| 00:12.15 | MonMotha | ds2: for low speed stuff that works, though getting sufficiently realtime performance out of, say, Linux is hard |
| 00:12.43 | MonMotha | what I've seen people do is use a PLD to clock the data into latches at a consistent rate then read it back off using the parallel port or whatever |
| 00:12.56 | MonMotha | with a sufficiently large, you can get like a 2-4 level FIFO in so that you have some jitter tolerance |
| 00:13.03 | MonMotha | *sufficiently large PLD |
| 00:14.04 | MonMotha | you can also use a micro with high-speed IO like an ARM7 or small ARM9 |
| 00:14.08 | MonMotha | then you can use USB or what have you |
| 00:14.20 | MonMotha | though you'd have to find one with high-speed USB |
| 00:14.29 | *** join/#edev linac (n=lin@121.41.38.239) |
| 00:14.34 | MonMotha | and isosynchronous mode would be recommended |
| 00:15.43 | MonMotha | for the parallel port method, EPP mode is helpful |
| 00:32.18 | ds2 | MonMotha: my complaint with that is that isn't very 3.3V or less compatible |
| 00:34.21 | MonMotha | well, what you do is put a bunch of comparators in front of the device, and use a DAC to generate the reference voltage |
| 00:34.48 | MonMotha | then you can configure the threshold, and, with sufficient power rails, handle weird stuff like ECL |
| 00:34.57 | MonMotha | also, if you want, you can then optoisolate things |
| 00:35.13 | MonMotha | though then all the power for the comparators has to be separated |
| 00:35.15 | ds2 | stupid question - where does one find a fast enough optoislator? |
| 00:35.24 | MonMotha | ok, use a magentic isolator :) |
| 00:35.31 | MonMotha | they have those good to 150MHz |
| 00:35.43 | ds2 | oh heh.. sounds like the price is getting up there |
| 00:35.54 | ds2 | a cheap 18 line USB dongle runs for about $300 |
| 00:36.06 | ds2 | and has (I think) the pin probe things with it |
| 00:36.35 | ds2 | optoisolators seems to be good to maybe 100KHz at best or so the specs suggest |
| 00:36.48 | MonMotha | I've seen some single-digit MHz ones |
| 00:36.51 | MonMotha | not cheap, though |
| 00:37.09 | MonMotha | you can isolate 16 data lines for about $30 using magnetic ones good to 150Mbps |
| 00:37.24 | MonMotha | isolation isn't as good as the optical ones, but it's sufficient for most purposes (several 100V) |
| 00:37.41 | ds2 | guess I will look into them |
| 00:38.02 | ds2 | my other hobby is the cnc stuff and I been looking for fast isolators for a while |
| 00:38.03 | MonMotha | remember that you can't cycle a parallel port much faster than 10MHz, if that's your plan |
| 00:38.22 | ds2 | yeah I don't expect a parallel port do better then about 3-4MHz |
| 00:38.33 | MonMotha | yeah, that's at least reasonably realistic |
| 00:38.46 | MonMotha | 10MHz in EPP mode using hardware handshaking on high-end PC hardware |
| 00:39.05 | MonMotha | I"ve heard of people hitting 25 on dedicated PCI parallel port cards using high end stuff |
| 00:39.25 | ds2 | but that means I have to look for a decent parallel port... got some old laptops I want to use |
| 00:39.34 | MonMotha | at that point, the PLD is necessary to handle the handshaking and also you really need a queue since there will be plenty of timing jitter unless you're running on DOS or QNX |
| 00:40.07 | MonMotha | one thing that also helps is to mux the clock sources so that you can actually clock on an external clock |
| 00:40.13 | MonMotha | great for spying on synchronous busses like SPI |
| 00:40.24 | MonMotha | I guess you could call that SPIing |
| 00:40.32 | MonMotha | (I'll be here all night...) |
| 00:40.41 | *** join/#edev sjdsphreaker (n=sjdsphre@72.51.99.213) |
| 00:41.10 | ds2 | that is exactly the stuff I want to look at |
| 00:41.11 | *** part/#edev sjdsphreaker (n=sjdsphre@72.51.99.213) |
| 00:41.20 | ds2 | SPI, I2C, 1wire, RS-232 |
| 00:41.31 | ds2 | USB would be very nice i can do it |
