<Ansuel>
trying to find the best words to not be rude and say that he is proposing a very bad solution to the problem
<PaulFertser>
I'd say if the board features can't be autodetected run-time then different DTs are needed. But in this particular case, why not just enable both SATA and PCIe, one will have no link on boot, so what.
<Ansuel>
PaulFertset to me the fact that hw mod are required (the 4 capacitor tweak) mean that they are 2 different device based on the hw configuration
<Ansuel>
but yes I didn't understand where is the problem in providing boith SATA and PCIe
<Ansuel>
that's why i'm asking if someone have a clean idea on how to handle this
Tapper has quit [Read error: Connection reset by peer]
<PaulFertser>
btw, literally "moving" capacitors is a risky move, small caps like that get damaged easily during desoldering.
<Ansuel>
and also the trace goes rip with the wrong solder and temp
<Ansuel>
that is why my idea of that board is that there should be 2 version and the user should not tweak with that kind of stuff
Tapper has joined #openwrt-devel
valku has joined #openwrt-devel
<mrkiko>
Ansuel: yeah, definitely better; or why not use a GPIO or something?
<mrkiko>
I tend to have a strong opinion on those matters - since it would make things much morecomplicated for me if I find myself in the need to switch configuration
ldir has joined #openwrt-devel
m has quit [Remote host closed the connection]
m has joined #openwrt-devel
<PaulFertser>
mrkiko: switching high-speed buses isn't trivial, you'd need a relatively expensive IC to multiplex between them.
<PaulFertser>
So it would add price to each board. While with the solution they're using it's basically free.
hanetzer has quit [Ping timeout: 480 seconds]
<mrkiko>
PaulFertser: understandable and has it's own merits. Still I prefer to pay for that, since it would actually allow me to use the feature when I need it, without asking for help, potentially wait long time and so on
<PaulFertser>
mrkiko: my point is that _every_ customer would have to pay for a feature that very few would really need.
<mrkiko>
PaulFertser: and of those, we should hope _all_ of them have very good soldering skills :D
<PaulFertser>
mrkiko: again, no, because most often than not they'll choose the SATA or PCIe variant according to their needs.
<mrkiko>
PaulFertser: yes, I've seen that news :D
sorinello has joined #openwrt-devel
<PaulFertser>
mrkiko: btw, what ideas come to your mind which could be implemented on this Unielec board with an add-on PCIe x1 card?
<mrkiko>
PaulFertser: well, I continue to see it as a choice making things more difficult. But then, if they like it this way... it's fine:)
hanetzer has joined #openwrt-devel
<mrkiko>
PaulFertser: I'm not expert enough on specifics of PCIe cards specs and form factor - mind you, I need to ask for help even to connect a mini-PCIe LTE module to a router :D
<PaulFertser>
I see it lacks integrated 5 GHz wifi card, so probably that is the most obvious.
<mrkiko>
PaulFertser: and then there's always someone building nice things with openwrt I guess
<PaulFertser>
mrkiko: usually those modules are using USB, not PCIe. M.2 connectors for wireless devices can have both, but only USB is actually used by LTE modems usually.
<PaulFertser>
(same about mini-PCIe)
<mrkiko>
PaulFertser: sure, maybe the most modern ones also use PCI and so on ... but being blind, for me the problem tends to be in how I can connect things
m has quit [Remote host closed the connection]
m has joined #openwrt-devel
ldir has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
danitool has quit [Quit: Cubum autem in duos cubos, aut quadratoquadratum in duos quadratoquadratos]
GNUmoon2 has quit [Remote host closed the connection]
<Ansuel>
Mangix thanks for the friendly ping ahahha
<Ansuel>
i'm testing a qca8k and check the pr sorry
<mangix>
sure. I'm looking at replacing my RT3200 with that
<mangix>
I kind of regret buying the RT3200
<Ansuel>
mhhh why
<mangix>
I didn't know the wifi was going to be a downgrade
<mangix>
2.4ghz specifically
<Ansuel>
bad driver or hw limitation?
