cheers...
anyway i already recorded 17 minutes now and there is still one part left to record...
but this recording session showed me... the firmware works quite good ... i didn't had a crash or something and i was working with it over 6 hours...cool
so i'll continue next week... hopefully i can finish it then...
have a nice weekend...
so maybe tomorrow i do some videos... let's see...
well almost 2 1/2 year no update... sorry for that...
ultra303 is not dead as it looks like !
the problems which caused the delay was a) the all oled lcds died but b) the main problem was that i had problems with my 303... there was a loose connection which drove me mad... after i tried to fix it (and failed) 2 years ago i totally lost motivation to continue...
but after some distance from the project i recovered a bit and finally could fix the loose connection... ok almost some small thing is still not wrong but anyway that's ok to continue...
well very at the beginning when i got my 303 a speaker fell down on my 303 and the mainboard got some heavy fractures... every time i thought i finally fixed it it started again... a real nightmare
anyway... i was working quite hard in the last days to get some stuff working and looks good so far... so i hope i can upload very soon some new preview videos...
so stay tuned again...
nice to see some ppl still visiting my page after that long time... i tracked the counter a bit in the last days...
anyway... i've decided to try out another lcd... an older one...
and... ? #1 lives again:
as you can see the display doesn't look as good as the old one... and it's somehow far too slow... but anyway... i ordered today some new of course no oled displays...;)
even uploading of the firmware works so far... had to search my entire room where the isp uploader is ... so theroretically i could code a bit on the firmware..hm let's see
but somehow the 303 sounds strange...something is wrong there... maybe i'll trace it down if i get the new lcd displays
so ... oled displays look good as you can see on my page but do not last long... i really could puke...as you maybe know oled displays are quite expensive i bought a few of them ...
a nearly rip... for this project... really sad... invested quite alot time in this project... bla
http://sharc.ultrafex.de
no worries... i'll continue ultra303 soon ;)
well i've got this for free for my sharc synth project very nice... the special about this is not the development board itself... it's more the software which is included... you can code this fpga via vhdl or verilog ok nothing really new... but ! with this one you can code your fpga in c !... that rules totally ;)
so you can imagine i have to take a small break from the ultra303 project... this rc10 development board is just too interesting...;)
the leds are flashing to the current note and up/down/accent/slide looks cool ;)
i had a closer look today why the 303 runs out of sync when it synced by the midi clock... hm very odd... when i turn off led multiplexing it works so far... when it's switched on it runs out of sync... that's no good...
until now i could not figure out why this happens... the strange thing is... when don't sync to midi clock just to the bpm count it works fine without problems... there is no big change between midi clock and bpm... just the timer setup is a bit different...
anyway i'll have closer look later surely a smaller problem which i can solve later...
and number two lives ! ;) i build up the second prototype for my other tb... switched it on and worked... fine fine ...
hm the first 90% of a programm is fun... the last 10% is a real pain in the ass... sometime i have the feeling i need round about 5 times more motivation to handle the last 10%... mostly totally boring to code...
but i've decided not to include new features from now on... sure i will include new feature but only features i will really need...
beside this i started to code the sysex routines for pattern sending,,, but not finished yet...
i had an idea today... maybe it's a good idea that the firmware can output the sequenced data to midi out... means you can use your tb 303 as sequencer for another synth... or just to record it in cubase ...
i will implement the midi stuff in the next time... means online editor reading and writing ... pattern dump etc. i played a bit with the online editor hm a bit outdated i really have to code a bit on it ... hm beside the rebirth reader i will include a audio realism bassline pattern reader... maybe a good idea...
finally i included the led multiplexing for the tb... (yes i was shifting this like always... last time i included it for the ml 303) but works fine so far... but still absolutly boring to code this... i really had to force myself to do this...
i'm really happy i ordered that prototypes...it seems it's the final version of the pcb... so now i can concentrate to finish the firmware itself...
