| + alex4nder (~alexander@ip98-182-18-230.sb.sd.cox.net) | 00:03 | |
| chartreuse | Alright got the LARID antennta now, now to restart and install it. And debate on if I want to swap the audio caps again, really don't want to damage the pads | 00:54 |
|---|---|---|
| - chartreuse (QUIT: Quit: leaving) (~chartreus@S0106f0f249dfd9c3.cg.shawcable.net) | 00:56 | |
| - alex4nder (QUIT: Ping timeout: 268 seconds) (~alexander@ip98-182-18-230.sb.sd.cox.net) | 01:46 | |
| - nio (QUIT: Ping timeout: 252 seconds) (~nio@2001:470:69fc:105::172d) | 01:54 | |
| - cryptix (QUIT: Ping timeout: 252 seconds) (~cryptxxma@2001:470:69fc:105::94a) | 01:54 | |
| - indefini[m] (QUIT: Ping timeout: 268 seconds) (~indefinim@2001:470:69fc:105::1e2a) | 01:55 | |
| - scops (QUIT: Ping timeout: 272 seconds) (~scopstchn@2001:470:69fc:105::8da) | 01:55 | |
| - jryans (QUIT: Ping timeout: 268 seconds) (~jryans@2001:470:69fc:105::1d) | 01:55 | |
| + indefini[m] (~indefinim@2001:470:69fc:105::1e2a) | 02:12 | |
| + darth-cheney (~user@2603-7000-8d00-1f72-0000-0000-0000-1c80.res6.spectrum.com) | 02:23 | |
| + chartreuse (~chartreus@S0106f0f249dfd9c3.cg.shawcable.net) | 02:31 | |
| chartreuse | Alright 330uF caps installed instead of the 100uF ones, it's a nice improvement, especially with 16 ohm earbuds. I think 100-220uF is probably the sweet spot for most people as the caps a bit smaller. | 02:34 |
| chartreuse | Also got the LAIRD antennas installed, since I had a 3 channel I bought a 3rd single antenna for a 3x3 MIMO setup, though mistakenly got one with the wrong connector, had to desolder it and replace with the coax and connector from the cheap ones I had before | 02:35 |
| chartreuse | I think this is the nicest sound I've had out of a laptop. The nice Wolfson dac and the improved dc blocking caps are great | 02:36 |
| + scops (~scopstchn@2001:470:69fc:105::8da) | 03:13 | |
| + nio (~nio@2001:470:69fc:105::172d) | 04:08 | |
| + jryans (~jryans@2001:470:69fc:105::1d) | 04:11 | |
| - wagga (QUIT: Ping timeout: 246 seconds) (~wagga@node-1w7jra22ildhx3flif1ae7one.ipv6.telus.net) | 04:12 | |
| + cryptix (~cryptxxma@2001:470:69fc:105::94a) | 04:14 | |
| - erlehmann (QUIT: Ping timeout: 272 seconds) (~erle@dynamic-046-114-035-003.46.114.pool.telefonica.de) | 05:39 | |
| + erlehmann (~erle@dynamic-046-114-036-002.46.114.pool.telefonica.de) | 05:51 | |
| + alex4nder (~alexander@ip98-182-18-230.sb.sd.cox.net) | 07:38 | |
| - alex4nder (QUIT: Client Quit) (~alexander@ip98-182-18-230.sb.sd.cox.net) | 07:38 | |
| + alex4nder (~alexander@ip98-182-18-230.sb.sd.cox.net) | 07:39 | |
| - alex4nder (QUIT: Client Quit) (~alexander@ip98-182-18-230.sb.sd.cox.net) | 07:39 | |
| + wiedi (~wiedi@37.228.191.159) | 07:42 | |
| - darth-cheney (QUIT: Ping timeout: 245 seconds) (~user@2603-7000-8d00-1f72-0000-0000-0000-1c80.res6.spectrum.com) | 07:43 | |
| + natalie- (~natalie@user/natalie) | 08:36 | |
| - natalie (QUIT: Ping timeout: 268 seconds) (~natalie@user/natalie) | 08:37 | |
| * natalie- -> natalie | 08:37 | |
| mntmn | chartreuse: that's cool, i will have to play with that @ caps. i was pretty happy with the sound already though | 09:27 |
| chartreuse | With speakers it probably sounds fine as they'll have a higher impedence, but with headphones I found it quite lacking in the low end | 09:34 |
| chartreuse | With the stock ones it's a high pass filter with a -3db point of 100-220Hz for standard headphones | 09:36 |
| chartreuse | The 47uF you have seems to be the largest MLCC cap you can get in 0603 that'd fit right on the pads so it's not as clean of an upgrade | 09:37 |
| mntmn | yeah, noted | 09:39 |
| chartreuse | Still have to get myself sorted to try looking into getting that feature added to the video driver for WebRenderer. Probably will give it a shot on the weekend. At least give it a shot while I still have the system booting from SD so trying the proprietary driver is easy | 09:43 |
