Showing posts with label SID. Show all posts
Showing posts with label SID. Show all posts

Thursday, 25 February 2016

Commodore 64 Pseudo Stereo Dual SID Questions

This is an old post, preserved for reference.
The products and services mentioned within are no longer available.

Just a quick update on the Commodore 64 Psuedo Stereo Dual SID boards, see the previous article for more info and the buy it now links. This is to answer a selection of questions along the lines of 'will it work with...'.

Will it work in a Commodore 128?

Yes, it will. The Commodore 128 has a 6581 or 8580, driven in a similar way to a Commodore 64. When ordering, go for the normal height (i.e. raised up by one socket height) version.
The C128 does have a metal shield which will no longer fit if the board is installed, so you would need to cut out a section, or remove it. If you remove it, may be an idea to fit heatsinks on some of the chips which use to sink to the metal cover. The C128D and DCR should both be fine, as long as there is space enough for the board to fit.
Power isn't as easy to tap as the C64, the best place I found was on pin 1 of the 7812 regulator. Be careful when clipping on there to avoid touching the centre pin. Maybe add a drop of hot melt glue or bluetak to hold it in place and away from the other pins.

Will it work with a Commodore 64 Reloaded?

Sort of. The power input to the Reloaded boards is 12V. This is filtered and fed to the SID, or can be fed via a 3V zener diode to give 9V for an 8581 SID. The normal dual SID boards have onboard voltage regulators to generate 9V and / or 12V for an unregulated DC of 15V or more, which is present inside the original Commodore machines.
This will need a special build of the dual SID board without the regulators. There is a jumper on the Reloaded board which shorts out the 3V zener to select 9V or 12V. I think I can remove the jumper and connect a 2 wire lead there to tap off the 12V and 9V. This is currently untested, but it should work.

Will it work with a Nano SwinSID?

Yes. The Nano SwinSID can be used, but with a few caveats. Firstly, the board is a bit higher than a standard SID, so will cause height issues in a C64C. I could solder the Nano SwinSID directly to the board and reduce the height if required.
The sound level output seems to be lower than a normal SID, so bear that in mind; you may need to adjust the levels on whatever you connect to.

Can it be used as dual individual SIDs?

Yes. The default setting is both SIDs are fed the same inputs and generate pseudo stereo effect. But, remove the CS#2 jumper and connect to one of the decoded IO lines on the cartridge port and if your software supports it, you can drive the two SID independently.

If you have any further questions, let me know, or see the previous article for more info and the buy it now links.

Saturday, 28 November 2015

Commodore 64 Pseudo Stereo Dual SID boards

This is an old post, preserved for reference.
The products and services mentioned within are no longer available.

One of the strong points of the Commodore 64 is it's sound chip, the SID (Sound Interface Device). There are all sorts of things use can do with these, and an interesting one I saw recently in this video from GadgetUK164, was the idea of using two SID chips, wired in parallel, to generate a sort of stereo signal. Both chips would be driven by the same digital control signals, but the outputs of SID chips do tend to be a little different, and so create a sort of stereo effect. This different is more marked if you use the version from the C64, the 6581 on one side, and one from the later C64C / C128, the 8580, on the other side.
I was concerned with wiring them directly in parallel, as the SID can be both written to and read, so when a read request is made, both chips will be trying to write to the data bus, which is not a good thing. We had a discussion in the comments on that video, and I suggested using an OR gate to generate the chip select signal for the second SID, most easily achieved in this case using a diode OR. In the followup video, he showed the modified board, with diode OR and pull down resistor.
The inverse logic is sometimes tricky, the chip select line on the SID is active low, so needs to be low to be enabled. The read/write line is low when writing, high when reading. Here we only need the second SID to be selected when there is a write request, so CS2 is low when CS is low AND R/W is low. Or to think of that the other way around, CS2 is high if either CS is high OR R/W is high, hence the OR gate. A variation on DeMorgan's law.
Or as a truth table:
CS
R/W
CS2
SID1
SID2
0
0
0
Write
Write
0
1
1
Read
-
1
0
1
-
-
1
1
1
-
-

