^ Sadly Lethal Xcess seem to use a custom format and I can't locate a MOD version of it's theme...

About price, I don't think that removing the dip switches or the 4-5 caps would make a considerable impact in it , most of it is manufacturing/assembly cost I imagine and for such a small item a couple of components addition/removal is not something to drastically change the cost. *IMHO*

Just a heads-up for anyone interested, CVX-4 preorders are now open: http://www.serdashop.com/CVX4

Keropi, can you please give a small summary how the CVX4 compared to the Original.
Did we eventually figured out what resistors were in the original and how quiet is the result in comparison?

- anyone correct me if I am wrong - but we determined the original uses 100K 0.1% resistor ladder with a divider at the end and no filter capacitor (maybe the covox speaker have it onboard, maybe not). CVX-4 lets you choose if you want a filter cap from several values, if you want to remove the DC bias or not and even if you want the divider or not. AFAIK if you match the original covox settings the result is 99% identical ... sadly I don't have the real thing to compare directly. You can hear some mod playback in the previous page - recorder through an Auzentech X-Fi Forte (had to apply the amplify effect on audacity though since the card records with a low volume from older sources...)

If you want any extra recording I will provide it 😀

I'm a bit surprised by the price, 3/4+ of dreamblaster's price is something I wasn't expecting for an lpt dac.

keropi wrote:

- anyone correct me if I am wrong - but we determined the original uses 100K 0.1% resistor ladder with a divider at the end and no filter capacitor (maybe the covox speaker have it onboard, maybe not).

I will clarify a bit; yes the ladder is made with 100K and 200K resistances and there is a 25K resistance from ladder output to ground.
What we don't know is the original ladder tolerance as it is a custom resistor array part, but most likely due to manufacturing the resistances and their ratios are very closely matched between each other, so maybe the tolerance between resistances is even better than 0.1%, while the absolute values of the tolerance might be 1% or 2% as with standard resistor arrays.

So, for a 5V parallel port, this will give out 1V output signal (when unloaded), with output impedance of 20Kohm which is pretty high so even audio cable capacitance acts as a low-pass filter.
And as soon as it is connected to something that loads the output like sound card or amplifier input, the amplitude will drop.

Jepael wrote:

I will clarify a bit; yes the ladder is made with 100K and 200K resistances and there is a 25K resistance from ladder output to ground.

Hi thanks all, for your support. The CVX4 has 100K 0.1% resistors instead of 100k + 200k resistors, this means all resistors are from the same reel so expected to be very close.
It is a small batch and I tried to make it the best and most featured. I cannot match the price of the cheapest clone, so I hope the good sound makes up for it.

Idea, but not sure if it would be worth the hassle : if someone outside Europe (for example in the US) wants to organize a group purchase, I could ship to him in one larger packet (10+ units), at lower unit price, without VAT. PM me if you want to do this.

a look at the geode GXm 233 Mhz thin client pc running cronologia demo with CVX4
https://youtu.be/8DWWFT0zFbE?t=1m5s
around 1m12 i show the pc and the CVX4 plugged into it

It is this pc : http://www.ebay.com/itm/272625508592
CPU: National Semiconductor Geode GXm CPU, PGA Socket 7, 233 Mhz, Enhanced MMX
Chipset: NS Geode CS5530A

The geode board is quite glitchy on the VGA visuals and sound, probably it is still too fast
for these 286 era demos.

The 533 Mhz Celeron 'Pc Punaise' works much better for demos
Here's the overload demo, played through CVX4 on this pc : https://youtu.be/WhvBbs3OC5U

why not make a Chip DAC ?

matze79 wrote:

why not make a Chip DAC ?

Because it needs external power?

There are options for harvesting data off the parallel pins. A powered parallel port DAC would make a very interesting project.

gdjacobs wrote:

There are options for harvesting data off the parallel pins. A powered parallel port DAC would make a very interesting project.

What would the benefit be over a resistor pack or discrete ladder?

Scali wrote:

What would the benefit be over a resistor pack or discrete ladder?

The simplest upgrade would be an opamp buffer on the output which would stabilize the performance of the DAC and present consistent output levels regardless of the characteristics of any downstream input you use. You wouldn't want to do much more, though, in order to maintain low power requirements for the device.

I will make a Active Covox with Chip DAC and simple Amplifier when i'm done with this PS/2 to Microsoft Serialmouse Adapter Converter / Updated First Post / Firmware Update added ...
Parts ~6-7 €

gdjacobs wrote:

The simplest upgrade would be an opamp buffer on the output which would stabilize the performance of the DAC and present consistent output levels regardless of the characteristics of any downstream input you use. You wouldn't want to do much more, though, in order to maintain low power requirements for the device.

But the output of the DAC would still be too low to drive a line-out, headphones or speakers, right?
In which case you'd need another amp for that.
But if you were to connect the resistor ladder to such an amp, you'd get the same advantages of buffered output and stabilized/consistent output anyway.

and you also need additonal power, because not all lpt's are have enough left to power this solution.
maybe you can get some power from status line etc with low schottky diodes..

EDIT:
http://320volt.com/en/paralel-port-dijital-an … -cikis-tda7050/

matze79 wrote:

and you also need additonal power, because not all lpt's are have enough left to power this solution. maybe you can get some pow […]
Show full quote

and you also need additonal power, because not all lpt's are have enough left to power this solution.
maybe you can get some power from status line etc with low schottky diodes..

EDIT:
http://320volt.com/en/paralel-port-dijital-an … -cikis-tda7050/

Well, I'd go for external power. Just put a connector on it. Then you can wire it up however you like: A mains adapter, tapping +5v from a COM port, tapping +5v from some USB, or from a floppy connector or whatnot.
Getting +5v from a PC shouldn't be too difficult 😀

Scali wrote:

But the output of the DAC would still be too low to drive a line-out, headphones or speakers, right?
In which case you'd need another amp for that.
But if you were to connect the resistor ladder to such an amp, you'd get the same advantages of buffered output and stabilized/consistent output anyway.

Certainly not speakers, it would be for a line in or perhaps headphones (with a high enough impedance). Much would depend on how much power you can harvest from the LPT lines. Of course, sourcing external power removes a lot of complexity in engineering something like this.

In an ideal world, you could depend on any downstream amplifier input being deterministic, but impedance characteristics just aren't the same between designs.

gdjacobs wrote:

In an ideal world, you could depend on any downstream amplifier input being deterministic, but impedance characteristics just aren't the same between designs.

Yea, but my point has always been to integrate the amp on the PCB, so it becomes part of the design.
Because imho, the Covox is pretty useless by itself, because the output is too low for any of the normal things you'd want to connect to it (headphones, line-in of some amp, unpowered speakers... powered speakers may work, but might be too noisy and not very loud, because the input is much lower than what they normally expect).
I would much prefer it to have a 'normal' output level, like any common sound card.
The real Covox solved this by having custom-designed powered speakers, so they modified the pre-amp in the speakers to work with the low output of the Covox.
Many clones used lower resistance to get more gain on the line-out, but as we found, you trade in linearity in the DAC, so the sound quality suffers considerably.