Wilba
-
Posts
3,310 -
Joined
-
Last visited
-
Days Won
2
Content Type
Profiles
Forums
Blogs
Gallery
Posts posted by Wilba
-
-
It should only do cloning when the menu button is down on startup. Also it shouldn't start on the oscillator screen. This implies it is sensing switches being pressed when they aren't, so something is wrong with your control surface, i.e. a short or something, either on the CS PCB or the connections to the base PCB, or even on the 74HC165 ICs. Check there for shorts. Also, did you install the resistor networks the right way, with the dot marking matching the end with the part number i.e. R30?
Also, it should clone up to slave 03 before restarting. Do you have all four PICs installed? Did you upload MIDIbox SID V2 firmware on each PIC? Check R80 is installed too (but I think it must be if it's doing any cloning).
You might want to try taking out the slave PICs until you sort out the control surface issues (i.e. get it to a point where the menu button isn't "stuck"), and then upload the firmware to each PIC separately in the far right IC socket before putting them in their right place. You won't need to test cloning until you have everything else working.
-
I've done the layout according to the schematic... the row of extra buttons at the bottom was pretty tricky, but I guess I complicated it by putting ludicrous amounts of mount holes (much like my MB-SEQ PCB) to give full flexibility in construction, i.e. mounted above to aluminium front panel, mounted below to base like sammichSID, plenty of holes for mounting a "spacer" like the Livid ones, or securing the rubber pads separately to the mounts, etc.
So as I see it, the most optimal solution is one PCB with 4x16 buttons + 4 on the left (i.e. just like the schematic). That layout also supports the extra 1x16 row + shift button, as either a separate PCB or as part of a 4x16 PCB. So in theory, if this went to a mini bulk order stage, you could structure it as two different PCBs, either with the extra 1x16 row attached or separate.
I chose this arrangement over the 8x8 because it simplifies the "scalar"-ness of the design... I couldn't find a nice way of making an 8x8 board stack vertically and horizontally without creating four different 8x8 layouts, each with different placements of the J3/J4 connectors, and then there would still be the extra 1x16 row to deal with. Also, the aim was implementing BLM16x16, not make a monome clone... which is easily done with the breakout boards already available by Livid (but not with bicolor LEDs though). I'm also recalling discussions with TK regarding this:
- we shouldn't consider too many options, which people probably will neveruse, e.g. RGB LEDs are too expensive, and a 8x8 momome-like matrix
is not really useful for a step sequencer
However, this 4x16 PCB should work as an original BLM 4x16 extension to MB-SEQ, a much more useful "reuse" than trying to be all things to everybody. Would the original BLM 4x16 extension to MB-SEQ benefit from the extra row of buttons at the bottom? Or would people prefer just the 4x16 plus extra 4 buttons (aka. Track Group select)? If the latter is true, then maybe it is worth putting another J3/J4 connector on the left so you could connect another BLM scalar driver board for the extra buttons.
-
I'm OK with it, although I would like to know who buys it so they can continue to get support from me.
-
I haven't found SPDT switches like you describe, and I've looked very hard.
They must exist, because some controllers like the Axis-64 use them, I think.
"Piano" type keyboards that have velocity sensitivity typically use two discrete switches... sometimes this is as simple as two rubber membrane switches that will close at different times of the keystroke, with different length actuators above it which are connected to the key.
The Axis-49 uses an interesting method of a rubber membrane with two conductive rings at different heights:
http://a2.vox.com/6a00cd972aa36b4cd5011015f5dc82860b-pi
http://musicscienceguy.vox.com/library/post/the-axis-49-reveled.html
I'm using a very time-consuming method to add extra contacts to Cherry MX keyswitches:
-
No, are you referring to the audio socket above the LED?
-
Looks good, I like the yellow.
If you are missing a switch, I will send you a one, or get nILS to send you one. Just email me your postal address.
-
Speak for yourself, I can't get any work done with all the noise.
-
That's the LED I was looking at. There is another one: http://cgi.ebay.com.au/ws/eBayISAPI.dll?ViewItem&item=280516196821 but I don't know the brightness.
BTW Don't forget to add the price of switch caps for the E-switch tactile switches, which are only available from Digikey AFAIK (part 401-1152-ND), another 48.12 USD for 289.