i will do soon a new video of my firmware with the tb 303 to show the new features... when i build up another prototype for my second tb 303... maybe i can do a video with two tb's running in real time editing synced to a sequencer... hm never tried this... but i doubt it will sound great hm anyway
time for a backup ... cheers !
inside i found this:
these are my very first real pcbs ... cooool ;)
so after some time at work i went home to solder ... finally it looked like this:
ok not fully build up i didn't soldered in the dac stuff... anyway ... so hm will it work ? can i upload the firmware? hm when i tried to connect to the cpu... arf "time out error" hm slowly a nightmare started... i already considered with the thought fuck it still don't work... hm i started to check the voltage... hm there was no...hm maybe a short circuit...
hmm but in the end ... hm lol ;) in the excitement i forgot to switch on the tb 303 ... so no wonder i couldn't connect to the cpu ... hell i'm a noob ;)
anyway after i switched on the tb 303 it was just like how it should be... uploaded the current firmware... reset and all was working fine...
next cool thing is... because of the smd and the new connector to the tb 303 it fits just perfect into the tb 303 now without problems...
the only annoying thing was... it took 2 weeks to get the pcbs... bla... pcbpool sent the pcb to a wrong adress...
so after this perfect day... it's time for a good beer...
cheers!
i will check it again at weekend so maybe i will order them at monday...
hm i tried today to breath life into the new prototype...hm no chance i have no idea why i can't upload my firmware to the cpu... the rs232 thing works ... i tried it with another pcb... bla!
the prototype itself works but upload don't work... i checked now all i could check and decided to skip this prototype.. and i'll create a new one... maybe the cpu has a problem dunno... to desolder it is nearly impossible because of smd and the thin tracks on the pcb... the pcb will surely be broken after it...
it's damn hard to create this pcb here at home...and much of work... because of many vias and drills with 0,5 mm... and beside this the pcb is really small... one small mistake and the pcb is unusable...
this time i'll order a prototype board from pcbpool... the minimum pcb must be 1 dm2 so i have to order a bit larger pcb...anyway advantage is of course i have 4 prototypes then...
maybe a better way to continue... and much more stressless...
next will be to implement the code i did for the ml 303 for the tb 303... a bit boring... but anyway
today i could figure out why... hm just forgot to save a register during an interrupt...bla what a noob bug... hmm now i'm wondering how could this work ;)
so finally all is working really good and it's time to send out the first version to an alpha tester...
hm in the end it was just an hour of work... he unbelievable ...would have been much less work to just to do it than to shift it since a year... anyway ... but anyway the led stuff is only implemented for the ml 303 at the moment i still need to implement the new multiplexing for the tb... hm maybe i need to shift it another year !? ;))
and of course i did some sounddemos... ok not very impressive just recorded a bit during i tested the features a bit...
filter frequence modulation (fm)
yes doesn't sound like a 303 anymore i know but this was the reason why i've included this feature...;) you can hear the filter fm feature with a synced lfo ... goal was to create a sick as possible pattern... there is no distortion or something.. the sound comes as you can hear it our of the 303... on the sound was only delay and reverb added...the old osc is still active no dco or something... this is done only via the filter cv
filter envelope
this one is just a small demonstration of the new filter envelope with attack and decay and stuff... at the beginning i'm slowy start to modulate the filter with the envelope then i'm changing the attack and decay time etc...
filter lfo
in this demo i'm playing with the lfo... starting with increasing the lfo filter modulation... then i'm changing the lfo frequence a bit...in the end i'm changing the lfo syncing...
all sound demos were recorded by using the ml 303 clone... i think that filter fm thing sounds funny... to be honest i didn't expected that strange noises ;) that bassdrum like sound seems to be the accent or so...
filter envelop is working now... added lfo pattern syncing ... what should i say ?