| mntmn | i haven't tried the proprietary driver on reform though, only on the imx8m evk. i believe you need a vendor kernel and slightly tweaked devicetree to make that work i think. i'm not sure if the blob driver's kernel interface can be plugged into the mainline kernel. | 09:52 |
| mntmn | by vendor kernel i mean a linux fork by nxp / boundary devices. | 09:53 |
| chartreuse | Ah okay, yeah might be a bit over my head at the moment, but I'll look into it. Maybe see if it could be insmodded into a running system | 09:56 |
| chartreuse | Kinda another silly mod idea, but was thinking I could swap the lights under the trackball switches to be 365nm UVC leds and have that slight glow effect to the caps. Too bad the Kalih switches don't just use 3mm through hole leds or it'd be a bit easier of a mod | 10:17 |
| - wiedi (QUIT: Read error: Connection reset by peer) (~wiedi@37.228.191.159) | 11:24 | |
| mntmn | chartreuse: haha that's creative | 11:33 |
| marex | vivante kernel side vendor driver is ... uh ... code quality could be improved | 11:57 |
| marex | chartreuse: what are you missing /wrt the webrenderer ? | 11:58 |
| mntmn | chartreuse: TXF operation | 12:00 |
| mntmn | sorry | 12:00 |
| mntmn | marex: TXF operation | 12:00 |
| mntmn | marex: Unhandled NIR tex src type: 3, Unhandled NIR tex type: 4 | 12:01 |
| mntmn | that's nir_tex_src_offset and texelFetch | 12:01 |
| marex | mntmn: so probably better discussed on #etnaviv ? | 12:02 |
| marex | on oftc | 12:02 |
| mntmn | yeah | 12:02 |
| mntmn | just asked there | 12:04 |
| - S0rin (QUIT: Ping timeout: 268 seconds) (~S0rin@user/s0rin) | 12:06 | |
| + S0rin (~S0rin@user/s0rin) | 12:07 | |
| scops | hmmm after about not using the reform for about a week my batteries now died x.x is there a recommentation where to buy new ones in germany? | 12:08 |
| scops | it's really problematic that the batteries could just die if someone dont uses the reform for a few days... i hope there will be a solution with the deep sleep functionality you mentioned mntmn in this thread ( https://community.mnt.re/t/battery-is-not-charging/270/10) soon | 12:11 |
| mntmn | scops: i know! | 12:11 |
| mntmn | scops: 1800mAh are out of stock at the moment but you can get 1600mAh https://www.nkon.nl/de/rechargeable/li-ion/18650-size/jgne-18650-1600mah-4-8a-lifepo4.html | 12:12 |
| mntmn | quite cheap | 12:13 |
| scops | thanks :) | 12:13 |
| mntmn | scops: so far everyone i asked for assisting me with undervoltage lockout for the battery packs has kind of ghosted me on the topic so i will have to do it myself at some point | 12:14 |
| mntmn | :D | 12:14 |
| scops | i wish i could help you with that ... there are is so much around building something like the reform... your company must grow up x) | 12:16 |
| mntmn | yeah, first we must ship. but today marks the 150 units shipped milestone | 12:17 |
| mntmn | chartreuse: https://github.com/etnaviv/etna_viv/blob/master/rnndb/isa.xml#L500 | 12:17 |
| mntmn | chartreuse: you might not need to run the blob after all... | 12:18 |
| + wagga (~wagga@node-1w7jra22ildhyczqx84usquiq.ipv6.telus.net) | 12:38 | |
| chartreuse | For the undervoltage thing, would it be possible to sense the voltage going too low, and shut down the everything such that the only way to power on is to open the case and hit the reset button? | 12:55 |
| chartreuse | Would be at least better than draining the batteries completely | 12:55 |
| chartreuse | Basically disable interrupts and halt the processor in whatever lowest power draw state it has | 12:56 |
| mntmn | chartreuse: maybe possible, but also this needs to be signalled to the keyboard, that it would shut down as well. luckily we have serial comms between lpc and keyboard | 12:57 |
| chartreuse | Ideally there'd also be an IC to just cut power completely if the cell voltage ever goes below a certain state, and only resume power when it goes above from the charger | 12:57 |
