VOGONS


Beyond a replica - the Covox Blaster

Topic actions

First post, by dionb

User metadata
Rank l33t++
Rank
l33t++

After a year or so of building other peoples' replica sound cards, the inevitable has happened - I've started one (or actually two) myself.

Given it's my first (and second) PCB project, I wanted to keep it simple. My inspiration came when it turned out my beloved Aztech NXPro wasn't as 100% Sound Blaster compatible as I thought and I needed to replace it with my Snark Barker SB1.0 clone. Works fine, but I lost Covox support. And Covox is arguably the simplest sound device out there. The easy solution would be to go for one of the many external dongles out there, but I hate dongles. And this machine has an internal parallel header. So an idea started to form...

The first iteration of the idea was simple: mount a small PCB with 3.5mm jack and the resistor ladder on an external bracket and connect that with a flatcable to the parallel header. Thus the design I'll call Internal Covox was born. I was inspired by Serdaco's CVX4, both in the use of identical resistors to minimize variation and in the addition of a dipswitch to enable different filter/volume options. Note that I didn't copy any of his schematics, I just went to the Speech Thing patent with its very convenient detail and used that. Any errors are entirely my own, and errors there certainly are. I quickly cobbled together a design for this and shot of a PCB order and then started to be confronted by my many mistakes. First up I should have used KiCad right away, instead I used EasyEDA. It seems very user-friendly, but not having as much focus on footprints leads to mistakes (at least in an absolute beginner). It should - hopefully - still work, but even buying the smallest resistors I can find they won't fit properly and the size means I have to use 1% instead of 0.1% resistors. Not good - but in a week or so we'll find out how not good.

Other things aren't mistakes as such, but realizations after the fact. The PCB is cheap as dirt, but two parts are pricey: the Keystone 9202 ISA bracket and the (connectors for the) flatcable. Also, having *yet another* flatcable draping through the system is very '90s style, but not my preferred situation. So how could I avoid cables and the need to buy a bracket... enter the Covox Blaster!

Covox_Blaster.png
Filename
Covox_Blaster.png
File size
162.16 KiB
Views
929 views
File license
GPL-2.0-or-later

(note: male header and IC at SW1 are incorrect, but holes are)

What is it?
A Covox resistor ladder DAC on a daughterboard that can be plugged straight into an internal parallel header. The other end has a standard MPC-2 'CD Audio' connector that can be fed into a sound card, or easily connected to something else, like a 3.5mm jack socket at a convenient location. By using either a straight or right-angle female 2x13 connector it can interface with a header on a card or on a motherboard respectively. The dipswitches let you mess around with the filter, with capacitor values from 0.1nf to 10nf. The resistor in the patent is supposedly 15k but it has been measured at 25k, so I let you select both (and a few higher values too, using 220k base and resp 57k, 57k and 37k on the dipswitch - all together in parallel they add up to 15k)

Everything uses THT, mainly because I don't want my soldering skills (and equipment) to be the weakest link. Because it's still electrically connected to a parallel port, compatibility should be identical to the original. I realize I/O cards with internal parallel headers are uncommon, but they certainly exist - and there are LOT of motherboards out there with the headers. Still, the image I have in my mind is a big I/O card with parallel header with daughterboard, connected to a sound card with daughterboard too. Practically useful? Nope. But why not? 😜

So, what do people think?

Reply 5 of 24, by Benedikt

User metadata
Rank Oldbie
Rank
Oldbie

If it's internal, already, you could add amplification, as well.
The lack of an easily accessible power supply is the biggest shortcoming that external dongles have to deal with.
With an internal device, however, getting 5V somewhere is trivial.

The SoundJr happens to work well without dedicated power supply, but has to jump through a lot of hoops to get there.

Reply 6 of 24, by dionb

User metadata
Rank l33t++
Rank
l33t++

After making a cardboard mock-up and playing around with some cards and motherboards I've come up with one simple improvement: rotate the 4p header by 90 degrees so the MPC-2 cable goes in by the side. Doesn't make much difference with motherboards, but if mounted on a card, the cable goes straight down towards the motherboard, and clearance can be low as it is (card is 7.7cm long). Depending on the results with the extra resistors and caps I might decide to drop a few of those and shrink or get rid of the dipswitch. That might save quite a few mm.

