MRE

As with everything show related, timing is everything. Let me use an analogy to the lighting world: I know that when Ethernet DMX came out, there was a lot of fussing before hand about how to assure that the packets arrived on time, and in sequence, as TCP/IP/UDP can and will send them out of order when collisions occure. The end result was: Yes, EthernetDMX works fantastic. It is fast enough that even with dropped packets comming in last, you didnt notice it much... So long as you 1: Have a dedicated network, and 2: only have one 'Host." I.e the DMX dimmers acted in a way like slaves, in that they would not generate their own traffic without requests. So, having one PC (managing the network), one or two 'Desk consoles' and a bunch of dimmers, and everything is perfect. Change things around, add more desktops.. and latency started to become apparent. At least, on the DMXnet I worked with that was the case. Obviously replacing hubs with switches fixes just about all the problems. Again though, cant stress enough how important it is to keep it a dedicated network. As soon as lil' Billy goes browsing for porn, the lights start acting up. How that translates to MIDI over Ethernet, I dont know. Scale is different. Three or four MIDI devices is not like 400+ lights (each a single ethernet node). However, lights only need one parameter (in the case of intelligent lights, no more than 16 usually) while MIDI can transfer a lot more information than just one byte to a single device. I would also assume you are not attempting to download the latest Jenna Jameson video while making music ;) Honestly, Ethernet DMX is a LOT of overhead for very little information. Remember that TCP has a minimum packet size. a single DMX channel only needs a single byte to describe its setting. A bank dimmer may have 8 or even 16 lamps on it, but still, at 16 bytes, there is still a lot of padded zeros in an EthernetDMX packet. Waste of overhead and processing power in my opinion. The only thing it solves is the 'world' system in which a DMX line is limited to 512 devices. An intelligent light might take up several 'devices'. It is easy to see that no more than 16 smart lights can exist on a dmx512 system. In large spaces, you might have to run several DMX systems if you have a lot of intelligent lighting. Thus ethernet comes into play in large arenas, since it is virtually unlimited as to how many devices can play. Not to mention, if one dimmer dies, the whole world doesn't fail (as is the case with DMX). Again, how that translates to MIDI... well really it is just a lot of overhead and wasted power on relatively small timed packets of data. The logical progression to me is more along the lines of USB. Much less overhead, still lots of device room. Simple interconnect. Simple protocol. I say leave Ethernet to what it is designed for: Connecting desktops to desktops .. transfering large chunks of data quickly, but without regard to packet order. Serial to USB adaptors are incredibly cheap, and offer the data from the serial port side to the PC as if it were a REAL serial port. Essentially NO extra programming needs to be done, nor is there much overhead. A simple driver install fools every piece of software you have into thinking it is talking to a serial port.

It would certainly work.. with a bit of programming on the PC side... Honestly I think it makes more sense though to integrate USB using a Serial/USB adaptor. Literally plug and play.

Yeah, Happ does a really good job of gouging the gaming industry.. but when reliable distributors for those parts are few and far between... Those flex keyboards are great. You can even stand on them and expect them to still work when you step off of it.

Along those lines (perhaps this should be a hardware section link): check out Happ controls. They do everything for video games and arcades. Those big buttons are cheap. They also have a lot of other styles including large illuminated buttons such as these: http://www.happcontrols.com/amusement/gameparts/midway_ipbs.htm http://www.happcontrols.com/pushbuttons/pushbuttons.htm heh.. this would be a huge but incredibly cool pitch bender: http://www.happcontrols.com/driving/50008300.htm Ohh.. absolutely do NOT buy pots from them. They are WAY over priced.

shneikies.. that does look good. Well.. eventually someone will.. its all a matter of time.. and need.

http://www.noritake-itron.com/Specs/CU-U/CU20025ECPB-W1J-03.pdf Dont know the difference between the W1j and the U1j but I cant find the U1j. From what I remember, the datasheet looks nearly identical.

Official count is 31. Noritake Itron P/n CU20025ECPB-U1J Connector is a 2X7

Heh.. I suppose I havn't been keeping up on my display technology...

