<azonenberg>
so i bought 50 edge launch SMAs around this time last year
<azonenberg>
when i'm done assembling all of the in-progress prototypes, i think they'll be just about gone
<azonenberg>
I think i have a problem :p
<monochroma>
azonenberg: stop using them as power connectors too just because you have a lot of them :<
<azonenberg>
lol
<azonenberg>
seriously, the AFE prototype has 9
<azonenberg>
the HMCAD prototype has 10
<azonenberg>
the probe characterization board has 9
<azonenberg>
so that's 28 just in boards i'll have built yesterday and today
<azonenberg>
(probe test board just came in btw, assembling when i get back from grocery shopping)
<lain>
:D
<monochroma>
12 on the inline HDMI probe, 8 on the inline ethernet probe
<azonenberg>
Yeah
<azonenberg>
So you can see how burning through 50 in a fairly short time isn't that hard if you do lots of T&M projects
<monochroma>
8 on the ethernet cross talk characterization board :P
<azonenberg>
Yeah. fwiw, i bought that 50 around this time last year and i was not out at the time
<azonenberg>
But once i build all these prototypes, i'll have enough to assemble a handful of the new probes
<azonenberg>
and that's it
<azonenberg>
i should throw some on my next digikey order
<azonenberg>
(for reference, that's $181 of SMAs lol)
<azonenberg>
At $3.63 each
<azonenberg>
$4.33 @ qty 1 but they drop down to $3.02 @ 250 and $2.94 @ 500. too bad i dont have the cash to spare or i'd just buy a couple hundred and save
Degi has quit [Ping timeout: 264 seconds]
Degi has joined #scopehal
<azonenberg>
lain, degi: probe cal board assembled, looks good
<azonenberg>
adc board back side just came out of reflow, waiting for it to cool
<azonenberg>
will probably not do front until tomorrow since i'm getting tired and dont want to make an expensive mistake
<azonenberg>
but tonight i wanted to do some xaVNA measurements of the probe on the cal board to see if i can get more stable results
<azonenberg>
probably using the crystek hand formable stuff
<azonenberg>
ok so xavna measurement don't really tell me a whole lot i didnt already know
<azonenberg>
given how much i trust the details of its measurements
<azonenberg>
it did validate that my test board was soldered right which i guess is a start :p
<azonenberg>
it showed some rolloff around 2.5 GHz, but i don't really trust the cal algorithms on the xavna at all
<azonenberg>
they've created massive nulls that didn't actually exist
<azonenberg>
so i basically consider it garbage at this point and i'm waiting until i get the pico :p
Implant has joined #scopehal
<Degi>
Heh neat
<balrog>
azonenberg: have you reached out to them about these issues?
<azonenberg>
balrog: no, not bothering because the pico is coming in a week-ish
<azonenberg>
and the XA has fundamental b/w limitations that also make it problematic to me
<balrog>
I mean, might be worth reaching out so that they are aware there is an issue
<azonenberg>
like the ~130 MHz floor, which makes it totally worthless for a 100 MHz AFE :p
<azonenberg>
i'll wait till i get side by side comparison data
<balrog>
alright
bvernoux has joined #scopehal
<bvernoux>
hello
<bvernoux>
I have just received my soldering Station AOYUE Int968A+
<bvernoux>
very nice even better than the previous 968
<bvernoux>
with solder tweezer it is amazingly fast to remove smd capacitor...
<bvernoux>
I have also different Solder SAC305 will test that as it is what is recommended now
alexhw has quit [Ping timeout: 265 seconds]
<Degi>
Huh cool soldering iron with smoke sucker
<bvernoux>
yes
<bvernoux>
the smoke sucker is very nice
<bvernoux>
I have never found that on other soldering station
<bvernoux>
especially to have all in one for a very good price and good quality product
<azonenberg>
i dont like extra stuff hanging off my irons
<azonenberg>
i'd rather have a fume extractor on the bench
<bvernoux>
yes a fume extractor on the bench is probably better as it will extract more fume
<bvernoux>
but when it is integrated it is nice too as anyway it can be switched off too
<Degi>
Huh outside smells really awful like somebody burned a bunch of tires or so
<bvernoux>
in fact my old soldering station Int968 still work fine after about 10years of use ;)
<bvernoux>
but I have bought the new one to use Soldering Iron Tweezer and also because Iron is 70W vs 35W on the old one
<bvernoux>
Also the new Int968A+ is better with less noise ;)
<bvernoux>
hmm I'mnot lucky the received AD1-10S
<bvernoux>
is dried
<bvernoux>
Mfg Date is 200108
<azonenberg>
2001??
