|
|
Post by solipsistnation on Jul 19, 2021 0:26:49 GMT
Okay, I've just finally solved a problem that had been bothering me for weeks (I haven't had time to sit down and figure it out...).
I am attempting to sync my AE rig to MIDI, but I've been running into problems where it would be tight for about 3 bars and then sort of skip a step and get all my AE stuff out of sync by a click or 2. I've been doing this with a bunch of sequencers (2 SEQ16s, 2 SEQ8s, a DIV-MM, TOPOGRAF, and 2 TRIQ164s) all connected through my giant breadboard mult. Signal path was MASTER or TOPOGRAF's clock output -> BEAT DIVIDER -> mult -> clock ins of everything that wants clock. I tend to leave them all hooked up all the time to save having to set it all up again every time.
I did a ton of midi troubleshooting (and fixed some other stuff while I was at it) and finally got to where I could be sure it wasn't the computer/MIDI stuff.
Finally I tore all the AE patching down and am now running MIDI -> MASTER -> BEAT DIVIDER -> TRIQ164 -> KICK. Nothing else patched into the clock at all, just the most basic possible thing.
It's been sitting here happily synced with a loop on the computer and a MIDI-clocked arpeggiated synth for like 5 minutes now. Turns out all that multing was most likely causing the problems.
So, what's up with this? Clearly there's something going on with trying to drive too many things from a single +5V clock, but what is it? It seems like I was trying to do too much with that one clock.
Would a buffered mult help with this? I'm happy to swap out my simple breadboard with a 4BUFFER if that would fix this for me. Can I cascade a 4BUFFER into another 4BUFFER for 1 input to up to 15 outputs?
Thanks.
|
|
|
|
Post by robertlanger on Jul 19, 2021 6:11:44 GMT
You are right, driving too much stuff from one clock source might put too much load on the output, causing signal level reduction that might lead to the dropouts. Also, a too long chain of Patchwires might be an issue, and maybe also slight contact problems with the Breadboard mult too. Indeed, a buffered mult should help for this, best with a "star-topology", meaning you go into the buffers first and from there to each input. No problem with one mult output driving two inputs, IMO.
I'd be very interested to hear if this works!
|
|
|
|
Post by solipsistnation on Jul 19, 2021 13:59:06 GMT
This was the last thing I had to sort out before doing some actual recording (...my last excuse, I guess 8) so if it works I'll post here! I ordered a 4BUFFER from noisebug and those usually show up very quickly.
|
|
|
|
Post by arti on Jul 19, 2021 17:58:32 GMT
This was the last thing I had to sort out before doing some actual recording (...my last excuse, I guess 8) so if it works I'll post here! I ordered a 4BUFFER from noisebug and those usually show up very quickly.
Thank you for bringing this up. I occasionaly have similar issues but never bother to actualy dive deep to the root of the problem. Time to put 4BUFFER on the list and be more careful with clock via multi connections for now  |
|
|
|
Post by lukylutte on Jul 19, 2021 18:59:30 GMT
Kina confused about the 4buffer would help? It's not a splitter nor mult or mixer and each buffered output is only double? Or am I missing something?
|
|
|
|
Post by solipsistnation on Jul 19, 2021 19:45:39 GMT
An unbuffered mult is just basically a bunch of jacks connected to each other. It's entirely passive-- think of just taking a bunch of wires and twisting them together. Put a voltage into one wire, it shows up on all the others. The problem comes when you hook the outputs up to a bunch of active modules and handwavey electronics stuff happens and the voltage the outputs see drops. Depending on what you're doing, it could be a little unpredictable. This discussion is a bit Eurorack-specific, but it goes into detail: learningmodular.com/when-do-you-actually-need-a-buffered-multiple/What I was doing was sending a +5V clock pulse into a breadboard with a lot of connections-- so 1 +5V signal got divided up into something like 10 outputs. Each of those (depending on what the module is and what it's doing, I think) drops the total voltage the outputs see by some tiny amount. 2 or 3 outputs would be fine, but my ridiculous "here is a clock signal for 8 sequencers and some other stuff" setup means that +5V occasionally drops below what the sequencers see as an actual clock pulse and some of them would skip a pulse and get out of sync with everything else and break my rhythms. If I had been sending a single pitch CV into this giant mult and sending that to a bunch of oscillators, what you'd have heard is some of those oscillators getting out of tune with the others in a not very predictable way. That might work for pitch (a little detuning sounds good, right?) but with a clock signal it sounds gross and arrhythmic. The buffered mult is, unlike a passive mult, a powered electronic circuit that sticks op amps on every output in such a way that if you put in +5V you are guaranteed to get +5V out of every output. (Or if you put in +3.2V, you get +3.2V out, and so on-- because the outputs are active, they avoid the voltage drop you get from just sticking a passive splitter in there). Here is a real simple circuit diagram that shows you more or less what's going on there: www.haraldswerk.de/Utilities/Buffered_Multiple/Buffered_Multiple.htmlThe 4BUFFER has 4 inputs, each of which, by default, is routed to 2 buffered outputs. It has switches that let you cascade input 1 into the other 3 sets of buffered outputs: wiki.aemodular.com/pmwiki.php/AeManual/4BUFFERWhat I am going to do once my 4BUFFER module arrives is pull out the strip of breadboard I've glued to a blank panel (SUPERMULT!!!! This ridiculous thing: forum.aemodular.com/thread/1026/supermult ) and replace it with the 4BUFFER, set the cascade switches so input 1 goes to all outputs, and run my clocks through that instead of a passive mult. That should solve my sync problems. |
|
|
|
Post by solipsistnation on Jul 19, 2021 19:46:48 GMT
This was the last thing I had to sort out before doing some actual recording (...my last excuse, I guess 8) so if it works I'll post here! I ordered a 4BUFFER from noisebug and those usually show up very quickly.
Thank you for bringing this up. I occasionaly have similar issues but never bother to actualy dive deep to the root of the problem. Time to put 4BUFFER on the list and be more careful with clock via multi connections for now The thing that bothered me was that it worked fine for 2 or 3 clocked modules, but once I hooked up a pile of them things started getting weird. I learned a new thing yesterday. 8) |
|
|
|
Post by lukylutte on Jul 20, 2021 7:15:37 GMT
Totally forgot about the switch 🤣. Sorry. So 1 to 8 buffered splitter 😅👌. Guess would need to look into this myself for this exact problem 🤣...
|
|
|
|
Post by solipsistnation on Jul 20, 2021 17:14:56 GMT
Totally forgot about the switch 🤣. Sorry. So 1 to 8 buffered splitter 😅👌. Guess would need to look into this myself for this exact problem 🤣... And I suspect (since it's buffered) you could add a second 4buffer so it could become a 1 to 15 buffered splitter, increasing in increments of 7 until you have a whole 2x20 rack with 1 input and 281 buffered outputs. (...assuming it's powered from somewhere else) Anyway, I think my 4buffers will be here tomorrow so I can hook some stuff up and get it all ticking along together. |
|
|
|
Post by solipsistnation on Aug 27, 2021 17:52:48 GMT
If anyone is watching this, yes-- a 4Buffer did the trick here.
There are still timing problems (I'm pretty sure Ableton Live running under Rosetta on an M1 Mac Mini is not as stable as they'd like) but clocking from something like an LFO or the clock output of the RBSS is nice and solid now. (I am not convinced the clock output you get from one of the topograf output modes is as stable as I'd like, but there are lots of other options.)
|
|