| 00:42.20 | MonMotha | grab a USB1T20A from Fairchild and you can do just that |
| 00:42.45 | MonMotha | though it's fast enough that you may have to implement your own timing recovery |
| 00:42.52 | MonMotha | since oversampling enough to get useful data off of it would be tough |
| 00:43.19 | ds2 | seems overkill.. just 1.5M/12M modes are sufficient |
| 00:43.48 | MonMotha | well, you'd have to sample at 24MHz just to get all the transitions |
| 00:43.57 | MonMotha | and then you have no indication of timing on the line, just some bits |
| 00:44.03 | ds2 | this is very tempting: |
| 00:44.06 | ds2 | http://www.rockylogic.com/products/ant18e.html |
| 00:44.12 | MonMotha | if you recover the timing, you only have to clock bits out at 12MHz |
| 00:44.19 | ds2 | true |
| 00:44.31 | MonMotha | not bad for the price |
| 00:44.41 | MonMotha | no linux support, I'm sure |
| 00:44.46 | ds2 | yeah, unfortunately |
| 00:44.50 | MonMotha | that's one of the reasons I'd been tempted to build my own |
| 00:44.53 | ds2 | for linux support I need to goto the bitscopes |
| 00:45.01 | ds2 | then it is only 8 channels |
| 00:45.02 | MonMotha | NONE of them support Linux until you get into the really high end stuff that tends to speak GPIB |
| 00:45.33 | ds2 | http://www.bitscope.net/store/?p=view&i=item+3 |
| 00:45.36 | ds2 | that supports linux |
| 00:46.02 | MonMotha | gotcha |
| 00:46.19 | MonMotha | I wish there was something like the EZ-USB series but for IEEE1394 |
| 00:46.31 | MonMotha | i.e. micro with 1394 interface and DMA to glue things together |
| 00:46.38 | ds2 | not enough volume I'd think |
| 00:46.54 | ds2 | ezusb is the cypress stuff right? |
| 00:47.05 | MonMotha | yeh |
| 00:47.11 | MonMotha | though there are several similar devices |
| 00:47.19 | MonMotha | Atmel makes a few based on both 8051 and AVR cores |
| 00:47.21 | MonMotha | EZ-USB is 8051 |
| 00:47.32 | MonMotha | Microchip has PICs with USB |
| 00:47.38 | MonMotha | TI has various micros with USB |
| 00:47.40 | ds2 | ezusb is nice in that you can pickup a board at the local computer store |
| 00:47.49 | MonMotha | and of course most ARM7s have USB, too |
| 00:47.49 | ds2 | ezusb2 is 68K based |
| 00:47.54 | MonMotha | really... |
| 00:47.57 | MonMotha | that's a HUGE change |
| 00:48.12 | MonMotha | the old EZUSB FX series was still 8051 |
| 00:48.25 | ds2 | well, the 68K version does USB2 HS |
| 00:48.36 | MonMotha | they had 8051s that did, too |
| 00:48.37 | ds2 | bit hard to do on the 8051, even the SiLabs version |
| 00:48.48 | MonMotha | you had to do DMA, obviously |
| 00:49.47 | MonMotha | I'll settle for just about any micro core, I just want IEEE1394 in place of the USB |
| 00:49.53 | MonMotha | USB2.0 High Speed SUCKS |
| 00:50.14 | MonMotha | USB in generall sucks. One hack on top of another |
| 00:50.16 | MonMotha | you can tell who made it |
| 00:51.24 | ds2 | Xilinx |
| 00:51.34 | ds2 | IEEE1394 IP + picoblaze |
| 00:51.38 | MonMotha | oh |
| 00:51.39 | MonMotha | well, |
| 00:51.41 | MonMotha | that's not the same |
| 00:51.49 | MonMotha | yeah, I could bust out my Virtex 2 Pro |
| 00:52.02 | MonMotha | heck, then you don't even need the 1394 PHY :) |
| 00:52.05 | ds2 | heh... your wallet becomes so much lighter then there is the 4th and 5th mortgages.... |
| 00:52.21 | MonMotha | I do have a V2P dev board |
| 00:52.38 | MonMotha | been thinking of getting a Spartan3E one, too |
| 00:52.45 | ds2 | not the board... the whole tool kit for the picoblaze stuff is rediclous |
| 00:53.01 | MonMotha | oh, I just use my own IP :) |
| 00:53.13 | MonMotha | but still, yes, EDK and ISE are fucking expensive |
| 00:53.22 | MonMotha | I have fairly old versions at this point |
| 00:53.24 | ds2 | for someone who don't speak Verilog/VHDL.... |
| 00:53.39 | ds2 | more then that, the Picoblaze has another toolkit on top of it |
