<sb0>
panic at src/libcore/result.rs:860: cannot initialize AD9154: "bad SYNC"
<sb0>
sometimes ...
<GitHub2>
[artiq] sbourdeauducq commented on issue #860: It was due to incorrect settings of my 100MHz source. The HMC830 locks fine and then both AD9154 DACs usually work, including JESD lane startup (but intermittenly I get a "bad SYNC" error). https://github.com/m-labs/artiq/issues/860#issuecomment-348077133
Ultrasauce has quit [Quit: No Ping reply in 180 seconds.]
<davidc__>
sb0: I picked up an ebay RGA probe (quadrupole, detector and filament). Its probably garbage, but it was only $25 and looked visually OK. let me know if you want help with designing / testing spectrapak
<sb0>
davidc__, are you able to get an ion beam with a few lab PSUs?
<davidc__>
sb0: from a theoretical perspective, yes I can get the necessary bias voltages + etc
<davidc__>
and I should be able to pull a suitable vacuum
<davidc__>
whether that translates to a working probe is TBD :)
<sb0>
there are only so many things that can go wrong with those things - vacuum leak/contamination, broken filament, misaligned elements
<sb0>
if you already get an ion beam it's a good sign re. those three things
<davidc__>
sb0: I can theoretically replace the filament (I have a few used SEM filaments kicking around that I can repurpose)
<davidc__>
but anyhow, when the probe arrives, I'll check back in
<sb0>
yea sure, in theory you can just stick some cheapo piece of tungsten wire in there ...
<sb0>
also you should ground the quadrupole electrodes for that test
<davidc__>
and if one is DIY'ing the filament drive, you can adapt it to the characteristics of the jury-rigged filament
<sb0>
davidc__, what does the ion source look like?
<sb0>
there are several different designs that require different bias voltages
<sb0>
if you do things like create the electron beam between the filament and the repeller, it won't work ;)
<davidc__>
but for $28, I'm willing to gamble :)
<davidc__>
I'm not terribly optimistic that this one is in good shape as it looks like it might have had a rough life
<sb0>
KF flange?
<davidc__>
could be. Pics aren't good and I haven't found any kind of docs
<sb0>
looks very dirty. you probably want to ultrasonic-clean this thing first...
<davidc__>
sb0: yeah; I was going to follow a full UHV clean prep from the LIGO folks
<sb0>
no need to overdo it
<sb0>
that flange is not suitable for UHV anyway
<davidc__>
(If you haven't taken a look at that: dcc.ligo.org , tons of useful info re materials, vacuum compatibility, schematics for laser drivers/physics/etc for the ligo experiment)
<sb0>
a few passes in the ultrasonic bath should be enough
<sb0>
with solvent/detergent (but be careful that some detergents attack aluminum) then distilled water
<davidc__>
(its labelled as "new in original packaging", but thats a load of crap)
<sb0>
also remove the dust with a compressed air can first
<davidc__>
I have dry oil free compressed air here, no need for a can :), but yes
<sb0>
hard to tell the ion source construction from those pictures
<sb0>
but anyway, if the feedthrough ceramics aren't broken, it seems like it should work
<davidc__>
sb0: yeah. Once I get my hands on it I'll take a set of nice high res well lit in focus photos; get a pinout, and probably some parameter measurements (rod size, r0, length, ion source design, etc)
<sb0>
don't bother with rod size, you can scan RF/DC and deduct it from the measurement results
<sb0>
length doesn't matter
<davidc__>
sb0: I'm more interested in it from a design perspective and relating it to the math in a few papers/books
<sb0>
the only things you need to deduct from visual inspection are pinout and how the ion source should be biased
<sb0>
yes, but you can compute r0 from the dimensions of the mathieu stability region
<sb0>
to measure it mechanically, you need to remove the shield, which doesn't go away easily (with one of the tubes I have you need to disassemble about everything)
<sb0>
and it's really just four steel rods. not much to see there ...
<sb0>
davidc__, also, to test the filament and avoid burning it, you probably want to put a window in front of it and look at it glowing
<sb0>
then apply, say, 70V to the electron collector plate and look at the current
<sb0>
then bias the other plates, and you should get some nanoamperes on the collector electrode
<sb0>
(you will need a good TIA, shielded cables, and no ground loops to measure that small current correctly)
<davidc__>
I'll need to build some instrumentation to be able to measure that, but it should be fairly straightforward
<sb0>
davidc__, and spare yourself some time and add a gas discharge tube plus series resistor. otherwise you will blow half a dozen op-amps while doing that.
<davidc__>
Ok, thanks for the heads up
<sb0>
also the failure mode of this particular op-amp from excessive input voltage is a battery-draining short circuit on the power supply. you can guess how I know that...
<davidc__>
heh, I think I have all the parts needed except the 1G resistor . I'll throw that together one of these days
rohitksingh_work has quit [Read error: Connection reset by peer]
early has quit [Quit: Leaving]
early has joined #m-labs
<_florent_>
sb0: about the gearbox: for hdmi rx on 7-series, you need a 8 to 10 gearbox since we are using 2 iserdese2 for phase detection and so cannot use 10 bits mode bit (master & slave) but only 8 bits mode.
<rjo>
_florent_: and you can't do 5 to 10 to free up the slave?
<_florent_>
rjo: we probably could with a frequency a bit bit higher at the input of the gearbox
<rjo>
sb0: there are rtms on sayma1 and sayma2? i can't get either of those to respond on jtag.
<rjo>
_florent_: yes. but an integer 5:10 gearbox is simpler than a 8:10.
<rjo>
sb0: i did the power toggle and now all three are down.
<_florent_>
rjo: yes, i should try. I don't mind is you simplify the gearbox in misoc if you think only integer ratios will is needed. I'll keep my gearbox in litex for now and will check if i also can only use integer ratios.
<rjo>
_florent_: i haven't even looked properly at your gearbox. i am just wondering.
<_florent_>
rjo: ok, but yes i think you are right, i just have to see all cases where i'm using it to see if i can always use integer ratios
FabM has quit [Quit: ChatZilla 0.9.93 [Firefox 52.5.0/20171114221957]]
<GitHub37>
[artiq] jbqubit commented on issue #854: Debugging this is top priority. The Sayma hardware and lots of M-Labs gate ware is ready to test. Getting Ethernet up and running is the bottleneck to forward progress right now. https://github.com/m-labs/artiq/issues/854#issuecomment-348358294