<terpstra> mwalle, yeah, i had it sort-of working. i could see the register contents anyway. however, it couldn't step the program or resume.
<terpstra> and the transfer rate was ... 9 bytes/second!
<terpstra> when i tried to use it to load firmware
<lekernel> http://makerfaireuk.com/ ...at the US event last year they had a few good spots
<wpwrak> lekernel: i see a lot of kids in the pictures ... which seems oddly appropriate :)
<kristianpaul> make fair? bah..
<kristianpaul> really thats the top
<kristianpaul> makeR
<kristianpaul> and i own reprap (just for fun)
<kristianpaul> north americans and his faires ..
<lekernel> wpwrak: I would indeed say that 98% of the makerfaire is arduino led blinkers and similar crap. but at the US event they still had some good stuff like someone messing with a stirling cryocooler to make liquid nitrogen and Jeri's DIY transistors. and unfortunately, those remaining 2% are hard to find otherwise
<lekernel> maybe I should organize a festival with only those :)
<lekernel> funny, xilinx has released ise 13.1 with support for their vaporware 7-series devices you can't buy
<roh> lekernel: i think you are wrong there.
<roh> you can never get those 2% without the 98% stuff which is 'NIH'
<roh> or 'boring' for our level of understanding technology. others need and do that to learn.
<roh> humans do not have a rectangular learning curve
<lekernel> NIH
<lekernel> ?
<roh> Not Invented Here
<lekernel> well, ok, maybe not only those 2%
<lekernel> but at least raise the bar a bit
<Fallenou> lekernel: I don't know if I will go
<Fallenou> it looks great
<Fallenou> but it's far away from london
<Fallenou> it might be expensive
<Fallenou> I will have a look on the train ticket price
<Fallenou> is this good for 200 EUR ?
<lekernel> london is expensive as shit, so trains should be cheap by comparison :)
<roh> Fallenou: doesnt sound that bad.
<lekernel> 50 MHz...mh, a bit weak I'd say
<roh> i mean.. its 200E for a dso. should be ok for a hobbyist.
<Fallenou> yes but it looks cheap
<Fallenou> compared to brand new scope
<lekernel> but it has a good amount of memory
<Fallenou> i've seen 50 MHz scope for like 400 eur :x
<Fallenou> scope is amazingly usefull and amazingly expensive at the same time
<lekernel> I have a 80MHz scope I paid 50E for. but analogue and no storage
<roh> i paid 550E for my welec which istn thaat much better in reality
<roh> but has 4 channels
<Fallenou> oh must be heavy
<roh> also got a hameg HM204 (20mhz analog 2 channel) but thats not useful mostly
<Fallenou> 2 channels + 1 trigger is nice too, you can do stuff with that
<Fallenou> lekernel: where did you find your analog scope ?
<lekernel> at an amateur radio second hand market
<Fallenou> I might try to contact my old old school which might have old analog scope
<lekernel> actually, i'm better off with that analogue thing
<Fallenou> personnally I don't care that much, analog or not
<lekernel> I had a DSO before and it went berserk when you exceeded it bandwidth a little
<Fallenou> as long as I can see my signal :)
<roh> analog scopes are dead cheap now.
<roh> i got mine all free
<Fallenou> when you say with memory, it means i can "stop", right ?
<lekernel> this analogue thing behaves when you exceed its maximum frequency
<Fallenou> and scroll a little bit
<roh> leftovers from repair shops using proper hw
<roh> Fallenou: storage oscilloscope. jap.
<lekernel> and when I need something beefier, I can always borrow one...
<roh> analog storage was very complicated, heavy and sucky. dont get that. its not worth the pain.
<Fallenou> humm ok
<roh> so in the end everbody out there doing digital stuff has a dso as 'new world device' to complete what you can't do with a voltmeter alone
<Fallenou> we have a 4 channels analog with storage scope at the club at school
<Fallenou> but it sucks
<Fallenou> 3 channels are dead ...
<Fallenou> and so on
<lekernel> Fallenou: is that the scope we played with at minet?
<roh> analog storage was ... not nice to use from a HIG pov.