Good news: my Mouser components order just arrived, so as soon as the PCBs get here I'm good to go 😀

Benedikt wrote on 2020-06-08, 08:51:
If it's internal, already, you could add amplification, as well. The lack of an easily accessible power supply is the biggest sh […]
Show full quote

If it's internal, already, you could add amplification, as well.
The lack of an easily accessible power supply is the biggest shortcoming that external dongles have to deal with.
With an internal device, however, getting 5V somewhere is trivial.

The SoundJr happens to work well without dedicated power supply, but has to jump through a lot of hoops to get there.

Hmm, will look into that, at least if levels are low. If not I'd prefer not to - extra cost and complexity on the PCB is trivial (LM368N or similar would do the job I assume), but getting 5V may be trivial but means another cable and more mess in the system, which is the one thing I wanted to avoid in the first place. If I can get a 'line' level that plays nice with CD-in on a sound card it's good enough for me.

Last edited by dionb on 2020-06-08, 09:41. Edited 1 time in total.

Reply 7 of 24, by Tiido

User metadata
Rank l33t
Rank
l33t

You could save substantial space by usng 1/8W or even 1/16W resistors, they will be a lot smaller while not being any more difficult to solder compared to the bigger ones you use right now.

I like the idea of being able to plug straight into the LPT header ~

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 8 of 24, by dionb

User metadata
Rank l33t++
Rank
l33t++
Tiido wrote on 2020-06-08, 09:27:

You could save substantial space by usng 1/8W or even 1/16W resistors, they will be a lot smaller while not being any more difficult to solder compared to the bigger ones you use right now.

In fact I originally used smaller 0.4W resistors, but I couldn't find any with 0.1W tolerance at a price I was happy with. Mouser has 0.4W 100k 0.1% for EUR 0.423 each (dropping to 0.32 with greater quantities), the cheapest smaller option - 1/20W 100k 0.1% increases to EUR 1.07 (or 0.804 for higher quantities), so well over twice the price.

Mouser's not the cheapest or my first preference, I usually order via Reichelt, but situation's the same - big resistors are affordable (EUR 0.35 each), smaller ones less so. Given that there are 23 of these little monsters, already making 2/3 of the total costs (even including shipping costs for PCB and parts), unless you know a good, cheap source of little ones, I'm sticking with the big cheap(ish) ones 😉

I like the idea of being able to plug straight into the LPT header ~

😀

Reply 9 of 24, by Tiido

User metadata
Rank l33t
Rank
l33t

I didn't know there was that big of a price difference, but I haven't actually looked at non-surface mounted stuff expect for high power handling reasons and not for anything needing precision performance 🤣. (I get majority of my stuff from Mouser too, DDP is great !)

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 11 of 24, by dionb

User metadata
Rank l33t++
Rank
l33t++
Tiido wrote on 2020-06-08, 10:23:

I didn't know there was that big of a price difference, but I haven't actually looked at non-surface mounted stuff expect for high power handling reasons and not for anything needing precision performance 🤣. (I get majority of my stuff from Mouser too, DDP is great !)

One day I'm going to get my SMD game in order, but for now the added advantage of doing it this way is that if I can solder it, anyone can 😜

Also it feels a bit more 'authentic' late 1986 with big barrels, but I'll be the first to admit that wasn't the main reason.

Reply 12 of 24, by matze79

User metadata
Rank l33t
Rank
l33t

1% does not sound bad at all, at least if all are from the same batch.
It's more about the tolerance among all of them.

Is your Pin Header compatible with CD-Audio In ? 😀

https://www.retrokits.de - blog, retro projects, hdd clicker, diy soundcards etc
https://www.retroianer.de - german retro computer board

Reply 14 of 24, by dionb

User metadata
Rank l33t++
Rank
l33t++
matze79 wrote on 2020-06-08, 16:31:

1% does not sound bad at all, at least if all are from the same batch.
It's more about the tolerance among all of them.

