v, so a number of devices would have problems if someone suddenly changed it to 3.3v.
So I can't really imagine that this actually occurs.

It seems one of my VIA chipset rigs has still has a separate 5V super-I/O chipset from Winbond, but another has same features integrated into 3.3V VIA southbridge. Both Socket A motherboards, so yes, newer than Pentium so maybe a little too modern for running programs for Covox.

Some hobbyists who connect their equipment like homebuilt CNC machines to parallel ports are also saying they are having issues because they did not expect that it was 3.3V but 5V.

The parallel port used standard TTL logic levels when it was made, and modern IEEE 1284 ports that support EPP and ECP still use same standard TTL logic voltage levels according to specs. If the printer or whatever device does not use standard TTL logic levels then it may not work, but it's a problem of the printer/device. It's just that discrete LS TTL chips were powered from 5V, and modern CMOS multi I/O chipsets from 3.3V, and that's not against any TTL chip or ECP specs. So in essence the supply voltage should be irrelevant as long as the chips use correct voltage levels determining high and low states at their inputs, high being 2.0V or higher, low being 0.8V or lower. Typically, a logic high voltage from a 5V LS TTL chip driving small load will be perhaps 3.5V, which is just as good as driving it close to 3.2V with more modern 3.3V chip.

Reply 327 of 484, by Scali

User metadata
Rank l33t
Rank
l33t
Jepael wrote:

The parallel port used standard TTL logic levels when it was made

Well, the manual specifically states that they are +5vdc: http://www.minuszerodegrees.net/oa/OA%20-%20I … r%20Adapter.pdf
See page 14.
So I would expect any clone until the end of time to do exactly the same. Else it's not PC-compatible.

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

Reply 328 of 484, by Jepael

User metadata
Rank Oldbie
Rank
Oldbie
Scali wrote:
Well, the manual specifically states that they are +5vdc: http://www.minuszerodegrees.net/oa/OA%20-%20I … r%20Adapter.pdf See pa […]
Show full quote
Jepael wrote:

The parallel port used standard TTL logic levels when it was made

Well, the manual specifically states that they are +5vdc: http://www.minuszerodegrees.net/oa/OA%20-%20I … r%20Adapter.pdf
See page 14.
So I would expect any clone until the end of time to do exactly the same. Else it's not PC-compatible.

That was a good catch to refresh my memory cells. I had not considered the control pins at all because covoxes only use the data pins.
Now that you mentioned it, while the PC I have has 3.3V logic levels on data outputs, it does have this pull-up resistor to 5V on control pins so these are compatible as well.

Page 14 is not about the data output pins, but control output pins.
And if you flip to page 9, it has listed the standard TTL logic level output voltages there as well, and the video connector is even driven by a beefier buffer chip at 5V supply.

The logic chip on control outputs is open-collector type so it cannot push high, it can only pull low. The control outputs therefore need to have pull-up resistors to 5V, otherwise nobody would see a logic high voltage levels. So when they are not pulled low by the logic chip, they float right at 5V when unloaded, but can keep output voltage just barely at 2.4V while sourcing only 0.55mA of current. So it's very weak as a high output, even a short circuit to 0V (ground) would not give you more than 1.1mA through the resistor.

The data outputs do not have pull-up resistors to 5V, so they don't go much higher than 3.5V when driven high, but they can keep output voltage above 2.4V, typically at 3.1V, while sourcing out up to 2.6mA of current. So it's able to drive inputs of quite many logic chips.

Reply 329 of 484, by dreamblaster

User metadata
Rank Oldbie
Rank
Oldbie

PC punaise running fasttracker 2 , playback on CVX4 : https://youtu.be/CQNUzXAutNs

Visit http://www.serdashop.com for retro sound cards, video converters, ...
DreamBlaster X2, S2, S2P, HDD Clicker, ... many projects !
New X2GS SE & X16GS sound card : https://www.serdashop.com/X2GS-SE ,
Thanks for your support !

Reply 330 of 484, by dreamblaster

User metadata
Rank Oldbie
Rank
Oldbie

...

CVX4_TOP.jpg
Filename
CVX4_TOP.jpg
File size
349.56 KiB
Views
1712 views
File license
Fair use/fair dealing exception
CVX4_BOTTOM.jpg
Filename
CVX4_BOTTOM.jpg
File size
192.38 KiB
Views
1712 views
File license
Fair use/fair dealing exception

Visit http://www.serdashop.com for retro sound cards, video converters, ...
DreamBlaster X2, S2, S2P, HDD Clicker, ... many projects !
New X2GS SE & X16GS sound card : https://www.serdashop.com/X2GS-SE ,
Thanks for your support !