This really sounds like every crackpot flop Nentendo ever came up with, all rolled into one machine. Pay particular attention to the fact that for the controller to work, you have to have a sensor bar wrapped around the TV. It says so in the Wiki. Sounds like the power glove doesn't it? Also, for REAL gaming (and by that, I am just refering to the multitudes of games that dont require oddball or specialized controllers) this thing is all but useless. Its going BACKWARDS in gaming controllers. Turn it on its side? so you have a D pad on the left, and a couple of buttons on the right... what? so the only ANALOG input is if you lean the controller? And whats with the left hand controller being called a nunchuck? Heres the thing.. I love Nentendo.. they always push the boundries.. WAY over... They have the cash to do something totally revolutionary, and watch it flop. (Power Glove, Virtual Boy, that silly robot that came with first gen NES.. DS is in some ways barely hangin on. Its been months since I have seen a commercial for it). This is all well and good for hacker types though. They will push it on the market and flood production. But with a controller like that, it is very likely to flop big. End result? Hackers get a fun toy cheap. (eventually).

What are some of the options for LCDs. To be more specific: I have 20X2 displays laying all over the place. I know the best setup is with 2 x 40x2 LCDs.. so.. can two 20x2 displays be chained together to work like a 40x2, and thus use FOUR 20x2 displays? Or is the hardware limitation TWO displays, regardless of their configuration? The data lines on the displays can all be daisy chained, so the only real limitation here is that 4 pins need to be used from the core to handle the enable for each display. Is this a supportable feature or would I have to write some crazy MIOS code to make it work? I might consider a seperate pic 16F84 or some such to 'fool' the core into thinking its connected to two 40x2s, and stream out data to 4 displays as an easier option. How does the Core handle lcd messaging. Awareness is set as a user panel option or is it set when you compile?

Have you considered Plastidip.. or tooldip by its generic name? It comes in several colors. White, black, red, blue, green... etc. It's latex based, so you could color mix. Some of the better tool supply stores stock it in smooth and rough finish as well. I havn't tried it for this purpose but it could be a good idea? Another nice idea might be powder coating aluminum knobs.. expensive unless you 'know a guy' but damn nice custom touch.

If you can afford them, VFDs are the best. Viewable from any angle, low light or bright light. Unfortunately there isnt as much of a selection of font sizes without a tremendous cost difference. Even comparing a backlight 20x2 lcd ($10 to $20) vs the same as a VFD ($50 to $90)... In a related thread, I mentioned I have a load of 20X2 VFDs... if anyone is interested..

Well shizzle.. I wasnt expecting that, but sure. I havn't counted lately, but if memory serves, I have like 20 or 30! ;) They are all only slightly used, as the undisclosed location from which I got them had done a major upgrade to 500 machines. About 200 of them needed side cars because they were way too old to install the imbedded version of the upgrade. Within 5 months, most of those machines were subsequently replaced with the imbedded units already installed. So.. long story short, 200 side cars went in the trash. Dumpster diving never hurt anyone! (I'm just kickin myself for not saving the 10 digit keypads and card readers!) Last I checked, the street price for a 20x2 VFD was between $50 and $90. Naturally I would not demand nearly that. How does $25 plus shipping sound? I can give a link to the data sheet. I have used the standard Hitatchi startup routines (and ran both 4 bit and 8 bit modes) with no problems but check the special fuctions and charactor sets to be sure. Comes with a short cable (1.5 inch?) that has a connector on both ends. The pins are two rows of I think 6 rather than one long row.

yah, I got my samples friday. It looks like they would be a bit of a biatch to make work. A sample and hold circuit for 30 pots might be a bit excessive and contrary to the whole "Midibox philosophy" anyway. I was considering perhaps using the straight linear for a modulation wheel, but it just all sounds too complicated since its not mechanical.. (with an unknown point contact when you dont hit it right dead center to start..) Ohh well.. they were a sweet dream for a moment. Now, I also got some nice button pads. Those I plan on using for the menu buttons. and, worst case scenario, I can chop them up and re-arrange the button layout.

I have a boatload of the same unit as a 20X2.. Noritaki-Itron makes some nice VFDs.. I have used them quite a bit, and they generally conform to the Hitachi standard as much as possible.

Looks like they have recently limited their samples section to a button, a pot, and an overlay. On a side note, the 'rotary' pots look sweet. They look like they would act just like the control on an Ipod. I could imagine a several of these circles used to make a totally 'flat' midi controller.