<Fallenou> humm actually it's at INTech
<Fallenou> maybe it's the one you are talking about
<Fallenou> since we only have one scope
<Fallenou> it's a big big fat heavy old scope
<roh> have only used that on a analog modular tek once.. but it was something like 50 or 60cm deep and i couldnt lift it alone
<lekernel> well, that day when I came over and we played with the e10k and such
<Fallenou> with a big keayboard to do mathematical operations
<Fallenou> FFT and such
<roh> fft.. such modern toys *ducks*
<Fallenou> yes I remember, but I don't remember having used the scope
<Fallenou> but if it was an analog scope yes it would be the INTech one
<Fallenou> hesitates to buy this scope
<Fallenou> 50 MHz is not very high frequency
<roh> Fallenou: its ok for any arduino hacking and similar microcontroller stuff.
<lekernel> Fallenou: i'd check with your local hamradio club, they sometimes have very good stuff
<lekernel> often old school analogue though :)
<roh> eeeek.
<lekernel> well, I happen to like analogue scopes.
<roh> dont buy used analog non-dso stuff.. usually not worth any money
<Fallenou> roh: yes for arduilol 16 MHz is enough :)
<lekernel> oh, come on
<Fallenou> but I don't want to just be able to scope arduilol
<roh> Fallenou: nope. remember.. nyquist?
<Fallenou> oh oh ok
<Fallenou> then 32 ?
<lekernel> you can do lots of stuff with a decent analogue scope
<lekernel> much more than with a crappy dso
<Fallenou> oh sure lekernel
<roh> Fallenou: actually they should already sample atleast twice per hz.. but reality.....
<roh> and marketing...
<Fallenou> yes that's what i wondered
<Fallenou> if the scope says 50 MHz
<Fallenou> it should mean i can show 50 MHz signals
<roh> lekernel: even a decent analog scope is worthless if you dont have repetitive signals. wich i have all the time.
<roh> means i cannot do the stuff at all with a non-storage device.
<roh> Fallenou: should.
<Fallenou> yes :x
<roh> in that case (rigol) i tend to believe them.
<Fallenou> the first I linked does 1 Gig samples/sec on 1 input
<roh> on tek and hameg, lecroy you can also believe them. on the china and noname stuff.. be careful ;)
<Fallenou> and 500 Meg samples /sec for 2 inputs
<Fallenou> so I guess it's really ok for it's 50 MHz
<roh> its a combination of 'sampling frequency' and 'analog stages'
<Fallenou> arg
<Fallenou> ok
<roh> the latter getting more complex with rising range
<Fallenou> and since I want something cheap I will definitely end up with some chinese stuff :)
<Fallenou> i gotta be careful
<roh> e.g. i know that my welec has a not really great analog stage.. but hey.. i never needed to measure the 200mhz it can in theore (does 1gsamples)
<roh> also its analog absolut measured voltages arent really precise. also the adc are usually 8bit on non-highend devices. still thats all ok as long as a device helps you solve your questions fast.
<roh> if i want absolute levels.. i dont use the scope.
<roh> i want to know e.g. what baudrate has a serial.. or find it at all first on a unlabeled connector
<Fallenou> that's more to verify what is going on on digital lines and do some triggering
<roh> exactly.
<Fallenou> I am not really fine with logic analyzers
<Fallenou> usually you need to connect it to a PC
<Fallenou> it samples in real time through USB at a low rate
<roh> and usually i just capture some random bytes from the stream to look at the small and wide timing spaces to guess the baudrate. then connect the serial adapers/levelconverters/inverters and go from there.
<Fallenou> I prefer a real scope
<roh> i have a logic analyzer here.. but i dont use it much. it needs the windows notebook stored in the same box with it to work
<roh> thus i use my dso. much faster
<Fallenou> yes
<Fallenou> I am a logic analyzer too
<Fallenou> but the software sucks
<Fallenou> it's in Java
<roh> hehe.. for ours there is no !win32 sw
<Fallenou> s/am/have/
<Fallenou> first the probes are crap, the grip grips nothing
<Fallenou> and the software is really bad
<roh> sigrok just istn evolved and complete enough by far
<roh> but the manuf isnt cooperative due to stupidness
<Fallenou> oh I didn't know this project
<Fallenou> oh it supports the Saleae logic
<lekernel> Fallenou: heh, what to expect from hackaday stuff :)
<Fallenou> I used that, it works , but only windows
<Fallenou> lekernel: at least it's not lolduino powered :)
<Fallenou> there is a beautifull fpga on it
<lekernel> send them something good, they ignore it. they want AMATEUR looking stuff so they can identify with it and not feel ridiculous
<Fallenou> roh: I could try sigrok !! my logic analyzer seems to be supported :)
<Fallenou> lekernel: lol I guess you asked them if they wanted to be a Milkymist sell point ?
