shock__ wrote:
If one of you could gather all info and send it via PM or E-Mail that would be largely appreciated as I'm currently only able to follow the thread on my smartphone.
I've just gone through the last 20+ pages to sum up the hardware-related feedback/issues/suggestions/ideas, so please find below the summary (with my personal comments in brackets):
JP4-JP5: ARGUS Project thread - Now with gerbers and documentation in initial post!
640K!enough wrote:
This relates specifically to JP4 and JP5. These are used to connect RA20 and RA21 to the ROM, as well as provide mode selection functionality at the trailing edge of RESET#. They are bi-directional pins, yet we are talking about bridging them to 5V or ground. It seems that this would be an invitation to undefined behaviour, instability or eventual damage.
As it is, JP4 should be easy enough to deal with: just use a 4.7K (or 10K?) resistor instead of bridging the pads. JP5 is a little more complicated, due to the differences between the requirements of a standard SOP44 flash ROM and the InterWave. For those using the original masked ROM or who plan to use a ROM module, the solution isn't too bad: use a resistor, as just mentioned, to connect pins 3 and 4, then connect a bodge wire from pin 2 to pin 4. If you're using the 29F800, use a resistor for pins 3-4, and connect 1-2 however you like (but a resistor probably wouldn't be a bad idea). If you've adopted a different SOP44 chip, then it will depend on the requirements of that particular chip.
It's worth mentioning that it may be inadvisable to follow these instructions before we have had a chance to discuss them further.
Socketed OPAMPs: ARGUS Project thread - Now with gerbers and documentation in initial post!
cyclone3d wrote:
What about socketed opamps so we could change them out if we want to?
Space between C725, C718, C721: ARGUS Project thread - Now with gerbers and documentation in initial post!; ARGUS Project thread - Now with gerbers and documentation in initial post!
keropi wrote:Capacitors C725 C718 C721 are too close together and C718 does not have enough space, had to use another taller one with a smaller diameter to have them fit OK. Sure this is dependent on the capacitors used but a little more space would be more convenient.
Phreeze wrote:have you also considered giving the caps a little bit more space ? around C718
Additional Audio-In: ARGUS Project thread - Now with gerbers and documentation in initial post!; ARGUS Project thread - Now with gerbers and documentation in initial post!
matze79 wrote:
i wanted to ask if you can add another Audio In with a Header.
So we can Design Addon Board which will connect to the ARGUS ISA Bus Header, maybe a OPL3 Board or OPL2, Tandy Audio Board etc.
matze79 wrote:thats why i asked for a additional audio-in header 😀
Adding OPL2/3 and maybe SN76489 etc to a daugtherboard, mixed into the gus.
(InterWave datasheet mentions 4 stereo inputs: MICL/MICR, LINEINL/LINEINR, AUX1L/AUX1R, AUX2L/AUX2R; as far as I can guess, the former two are linked to the MIC-IN and LINE-IN on the bracket, and the latter two can be found as two onboard headers, MPC and P45, thus all available inputs are in use already - please correct me if I'm wrong; however if that's the case, still the pinheaders for the bracket audio ports could be added in order to enable their use by any internal device like a daughterboard)
640K!enough posts with 7 suggestions: ARGUS Project thread - Now with gerbers and documentation in initial post!; ARGUS Project thread - Now with gerbers and documentation in initial post!
640K!enough wrote:I had a little more time to test the card, and have additional comments to make. I will include comments about possible design […]
Show full quote
I had a little more time to test the card, and have additional comments to make. I will include comments about possible design changes as well:
- Some through-hole parts needed a higher soldering temperature than I generally use to get a neat joint. It may be worth adjusting the thermal relief settings slightly, so that these parts are easier to solder at more reasonable temperatures.
- X3 has no allowances for the board guides or other supports that seem fairly common on SMD receptacles (I had to snap them off). Given that it is such a small receptacle, and will be connected as part of a larger module, one has to be exceptionally careful when mating and de-mating the module to avoid damage. If possible, holes for board guides should be considered, or change to a through-hole receptacle, which should provide greater strength. The other receptacles fitted to the ROM footprints don't seem to present as much of a problem. Another option would be to add the additional pins to the end of the ROM footprint, as was done for the SOJ DRAMs.
