jaytee

Ok, took it out of the breadboard, soldered everything up, double checked my connections, fired it up....nothing. U solder some wires to try out 4-bit mode, nothing. But at least now it's a consistent nothing rather than constantly changing gibberish, which indicates to me that the breadboard *was* an issue, just not the only issue. So the "nothing" I'm getting is two full lines of completely blacked out characters. Contrast knob does nothing. Backlight turns on fine, but as before, brightness knob does nothing. Checked the input voltages and they come out to 5V as expected. Not sure where else to start checking. any ideas? edit: just found the LCD connection test app. trying it now... double edit: I ran the app, everything checked out, felt *really* stumped, then realized I've have it connected backwards the entire time somehow, even though I double and triple checked before hooking it up every single time. WTF.

Already have your CS thread bookmarked! ;) It's one of the best pieces of documentation for the MB-6582 around (and I almost never found it because it's hidden in some other forum!). I'm 90% sure I got the cable mapping right (using both your guide and my datasheet), but I'll double check as I solder it.

I actually looked into OLEDs, but was mostly put off by the fact that the MB-6582 is built specifically for LCDs in terms of depth. Seems like an OLED will be really sunken in or require a bit of jiggery to mount it on top of the CS PCB. I am curious if that Crystalfontz I linked to will look as nice in person as it does in the picture. It's got beautiful contrast and almost looks like an OLED. I am skeptical, but otoh, their other photos of their LCDs look pretty much how I'd expect LCDs to look... Regarding soldering jumpers for 8-bit mode. I saw and read through that thread, but I am still left not really understanding the *why* of it all. It seems like some folks leave the MB-6582 in 4-bit mode but build the LCD cables according to 8-bit mode instructions and things turn out fine. That still leaves me wondering if anything needs to be adjusted for using a 16x2 instead of the standard 20x2? Will the 16x2 automatically truncate those last 8 characters, or do I need to change the firmware to adjust for the smaller test display? I haven't even uploaded the MB SID app yet, just vanilla MIOS.

Good to know. Maybe the blurb in the wiki about using DXFs at your own risk should be revised? It definitely had me under the impression that I was gonna foul everything up by using an exported DXF. Still gotta figure out how to place the corner screw holes...

Isn't DXF just a vector file? Does it maintain accurate measurements in the conversion somehow? The MB-6582 warns against this. Sorry, I'm really new to anything nearing graphic design.

What's the best way to get accurate measurements for a front panel? Measure the PCB with calipers? Or can this information be extracted or gleaned somehow from the available FPD files? I am very keen on making my own panels but FPE is way too pricy. The other problem is that I don't really want to deal with the rigmarole involved in eliminating external screws (though props to Wilba for figuring that out), but the file that includes screw locations is a dead link now.

What's the best way to get accurate measurements for a front panel? Measure the PCB with calipers? Or can this information be extracted or gleaned somehow from the available FPD files? I am very keen on making my own panels but FPE is way too pricy.

I've read all the relevant documentation that I could find on LCD character displays—mostly this, this, this, and this, along with various sections of the MB SID documentation and dozens of forum posts—and I'm still left with a few questions... First off, where I am in my project. Questions are down below. I have my MB-6582 power section fully built. I went ahead and built all four cores (I was going to do one at a time, but after you solder in all the resistors, capacitors under sockets, sockets themselves, and headers, there's only like three more components per core to add). I checked all my voltages and everything checks out. I inserted my first PIC and successfully uploaded MIOS (though not the SID application yet). So everything seems good so far. I haven't ordered my 20x4 LCD yet (more on that later), but I have a 16x2 LCD handy (this one) and figured I would use that to test while I wait on more parts. The MB SID manual says that 16x2 is supported, just not recommended because of its size. I read the above documentation, get a cable all ready, connect it to the LCD, and...nothing. Fiddling with the contrast trimmer and the cabling, I manage to get a few weird things to display, but nothing that indicates things are hooked up properly. Now, I strongly suspect that it's either a mechanical issue with my connection (its hooked up somewhat haphazardly through a breadboard, rather than soldered) or a mix up with pin connections. But! before I go chasing phantom soldering issues, I am hoping to clear up a few questions I have, just to rule stuff out and gain a better understanding of how everything works. HERE ARE MY QUESTIONS First, does anything need to be changed in the code to support a given size of LCD? Like right now, I just have basic MIOS installed on my PIC, v1.9 as available as the top download link on this page; should I expect it to display startup information on any size (2x16 or 20x4, for instance) of standard LCD, or do I need to tell it what size to expect somehow? How about once I install the MB SID app? Second, 4-bit mode and 8-bit mode. My understanding is that the MB SID app wants to run LCDs in 4-bit mode. Some documentation indicates that a special cable wiring setup is required for this (seen here), but while trawling the forums for information, a lot of folks suggest not to worry about special wiring. Not really sure what's what here. Third, can anyone offer a wee bit of guidance wrt choosing an LCD? I was looking at Crystalfontz and this LCD looks particularly striking. Green on black! Looking at the specs, I'm 99.9% confident that this part is compatible with the MB-6582. Only thing holding me back is the price: $25 once you add shipping. Meanwhile, there are 20x4 LCDs from China on eBay for like $6. I understand that you get what you pay for, but I'm wondering if anyone can comment on where that extra $20 goes on a Crystalfontz. Nicer colors? Less likely to break? Better contrast? More responsive? All of the above? Last, my backlight trimmer doesn't seem to affect anything. Backlight is on and at constant brightness no matter how this trimmer is set. I think I'm missing something obvious here. Now, like I said, I understand that the simplest explanation is that I just don't have things hooked up securely or properly. I just figured this would be a good chance to stop and ask some questions that were bothering me (and probably only still bother me because things didn't work on the first try).