This is why I stopped working on a design using E-switch, it's still a lot to pay for switches/caps and they wouldn't be illuminated. TK's DIY approach is better if you want to make it cheaply, and have illuminated buttons. The Livid pads aren't ideal, they're quite tall compared to the ones used on monome, making the overall height of the box at least 45mm, but the 0.8125" spacing keeps things smaller than the Sparkfun button pads, which means saving a bit on PCBs, panels, etc. It's still going to cost a lot just to make one set of prototype PCBs, but in a mini bulk order (20 orders) it should work out at under 100 USD for all the PCBs required, including driver boards... I'm guessing though... my mini bulk order of SEQ PCBs came to 32 AUD each, and this is over four times the board area.... but with a design using 3 pcs 16x4+4, 1 pcs 16x4+4+16, 5 pcs driver boards, that might bring the board cost down.
-
I was already planning to use 3mm LEDs. It also would require some slight twisting of the leads so the lead spacing is smaller, 3mm bicolor LEDs typically have 0.1" spacing, meaning the outer leads are 0.2" apart, plus width of leads, this is too close to the inner edge of the conductive rubber ring. My current LED footprint has 0.0625" spacing, so outer leads are 0.125" apart. Pads are 36mil x 56mil round. No conflict with plated contacts on PCB which has inner diameter of 0.2" (same as conductive rubber ring on the button pad).
While I haven't proved this yet, I think you would need some ultrabright waterclear bicolor 3mm LEDs, like found on eBay... and not the weak diffused kind like you find at Mouser (and what I used on my MB-SEQ). SMD LEDs might work but it's too hard to find them with good brightness and low cost, plus they might not illuminate the cap as well as a 3mm LED which is more in the middle of the hole. 3mm LEDs are cheap and available from many suppliers so that's good. It's bad enough to lock into one design/supplier of the rubber button pads.
Even with 3mm LEDs, I still needed to do some tricks with the layout... like taking off the top layer pads, because 3mm bicolor LEDs are typically not soldered so close to the PCB, as the leads are so close together and have stupid right angle bends in them, the outer leads would touch a top layer pad around the middle lead. But when you think of the history of 3mm LEDs, they came from a time when single layer PCBs with wire jumpers everywhere were the norm... :wink:
-
I didn't like any of the bezels and LED holders available at the time of designing the back panel, and I wanted the LED to look the same as the LEDs on the top panel, i.e. just poking through a hole in the panel. So I just made the hole 3mm diameter and glued the LED from the back using superglue. In case it's not obvious, header J2 (next to power switch) is where you connect the LED directly, resistor R81 does the current limiting.
-
BTW don't get confused by the 0.813" spacing dimension in the 8x8 button layout: http://www.lividinstruments.com/media/64griddimensions.jpg
Everywhere else in the Livid documentation, it's defined as 0.8125"
I'm going to do a layout just for fun... matrix layouts are pretty easy and lots of copypasta anyway.
Don't count on me to finish before anyone else or do bulk orders next week etc. :tongue:
After layout of the button contact footprints, using a rough sketch of a circular contact, there seems to be enough room to route it all with two layers, it's a little bit easier since there's not so many through-hole pads getting in the way of the bottom layer tracks. The only thing I'm worried about now is whether the wider spacing of the bicolor LEDs would mean the outer pads might make contact with the conductive rubber of the button.
e.g. look at this:
Does that dotted ring area around the LED represent the inner hole and the conductive rubber is outside that?
If so, maybe even 5mm LEDs would fit in the hole.
It looks like phunk used double the normal gap between button caps (0.250") so spacing between centers of the first row/column of buttons and centers of the extra left/bottom buttons is 0.8125" + 0.250" = 1.0625" I'll use that for now (unless phunk wants to correct me? :wink:)
-
Possibly your C64 PSU is unable to supply enough current to all the SIDs plus the LCD.
I wouldn't call this a common problem, but I know of it happening with a few people.
Essentially each SID will draw 100mA on the 5V supply.
How many SIDs did you insert before you observe the 5V drop?
-
Looks like SMD LED will fit - here's a screenshot of a comparison between a few different layouts/button options on a 19mmx19mm spacing grid. The rectangle inside the Livid pad footprint on the left is the dimensions of an SMD LED.
Wouldn't it be better if the pad contacts were circular like the SparkFun ones? Surely there would be enough room for a 3mm bicolor LED.
The Livid ones are my current favourite.