<azonenberg>
where the heck did you buy it from
<Degi>
YYMMDD?
<Degi>
Maybe its 2020 january 8?
<azonenberg>
20-01-08 would make a lot more sense. It's a very high viscosity glue
<Degi>
At least on my flux pen it says 200220 in YYMMDD
<azonenberg>
it's easy to think that it's solidified when it's really just very thick
<Degi>
Doesnt it have like a year of shelf life
<bvernoux>
yes January 2020
<bvernoux>
so a bit old for something new ;)
<azonenberg>
i've never had problems with that glue curing during normal ambient temps
<bvernoux>
I'm pretty sure it was exposed to temperature > 30°C which will explain it is dry and cannot go outside ...
<azonenberg>
i think mine is >1 yr old
<azonenberg>
short exposures to higher temps won't hurt it unless it's like 100+
<bvernoux>
ha ok anyway there is clearly something wrong with mine
<bvernoux>
I will ask mouser to send me an other for free as this one is totally dry
<azonenberg>
are you positive?
<azonenberg>
it's very thick and pasty, thicker than solder paste
<azonenberg>
this is normal
<azonenberg>
it's not like normal epoxy
<azonenberg>
the supplied needle is imo too small to get good flow
<bvernoux>
it is totally dry
<bvernoux>
inside
<bvernoux>
I have remove a bit paste and it is dry
<bvernoux>
using a nail
<azonenberg>
New probe shells shipped
<bvernoux>
azonenberg, ha great
<azonenberg>
electronic_eel: with C15 removed, 2V5_REF is stable
<electronic_eel>
ah, nice
<azonenberg>
So that's easy fixes for 2 of the 3 known bugs
<azonenberg>
now we have the 1V8 issue
<electronic_eel>
I suggest trying to replace it with increasing caps (like 1nF, 10nF, 100nF) to see how much margin we have for stability
<Degi>
Neat
<Degi>
Hm could bigger caps make noise worse?
<azonenberg>
interestingly i measure 1.244V on 1V8 now
<azonenberg>
i swear it was less before
<azonenberg>
maybe the oscillation was causing problems there?
<electronic_eel>
check the diodes, the real ones
<Degi>
You said 0.6 something yesterday
<Degi>
Hm yeah are the output diodes correct? Are there weird voltages on the output rails?
<azonenberg>
the left two diodes have the bar on the signal trace
<azonenberg>
right two have bar on the via
<Degi>
HMm ill download the project wait a sec
<azonenberg>
on semi datasheet says cathode is pin 1 which is indicated by the bar
<azonenberg>
so we have one set with anode on ground and cathode on signal
<azonenberg>
and one set with anode on signal and cathode on 1v8
<azonenberg>
that sounds right
<Degi>
Are there weird voltages on the outputs relative to gorund?
<azonenberg>
i'm checking that now
<azonenberg>
i see 1.255V on out-
<Degi>
Positive relative to ground?
<Degi>
(Did you measure 1V8 with a scope yet?)
<azonenberg>
and a ~180 kHz waveform centered at ~915 mV on OUT-
<azonenberg>
at OUT+*
<Degi>
With what amplitudeß
<azonenberg>
~140 mV
<azonenberg>
backing up to before the final gain stage
<Degi>
Hm wait even if it was all wrong, 1V8 being less than 1V8 is still weird...
<Degi>
(As long as the diodes are correct, which they should be)
<Degi>
(On the other hand if the diode on the 915 mV thing was wrong, then 1V8 should be 1.2V approx)
* monochroma
dual wields irons instead of solder tweezers
* Degi
has a butane torch for the harder to solder components
<Degi>
When that 16 mm² cable refuses to be warmed up by your TS100
<electronic_eel>
monochroma: yes, that works too. I do that sometimes too if I have the second iron plugged in the station and are too lazy to switch over to the tweezers for just one part
<azonenberg>
I still see oscillation on TP3
<electronic_eel>
is it the same freq/amplitude or changed?
<monochroma>
electronic_eel: i havn't had the chance to use a tweezer set, they look nice for TSOP removal, thugh most of the ones i have seen look a bit bulky
<azonenberg>
electronic_eel: right pad of C70, when DNP'd showed 172 mV p-p
<azonenberg>
now it's ~90
<azonenberg>
a bit noisier i think
<Degi>
Did the frequency change
<Degi>
Wait what is R63 and C68?