| 00:53.51 | MonMotha | which is why I don't use it :) |
| 00:54.09 | MonMotha | easy enough to just chuck some other core on there from opencores, in most cases |
| 00:54.27 | ds2 | as long as you know enough to do that |
| 00:54.36 | MonMotha | I do know Verilog, though, so I can write any glue that's needed |
| 00:54.46 | MonMotha | heck, I've been known to actually write real hardware before |
| 00:54.56 | ds2 | i just don't feel an incentive to learn it... maybe when they have fpgas that can run off 2AA's for a few hours then I might be enticed |
| 00:54.58 | MonMotha | all of my PLD designs are from-scratch, usually |
| 00:55.13 | MonMotha | you can use Verilog on CPLDs, too |
| 00:55.21 | MonMotha | and they certainly have CPLDs that'll run off a couple AAs for a rather long time |
| 00:55.33 | ds2 | if someone would write a free synth tool for a 22V10... |
| 00:55.36 | MonMotha | now, you won't be putting a micro on most CPLDs (though you'd be surprised) |
| 00:55.52 | MonMotha | yeah, that'd be nice, but you do know that the free version of ISE targets most of the 9500 and Coolrunner series, right? |
| 00:55.55 | ds2 | problem is the smaller devices lack free tools |
| 00:56.15 | MonMotha | XC9536 and 9572 are both in ISE Webpack, I thought |
| 00:56.22 | ds2 | those are relatively big parts; would be very very nice to be able to relate to my palasm experience |
| 00:56.29 | MonMotha | XC9536 is tiny |
| 00:56.34 | MonMotha | it's twice the size of a GAL22V10 |
| 00:56.42 | ds2 | oh |
| 00:56.43 | MonMotha | heck, it barely qualifies as a CPLD |
| 00:56.57 | MonMotha | actually, I guess it's 3.6 times the size |
| 00:57.00 | MonMotha | still, not big at all |
| 00:57.09 | MonMotha | 36 macrocell (so 36 flip-flops) |
| 00:57.14 | MonMotha | split into two halves, 18 each |
| 00:58.41 | ds2 | that's something I can relate to |
| 00:58.46 | ds2 | orther problem is time |
| 00:58.54 | MonMotha | I have a project right now where I'm using a 9572XL |
| 00:58.59 | MonMotha | it's actually not that big |
| 00:59.10 | MonMotha | basically, it's a 72 bit shift register with some combinational glue :) |
| 00:59.14 | MonMotha | (at least for what I'm doing) |
| 00:59.27 | ds2 | do they expose all the registers like on a GAL? |
| 00:59.31 | MonMotha | I could do it in 74 series using about 10-12 chips, if I wanted to |
| 00:59.52 | MonMotha | I've only ever used Verilog to design for them, so I don't know what they expose |
| 01:00.03 | MonMotha | basically, I just hit the synthesize button and hope it says "Design fits!" |
| 01:00.12 | MonMotha | if it doesn't fit, I see if I can make my logic smaller |
| 01:00.39 | MonMotha | the two magic phrases everybody likes to see: "Design Fits" and "All user constraints met" |
| 01:00.39 | ds2 | that's a different way of thinking then with palasm |
| 01:00.47 | MonMotha | yes, yes it is |
| 01:00.52 | MonMotha | very different |
| 01:01.09 | MonMotha | if you want, you can use only primitives, and then you pretty well know what's going to get inferred |
| 01:01.18 | MonMotha | but always blocks are SO handy |
| 01:01.22 | ds2 | if I only managed to flunk enough classes to delay things by 2 years I could've gotten the verilog version of the class |
| 01:01.44 | MonMotha | yeah, pretty much nobody uses palasm these days |
| 01:01.49 | MonMotha | at least not people who haven't been using it for years |
| 01:01.55 | MonMotha | I don't really know much about it |
| 01:04.15 | ds2 | it is too bad there isn't more uniformity; ADI makes a nice processor with a small GAL on the outputs but you have to compile the maps yourself |
| 01:05.55 | MonMotha | Atmel has something like that too, but you need to use their dev tools for the FPGA, which are expensive |
| 01:06.04 | MonMotha | it's a full blown FPGA with a built in micro, though |
| 01:06.25 | ds2 | so does Xilinx |