If I ever get time to build my own BLM16x16, I would probably do a sort of "modular" PCB using the Livid pads, 8x8 plus extra 8+8+1
X XXXXXXXX
X XXXXXXXX
X XXXXXXXX
X XXXXXXXX
X XXXXXXXX
X XXXXXXXX
X XXXXXXXX
X XXXXXXXX
X XXXXXXXX
[/code]then chop off the extra buttons on three of the boards, join together to extend it up to 16x16. Maybe each board has mounts along all four edges, so after chopping, you'd still have mount points for the top/right edges of the joined PCBs.
It might need a bit of extra wiring here and there to make it work, and something like the "stitching" method on my SEQ CS PCB proto.
[i][color=#FF0000]* Wilba scratches beard[/color][/i]
hmm....
-
sealion: You have a dead SID. Get a replacement. This should be no trouble, I know who you got them from, but if it is, email me.
-
Just noticed this... Bliptronome kit... convert Bliptronic 5000 to monome clone.
-
That happens to me too, with some LCD/PSU combinations. Unfortunately, sammichSID design is a compromise between overall size, heatsink size, overall current draw, LCD current draw (brightness), PSU voltage/max current, parts cost, etc. Powering the LCD from the (potentially) unregulated input voltage is a big hack, but it doesn't generate more heat out of the voltage regulators and heatsink.
If SIDs didn't need dual voltage, it would be a lot simpler *sigh*
-
Send the clicky 6582A back, I will send you a new one.
-
The oscillators are digital. Frequency is set by a 16-bit register, which is relative to the system clock (standard 1.0 Mhz). The waveforms are then generated by an accumulator, which produces the sawtooth waveform. Some other digital tricks are done to produce the other waveforms. In case you're interested, here's an article with Bob Yannes, creator of the SID: http://www.joogn.de/sid.yannes.html
-
If you want to reproduce this same sound in a patch (after startup), this is what the patch is... converted from the registers being set:
Volume set to 15
Filters off.
OSC 1 Pulse waveform, Attack 0, Decay 0, Sustain 8, Release 8, Pulse Width 800, everything else zero/off.
OSC 2 and OSC 3 are off.
This is almost identical to the init Lead patch... looks like only Sustain and Release are different.
It would be interesting if you could reproduce the click in a patch, or not, and then know if it's a startup-only issue or your SIDs being clicky.
So the 2x 6582A that came with the sammichSID don't make this clicky sound?
-
So it could have been bad contacton the P1 wiper, or current draw close to current supply limit. If you're happy with the brightness where it is, then it ain't broke, don't fix it, nothing to see here move along.... ;)
-
Something does sound a bit wrong there... and also, when you look at it in a sound editor, you've got lots of strange stuff going on. Your left and right channels aren't matched, and it looks like it's almost clipping it's so loud. Compare it with the sample I made and tell me what you think.
-
MY LCD gets occasional flicker and dimming. It's slight, but noticeable. Is this normal or is there something I should start looking at as the cause?
Flicker/dimming is not normal.
It could be a bad solder joint in the backlight circuit.
Check solder joints in that area. i.e. J16 pins 15,16 (or equivalent on LCD), T1, JR4, R4, P1, JBL.
Maybe the P1 pot wiper is making poor contact. Try turning it a little. Also JR4's shunt might be making poor contact.
It could also be a current supply issue, but I think you posted a link to a good PSU so it's probably not that. Check if it still happens when P1 set to 1/3rd brightness (instead of 1/2).... or if you remove the SIDs.... or if it happens when sammichSID isn't doing anything (i.e. take out the PIC).
-
Now you're just showing off... :rolleyes:
SRSLY :drool:
-
WAVs or it didn't happen :D
MBBLM schematics and PCB discussions
in MIDIbox BLM
Posted
As I said earlier, I'm just playing around with a layout using the Livid buttons, since they look pretty nice and I've been told they have good action. Other switch/LED/cap combos won't give much of a saving, given that PCBs, panel, LEDs, etc. will cost the same with any switch/LED/cap combo. So it comes down to personal preference, and the total price on one of these things is going to be huge. You would be better off trying to save on the panels and case.
Example: To get a 15"x15" 3mm aluminium panel from Front Panel Express with 289 holes is 246.48 USD. I haven't even added reverse side threaded blind holes yet! Cha-ching! Ponoko can cut a similar size acrylic panel (15"x15"x3mm) for much less, but then you will probably need double thickness to stop it being bendy in the middle, or some spacers, it's not a direct substitution for aluminium panel, and you couldn't mount PCBs to it like you can with aluminium. Maybe a complete case made out of 5.5mm bamboo sheets would be possible, the top panel alone would be cheaper than FPE, but then it might need a chunkier case.