<Degi>
Are they both DNP'd?
<azonenberg>
Still about the same i think. But the kHz frequency is wrong, it was apparently measuring spikes
<azonenberg>
not the fundamental sineweave
<azonenberg>
zooming in more, my scope is saying more like 60 MHz
<Degi>
(Wait that was my suggestion to put it there to shunt the RF away, but a smaller cap there may work and help with preventing oscillations)
<Degi>
Oh 60 MHz?
<azonenberg>
let me try and use a cursor measurement
<Degi>
I dont think that thats the OP07
<azonenberg>
the waveform is noisy
<electronic_eel>
60 MHz, that can't be the OP07
<azonenberg>
ok no
<azonenberg>
the 60 MHz is noise superimposed on it
<azonenberg>
the actual oscillation is still the same about. 194 kHz
<Degi>
Hm just count the divs
<azonenberg>
(it kept jumping all over the place and was hard to get a good read on)
<electronic_eel>
194 kHz is something in the range of the OP07
<Degi>
Hm yes maybe C68 isnt the best but its kinda required for very good RF performance but might even work without it
<Degi>
Cant you freeze the screen
<azonenberg>
yeah uploading screenshot now
<electronic_eel>
maybe remove R63 to see if the cap at the output is the problem
<Degi>
The idea behind that cap is to absorb RF going thru R62 but I think if we omit it the mismatch at the input at higher frequencies is still only minimal
<electronic_eel>
I think we should get it stable first
<azonenberg>
measured at right side of C70
<azonenberg>
with a 1M passive probe, ground at TP8. long alligator clip wire, not a high freq ground
<electronic_eel>
that doesn't look nice with all that noise
<azonenberg>
Let's make it stop oscillating
<azonenberg>
then we'll worry about noise
<electronic_eel>
I guess the noise is something else that is oscillating a bit
<azonenberg>
fwiw, before i added the cap there was a lot smoother waveform but it was bigger
<Degi>
Is TP8 connected to anything?
<electronic_eel>
GND?
<Degi>
Remove R63 I suggest
<azonenberg>
Degi: tp8 is the ground clip southwest of the op07
<Degi>
Hm I downloaded starshipraider and there the pad says GND but isnt connected to anything
<Degi>
Oh wait
<Degi>
Theres a via lol
<Degi>
Totallym issed that
<Degi>
Anyways removing R63 could stop the oscillations, since that would disconnect the apacitive load
<azonenberg>
Removed, let's see
<azonenberg>
let me switch to a better ground on this probe
<azonenberg>
ok so with R63 gone it looks to have mostly been eliminated
<electronic_eel>
is the 60MHz noise still there?
<azonenberg>
There's high freq noise that i think might be due to my scope
<Degi>
Hm we need to figure out a better RF filter network which shunts everything above 100 kHz to ground with < 10 Ohms and lets DC pass (below 100 KHz the OP amp should be able to handle it)
<azonenberg>
it was present when i floated the probe too
<electronic_eel>
does the probe catch something from one of the switchers maybe?
<azonenberg>
possible. Going to switch to an active probe and compare
<azonenberg>
20 Gsps is probably too fast for this anyway. let me turn the scope down a bit
<Degi>
Hm as long as you limit the bandwidth..
<Degi>
Scope frequency measurements would be so much better if it just determined the peak of the FFT
<azonenberg>
let me do an fft then
<Degi>
Nah that isnt really important, just seeing the waveform is enough tbh
<electronic_eel>
the question is where does it come from
<azonenberg>
ok so i did a bit more analysis
<Degi>
And what kinda input network will go to the OP amp, because now is the problem of the impedance at 0 Hz being nearly 10 ohm lower than at a few MHz
<azonenberg>
swapped the passive probe for a tetris
<azonenberg>
and applied a 3-bit noise filter
<azonenberg>
Which gives 3 extra bits of resolution and a -3 dB bandwidth of 160 MHz
<Degi>
Wait does that rail have a capacitor anywhere?
<Degi>
Hm yeah C69
<azonenberg>
You think the dac output buffer is oscillating?
<electronic_eel>
I don't think the DAC outputs that willingly
<Degi>
(Can DAC outputs oscillate too?)