<roh> Fallenou: please do. these guys need support.
<lekernel> no... they don't sell stuff
<lekernel> and even though I did send them milkymist links, it's not the only thing they ignored
<Fallenou> lekernel: oh yes, I meant dangerous prototypes
<lekernel> even by submitting anonymously, which excludes the theory that they simply do not like me :)
<Fallenou> which sells those logic analyzers
<Fallenou> lol
<lekernel> Fallenou: the salae thing isn't too bad
<Fallenou> works on mac os X sigrok, nice
<Fallenou> lekernel: yes it worked well when i used it, just limited in sample rate
<Fallenou> but too bad the linux port is never released
<lekernel> it does work on linux... but binary-only software
<roh> must admint that he never had a piece of hw where the vendor sw was 'great'
<Fallenou> oh !!
<Fallenou> I didn't know
<roh> that was always something written by a 3rd party
<Fallenou> last time I had a look it was windows only$
<lekernel> well, the salae software isn't bad
<lekernel> it's very simplified, but it gets the job done
<lekernel> and zoom/scrolling has a good feeling and usability
<Fallenou> and the probes are good quality
<Fallenou> for saleae
<rejon> lekernel met with google gsoc manager...submitting app now
<rejon> had lunch with her
<Fallenou> carol ?
<rejon> yes
<lekernel> rejon: cool, thanks for taking care of this
<rejon> yah
<Fallenou> oh nice :)
<Fallenou> where did you met her ?
<rejon> so you guys can keep hacking
<rejon> Fallenou google sf office
<Fallenou> are you working nearbye ?
<Fallenou> nearby*
<rejon> sf
<Fallenou> hehe nice
<rejon> http://fabricatorz.com is based here.
<rejon> my co.
<rejon> my MM1 should arrive here on friday blat!
<Fallenou> so Milkymist would be a gsoc organization this time ?
<rejon> well, we hope
<rejon> i'm trying hard!
<Fallenou> great :)
<rejon> if going there for lunch qualifies ;)
<Fallenou> :D
<Fallenou> I hope you bought her lunch =)
<rejon> google free lunch
<rejon> i gave her our http://freesouls.cc book
<Fallenou> well ... what was I thinking =)
<Fallenou> oh
<Fallenou> what is this about ?
<wpwrak> Fallenou: the DS1052E is USD 400 new: http://www.tequipment.net/RigolDS1052E.html
<wpwrak> Fallenou: i have a DS1102CD, the previous generation. they're quite good for their price. the deep memory allows you to do things like debug USB.
<lekernel> USB is never debugged :)
<wpwrak> Fallenou: this is what the deep memory does for you: http://downloads.qi-hardware.com/people/werner/tmp/teaser1.png
<wpwrak> Fallenou: download to a PC and decode: http://downloads.qi-hardware.com/people/werner/tmp/teaser3.png
<wpwrak> lekernel: yeah, you only fix the worst bugs and leave the rest to the next guy who comes along ;-)
<Fallenou> oh nice
<Fallenou> what soft is teaser3 ?
<Fallenou> it' in the pc ?
<lekernel> wpwrak: what do you use for decoding like this?
<lekernel> and where is that clock coming from?
<lekernel> USB is asynchronous...
<Fallenou> i guess it's the sampling clock from the scope
<wpwrak> lekernel: the decodes comes from there: http://svn.openmoko.org/developers/werner/ahrt/host/tmc/
<wpwrak> lekernel: the clock is the reconstructed USB clock
<lekernel> so your software reconstructs the clock? nice
<lekernel> cool
<lekernel> when did you develop that?