that lfo and env feature is cool... really funny to play with esp. that lfo pattern syncing is nice... you can sync the lfo to a note, 1/1 , 1/2, 1/4 and 1/16 pattern and stuff... or filter envelope attack is nice too... so you can provide some more advanced filter sweeps ... sounds cool when you do not distort the 303 too much and creates more bleeps and bloobs somehow ;)
but at the moment the modulation of the envelope and lfo is always 100%... a bit too much sometimes... so next is to include the filter lfo and envelope modulation amount... at the moment is not possible to modulate for example the lfo with the envelope this is what i will next beside of the modulation amount... when all is working so far i will upload some demos of the new features... hm there is still a groove mp3 missing...
some stats of the source:
97 files
280 kb in all
27 kb object code... so enough space left to 64kb ;)
cheers!
this week i soldered a bit ... build up 4 of accenters distortion (http://beam.to/midi303) but the potis are missing to i can't finish them now... i did also a second pcb of the new prototype for the ml...
and finally i did some code yesterday...well i've continued to implement the filter cv stuff... filter cv control is now working... means i can now modulate the filter via a lfo... filter envelope is from coding side done but not called at the moment... i was too tired yesterday to do that ;)
but looks like this is going to be a nice feature... but at the moment a bit hard to use... i did not add a possibility to change the values for filter lfo and filter envelope so i always need to change them in the source code... this is of course a bit annoying so next step will and must be a implementation for the user interface...
at the moment i added 2 lfo and 2 envelopes and they can modulate each other... maybe i will add some sync or ring mod for the filter lfo... well have to do some test if it sounds interesting otherwise i will skip it...
hm skipping...i've decided to skip the sid implementation... i'm slowly running out of memory... i just need memory esp. for stack and stuff... well since i'm using c i need of course more stack... hm and the sid implementation uses much of memory...it uses round about 50% of the low memory(stack is placed there too) to change it to xram will be too slow so it's skipped... maybe i will do the sid stuff in an external cpu but not in the main cpu... same for the beatbox feature... but this one will be skipped totally...
well all was done very fast but while i was coding the filter cv control itself i recognised a bug... so since monday i was searching a bug in the firmware...and wasted 4 days... bla...
the bug produced an effect i could absolutly not explain... very strange... finally i found out i just accessed memory locations outside the available memory... even a read from that position forced the lcd to show totally crap...
i stripped down the entire firmware source...kicked off all unneeded stuff so in the end it was just a source of maybe 100 lines... damn much of work...and even with this 100 lines i could not trace down what the problem is... when i read from the $0800 mem this happens... when i read from $0200 it doesn't happens... all access are going to the internal ram so no external bus i used (no ports are effected because of the mem access)...
in the end i come to this conclusion... this must be a strange behavoir of the cpu... i can't see why a read to the mem produces such an effect...anyway
so the bug was just a *4 for a value...bla... however now i know about this problem and i can continue...but still annoying....
so far all is working fine ;) i really hope i can motivate myself to do some code today... instead of sleeping on the couch ;)
but it was real fun to build up this prototype... so i will do some more pcbs for my other ml 303 ...when i located the switch problem...
that new plate to plate technic i figured out works really good...and it's done fast... remembers a bit to sewing...but with this technic i'm even able to do plate to plate connections under smd pads... anyway i will post some fotos later how it works ...
but looks really good so far... this time i guess i found good way to place the upper and lower layout for the pcb 100% excat... after etching i did some test drills and wow never had such a excat upper and lower pcb side layout... no displacement at all... i mean the plate to plate pads have a diameter of 0,8 mm (hole 0,5mm) and the hole is 100% inside of the pad...that's pretty cool
the new prototype is based so far it was possible on smd parts only... that saved bit of pcb space... the pcb is now 105 cm2 instead of 131 cm2...that's 20% smaller... and 110 holes less to drill
i had some days ago a new idea for a new plate to plate connection technic... before i just used rivets from bungard... but a) this is not working very good... b) needs too much time and c) stressy d) too expensive... so let's see if this will work...