- The prescribed configurations for JP4 and JP5 are somewhat problematic, especially when using just solder blobs or 0Ω resistors/jumpers. They don't seem to cause trouble, as long as the ROM banks have a capacity of 1 MiB or less; with 2 or 4 MiB banks, one invariably ends up with garbled playback, because the voltage levels (or shorts!) present on those lines corrupt the addresses. Some re-working of this section would be advisable if support for larger ROM banks is desired for the final board.
- Similar to item 2, we may have trouble if the SIMM sockets we source once shock__ stops supplying them are significantly different. To use the upright socket, I had to clip off the board guides to get it to fit. I should have also filed off the remaining bumps, as it doesn't sit perfectly flat now.
For item 3 specifically, there is one approach that seems to fix the problem. For JP5, connect pins 3 and 4 with a 10kΩ resistor (0603 is probably easier to work with), and use a wire to connect pin 2 to pin 4. Of course, if there is room, there are ways to adjust the design to avoid having to use a bodge wire.
Because we are feeding 5V onto RA20, the case with JP4 may not be so simple. I think we may be able to get away with just using a 10kΩ resistor, but I didn't test extensively. The better solution seems to be using an appropriate circuit to effectively cut off the flow of current following initialisation. This seems to be the way it is done on the GUS PnP version 1. I tested this with a 100kΩ resistor in series with a 2.2µF capacitor. This appears to work, though the values may not be ideal; that's what I had on hand.
640K!enough wrote:In my rush to post the previous comments, I forgot a few items, so I'll continue my list here: […]
Show full quote
In my rush to post the previous comments, I forgot a few items, so I'll continue my list here:
- The InterWave datasheet indicates that CFILT should have two capacitors: "The Capacitor Filter input must be connected to analog ground through a 0.1 μF capacitor and a 10 μF capacitor.". Currently, we only have C722. Just to follow the datasheet, I added a 0.1 µF electrolytic capacitor on the back of the card, soldered to the ends of the leads of C722. I don't know if that has any effect on functionality or audio quality, but for the cost of one capacitor, I see no reason not to have it, if only to follow specifications.
- The GUS PnP seems to feed 5V onto both RA20 and RA21, if I'm not mistaken, whereas the ARGUS has RA21 connected to ground. I don't think this is really worth any further discussion; we are just selecting unused signals for additional external hardware, rather than suspend fundtionality. This means that the contents of on-board DRAM may be lost during suspend, if I've understood correctly, but I have yet to see a design that provides the necessary 32 kHz clock signal to support DRAM refresh under suspend, anyhow.
- The current design defaults to PNP-compliant mode, which is fine for most systems that have Pentium and newer processors and a BIOS that is reasonably Plug-and-Play-compliant. Is it worth considering another set of jumper pads to default to PNP-system mode? This, in theory, should make it easier to use the card on older systems without additional software. I can't test this, but it should be possible to get the card working on 286 systems, and maybe even 8086/88-era systems.
Complementary spin-off projects:
Brackets: ARGUS Project thread - Now with gerbers and documentation in initial post!
(and yes, I confirm my intention to finance the initial production if needed; when you contact the contract manufacturer, could you please ask them for the available options regarding material, finish, engraving, etc.? if we are to produce a special bracket for ARGUS, and if we are to finance it, let it look really nice! at least not worse than those golden Terratec brackets or the famous black one from MX300)
ROM DB: ARGUS Project thread - Now with gerbers and documentation in initial post!
MPU DB: ARGUS Project thread - Now with gerbers and documentation in initial post!
(and yes, I confirm that I'm absolutely interested in this endeavour! regarding the wavetable header: there is a way out, you can design the header to be 90-degrees upward (and solder it "inwards", i.e. to the "bottom" side of the MPU DB), so that the possible waveblaster daughterboard could be connected in parallel with the motherboard, or via a 26-pin extender IDC cable)
If anyone notices that I've missed anything, please feel free to comment!
P.S.
shock__ wrote:Some news:
-------
-17:13:26- (ZrX-NoMs) Sent two InterWave dies to a US guy who's gonna image them using automated setup.
ZrX a […]
Show full quote
Some news:
-------
-17:13:26- (ZrX-NoMs) Sent two InterWave dies to a US guy who's gonna image them using automated setup.
ZrX also agreed on cracking open an IW78C21M1 to check for the tech used to confirm whether we were lucky or if the chips are actually factory programmed.
Any news on this? 😀
P.P.S.
I remember that earlier in this thread some people suggested some ARGUS-related merchandise, keyrings, "retail" boxes, etc. - if you ever decide to produce all those, count me in, please 😀