<electronic_eel>
dac outputs are small opamps
<Degi>
Hmm if removing C69 helps maybe
<electronic_eel>
usually
<electronic_eel>
yes, C69 could be the culprit
<azonenberg>
VDAC1 looks flat
<azonenberg>
so i dont think its dac power or something
<azonenberg>
anyway i'm gonna go grab some lunch and get ready for a call for $dayjob
<azonenberg>
doing another training session
<Degi>
Hm if C69 is the culprit, we definitely need a better way to decouple RF from DC... I guess it was a bad idea to just stick a capacitor there
<azonenberg>
i was supposed to be doing an external network penetration test today but nobody ever told me what im testing so... :p
<electronic_eel>
buhh, bug hunting was fun
<azonenberg>
Will get back to this in a bit
<azonenberg>
This is progress though
<Degi>
^^
<azonenberg>
So far we have four known problems, two fixed (bad relay pinout and oscillating 2V5)
<azonenberg>
one not-touched (bad 1V8)
<azonenberg>
and are making good progress on the root cause of the input noise, now we know the offset voltage is oscillating
<azonenberg>
we just have to figure out why and fix it
<azonenberg>
C69 is certainly plausible
<Degi>
(After fixing it, we could try removing C70 maybe and yeah more complicated RF grounding filter needed...)
<electronic_eel>
azonenberg: about your external pentest - just check if they have a RDP port open, directly to their main file server
<electronic_eel>
a customer at my $dayjob had that, and got some passowr
<electronic_eel>
passwords stolen
<Degi>
OOf
<electronic_eel>
we warned them about the open rdp like a year ago, but they had it open until yesterday
<Degi>
Huh I think I just reverse-engineered the schematics of this PCB... Now to figure out why it doesnt do what it should do
<azonenberg>
electronic_eel: at least i havent seen telnet in a while
<azonenberg>
world facing at least
* Degi
just built a device with telnet..…
<azonenberg>
well thats the thing
<azonenberg>
there is nothing inherently wrong with telnet
<Degi>
I mean I just need it to connect my laptop to a UART...
<azonenberg>
It provides no security guarantees, but neither does USB or SATA
<azonenberg>
If you understand what it does, and only use it on a network that can never be accessed by anyone you don't trust to have full control of the device, then it's totally fine
<azonenberg>
Most people just aren't able to provide that guarantee :p
* monochroma
remembers telnetting into a customer's core cisco router that was a full rack tall back in ~2012 to turn up a circuit :D
<Degi>
I dunno its to control a pulse laser on my home network when that thing is ona
<azonenberg>
monochroma: When i was at rpi, the dorm i lived in sophomore/junior year was ancient, it dated back to the early days of the school. Probably built in the late 1800s
<azonenberg>
and it had a network you'd expect. The edge switches were cisco catalyst 2924s
<azonenberg>
Which were EOL even back then
<azonenberg>
24x 10/100 ports and, in some versions, a 100baseFX uplink
<azonenberg>
They didn't support QoS so to keep bandwidth usage fair, all dorm ports were set to 10/half
<azonenberg>
This was in about 2009, mind you. In 2009 they were using 10/half
<azonenberg>
And sharing ~11 dorm rooms (22 wall ports) onto one or two 100M uplinks
<azonenberg>
anyway, these switches were also so old they did not support ssh
<azonenberg>
as in, cisco did not *have* a crypto capable firmware image for them
<azonenberg>
the only admin interface they had was telnet
<azonenberg>
i had many dreams of ways i could pwn them and get the campus switch enable secret, but was always too afraid of getting caught to try it :p
<Degi>
Well was the traffic logged? I mean if they only had telnet lol
<azonenberg>
Don't know
<Degi>
And whats 10/half?
<azonenberg>
my plan, which i never put into action
<azonenberg>
10 Mbps half duplex
<Degi>
oh
<azonenberg>
anyway my plan was to go to one particular classroom with a switch rack in it, and no security cameras covering the path to get in/hout
<monochroma>
(likely had flow monitoring at the router)
<Degi>
Well that way you sure can split 100 Mb/s over 20 users loll
<azonenberg>
the cabinet was never locked, maybe didnt even have a lock
<azonenberg>
door tended to pop open sometimes and present a tripping hazard :p
<Degi>
ol
<azonenberg>
anyway, so the idea was to stick a passive tap on the upstream, ideally by nicking the wires and sniffing without ever breaking the connection
<azonenberg>
then connect two ports to form a network loop and remove the loop
<azonenberg>
this would trigger bpduguard and shut down the port
<Degi>
Lol was it only 100 Mb/s too?