<wpwrak> lekernel: dec.py does clock reconstruction and stuff, then hands the digitized waveforms to the interactive decoder
<wpwrak> lekernel: hmm, the USB part must have been around december 2009
<Fallenou> humm so $400
<wpwrak> lekernel: the rest of the system is a bit older. i used it to chase weird bugs in openmoko's SDIO (for the wifi module)
<wpwrak> lekernel: here's an example with SDIO: http://people.openmoko.org/werner/wlan-spi/spi-poll.png
<wpwrak> lekernel: that uses the logic analyzer in the ds1102cd. the LA has issues with spurious transitions and crosstalk, but if you massage the data long enough, you can get useful results with it
<Fallenou> what did you use SDIO for ?
<wpwrak> lekernel: and it's a fraction of the price of an MSO from a more reputable vendor ;-)
<wpwrak> Fallenou: we had a WLAN module connected via SDIO inside the Openmoko Freerunner phone
<Fallenou> oh wlan :)
<Fallenou> is SDIO widespread in embedded systems ?
<wpwrak> Fallenou: SPI and I2C are more common. but you run into SDIO from time to time, too
<Fallenou> hum ok :)
<wpwrak> ah, here's the test setup i used back then: http://people.openmoko.org/werner/sdio-test-setup.jpg
<wpwrak> as one can see, i cleaned the table extra for the photo occasion ;-)
<lekernel> is downloading ise 13.1. let's see what's inside those vaporware chips.
<Fallenou> wpwrak: i'm a bit confused, can the DS11052E show a 100 MHz signal ?
<Fallenou> it says 50 MHz in the title
<Fallenou> DS1052*
<lekernel> ah, they even published technical docs: http://www.xilinx.com/support/documentation/7_series.htm
<lekernel> given the long time it took them between s3 and s6 and the mess s6 was, I expect the worst for the 7 which has been announced even during s6 wasn't really available
<Fallenou> :)
<wpwrak> Fallenou: it can show a 100 MHz signal, but it will be severely attenuated. normally, the analog bandwidth is given for a -3 dB attenuation.
<Fallenou> oh they use stacked interconnect
<wpwrak> Fallenou: if you go to twice the frequency, you'll lose another dB or such
<Fallenou> hum ok
<Fallenou> so the values I will see won't be the real values
<Fallenou> but the global shape will be kind of the same
<lekernel> hmm... seems to be largely a die shrink of the 6's
<lekernel> good, easier to support in llhdl :)
<wpwrak> Fallenou: the shape of curves will change. that is, unless they're already pure sinus
<Fallenou> well they added hard coded arm core + ADC
<Fallenou> and it uses stacked interconnect so maybe the format of the description files will change ?
<mwalle> one of the rigol scope can be upgraded with software
<mwalle> iirc it was the 100mhz scope, which can be upgraded to 200mhz
<Fallenou> humm 50 to 100 ?
<Fallenou> but I think they fixed that :x
<mwalle> btw we got some new nifty scope :)
<mwalle> terpstra: that juart protocol is rather slow.. esp if your jtag pod isnt fast, i archieved 900bytes/s iirc, you could use the dp version though
<mwalle> 'lm32 debug_protocol enable' iirc
<Fallenou> 900 bytes/s, reminds me of the wifi at my office :)
<mwalle> lol :)
<Fallenou> is tempted by the rigol ds1052E
<Fallenou> if I can upgrade it to 100 MHz
<Fallenou> it's awesome
<mwalle> although we have that high speed scope now.. there was no more money for more than one differential probe ...
<Fallenou> today I had to hold 3 probes with my hand
<Fallenou> how do you hold 3 differential probes ? :p
<Fallenou> is thinking about upgrading his body with new arms and hands
<mwalle> use pliers with some rubber around the handles
<mwalle> but be sure it wont squeeze to much :)
<mwalle> agilent sells some kind if this for 50 EUR ...
<mwalle> although we connect our diff probes usually like this: http://www2.tek.com/cmsreplive/psrep/img/13426/A000_7985-L.jpg
<lekernel> Fallenou: they didn't add the arm core
<Fallenou> oh?
<Fallenou> The Zynq-7000 family integrates a complete ARM Cortex-A9 MPCore processor-based system  <- ?
<Fallenou> what about that ?
<lekernel> it's just hot air now
<lekernel> it seems they changed the routing structure in kintex 7 too
<lekernel> there is this new freaky "VBRK" switch box i'll have to figure out
<lekernel> at least I guess it's a switch box
<mwalle> hm that tmc is nice..
<mwalle> i used vxi11 with a python wrapper for the same thing
<mwalle> gn8 :)