Reply 334 of 484, by dreamblaster

User metadata
Rank Oldbie
Rank
Oldbie

Hey,
a small update : I noticed I needed to swap 2 smd caps, to get all right filters :
The automated assembly had put 47nF and 100nF on there (my mistake in part nrs), but we need to do subtle filtering of course (exactly as tested)!
For the 6 units already shipped out : I will send you another CVX4 with corrected caps set, so you will receive 2 x CVX4 in total 😀.
All production boards will have the correct filter options, for tuning your sound to 8 bit perfection :

Dip switch legend for the production boards :
dips 1-5 : low pass filter
1: 1nF : significant low pass filter
2: 4.7nF : heavier low pass filter
3: 10 nF : very heavy low pass filter
4 : 0.1nF : very subtle low pass filter
5 : 0.47nF : subtle low pass filter

dip 6-7 : extra volume divider :
6 : 56k (1/2R)
7 : 27k (1/4R)

dip 8
8 : bypass coupling capacitor

Visit http://www.serdashop.com for retro sound cards, video converters, ...
DreamBlaster X2, S2, S2P, HDD Clicker, ... many projects !
New X2GS SE & X16GS sound card : https://www.serdashop.com/X2GS-SE ,
Thanks for your support !

Reply 335 of 484, by dreamblaster

User metadata
Rank Oldbie
Rank
Oldbie

Here's the production version of CVX4 parallel port sound card,
playing a complete song (picked a creative commons song!) using the cvxplay utility (thanks scali and mobygamer!):
https://youtu.be/meQ2L5S5ULU (no filtering)
https://youtu.be/nDQye6JKDcg (slight filtering)
and the complete test suite (some files are quite noisy conversion, also when played in audacity on my 24 bit soundcard): https://youtu.be/0EuefF9UI48
There may be a better algorithm than plain dropping the lowest 8 bit, when converting from 16 bit to 8 bit and reduce quantization noise ?

Visit http://www.serdashop.com for retro sound cards, video converters, ...
DreamBlaster X2, S2, S2P, HDD Clicker, ... many projects !
New X2GS SE & X16GS sound card : https://www.serdashop.com/X2GS-SE ,
Thanks for your support !

Reply 336 of 484, by Jepael

User metadata
Rank Oldbie
Rank
Oldbie
dreamblaster wrote:

There may be a better algorithm than plain dropping the lowest 8 bit, when converting from 16 bit to 8 bit and reduce quantization noise ?

Yes, for instance audacity offers dithering options; try triangular or noise shaped dithering when exporting into 8-bit file.

Reply 337 of 484, by dreamblaster

User metadata
Rank Oldbie
Rank
Oldbie
Jepael wrote:
dreamblaster wrote:

There may be a better algorithm than plain dropping the lowest 8 bit, when converting from 16 bit to 8 bit and reduce quantization noise ?

Yes, for instance audacity offers dithering options; try triangular or noise shaped dithering when exporting into 8-bit file.

Hey yeah, have listen at this page : http://www.audiocheck.net/audiotests_dithering.php
I'm looking into the option of audacity

Visit http://www.serdashop.com for retro sound cards, video converters, ...
DreamBlaster X2, S2, S2P, HDD Clicker, ... many projects !
New X2GS SE & X16GS sound card : https://www.serdashop.com/X2GS-SE ,
Thanks for your support !

Reply 338 of 484, by Scali

User metadata
Rank l33t
Rank
l33t

I've made a small update to CvxPlay: https://www.dropbox.com/s/d676aihs71fw2bk/cvxplay.zip?dl=0
It should now end properly at the end of a file, regardless of the length.
I also noticed that I accidentally did 'out dx, ax', where it should have been 'out dx, al'.
Some things I may add in the near future:
1) Commandline switch to select LPT port
2) Commandline switch to select sample rate

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

Reply 339 of 484, by dreamblaster

User metadata
Rank Oldbie
Rank
Oldbie
Scali wrote:

I've made a small update to CvxPlay

Wow thanks!

Visit http://www.serdashop.com for retro sound cards, video converters, ...
DreamBlaster X2, S2, S2P, HDD Clicker, ... many projects !
New X2GS SE & X16GS sound card : https://www.serdashop.com/X2GS-SE ,
Thanks for your support !