<azonenberg>
(yes)
<azonenberg>
then wait, a week or so and submit a trouble ticket saying this port doesn't work
<azonenberg>
someone would then log into the switch and reset the port, boom i have the password
<azonenberg>
(and of course, the uplink would also give me access to the management network so i could move laterally and own other switches too)
<Degi>
bpduguard auto disconnects the port till support does soemthing?
<azonenberg>
it's meant for ports without spanning tree
<Degi>
Lol nicep lan
<azonenberg>
when it sees a BPDU on the port, it shuts it down because it means there's a loop
<azonenberg>
Or somebody plugged their own managed switch into the wall port, but they assume nobody will do that
<azonenberg>
that was standard config at rpi, portfast and bpduguard for fast link startup
<azonenberg>
anyway i have zero doubt that it would have worked
<azonenberg>
i just don't have super high confidence that i could have got away with it long term
<Degi>
Wear gloves, dont leave DNA on the nicked lan cable... lol
<azonenberg>
in retrospect, i was pretty friendly with the head of netsec
<Degi>
And who files the support ticket?
<azonenberg>
(that was one of the ways that would be easy to get caught)
<azonenberg>
i probably could have got permission to test it as an authorizedpentest
<azonenberg>
but didn't think of that option until i had graduated
<electronic_eel>
that would have been a neat thing
<azonenberg>
like, one time i was in a classroom and nobody could get on the wifi
<azonenberg>
i noticed the DHCP server was giving out the wrong default gateway, it was a /23 subnet and they gave out .254 of the wrong half as gateway
<azonenberg>
i correctly guessed the actual gateway and sent an email to her from that subnet explaining the misconfiguration
<azonenberg>
and requesting a fix
<azonenberg>
it helped that i had been her student in several of the cisco academy classes so she was acquainted with me personally, in addition to being the head of IT for the school she taught the CCNA security and voice classes
<Degi>
Ugh there are company specific classes in IT?
<monochroma>
"company specific" ?
<Degi>
Well cisco
<monochroma>
everyone in the market emulates cisco
<Degi>
I meant the "cisco academy classes"
<azonenberg>
Degi: it was like 2-3 networking classes taught in the CS department (which really should have been IT imo)
<electronic_eel>
you get some certificate there, and if you have that, it helps you to get a job
<azonenberg>
they had good background knowledge on routing protocols, but it also qualified you to take the cisco certification exams
<azonenberg>
if you passed, that was considered your final for the class
<electronic_eel>
also if you know cisco ios commands, it helps you with any more advanced networking gear
<azonenberg>
if you didn't want to pay for the cert, you had the option of taking a practice exam from some test prep book as your final instead
<azonenberg>
iirc
<azonenberg>
i got the cert for lulz, but never ended up using it
<azonenberg>
and yeah
<Degi>
Huh wow there's 60 kg cabbage in Alaska
<azonenberg>
When i have time to work on my ethernet switch project, i will be getting back to the CLI design at some point
<azonenberg>
the planned interface is very much based on the IOS command set, however i freely deviate from them if cisco is dumb
<azonenberg>
e.g. requiring you to enter config mode and having some commands only valid in/out of config mode
<azonenberg>
so you can "show" at any time, no "do" required
<azonenberg>
miek: wow
<Degi>
lol @miek
<monochroma>
miek: that's super common
<azonenberg>
in cisco, "no service password-recovery" allows you to reset creds BUT it also wipes the switch config
<electronic_eel>
unfortunately
<Degi>
I think my router has a hardly accessible total reset button too...