| mntmn | chartreuse: so LPC could tell the keyboard to shut down | 12:57 |
| mntmn | chartreuse: yes, and this IC / circuit would need to be implemented on a replacement battery board | 12:58 |
| chartreuse | So when the reform is currently 'off' the LPC, the keyboard, and the trackball are currently running? | 12:58 |
| chartreuse | No wait the trackball is just over usb | 12:58 |
| mntmn | chartreuse: not trackball. only the 3V section of the keyboard | 12:58 |
| mntmn | chartreuse: the atmega32u4 has 2 power supplies. 5V for the usb section and 3V3 for the rest of the mcu | 12:59 |
| chartreuse | Ah okay | 12:59 |
| mntmn | chartreuse: if the 5V is offline then the usb is dead, but it respawns once it gets 5V supplied | 12:59 |
| chartreuse | Is the circle button anything special on the matrix? Like is it on a interrupt pin on the atmega? | 12:59 |
| mntmn | chartreuse: the 3V3 for this is a standby 3V3 rail that also powers the LPC and the battery monitor | 12:59 |
| mntmn | chartreuse: nope | 12:59 |
| chartreuse | Just was thinking maybe a happy coincidence that the column it was on would be an interrupt one | 13:00 |
| mntmn | ah, this idk | 13:00 |
| chartreuse | So leave the controller with just its row powered and enter standby waiting for a pin interrupt from its column | 13:01 |
| mntmn | yeah, that could work | 13:01 |
| chartreuse | Do you know what the row/column the circle key is on? The kicad schematic is unlabelled | 13:04 |
| mntmn | first row, last column | 13:05 |
| mntmn | ah wait | 13:05 |
| mntmn | it's SW78 | 13:06 |
| mntmn | col14 | 13:07 |
| mntmn | row1 | 13:07 |
| chartreuse | Alright thanks, just reading through the atmega32u4 datasheet to see what the interrupt pins are | 13:07 |
| chartreuse | ROW1 is on PB6, so need to see what that pins capabilities are | 13:09 |
| chartreuse | It's not one of the "proper" interrupt pins but it is PCINT6 so it can do a pin change interrupt | 13:12 |
| chartreuse | Now to see if that interrupt can be used to wake the system | 13:12 |
| chartreuse | I don't think you're using any of the pin change interupts at the moment so it shouldn't be wrong to take that vector to wakeup | 13:13 |
| chartreuse | > Pin change interrupts on PCINT7 ..0 are detected asynchronously. This implies that these interrupts can be used for waking the part also from sleep modes other than Idle mode. | 13:13 |
| chartreuse | Sounds like this could be done. Basically set col 14 as outputting high, set enable pcint6 interrupts, then go into a sleep mode | 13:15 |
| chartreuse | Yeah so PCI0 will fire when any of the PCINTx conditions trigger and even if it's a level trigger that can be used to wake | 13:16 |
| chartreuse | And the interrupts keep working in the "Power-Down" mode state | 13:18 |
| chartreuse | In this mode, the external Oscillator is stopped, while the external interrupts, the 2-wire Serial Interface, and the Watchdog continue operating (if enabled). Only an External Reset, a Watchdog Reset, a Brown-out Reset, 2-wire Serial Interface address match, an external level interrupt on INT6, an external interrupt on INT3:0, a pin change interrupt or an asynchronous USB interrupt sources | 13:18 |
| chartreuse | (VBUSTI, WAKEUPI), can wake up the MCU. | 13:18 |
| chartreuse | In power-down mode with the watchdog disabled it'd draw around 0.45uA at 3.3v | 13:22 |
| chartreuse | Though what I'm not certain on is if it can still output a voltage on a pin, or use the internal pullup to do so (probably more ideal) | 13:23 |
| chartreuse | From what I'm seeing the answer is yes, so a fairly low power state can be achieved with the keyboard controller | 13:26 |
| chartreuse | I guess the next would be what can be done with the LPC to have it in as low of power as possible state | 13:27 |