<mangix>
I think both
zatwai has joined #openwrt-devel
<mangix>
so the WAX202 uses mt7915 for 2.4ghz wifi. RT3200 uses mt7603
<slh>
the 2.4 GHz of the rt3200 is provided by the mt7622bv SOC, which only ships with an 'old' 802.11n implementation, compared to 802.11ax on 2.4 GHz
<Ansuel>
ouch that's bad :(
<mangix>
I completely understand why they did it but the result is bad wifi on 2.4ghz
<Ansuel>
and that is bad considering how 2.4 is still that important
<mangix>
yeah a lot of IoT stuff uses 2.4ghz
<mangix>
TV and printer too...
<slh>
to be fair, most of the clients that will need 2.4 GHz, wouldn't make use of any 802.11ax anyways, but still - if you have 802.11ax capable 2.4 GHz clients, it's a significant speedup
<Ansuel>
nha even range 2.4 is superior
<mangix>
yeah it is
<Ansuel>
fun thing with my ax3600 with 2.4 i can use my wifi from the stree and the router is at third floor LOL
<mangix>
where this laptop is 2.4ghz on mt7915 was usable. not with mt7603
<mangix>
Ansuel: I think the other issue is that with mt7603, mediatek did close to no development. mt7915 is full of them OTOH.
<mangix>
the one downside to the WAX202 is no 160MHz bandwidth
<mangix>
funny thing is, even on the RT3200 there's no speed improvement.
<mangix>
well, last I tested anyway.
<Ansuel>
aaaand i was trying to powering my r7800 with no power cable attached ...
<slh>
hehe, did pretty much the same thing earlier tonight - wondering why the USB wlan stick wasn't working, replugged it from the dock/ extension, before noticnig that the dock wasn't plugged into the computer...
zatwai has quit [Ping timeout: 480 seconds]
<Ansuel>
i was just getting angry with the poor usb serial thing that wasn't giveming me serial output... poor red card :(
<Ansuel>
giveming ???
<mangix>
Ansuel: I had a Linksys E7350 that wasn't giving me output on serial. Ended up frying it from hooking up the wrong pins.
<Ansuel>
vcc ^
<mangix>
yeah not the first time I've done it either.
<Ansuel>
i wonder what should be the use of vcc on these board
<mangix>
most recent thing of mine that fried was a monitor. A short appeared on the circuit board that I "fixed" by pouring alcohol. But now there's a random burned part which seems to be needed. I don't have the equipment to replace it.
<mangix>
Monitor was extremely interesting. There were two LVDS connections on it.
<mangix>
it's some 3 pin thing. I can't make it out.
<Ansuel>
that is a gamenexus mat ?
<mangix>
in the back yeah
<Ansuel>
ok found the burn't thing
<Ansuel>
also to me the fuse seems blown
<Ansuel>
did you checked that?
<mangix>
what fuse?
<Ansuel>
it's too yellow for me
<Ansuel>
on the left
<Ansuel>
resistor with 0 on it
<mangix>
top left?
<mangix>
oh that one
<Ansuel>
yep
<mangix>
forgot how to check if it's blown. continuity test?
<Ansuel>
yes
<Ansuel>
it should be shorted
<slh>
ssems to be a voltage regulator that has blown up
<Ansuel>
do you think it's a regulator ?
<mangix>
I think the 0 resistor is fine
<mangix>
says .000 ohms
<Ansuel>
to me it seems a transistor anyway it should have not compromised the pcb
<Ansuel>
ok it's good then
<mangix>
well, voltage regulator sounds important
<Ansuel>
well if we read the pcb
<Ansuel>
it's something related to 5v line or 12v
<mangix>
so the end result when I hook it back up is that the backlight works, but there's no OSD. There's also no image from the computer.
<Ansuel>
if that's a transistor then it's not passing voltage so it makes sense that nothing works same thing if it's a voltage regulator
<Ansuel>
also funny that the shorted component is just where the oem save some money with using lower tier component ahahah
<mangix>
oh the story of how it fried is even better
<Ansuel>
?
<mangix>
the stock PSU was 12V4A. I was using the turris omnia PSU on it, which is 12V3.33A. I noticed it was getting quite hot. Probbaly an efficiency thing. So I switched back to the cheapo stock china PSU, which ended up getting even hotter. Probably the reason why the thing shorted.
<Ansuel>
conspiracy time it just broke where they planned to :D
<mangix>
looking back on it, the turris omnia PSU is way better. the wire is thicker and there's a ferrite core.