<miek>
yeah, full reset i can understand - it's at least noticable. i had never seen a handy delete password button before :D
<electronic_eel>
total reset is ok, but just password reset is evil
<azonenberg>
electronic_eel: the other thing i will be doing differently is not having the concept of a "native vlan"
<azonenberg>
an interface is either a trunk, in which case it drops untagged packets
<azonenberg>
or an edge port, in which case it drops tagged packets
<monochroma>
iirc those credential delete buttons started showing up shortly after that one network admin in iirc california "held the network hostage"
<azonenberg>
no in between
<Degi>
loll
<monochroma>
it was a city wide municipal network, and he changed all the passwords, and also apprently set the configs on everything to only live in RAM, so a reboot would leave you with an unconfigured switch :P
<Degi>
haha
<monochroma>
so i suspect the buttons were primarly added for "regime changes"
<azonenberg>
monochroma: i would assume being held in the local jail for a while would convince him to change his tune
<Degi>
Is it normal for PMS7003 sensors to just jump up to 400% particle concenration and stay at that value? Or is my room just contaiminated for the past like 6 hours
<monochroma>
azonenberg: not really, iirc he said he would only tell the password to the mayor :P
<electronic_eel>
but if you need such a button for a regime change, then it means something went wrong with management for a long time
<azonenberg>
yeah, like not having backups
<electronic_eel>
usually upper management should have access to password lists, and that access procedure should also be checked every now and then
<monochroma>
yeah, but, most places don't :P
<monochroma>
so we get a button! :D
<electronic_eel>
that is not only necessary for regime change, but also if some accident happened to the it guys
<monochroma>
yup
<electronic_eel>
not only a button, but a button for the bad guys to sneak in
<monochroma>
a lot of places the "IT department" is yaknow, one person, and maybe a jr to do technician grunt work
<monochroma>
eh, if you can touch the hardware, it's already game over
<Degi>
Wow I opened the balcony door and holy shit the air just got cleaner
<electronic_eel>
depends on the time you can touch it
<Degi>
Theres a difference between pressing a button and unscrewing the whole device and resetting the BIOS battery I think
<monochroma>
if you have the unsupervised time to poke the reset button and plug into the console port it's not great
<electronic_eel>
if you just got 5 minutes, having a button or requiring opening the case and jtag and so on makes all the difference
<Degi>
Wow the VOC concentration just decreased by quite a bunch
<Degi>
Lol jtag... Thats quite overkill, a USB is probably enough
<Degi>
Depending on how good its configured
<monochroma>
most switches don't have any USB interface
<electronic_eel>
often you don't want the device itself, but the passwords or private keys that are in it. just a cli usually won't give you the private keys in cleartext, they are crypted
<electronic_eel>
so you need the device keys that they are crypted with, and that is where the jtag comes in
<Degi>
Hm oh geez the air indoors semms bad
<Degi>
losed the balcony doors, VOC and PM goes up
<Degi>
*closed
<monochroma>
Degi: close the 55 gallon drum of acetone :<
* Degi
accidentally spills it
bvernoux has quit [Quit: Leaving]
<azonenberg>
oook so back
<azonenberg>
So it looks like the DAC is oscillating
<azonenberg>
And we're suspecting C69 is at fault?
<azonenberg>
ok C69 removed, DAC output is stable now
bluezinc has quit [Quit: Do not go gentle into that good night.]
<azonenberg>
Now i'm going to try putting R63 back
bluezinc has joined #scopehal
<azonenberg>
electronic_eel: ok so... with C69 removed and R63 replaced, DAC outpiut is stable
<azonenberg>
things look good on that front
<azonenberg>
TP6/7 are now flatlined at 2.5V +/- scope noise
<azonenberg>
TP10/11 are at 2.3V +/- scope noise
<azonenberg>
Then the output is looking a bit wonky
<azonenberg>
OUT- is at 1.249V, OUT+ is at 909 mV
<azonenberg>
And i think i see the problem
<azonenberg>
it looks like the diode on OUT+ is shorted
<azonenberg>
probably too much solder paste, need a smaller footprint
<azonenberg>
OUT+ and OUT- are now both around 900 mV as expected for no signal
<azonenberg>
So at this point i think all known bugs are fixed
<azonenberg>
Bad relay pinout: already fixed in kicad lib, reworked on prototype
<azonenberg>
2V5 oscillation fixed by removing C15
<azonenberg>
DAC oscillation fixed by removing C69
<azonenberg>
1V8 error and incorrect Vout fixed by clearing shorted D6
<azonenberg>
need to design more paste reduction into footprint on that package for next board
<azonenberg>
Unless anyone has more action items for me on this board i plan to move to assembling the front side of the HMCAD characterization board
<azonenberg>
ADC board out of the oven and cooling off
<azonenberg>
Going to need a little bit of manual rework as I forgot to break out some FPGA GPIOs for the UART, so i'll have to solder some wires onto the pin header by hand. Easy enough to fix
<azonenberg>
I just want to have the fpga be able to talk directly to the stm32 on the AFE board as if it were the actual BLONDEL mainboard