<wolfspraul>
I will try to source a few more, directly from the manufacturer. for now I assume you want black, but I'm curious how strongly you feel for it - if I think here that blue looks better, would that be OK for you?
<wolfspraul>
then I have some mechanical feedback/questions
<wolfspraul>
one is the metal 'push' button on the DMX TX connector
<wolfspraul>
it's on the top side, but a bit hard to push down because of the acrylic over it
<wolfspraul>
cutting into the acrylic would probably look ugly, but what about finding a connector where the 'push' button is on the left side?
<wolfspraul>
then we could move the 'DMX' label a bit to the left, and it should be easier to press the 'push' button
<wolfspraul>
the other one is the internal microphone, whose reception is not good with a closed case.
<wolfspraul>
how about moving that to the side a bit, and letting it point outwards, cutting a hole into the acrylic between VGA and line in/out?
<wolfspraul>
that's all :-)
<wolfspraul>
need to get roh's feedback on push button and microphone hole as well
<xiangfu>
one little thing: in taobao.com, there are also a convert USB--> PS2 in picture. but when you get the package. there is just don't include that converter.
<lekernel>
wolfspraul: (keyboard) it doesn't matter much...
<lekernel>
this is cool: http://www.windfreaktech.com/ (or rather, the ad chip on it is cool... it's $13 on digikey btw)
<xiangfu>
stupid question(for understanding)can we send out TV signal by using the signal generator? then if the normal tv can receive this signal by search channel? :)
<lekernel>
no, TV bands are above the 70MHz limit you have on the VGA signals
<lekernel>
but you can generate CVBS which in turn can drive a cheap analogue modulator (you can find those e.g. in VCRs)
<wolfspraul>
what can you do with this kind of signal generator? (windfreaktech)
<lekernel>
electronics experiments/debugging
<wolfspraul>
leading eventually to which kind of applications?
<wolfspraul>
my imagination/knowledge is just too little to see the end application :-)
<wpwrak>
wolfspraul: you can use it to provide the RF carrier for modulators/demodulators
<wolfspraul>
ok but what types of applications are realistically achievable, say with the ADF4350
<wolfspraul>
data transmission? what distance/speed?
<wpwrak>
wolfspraul: well, it needs a bunch of other chips. but the RF generator is one of the building blocks of any RF system
<wpwrak>
wolfspraul: (distance/speed) naw, nothing like this. you have to think of it as an oscillator.
<wolfspraul>
if ADF4350 would be a building block in you wifi/802.15.4/gps/3g/lte system, would that eventually work?
<wpwrak>
wolfspraul: page 24 has an example circuit that adds a bit more meat
<wolfspraul>
I still don't understand the connection to applications, which ones this could be a building block for, and which ones not
<lekernel>
wolfspraul: it's a bit like asking what the end applications of a capacitor are... that's a fairly generic building block
<wpwrak>
it could be a building block for anything in the 137-4400 MHz range. again, it's "just" the oscillator. but the oscillator is a crucial piece. get that one wrong, and the rest goes down the drain as well.
<wolfspraul>
got it. 'anything' in that range, easy to remember :-)
<wolfspraul>
lekernel: yes but that is my question, and it was answered. thanks!
<wolfspraul>
I don't like to jump to conclusions if I know I know too little...
<wolfspraul>
so I rather take the blame for asking stupid questions :-)
<wpwrak>
it could be, for example, be a building block of a little SDR system that could do IEEE 802.15.4. something to consider for the ya-mm ;-)
<wpwrak>
since you'll ask about wlan: also possible, but more demanding
<wolfspraul>
you know I'm the last person to ask about wlan
<wolfspraul>
then I'd rather ask about 3g/lte etc.
<wpwrak>
well, wlan with SDR wouldn't be so bad. our biggest problem with wlan are the chips/modules.
<wpwrak>
of course, there's probably patent hell when you get near modern stuff like 802.11n and such.
<wolfspraul>
lekernel: how about the other two items I had 1) 'push' button on DMX TX on left side and 2) hole for microphone
<lekernel>
I don't care either. for me, the DMX pushbutton is reached easily enough to release the cable and the mike picks up sound well enough to react in loud (i.e. typical) environments.
<wolfspraul>
ok I agree that the importance of this is 'medium'
<wolfspraul>
but you agree that 1) both of those proposals are in fact improvements and b) you are not opposed to them?
<lekernel>
wolfspraul: see? with the new protection, 20V doesn't damage the board
<wolfspraul>
yes
<lekernel>
hmm... though 1.1A hold isn't enough
<lekernel>
(and yes for those items, if you want to spend time on that... though making more people give a shit about this project would be way more important imo)
<wolfspraul>
ok thx, that's settled then
<wolfspraul>
lekernel: what does it mean that "1.1A hold isn't enough"?
<lekernel>
the fuse is only rated to work continuously with a 1.1A current through it
<lekernel>
if we already draw 1A from USB that's only 100mA left for the rest
<lekernel>
so we'd need a 2A hold instead
<lekernel>
also, the 0.21 ohm resistance would drop 0.42 volts at 2A
<lekernel>
that's quite a bit
<lekernel>
mwalle: did you get your wm9707?