| chartreuse | I think I'll give this a try, I'm fairly comfortable with AVR microcontrollers, and it's pretty easy to access compared with the embedded controller on the processor board | 13:36 |
| mntmn | chartreuse: that sounds cool | 13:37 |
| mntmn | chartreuse: i've already tested putting LPC in sleep mode, that def. works and reduces power consumption a lot, but i haven't tried to wake it up yet | 13:38 |
| mntmn | // deep sleep test | 13:39 |
| mntmn | /*LPC_PMU->PCON = 0x1; | 13:39 |
| mntmn | SCB->SCR |= SCB_SCR_SLEEPDEEP_Msk; | 13:39 |
| mntmn | __WFI();*/ | 13:39 |
| chartreuse | Might be possible to set a wakeup from deep sleep to be one one of the serial lines from the keyboard controller | 13:39 |
| mntmn | yeah | 13:39 |
| chartreuse | And we just use the serial turning on to wake it | 13:39 |
| mntmn | but! i'm not sure how safe charging is with lpc sleeping | 13:40 |
| mntmn | so it has to wake up periodically | 13:40 |
| mntmn | like, something has to check if wall power is back | 13:40 |
| chartreuse | Should be a significant savings just with the keyboard, seems like it's always just running full blast 16Mhz, it'd be drawing close to 10mA, though I don't think it can run at 16MHz at 3.3v so not sure what it's drawing | 13:41 |
| chartreuse | Is the LPC the battery charge controller? I thought there was a second IC for that | 13:42 |
| chartreuse | Something that'd prevent overchange, though I guess the balancer needs to be controlled? | 13:42 |
| chartreuse | The LPC might have a watchdog timer that can be run even in its lowest power mode, which would wake it up after a delay | 13:43 |
| chartreuse | If you want to check on charging. | 13:43 |
| chartreuse | It'd be nice though to enter this low power mode any time the laptop is off, just so it doesn't drain the battery so much while off. Not just as an emergency "don't kill the cells" mode | 13:44 |
| mntmn | chartreuse: yep, this is what i would like @ watchdog timer | 13:44 |
| chartreuse | Having to draw a few mA every 10s of so for a fraction of a second is still a sigificant drop over running constantly | 13:44 |
| mntmn | chartreuse: and i agree @ low power mode any time the laptop is off | 13:45 |
| chartreuse | The keyboard should be able to run in power-down mode without the watchdog since the circle key will act as an interrupt | 13:45 |
| mntmn | yeah | 13:45 |
| chartreuse | I'll try and get that working since that alone would save quite a few mA draw on the cells | 13:46 |
| mntmn | awesome. i can look into the LPC again | 13:46 |
| mntmn | btw did you see that TXF opcode was already identified @ gc7000l? it "just" needs to be wired up in mesa | 13:46 |
| chartreuse | I'm wondering if you're overclocking the keyboard controller somewhat, it's got a 16MHz crystal but running at 3.3v? | 13:47 |
| chartreuse | Yeah, certainly worth a try to implement it, and hope that's all firefox needs | 13:47 |
| mntmn | chartreuse: i think there's a mode where you can use an internal oscillator instead | 13:47 |
| mntmn | chartreuse: if it's not all it needs, we'll know what it needs next. | 13:47 |
| chartreuse | Yeah, though I was wondering with how it's going currently. Even though the USB supply is 5v, the micro is running at 3.3v | 13:48 |
| chartreuse | And it's only rated at ~10-12 at 3.3v | 13:48 |
| mntmn | huh. | 13:50 |
| chartreuse | Page 386 of the datasheet shows the graph depending on vcc | 13:51 |
| chartreuse | It's 8MHz at 2.7v and 16MHz at 4.5v+ and linear between | 13:51 |
| chartreuse | It seems to be a relatively common overclock to do 3.3v at 16MHz though it is a fair bit out of spec | 13:54 |
| chartreuse | Should really be down at 8MHz or 10 | 13:55 |
| mntmn | ok | 13:59 |
| + wiedi (~wiedi@37.228.191.159) | 14:07 | |
| + mjw (~mark@83.162.59.96) | 14:26 | |
| - wagga (QUIT: Quit: Client closed) (~wagga@node-1w7jra22ildhyczqx84usquiq.ipv6.telus.net) | 15:36 | |
