-
Posts
241 -
Joined
-
Last visited
-
Days Won
5
Content Type
Profiles
Forums
Blogs
Gallery
Everything posted by mongrol
-
Ok. So the switch footprints can be Mattius (alps), 6mm tact and/or DT6. Nice. I just had a look at the Mattius page and they seem be pumping the backliight support. Other than being transparent it seems how to mount the LED's are left as an exercise. I guess leaning on the mechanical keyboard crowd will again be beneficial. :) What do we want from Fentek? I'm happy to send queries but until we have requirements nailed it'll be a bit vague. I do like the look of SigPlastics G20 keys. Flat and square but they don't do a translucent version.
-
What do you mean by 6mm ad DT6? Where? More importantly what's the plan for the Oled's? The diagram looks like they are above the switch PCB and up against the encoders. To support the switch PCB you need standoff's underneath and if you use screws to secure it, you'll foul the little angle feet things on the bottom of the oled PCB's. Can oled's be mounted upside down and made to display upside down? If so, there's much more room on the top flange with about 9mm clearance before you hit the little angley feet things. I've got one here in a bag, I'll take a pick later.
-
The "track" encoders that replace the track buttons could be clicky. Or clicky the datawheel.
-
I asked on reddit about "shorter" rectangular keycaps but didn't get much luck. Thread here. https://www.reddit.com/r/MechanicalKeyboards/comments/5ekerd/rectangular_keycaps/ Relgendable caps here (down bottom) http://www.fentek-ind.com/keys.htm
-
Nice work! To boost structure it could be worth digging into the mechanical keyboard scene to see how they fabricate rigidity. I expect to reduce the rear standoff (Mongbox was really tricky to arrange, even with just the Wilba CS) some kind of cut alu plate or perhaps just acrylic backplane woud be needed. This would make the height of a desktop unit even higher. Lowest I could get Mongbox was 42mm.
-
I'm not getting it. If a fourth row is possible then using encoder buttons as that 4th row is also possible. The DIN issue is on the existing CS. Surely it can be removed. I realise this would remove the press FAST functionality but it may be a beneficial trade off. I do agree that a 16x4 RGB button matrix with optional BLM mode would rock like a bulldog eating wasps but unless narrower keycaps can be found from a sustainabe supplier (important) it's probably not going to be 3U.
-
Madam Speaker, Keep the 3U Go back to 16*3 rows of MX with relegendable keycaps Hold breath for RGB MX's at some point next year. Use encoder buttons for top row. Since Menu+Encoderbutton is a bit clunky consider swapping some row functions Or rework the entire button layout. Embrace entropy! Change is good!
-
Ah yeah, of course. Brain failure. Cherry MX's are great switches and as you say there's a ton of action types and support from various communities. There's also loads of "custom" keycap styles available and a few with different profiles (signature plastics do low profile caps)
-
Also used on the Social Entropy Engine by the looks.
-
I understand that the encoders provide a visual reference as to what you're adjusting but that can work with my method too. I didn't intend the datawheel to cursor over to the step to select. I meant this; Press top row button (jelly lights up) Use datawheel to adjust the value. Click datawheel to cancel out or deselect or press the jelly button again. The light from the pad provides the visual reference, probably better than an encoder ever will (we could also colour every 4th pad with a different colour if not RGB. Also, swap the encoder row for pads. 16x4 BLM mode?
-
Oh, there goes a brain spark! Why do we have 16 encoders? You only have two hands and the only time I ever turn two at the same time is when playing with ramps. A single encoder could satisfy 95% of the use cases. Or worst case, two encoders (one at each end or in middle). Press and hold step (or perhaps toggle), turn datawheel to set value. There's an entire row away.
-
Ah yeah that is a bit smaller right enough. Maybe a bit more UI tweaking could rid us of the track buttons(or knobs) AND the datawheel as well chopping another 30mm off it. Of course, the display's still get in the way with that pcb surround. Hmm, I think my ideas warrant more of a major redesign and UI workflow more than current design optimisation. Still, with jelly cube buttons and led's your design would be a very nice modernisation.
-
Looks good. My desire would be to shrink the CS considerably (at least by a 3rd) and reduce the amount of buttons as well. However, button reduction would be a massive talkfest as it would require analysis and refactor of workflows which would impact a lot of code. Not being a modular guy I haven't considered Euro but I agree it would have to be a requirement. Rather than clutter the thread I'll go back into my cage and think for a while then write something up in design concepts. Note: Sparkfun button pads have largers spacing and also come in 2x2 grids. The pcb pad footprint is also available.