<carlobar>
hi, does milkymist linux kernel needs the fast memory link of ddr memory? im trying to run just the kernel on a digilent board. I implemented a wishbone compliant ddr controller, but neither the bios nor kernel are working, can someone help me?
<guyzmo>
EHLO milkymist
<guyzmo>
has just booted the MM for the first time ;)
<lekernel>
carlobar: it needs a working SDRAM controller, which as far as I know excludes all those which are on opencores
<lekernel>
hi guyzmo
<guyzmo>
lekernel - where can I get the lm32 gcc (as source) ?
<lekernel>
scrts: indeed... and they dropped the parallel port programmer at last :p
<scrts>
that fpga itself costs 100 usd :)
<lekernel>
yeah, that's obviously a marketing push...
<lekernel>
when I saw that price it was already clear for me they probably sell at a loss
<scrts>
strange that they also sell their IP suites for 99usd
<scrts>
dsp suite, video suite, etc...
<lekernel>
IP is easy to sell at whatever price since it costs nothing to duplicate
<scrts>
well yes, afaik video suite was 995 some time ago
<guyzmo>
lekernel - thanks for the script pointer ;)
<lekernel>
scrts: as a matter of fact, it's not too hard to get cheap software licenses or even boards from X & A either...
<scrts>
yeah.. even since it is possible to get the IP for free from A
<scrts>
:)
<lekernel>
DDR3-1333 memory on that lattice board?!
<lekernel>
mh
<scrts>
I've contacted my distrib for it.. waiting for response:)
<lekernel>
I'm wondering if the FPGA and toolchain work well at this frequency
<lekernel>
it's not really the case with spartan6
<carlobar>
lekernel: i got the controller from: https://www.das-labor.org/trac/browser/fpga/ddr. The clearBSS section of crt0.S (in bios) is made without problems, and even i can execute serialboot, but when it try to boot nothing happens
<lekernel>
this controller doesn't work either
<lekernel>
why did you buy a digilent board btw?
<lekernel>
scrts: how many LUTs are in there? 33K?
<lekernel>
4 input?
<carlobar>
lekernel: i bought it last year.. i thought it would be fine.. i hadnt experience with embedded systems
<carlobar>
lekernel: do you know about any controller for 16 bit ddr memory?
<lekernel>
ddr memory is always painful, especially when you need it to work really well e.g. to get Linux to boot
<lekernel>
if you want a working one, either use proprietary Xilinx IP, adapt HPDMC or try your luck with GRLIB
<lekernel>
or buy a M1 :p
<carlobar>
ill try first with GRLIB, i used it with the LEON processor... the problem is to adapt the AMBA interface to a wishbone interface
<carlobar>
lekernel: adapt HPDMC would be easier?? maybe just modify the controller to perform 2 reads f 16bit instead of 1 read of 32 bits?
<scrts>
lekernel seems like yes
<lekernel>
carlobar: no, make the memory bus 32 bit and modify the bridge
<lekernel>
Please note that this free Diamond for Windows is valid for use only with the LatticeECP3 Versa Development Kit. The free Diamond License will be valid for one year from the activation or July 31st, 2012 at the latest.
<lekernel>
mh
<lekernel>
scrts: do you know anything about the lattice bitstream format btw?
<scrts>
I know nothing about lattice, sorry
<scrts>
are You also interested in that devkit? :)
<wpwrak>
lekernel: btw, when do you think llhdl will be able to synthesize the mm1 ? or is it too early to ask ? oh, and https://github.com/lekernel/llhdl says 404. is this expected ?
<lekernel>
I don't know
<lekernel>
and yeah I renamed a lot of repositories lately
<lekernel>
"If you are planning a new design, please avoid using the AUTOBUF primitive"
<lekernel>
lol
<lekernel>
they were touting it as a major improvement a couple of months before this AR
<lekernel>
BUFG might work... if a BUFG is the correct clock buffer to use in your case (check the spartan 3e user guide...)
<carlobar>
ok, thank you
<mwalle>
lekernel: whats autobuf?
<lekernel>
mwalle: a netlist generic clock buffer that was supposedly turned into the right clock buffer by the p&r tool
<lekernel>
for some reason they never got it to work
<lekernel>
on spartan6 (which was introduced at the same time as AUTOBUF) it was a complete disaster since the p&r tool _also_ never placed the clock buffers correctly. using AUTOBUF there would also trigger extra bugs when you tried to manually place the clock buffers to workaround the tool problems...
<lekernel>
with all that crap i've spent a good week or so getting all the clocks to work on m1
<mwalle>
lol, ok ;)
<wpwrak>
lekernel: seems that the world needs llhdl, badly ;-)
<lekernel>
hmm... chrontel vga-to-cvbs chips have 9-bit internal DACs. so with nearly 10 bits on the m1 w/ oversampling it should work nicely