<mangix>
the breakage happened maybe 2 hours after I switched to the stock PSU
<Ansuel>
well you couldn't imagine it was that fragile
<Ansuel>
after all it's a monitor
<mangix>
sure.
torv has quit [Remote host closed the connection]
<Ansuel>
anyway for someone who repair pcb that, he should be able to fix that in less than 1 minute
<Ansuel>
and even understand the component without schematics
zatwai has joined #openwrt-devel
<mangix>
yeah haha need to find one
torv has joined #openwrt-devel
<mangix>
the whole reason I poured alcohol was to find where the short was, not to fry the shorted chip. Apparently alcohol is an alternative to using a thermal camera.
<Ansuel>
yes
<Ansuel>
also flux is a good alternative
<mangix>
haha I don't have any unfortunately.
<mrnuke>
When is alcohol not a good alternative?
<mangix>
would make router soldering easier
<Ansuel>
when you don't have thermal camera
<Ansuel>
and you have no idea where the short is
<mangix>
I tried to use the finger method
<Ansuel>
Mangix !! never solder without flux it's a suicide
<mangix>
Didn't work
<mangix>
Ansuel: what do you mean?
<Ansuel>
always use flux to solder
<mangix>
AFAIK flux is mostly for avoiding bridges. And getting the solder to flow better.
<Ansuel>
it does also lower the melting temp and produce less burned stuff in general you only get benefits by using that
<Ansuel>
just clean the board with alchool after the work is done
<f00b4r0>
Mangix: "burnt thing" is a SOT-23 package labelled "Q8" on silkscreen. That's likely a transistor.
<Ansuel>
yheee i was right
<Ansuel>
so it makes sense that it's not working no voltage is passing or even the wrong one (but by the looks of it it's not connected at all as the trace burnt to hell and is not connected
<f00b4r0>
given I also notice a "5V" label next to it, it wouldn't suprise me if it where a high-side switch
<f00b4r0>
on the 5V rail. It going toast could then be the result of a short.
<f00b4r0>
but ofc this this is pure late-night speculation from my side ;)
<mangix>
f00b4r0: better than nothing :)
<mangix>
hmmm how do I try continuity testing this transistor?
<Ansuel>
Your branch and 'net-next/master' have diverged, and have 23 patches... they will fk me so bad for sending a series this big
<f00b4r0>
also I read up that you poured alcohol, unless it was 99.9% IPA, that's not a very good idea. Commercial alcohol has water content.
<mangix>
f00b4r0: 91%
<f00b4r0>
yuck.
<Ansuel>
Mangix you can't you should use a dedicated tester... but that is 100% broken anyway
<f00b4r0>
yeah it's dead Jim.
<mangix>
so Black side bottom red side to the ride shows a resistance of .015
<mangix>
s/side/right
<mangix>
s/ride/right
<Ansuel>
thins is that to me it seems gate is 5v
<Ansuel>
right is GND
<Ansuel>
and left goes to via... rip no idea what is the use of that...
<f00b4r0>
if it is a high side switch it's gonna be a mosfet of sorts. But anyway, it's gone.
<mangix>
the component or the board?
<Ansuel>
component
<Ansuel>
board is good
<f00b4r0>
and I'm gonna try to catch some zzzs while the outside temperature allows; forecast is another day of 40+ tomorrow :(
bluew has joined #openwrt-devel
<f00b4r0>
yeah board is probably good.
<mangix>
Ansuel: board is important. It gives me 1440px165Hz
<Ansuel>
the fuse got some damage but it's not dead so.... (it's not normal that is that yellow)
<mangix>
wait...the 0 is a fuse?
<Ansuel>
yep
<Ansuel>
got some stress but nothing bad
<Ansuel>
the bad power suplly rip the transistor
<mangix>
not sure how to understand that. so it should have blown instead of the regulator?
<Ansuel>
if it was something really bad the transistor should be exploded and the fuse open
<f00b4r0>
a 0-ohm resistor is not a fuse
<Ansuel>
i mean it's a bridge right? but it should open with some stress condition
<mangix>
forgot what 0 ohm resistors are for. detecting power IIRC
<f00b4r0>
especially when used in a spot labelled with "Lnn" which is an inductor label
<f00b4r0>
no, they serve as jumpers
<f00b4r0>
anyway, I'd be happy to help, only later :)