True, but they should be tighter too with tighter overall tolerance. My first prototype will be with 1%, we'll see (hear) how that works.

Is your Pin Header compatible with CD-Audio In ? 😀

That's the idea 😀

That means you can also use an ISA CD card for audio out if you don't have a suitable sound card.

Reply 15 of 24, by imi

User metadata
Rank l33t
Rank
l33t

I've recently built an audio summing cable and only had 1% resistors on hand, and while they both were off by about 0.6-0.7% from the stated value (I only measured after making the cable) they both had the exact same value as far as my multimeter goes.

Reply 16 of 24, by dionb

User metadata
Rank l33t++
Rank
l33t++
imi wrote on 2020-06-09, 08:34:

I've recently built an audio summing cable and only had 1% resistors on hand, and while they both were off by about 0.6-0.7% from the stated value (I only measured after making the cable) they both had the exact same value as far as my multimeter goes.

Well, we'll see. First prototypes will be with 1% resistors, all straight off one ammo belt. Afterwards I'll do at least one with 0.1% resistors. It'll be interesting to hear the (lack of) difference. I'll also measure them as well as I can with my crappy multimeter.

Reply 17 of 24, by johnnycontrario

User metadata
Rank Newbie
Rank
Newbie
dionb wrote on 2020-06-09, 13:58:

Well, we'll see. First prototypes will be with 1% resistors, all straight off one ammo belt. Afterwards I'll do at least one with 0.1% resistors. It'll be interesting to hear the (lack of) difference. I'll also measure them as well as I can with my crappy multimeter.

Have you considered using a resistor network instead of using discrete resistors? If you can find one with the right values, it could be a significant cost, time, and space savings.
https://www.mouser.com/datasheet/2/54/r2r-1018811.pdf

Reply 18 of 24, by matze79

User metadata
Rank l33t
Rank
l33t

You also could hook up 9 Pin Joystick Port to LPT 😀
Adding a additional Header and few Diodes, here you go.
As Covox is using only one Port, the other free pins can serve Digital Joystick.

https://www.retrokits.de - blog, retro projects, hdd clicker, diy soundcards etc
https://www.retroianer.de - german retro computer board

Reply 19 of 24, by dionb

User metadata
Rank l33t++
Rank
l33t++
johnnycontrario wrote on 2020-06-09, 17:37:
dionb wrote on 2020-06-09, 13:58:

Well, we'll see. First prototypes will be with 1% resistors, all straight off one ammo belt. Afterwards I'll do at least one with 0.1% resistors. It'll be interesting to hear the (lack of) difference. I'll also measure them as well as I can with my crappy multimeter.

Have you considered using a resistor network instead of using discrete resistors? If you can find one with the right values, it could be a significant cost, time, and space savings.
https://www.mouser.com/datasheet/2/54/r2r-1018811.pdf

Interesting. There is one with exactly the right values (damn, Bourns really, really needs to make their site more informative and intuitive), 4116R-R2R-104(LF - with a factor 3 premium for no lead. Still a lot cheaper though). Tolerance is 2% though, which looks a lot less attractive than 1% or 0.1%.

This would allow an extremely minimalistic design - assuming I find one filter resistor and cap value I'm happy with, the entire thing would just consist of a 26p F header, this DIP-16 IC, one cap, one resistor and a 4-pin audio header. PCB could be tiny. If I can get this RN locally (which doesn't look likely) or I do an order big enough to make Mouser an interesting option, I'll get one and try it in a breadboard 😀

matze79 wrote on 2020-06-09, 18:09:

You also could hook up 9 Pin Joystick Port to LPT 😀
Adding a additional Header and few Diodes, here you go.
As Covox is using only one Port, the other free pins can serve Digital Joystick.

Er, yes, theoretically possible. However for the same reason I don't want to run in 5V to the board (more cables), this also sounds iffy. It might be an idea for the other bracket-mounted design, although it would need a custom bracket most likely. Metalwork isn't exactly my thing. I think I'll pass for now. Out of interest, how do you get software to use a digital joystick hooked up to the parallel port?