-
I'd love to do some design work around a new CS. Is there an archive of previous discussions? I presume a lot of talking was done to arrive at the current layout.
-
Yeah fair enough. I'm having a grumpy day. :) Pulled my old Midirex out to get some jamming done and it kept hanging on me too. Just one of those days where nothing goes right. Although I do like these days in that they put things in perspective in how much "stuff" I have to maintain. Will get some cheapo buydisplays and see how we go. Least I get 6 weeks piano practise waiting on them arriving. :D
-
One of them broke when I was removing cable. My cables are on angled pin headers to fit my enclosure and the little copper coil thingy close to the pins broke off as it was tight to the IDC socket. Tried to repair and replace it but too small and fiddly for my big ol'beef fingers. The other broke when I was changing the J80 jumper back to J68 and the pad came loose. Again, big ol'beef fingers. So right now I have a choice. Go with some LCD's which isn't exactly plugnplay as I have backlights to deal with, or go with more Oled's. I've found RS sell Vishay brand for about $30 USD each (cheap or what!) and again, appear to be identical to Raystars. What I don't get is how I had them working flawlessly before. 1 Raystar and 1 Midas, running off 3.3v with the jumper at the the default J68. 0x02 in bootloader and both displays worked everytime. Now I've replaced the core with a newly flashed one it all stopped giving me joy. I have another choice. Sell the SEQ parts off and be done with it. :|
-
Hmm, my Winstars also have a jumper set to J68 (the pcb is exact same as the Raystars) but previously I had both working fine (after a lot of kerfuffle). Off to dentist, will change jumper when back. Buffoon Edit: I don't have Winstars at all. I have one Raystar and one Midas. UltraBuffon Edit: Now I don't have any oled's. Changing to J80 gave me the same odd chars as Jarvis but I broke one in the process. Then doing more elimination broke the other one. Bums, cheapo LCD's for me from now on. Oled's are too much hassle.
-
@tk. Yep, these Oled's were perfect before I b0rked by discovery board. I'd forgotten about the 0x02 though... so... Uploaded boatloader_update app Set lcd_type 0x02 Store Reset Looked back in my enclosure thread and checked my previous voltage for my oled's (it's clear it's 3.3v on the pic). Both screens still won't work at the same time. I remember having a helluva time when building this first time around and somehow in my build thread I came up back to 5V and 0x02. However, every photo I look back it it's clearly set to 3.3v. Time for more oxygen I think.
-
What a timely coincidence. I'm repairing my SEQ atm and have the exact same issue. I have Winstar OLED's which are the same component as the Raystars. Here's the datasheet. I've replaced a fried MB997C Discovery board with a brand new one labelled MB997D. I haven't changed any jumpers on the core board. OLED's are set to 3.3v. The result is the same if I run over USB or 5v direct power. One OLED attached shows fine. Two just come up blank or the left one shows a few odd characters. Wasn't there a trim pot contrast adjustment thing at some point in the build or am I thinking of a differenet project?
-
Muha! Language is a beautiful thing. The new discovery board I received was a MB997D. I was going down the Core page verifying my build when I noticed there was a firmware update to make the new model D run from USB. Downloaded from ST.com, flashed and it works! TK: On the core page it states the new firmware is untested. This isn't true now. :) Now to put this bloody orange box back together for the LAST time.
-
So my new DISCO arrived today and it does the same thing. Disconnected everything (OLED, MIDIIO, CS) and left it on the Core board and the same. No response from it. Remove Disco from the Core and it appears correctly in MIOS Studio. It appears a component on my Core board is fried. Time to follow the schematics. :)
-
It's toast. Good job these things are only $30.
-
I've been moving my battlestation back into my Cave this weekend after it was vacated by a "visitor". Unfortunately in my rush I plugged my SEQv4 into a 7.5v supply which I use to power the majority of my kit. Every light came on with no display. After realising my error and going back to the proper 5v supply I have the same result. Every LED on the CS lights up, no display or response otherwise. Have I fried my Disco board?
-
#7 Final Assembly Oh dear. Too many reserved posts. This one can get deleted.