I feel like I've had this issue before as well. IIRC, I always just worked around it by assigning LFO depth to a spare envelope. I always assumed it was down to user error. :o

@Molom

Just wired this up this afternoon. Seems to work like a charm!

My understanding is that only the Lead Engine is supported by super-poly mode. So four voices only, but you get super flexible voices. You can do six voices in the Multi Engine, but those are paraphonic voices (only one filter per three oscillators), and as above, every three voices will switch from R to L. If you want absolute maximum polyphony, you could get 24 voices by using four instances of the same Multi Engine patch on different channels with some external MIDI trickery (software or hardware that can switch to the next channel every six notes), but I would suggest that if you need 24 voices, you're starting to get well outside the strengths of the SID chip.

This is a hardware limitation. Each SID has a separate output, directly routed to either L or R, so voices that only sound from one SID will be panned by necessity. Best way to adjust this is in a DAW or on your mixer. You could wire up a passive mixer inside your MBSID that allows you to adjust pan, but it's a lot of work for not a lot of payoff.

...

I ask a couple questions about a part and now I'm "obsessed." Getting the part intended by the designer, in this case, saves me from doing extra work: shopping through digikey's enormous selection of switches, checking datasheets to make sure it's a good fit for the available space, re-jiggering the back panel cutout. It's all minor stuff, but not as minor as spending two minutes to ask around a forum for spare parts (which paid off, I might add, as a friendly member sent me the part for only the cost of shipping). In other cases, the effort to use the intended part might *not* be worth it. For instance, I won't be using the C64-style DIN7 plug for power. I'll have to re-jigger the back panel to fit my panel-mount choice, but it saves me from having to assemble a power supply with a DIN plug. Do these tiny calculations of worthwhile effort make me obsessed? Perhaps. But I'm building myself a synthesizer; why shouldn't I make it to my very own specifications?

...holy shit!

I mean that half the community has started projects that never got off the ground in any meaningful way. I don't expect that pointing out the terms of the TAPR will result in any previous projects getting a public release. Rather, I just think that going forward, it's important that folks understand what license all this IP falls under, and that it does indicate that documentation should be shared.

The digital section is supposed to be cycle-for-cycle accurate once all is said and done. The analog emulation will be tougher, but the sound demos are very encouraging. The plus side is that you could theoretically "fix" any of your least favorite SID issues. Want to use the volume envelope without the infamous bug? No problem. Want to boost the resonance a bit to get a really nice acid bassline? Do it.

Hawkeye nailed it. Announcing a project to the community doesn't obligate someone to follow through; if that were the case, half the MB community would be deep in unfinished project debt! ;) Rather, it would just be nice if documentation were more freely shared, as the TAPR license indicates, instead of hoarded.

I don't have feedback pots installed, so I can't comment on those. On my sammichSID, the external-ins are tied to ground by default (via switching jack) and there's very little noise. As for the volume levels of patches, I can't really say either way without hearing an example....but it's certainly possible to set up patches to have very different volume levels...

As far as I know, the software is the same for different SID chips; the different hex files just specify certain default settings for the filters to save you the trouble of setting them. The Bankstick configuration shouldn't matter either. Patches designed for different SID revisions won't sound quite right on the "wrong" chip, but they're not incompatible. They can be assigned to any core from any Bankstick. Hopefully @TK. can come confirm what I'm telling you—only about 90% confident on my answers, having read the MB SID documentation a couple dozen times but not having finished a multi-core SID project yet—but I'm pretty sure you're overthinking things. I can definitely say that many folks have completed MB SID projects with mixed SIDs before, so it is 100% possible, no problem. edit: As for the difference between SID revisions... I personally prefer the 8580/6582. Between the additional waveforms, the noise reduction and the improved filter, it makes the most sense to me for building a musical instrument. That said, it seems like folks who grew up on C64 music prefer the 6581; I'll certainly admit that despite its flaws, the 6581 does have a really nice sound.

I don't think I'd build a new MIDIbox specifically to put these in, but I'll happily grab a full set to have when real replacement SIDs finally become impossible to find.

Thank you @ChinMuzik. I'm just waiting for the package to arrive now.

Nope, not even close. The proper switches are 90° mount.