| + wagga (~wagga@node-1w7jra22ildhyczqx84usquiq.ipv6.telus.net) | 15:56 | |
| wagga | mntmn: Thank you for helping me sort out my email address issue yesterday. I now have a Crowd Supply account and access to my order. I purchased the mnt-reform DIY kit, but forgot to order a wifi-card and trackball module. I can add the trackball module to my cart, but not the wifi card. How do I create a ticket with crowd source about this? | 16:05 |
| mntmn | wagga: there is some "change my order" form, google "crowd supply change my order" | 16:06 |
| wagga | Thanks for googling that for me. Sorry, I didn't think to do that first. | 16:07 |
| wagga | The ath9k wi-fi card says "Preview only: check back soon!". The pre-order button is there, but I bet the fact that it says preview only explains why I am not able to add it to my cart. | 16:12 |
| wagga | I hadn't noticed that before. | 16:12 |
| mntmn | oh, strange | 16:21 |
| mntmn | maybe they haven't decided on where to get that card from, as they didn't buy any from us | 16:21 |
| mntmn | i can bring this up internally soon | 16:21 |
| wagga | Okay, thanks. Is the card available directly from you guys? | 16:26 |
| mntmn | not yet at least | 16:31 |
| mntmn | wagga: if you google for "Compex WLE200NX" there are some stores that have it | 16:33 |
| jackhill | yep, I got one on ebay in the USA. It was alledgedly new old-stock; pulled out of some unsold connected devices. | 16:36 |
| jackhill | I kind of wish I had purchased directly form mtn research, but I was sure I already had a card onhand (I didn't). I as happy with the card I got, but the antennae were another goose chase for me. | 16:37 |
| - mjw (QUIT: Ping timeout: 268 seconds) (~mark@83.162.59.96) | 16:40 | |
| + mjw (~mark@83.162.51.9) | 16:41 | |
| - mjw (QUIT: Remote host closed the connection) (~mark@83.162.51.9) | 16:42 | |
| wagga | mntmn: Thanks! | 16:51 |
| + darth-cheney (~user@2603-7000-8d00-1f72-0000-0000-0000-1c80.res6.spectrum.com) | 17:40 | |
| darth-cheney | hey gang, what's the correct way to put the reform into sleep/suspend/hibernate modes? | 17:40 |
| mntmn | darth-cheney: reform-standby | 17:41 |
| mntmn | darth-cheney: but there is a problem with waking it up | 17:41 |
| mntmn | darth-cheney: like, reliably. sometimes it works, sometimes it doesn't. | 17:41 |
| darth-cheney | aha I see | 17:41 |
| darth-cheney | mntmn: where would I find the script? It does not appear to be on my path | 17:42 |
| eery | sounds similar to linux suspend/resume support in general IME :( | 17:43 |
| mntmn | darth-cheney: su; which reform-standby | 17:43 |
| mntmn | eery: well on imx6 it was working quite well | 17:43 |
| mntmn | darth-cheney: it's in /usr/sbin | 17:44 |
| darth-cheney | yeah aha I was looking in /sbin for whatever reason | 17:45 |
| darth-cheney | One day I'm going to be able to figure out the unix directory structure meaning | 17:45 |
| mntmn | darth-cheney: apt-file show reform-tools | 17:55 |
| - darth-cheney (QUIT: Ping timeout: 252 seconds) (~user@2603-7000-8d00-1f72-0000-0000-0000-1c80.res6.spectrum.com) | 18:56 | |
| - reform3179 (QUIT: Ping timeout: 245 seconds) (~jamin@2603-8000-b400-1b02-06f0-21ff-fe91-0625.res6.spectrum.com) | 19:36 | |
| + darth-cheney (~user@2603-7000-8d00-1f72-0000-0000-0000-1c80.res6.spectrum.com) | 19:40 | |
| - darth-cheney (QUIT: Ping timeout: 272 seconds) (~user@2603-7000-8d00-1f72-0000-0000-0000-1c80.res6.spectrum.com) | 20:29 | |
| - wagga (QUIT: Quit: Client closed) (~wagga@node-1w7jra22ildhyczqx84usquiq.ipv6.telus.net) | 20:54 | |
| + reform3179 (~jamin@2603-8000-b400-1b02-06f0-21ff-fe91-0625.res6.spectrum.com) | 21:34 | |
| - reform3179 (QUIT: Client Quit) (~jamin@2603-8000-b400-1b02-06f0-21ff-fe91-0625.res6.spectrum.com) | 21:35 | |
| + mjw (~mark@83-161-179-12.mobile.xs4all.nl) | 23:02 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!