Wilba

sammichSID Final Batch! After 350 kits sold (including 556 SIDs), I have finally got to the end of the waiting list. Woot! There will be a final batch of 50 kits in November. I don't have plans to keep lots of pre-packed kits in stock, therefore, once the last batch is sold, YOU CANT HAZ SAMMICHSID! There are still plenty of kits available in the final batch. Yes, this is some shameless self-promotion to boost sales so I can finish with my routine 50 kit batch and sell off all the PCBs and parts I have in stock. If you're interested in getting a sammichSID, or want to find out WTF this thing is, all the information is here: http://www.midibox.org/dokuwiki/sammichsid

Also... you can get your 5V DC for testing from the sammichSID itself... use J2 in the bottom right corner on the base PCB, or even J2 on the control surface PCB (Vd=5V, Vs=ground).

The LEDs on sammichSID are multiplexed... they are in an 8x6 LED matrix, but driven by the PIC using the default MIDIbox SID 8x8 LED matrix code, so in reality they are really in a 1/8th duty cycle. In my experience, if you are testing what resistor will work, testing one LED with constant 5V will be about the same brightness as when it is in the matrix, it won't be much less bright when put in the matrix. It depends on the LED type, though. But as a rule, it's OK to pick a resistor using constant 5V as a test. Alternately, solder the LEDs before soldering the resistors, run the sammichSID CS test app (I really should put this somewhere better): and then touch resistors in the pads to see actual brightness. You could in theory put all 8 resistors in the pads (without soldering) and trim/bend the leads to make good contact. As a general rule, ultrabright LEDs (especially white or blue) will need at least 1K resistors... perhaps even as much as 3K.

It could be PSU related, but also, 6581 SIDs are notoriously noisy. How exactly are you powering the SID module (i.e. circuit)? I'm assuming you are using an 78L12 voltage regulator on the module (for 6581 SIDs). What is the voltage going into the 78L12? and what voltage going out? (should be 12V). Test these on startup, when you are experiencing low volume... maybe you can observe the voltage being lower than expected. Have you connected the audio in to ground on the SID module? This is a common cause of noise, if you're not using it, ground it! If you haven't done that already, do that, and record new audio clips.

If I was still selling SIDs, the current price would be 260 AUD (~255 USD). It's also built by an old hand at soldering so it's a real bargain. Did you paint the engraving yet?

Damn! Missed opportunity for a beer donation... oh well... buy nILS one instead. :thumbsup:

A 30K pot won't help, since you're using the pot as a voltage divider, the range would still be 1.5V - 3.5V (assuming 0V-5V, 40% range at middle).... I'm guessing, you might have one that is only the starting 40% so the voltage is 0V - 2V. Converting 40% of a 10 bit value into 7 bits would work, but you'd be bypassing the built-in jitter handling. It's much better to use 0V-5V inputs into the Core. You could try shifting and amplifying the voltage, but that would involve opamps. I'm not the best person to explain how that would be done... but in theory you could do it with a rail-to-rail opamp... assuming the middle of the range is 2.5V, it could amplify 1.5V-3.5V to 0V-5V. You could also try modifying the pot so the full range of resistance is only in the middle 40%. That would involve opening the pot and "shorting" the resistant strip at each end... you could use a silver conductive pen (eg. this: http://www.amazon.com/CircuitWriterTM-Precision-Pen-silver-based-grams/dp/B0002BBVQO ). You would have to carefully measure the pot's resistance at the ends of the modwheel's range, then with a multimeter, work out where that corresponds on the resistant strip, and "paint" from those points to the ends of the strip. Again I'm assuming it's the middle 40%, if it's only the starting 40%, you only need to paint one end.

First, make sure you are not using a heatsink that is connected between the voltage regulators, as this will not work for PSU Option B (the voltage regulators have different "grounds"). Uploading should take less than a minute. Something is wrong here, so you should remove the SIDs and solve the uploading problem, as well as the voltage problem. Even if uploading did work, you say the 12V supply is not correct, so 6581s will not be making any sound anyway, even assuming you have set the jumpers so they were connected to the 12V supply (another possible point of failure). You cannot expect to get anything audio-related working until you resolve issues with voltages, check the PIC is working, check MIDI In/Out is working correctly (J11 jumper set correctly, for example), get uploads to take only a minute, with "use feedback from core" checked (if it's the old MIOS Studio).

