i’m afraid we need @tehn’s help at this point… i’ll message you.
got your email @esquilofreniko
sorry to hear this. we should make a bigger warning. i wouldn’t have expected the 5v regulator to kill itself (this is certainly what happened).
but luckily it’s an easy fix (one SOIC8 IC). you’ll need to send it in.
the real question— is your grid ok? if not that might suggest further damage to the TT.
yep, updating all the relevant threads and the wiki.
Thank you for fast response, got your email!
The grid is fine, i tried connecting it to ansible after that and it all worked as expected.
since we now have a confirmed case that grid can indeed damage teletype when connected directly i’m going to suggest adding an explicit action that a user has to perform in order to enable grid integration.
basically, teletype will not output anything to grid by default. in order to enable it you will need to use a special shortcut (like Alt-Shift-Ctrl-G or something similar) which will take you to a screen that will reiterate the warning and ask you to confirm. this confirmation will not be saved and will need to be done each time a teletype is turned on (and will not be available as an op for obvious reasons). grid visualizer will work regardless of this setting.
Maybe showing my ignorance, but how could this prevent damage to the TT if the voltage is always available on the port? Overcurrent is overcurrent. Unless there is a MCU-controllable relay/FET/etc?
This is probably best handled with a hardware revision, unfortunately. Are there polyfuses on the USB power lines? I know that’s a popular method to protect PC USB ports.
grid draws more power when LEDs are on. i think (hope) that there are no circumstance under which a non lit grid would cause damage to teletype. so the idea is to not have any LEDs on until an explicit confirmation is received.
Oh that is indeed a good point!
I’m battling with myself on a response because I totally get the point for this - but it feels naggy to have it appearing on the first grid access after every TT restart. Making it a one-time-only kind of thing seems less painful (only the first time you use the grid op ever). Though, I know that opens up risk for units that are purchased on the used market. And, yes, I know I could comment the thing out in the source and build my own firmware version without it.
Regardless, I think my vote is for a one-time-only warning - if there is room in flash to store the status.
EDIT: Oh, and @esquilofreniko , super-sorry to hear about your TT. Hope it gets up and running soon…
yeah, my thinking behind not storing the confirmation was teletypes changing owners, and to a lesser extent people moving tt to a different case or something like that. it will be naggy for sure but just feels like a lesser evil - i feel really bad because of what happened to @esquilofreniko
and perhaps it doesn’t have to be all that naggy… since it only matters if a grid is connected then the way it could be done - as soon as it’s connected (or if tt is powered up with grid already connected) it will show the confirmation screen, and to confirm you’ll have to push 3 buttons together, something like that.
At this point i almost always have my grid and TT connected, far more often than a keyboard. I completely understand needing a protective barrier, but every power up seems… inconvenient.
Edit: Should we consider an “Expert Firmware” version, à la Batumi?
Really sorry to hear about your Teletype @esquilofreniko
However I agree that confirming every time would be a real drag.
I often just leave my grid plugged in, so having to remove it, add the keyboard, do the keystrokes, and plug the grid back in every time it powers up is a little bit of a drag. Not the end of the world though if it saves some people from frying a TT.
Could it avoid the keystrokes in this situation, given that you must have done it once in order to connect the grid in the first place? Or have some secret key combo on the grid itself, to wake it up? To avoid the swaparama?
I had no time to dive into the grid ops by now and I am pretty sure that I would have ended in the same situation (already fried a PSU by accidentally plugging in the grid into one of the other modules without switch power ).
Nevertheless I would second something like a grid gesture as a safety check when starting up or connecting a grid.
Also I wonder if @tehn could suggest a hardware mod to make it safe? In case it is possible and for those who feel safe with some soldering on the tt.
I have a silly question. I started exploring the grid/ops before these warnings got so explicit, and tried plugging Grids directly to the TT. My case is powered by the largest Intellijel bus board, which I believe is one of the more robust boards out there. At any rate, I’ve just blithely kept plugging in directly this whole time. Is it possible that the bus board one is using would make a difference, or is this really a matter of some limitation of the TT itself, regardless of board? I do have a Switch and also some of the earlier power thingies lying around, so I should probably start using ‘protection’.
Is the current consumption on the USB port available in ASF? Possible solution might be for TT to not issue any grid LED on commands if port current is too high. I’ve been poking around in that code (slowly) thinking about hub support, I’ll take a closer look.
leaving other aspects of having multiple versions aside (increased maintenance, each time we release we’ll have to do 2 versions etc) it will not solve the problem, since somebody could install the “expert” version, and then sell it to somebody else who doesn’t have their grid powered externally.
the ideal scenario is if tt can detect grid drawing too much power automatically - if it’s doable (i asked @tehn) then that’s exactly what i’ll do.
assuming it’s not possible i still think an explicit confirmation is needed. i don’t know if it comes down to case power, and i had occasions where i had grid powered by tt result in tt freezing when too many LEDs were lit, but then it recovered after it was power cycled. so it’s not a rare scenario, and i don’t think we want to find out that it’s not always recoverable by having more busted teletypes out there.
but: i don’t think it needs to be that intrusive. here is how i think it could work: first of all, it’s not going to be a keyboard shortcut (that doesn’t really makes sense as you’re right, forgetting to do it would necessitate switching back to keyboard, doing the confirmation, switching to grid - really intrusive).
instead it will ask for a confirmation whenever grid is plugged in for the first time after teletype is powered on. once confirmed you won’t have to do it again, until tt is powered down. the confirmation will be done on grid itself, tt will show a screen with the warning and ask you to press to confirm (i’m thinking for confirmation it could be pressing 4 corner buttons simultaneously).
I thought about this exact scenario this morning.
I think that this sounds like a wonderful solution.
One alternative possibility might be to try setting the max USB current limit differently for the teletype. IIRC a single setting in the 500-600 mA range is used across all the modules. Setting it lower for the teletype would then likely trip the over current protection; shutting off the +5v for the current device, if a grid is plugged in and the leds are lit up…
that’s a great idea and easy to test, i’ll give it a try! if this indeed works then it could even make it okay to connect grid directly if your scene doesn’t use a lot of LEDs.