I've designed some boards which include that additional logic. The digital side of the SID remains all connected in parallel, apart from the CS line on the second SID. I've used a single gate 74HCT32 here to OR the signal. There are four parts to the analogue side. Firstly, and least important here, are the two analogue inputs, these are used by the paddle controllers on the C64. These are wired to the first chip, and there values are read when using read commands. When read commands come in, the second chip is not selected, so ignores them. The second chips analogue inputs are tied to ground. There is a jumper to also supply the second chips CS line from another source, such as one of the decoded IO address select lines on the cartridge port, for software which supports dual SIDs. Even switch between the two options should you so desire.
Second on the analogue side is the power supply. This is 12V for the older 6581 chip, and 9V for the later 8580. That confuses things a bit. The power is supplied from a 7812 regulator on the C64, and a zener diode / transistor regulator on the C64C. What I've done on these boards is ignore the supply voltage from the board, which may be 9V or 12V, and generate dedicated 9V or 12V rails for each chip with separate 78M09 and 78M12 regulators. The power for these comes from clipping onto the positive lead of reservoir capacitor C88 on the C64 board.
The C64C doesn't have that capactor easily accessible, but you can get a suitable supply by clipping onto the leg of CR3. All the chips I have tried draw around 60mA for a pair of SIDs, so this needs to be clipped onto anything that can supply 60mA at 15V-24V. When tapping power like this, it is always worth checking the voltages with a meter before using. It is also good to power up the system with the dual SID board installed and powered but no SID chips installed, and double check there is 12V on pin 28 of the 6581 socket and 9V on the 8580 socket.
Third on the analogue side are two filter capacitors for each chip. Different values are required for the 6581 and 8580, so again I ignore what is onboard and have individual capacitors for each chip. Some of the dual SID boards require you to clip off or change the ones onboard, but here I just don't connect to the onboard ones, so you can go back to a single SID if you want.
Finally comes the audio output. Here the output from SID 1 is fed through to the board as previously, so you get sound from the A/V port or TV output (is anyone still using TV output?). The outputs from both SIDs are also separately amplified and fed to a 3pin header for the stereo output. The audio output stages on the C64 and C64C were slightly different, the C64 had a capacitor bypassing the transistor from base to emitter, the C64C from emitter to ground. The board is laid out to take either arrangement, and I've been matching those to the chip, so the 6580 has the C64 version, 8581 has the C64C version. This further differentiates the stereo image.
SIDs are in short supply, and are used in various music gear outside the C64, so it is a bit luxurious to use two SIDs in one computer. However, you can use partly faulty SIDs here, I have one which locks up when you try to read from it, that is fine as SID2 here, as it will never be read. I also have a couple which have one or more voices missing. This makes the stereo a bit more varied, if a bit biased to one side or another.
The Psuedo stereo dual SID boards can be configured so each SID socket is setup for each a 6581 or 8580, independent on what was originally fitted to the Commdore 64. So you could even use this to fit a signle 6581 to a C64C or an 8580 to a C64, and leave the second socket empty.
As with the previous post on the C64 PLA replacements, the SID is in different locations depending on the revision of the board. In some cases, the rear connectors get in the way, so you may need to stack one or two sockets to raise it up to clear the socket. I'll address that on the next revision. There are also some capacitors around, just gently lean those out of the way, or again use riser sockets.
Keyboard clearance is sometimes tight, it varies with the model, the C64C is most difficult. A SwinSID may just fit, but the extra height is right on the edge. You may have to cut or remove the metallised cardboard cover.
I'm now off to listen to more game music in glorious (pseudo) stereo.
My old favourite Spellbound (by David Jones - not that one) has great music on the C64.
Any questions? See Commodore 64 Pseudo Stereo Dual SID Questions.

UPDATE:
There is now a new revision of this board. I have moved the second SID chip select to a jumper to the edge of the board so you can more easily wire this to one of the IO selects on the cart port, or even wire in a switch to have both options. That option is there on the previous version, but it's hidden under the chip.

UPDATE 2:
A couple of reviews:
ORDERING:
These are currently out of stock. If you are interested, see contact details above.

Thursday, 3 September 2015

Upcoming Commodore 64 projects

This is an old post, preserved for reference.
The products and services mentioned within are no longer available.

Here's a quick preview of a couple of upcoming Commodore 64 projects.

Commodore 64 pseudo stereo dual SID
This was inspired by a video from GadgetUK, where he modified a dual SID board to run two SIDs in parallel. Normally the dual SID board is used to add a second SID and a different address, to generate actual stereo from suitable software. The idea with this mod was to place two at the same address, so they could be driven from normal software. You may think that there is no point in doing that, they will both sound the same? Well, there is quite a variation in the sound produced by different SIDs, and the idea here is to exaggerate that by using a 6581 from a C64, and a much later 8580 from a C64C.
I raised the issue of a possible problem with two parallel SIDs when it comes to reading, as both devices will be selected, and both would try to write back to the databus. I suggested adding an OR gate to the chip select of the second device, so it would only be selected when the chip was selected and it was in read mode (or since it's all negative logic, it would NOT be selected if either the chip was NOT selected OR it was NOT in read mode). I said this could be implemented with a simple diode OR, and in the followup video, that seemed to be working nicely once the pull down resistor value was sorted out. To avoid any issues with that, I've used a single gate OR gate here, rather than a diode OR.
There are a few differences between the 6581 and the 8580. The power supply, filter capacitors and audio output amplifier circuit are all different. So what I've done here is to ignore the C64's filter caps and fit these onboard. I also ignore the power supply and have onboard 9V and 12V regulators, with the power to be tapped from the main power supply filter capacitor. I've left the audio from the main SID going to the C64, but both audio signals are also buffered onboard and fed to the audio output jack.

Commodore 64 PLA replacement with EPROM
Replacing the commonly faulty Commodore 64 PLA with an EPROM is something I tried a few years ago, with mixed results. I also tried to make a replacement using two GAL chips, but didn't get very far with that.
Following another video from GadgetUK, he seemed to have more success using the EPROM method and adding a 68pF capacitor to the /CASRAM line. The issue here is that due to what could be considered a design flaw in the C64, the /CASRAM line needs to be delayed by about 25nS to allow the addresses to be ready in the multiplexer before they are loaded into RAM.
I've updated my board to have space to add a suitable capacitor. These probably needs to be tuned to each board, depending on the type of RAM used.
To make this more repeatable, I've added a quad OR gate to add a configurable delay. Each gate has a propagation delay of around 9nS, so with the jumper you can select to tap the signal after 1, 2, 3 or 4 gate delays. This will add 9nS, 18nS, 27nS or 36nS delay to the signal .These timings are approximate, but are typical values. I can also shorten or extend this by using a different logic family chip.