What I was suggesting as a "switch or loopback cable" was... 1. setup keyboard controller with MIDI Out port 2. setup MB-6582 PCB with MIDI In port 3A. use short MIDI cable between the keyboard MIDI Out and MB-6582 MIDI In, remove if you want external control or 3B. use a DPDT switch to internally connect the two "active" signals of the keyboard MIDI Out with MB-6582 MIDI In, in between the MB-6582's MIDI In socket and the MB-6582's MIDI In circuitry. i.e. the switch controls whether MB-6582 gets MIDI from a MIDI In socket, or internally from the keyboard's MIDI Out. (this is why I suggest DPDT, this way you can leave a cable connected to the MIDI In socket and it will be ignored when the switch is set to "keyboard"). Thus the switch acts like a virtual MIDI cable, and is very easy to implement. Note that both 3A and 3B mean you cannot have simultaneous keyboard AND external MIDI control of the MB-6582. If you want that, then yes, you should use a MIDI Merger like you describe... but you could ALSO could do the same thing by connecting the keyboard's MIDI Out to something else which routes it back to the MB-6582. Example: Keyboard MIDI Out connected to PC, PC MIDI Out connected to MB-6582 MIDI In, PC will "pass through" keyboard events to MB-6582 on channel 1 (perhaps), while also outputting on other channels via sequencer etc. My point in telling you all this is so you know you don't really NEED a MIDI Merger unless you really need simultaneous control of MB-6582 via keyboard and some MIDI hardware.

send to me for testing

Congratulations! Good to see another Aussie finish one... also shows how slack I've been, I still haven't finished the 2nd one like this (the 1st is in Munich now).

ahh... good to hear you worked it out. If you're in the mood, try replacing 22nF caps with something bigger, up to 100nF, this will reduce the max cutoff frequency but it will be better resolution (less steppy) at the bottom end... i.e. good for bass patches.

If you really want an 8xSID beast with SSM filters all in the one case, you should consider a rack-mount case (maybe 3U or 4U) and giving yourself plenty of room for all the modules and a bipolar PSU. Don't let the MB-6582 design constrain you too much... the base PCB can be put in something bigger than a PT-10 and you can make your own control surface on veroboard. It's not that hard to design and build a control surface better than the one I did... which was itself constrained to making it fit a PT-10, as there were no other DIY-friendly case solutions at the time. Or consider using the MB-6582 CS and sticking the lot in a bigger case. i.e. here are some random examples: CS5 by JuhaKauppinen, on Flickr MB-6582 frontpanel testing on enclosure by JoeLMutantE, on Flickr

I guess this means I have to pull my finger out and get a batch of those power switch/socket PCBs made up... :) Would it be more sensible to have them shipped with the cases? I like any idea that saves me from more packing and customs forms :frantics:

What version of MIDIbox SID are you using? I vaguely recall there was an issue with Multi engine and control of the filter... but I got TK to fix it a long time ago. It was an issue with how the filter cutoff depth for a SID was the sum of the filter cutoff depth for each single osc "voice"... thus each voice played in isolation could only get up to 1/3 the maximum depth. I may be remembering it wrong... I'll have to dig through my old emails. Perhaps this is related somehow... what's hard to determine is whether you are observing a difference in SIDs, a firmware bug or a hardware fault. I'll have a play on my MB-6582 and see if I can get a difference between a single-osc Lead patch and a single-osc Multi patch.

Also, I've never heard SIDs have any self-oscillating filter squeal unless I used feedback pots (audio out -> audio in, with filter "Ext In" param enabled). Are you using feedback pots? Are you using the typical filter capacitors (470pF for 6581, 22nF for 8580/6582)?

One thing that might be related to your issue... the 6581 and 8580/6582 have different filters (obviously), and also there's a difference in how the cutoff frequency registers work. TK has made allowances for this by making the filter range adjustable in the Ensemble. The default for all Ensembles is related to whichever version of the firmware was present when the Ensemble "bankstick" was formatted... so if you switch SID types, the Ensemble might have filter settings which don't match the SID. So check those settings and see if it makes a difference, if not, I'll try to help diagnose your problem some more.

EPIC LULZ THX GEEZERS I NEEDED A LAFF YOU CAN HAZ 8 SIDS

Yes. But I am busy packing kits, busy doing other things, and lazy.

I've started using these: http://www.sureelectronics.net/goods.php?id=1033 It's fairly standard pinout and dimensions. When people are requested to confirm their order for a batch, I suggest this one as an alternative: http://character-lcd-lcds.shopeio.com/inventory/details.asp?id=935&cat=Lcds&sub=Character%20Lcd You will need to check dimensions AND the header pinout AND the header position. i.e. I have bought one LCD off eBay that had the backlight pins 15 and 16 swapped. Also the Optrex STEP displays are awesome, but their header pinout has the backlight pins at the other end of the header, so 14 pins correctly align with the sammichSID base, but the backlight pins don't, so you need to do a bit of a fix. I can send you details if you plan to get one. Advice to everyone: let me check the LCD you plan to use.

Sorry it took so long, I should have suggested a short with LCD much earlier. You're not the first to have this problem magically disappear when I suggest detaching the LCD :wink:

The LCD could be shorting with the "Select 5" switch. Check that first. Then check the attached picture showing what is connected to that pin. Some vias are overlapped by encoders. It's unlikely but possible that an encoder is touching that via. AFAIK all the encoders I've seen have plastic feet so the metal base is not touching the PCB, but it's worth checking anyway.