| 01:06.38 | ds2 | but the ADI one is affordable... cheap enough they were giving away eval sticks at ESC |
| 01:08.19 | MonMotha | ah |
| 01:08.34 | ds2 | its a tiny ARM7 with some flash and RAM |
| 01:08.34 | MonMotha | yeah, my Virtex 2 Pro is not exactly an "affordable" chip |
| 01:08.40 | MonMotha | like $1000 in single unit quantity |
| 01:08.41 | ds2 | about the size of a Hammer board actually |
| 01:09.07 | ds2 | plus the vitrex will probally inhale a 40 pack brick of AA's in an few minutes |
| 01:09.32 | MonMotha | depends on what you do with it |
| 01:09.43 | MonMotha | they aren't noted for being low power, though, no |
| 01:09.52 | ds2 | they are those huge square, silver BGA packages, right? |
| 01:10.10 | MonMotha | yes |
| 01:10.15 | MonMotha | they are not available in leaded packages |
| 01:34.09 | *** join/#edev Crofton (n=balister@66-207-66-26.black.dmt.ntelos.net) |
| 02:28.36 | *** join/#edev samrobb (n=sam@dsl093-061-094.pit1.dsl.speakeasy.net) |
| 03:59.55 | *** join/#edev redblue (i=star@ppp087.108-253-207.mtl.mt.videotron.ca) |
| 04:24.03 | *** join/#edev DrZimmerman (n=theo@2001:5c0:8fff:fffe:0:0:0:6af1) |
| 05:32.50 | *** join/#edev catastrop (n=catastr0@d14-69-178-65.try.wideopenwest.com) |
| 05:33.25 | *** join/#edev DrZimmerman (n=theo@2001:5c0:8fff:fffe:0:0:0:6af1) |
| 06:57.15 | *** join/#edev linac (n=lin@121.41.38.239) |
| 08:56.21 | *** join/#edev linac (n=lin@121.41.10.118) |
| 09:11.34 | *** join/#edev lyakh (n=lyakh@dynamic-unidsl-85-197-26-24.westend.de) |
| 09:42.58 | *** join/#edev emil|k (n=emil|k@e179232033.adsl.alicedsl.de) |
| 09:52.05 | *** join/#edev linac (n=lin@121.41.52.184) |
| 10:33.28 | *** join/#edev wmat (n=btraynor@208.68.91.47) |
| 10:40.13 | *** join/#edev Off_Namuh56 (i=GPS@gateway/tor/x-ba7ee11e3ca5fa56) |
| 10:55.59 | *** join/#edev emil|k (n=emil|k@e179238150.adsl.alicedsl.de) |
| 10:57.17 | *** join/#edev ico2 (n=ico2@modem-2767.chimpanzee.dialup.pol.co.uk) |
| 11:29.24 | *** join/#edev ico2 (n=ico2@modem-2767.chimpanzee.dialup.pol.co.uk) |
| 12:35.38 | *** join/#edev steliosk (n=Stelios@ipa226.211.tellas.gr) |
| 13:04.48 | *** join/#edev GPSFan (n=kenm@12.10.255.248) |
| 13:04.48 | *** mode/#edev [+o GPSFan] by ChanServ |
| 13:28.10 | *** join/#edev ico2_ (n=ico2@modem-149.cheetah.dialup.pol.co.uk) |
| 13:30.37 | *** join/#edev FireEgl (i=FireEgl@Sebastian.Atlantica.IRCNut.Com) |
| 14:37.24 | *** join/#edev redb1ue (n=star@ppp028.108-253-207.mtl.mt.videotron.ca) |
| 15:29.02 | *** join/#edev ico2__ (n=ico2@modem-2212.chameleon.dialup.pol.co.uk) |
| 15:54.45 | *** join/#edev DrZimmerman (n=theo@2001:5c0:8fff:fffe:0:0:0:6af1) |
| 16:05.39 | chouimat|ibook | morning |
| 16:12.05 | *** join/#edev chouimat (n=dieu@kde/developer/chouinard) |
| 16:19.11 | *** join/#edev lyakh (n=lyakh@dynamic-unidsl-85-197-26-24.westend.de) |
| 17:04.13 | *** join/#edev chouimat (n=dieu@kde/developer/chouinard) |
| 17:29.25 | *** join/#edev ico2_ (n=ico2@modem-1685.cheetah.dialup.pol.co.uk) |
| 18:16.19 | *** join/#edev chouimat|busy (n=dieu@r2351064.cidc.net) |
| 19:09.44 | *** join/#edev davidc__ (n=davidc__@c-24-8-151-128.hsd1.co.comcast.net) |
| 19:14.40 | *** join/#edev Crofton (n=balister@hc6521406.dhcp.vt.edu) |
| 19:16.35 | *** join/#edev joe_bleau (n=joe_blea@adsl-75-13-178-169.dsl.fyvlar.sbcglobal.net) |
| 19:50.24 | *** join/#edev redb1ue (i=star@ppp133.109-253-207.mtl.mt.videotron.ca) |
| 20:08.47 | *** join/#edev joe_bleau (n=joe_blea@adsl-69-152-205-210.dsl.fyvlar.swbell.net) |
| 20:32.19 | *** join/#edev chouimat (n=dieu@kde/developer/chouinard) |
| 21:09.32 | *** join/#edev chouimat (n=dieu@kde/developer/chouinard) |
| 21:21.18 | *** join/#edev samrobb (n=sam@dsl093-061-094.pit1.dsl.speakeasy.net) |
| 21:44.07 | *** join/#edev Crofton|work__ (n=balister@ossie.wireless.vt.edu) |
| 22:08.11 | *** join/#edev Crofton (n=balister@66-207-66-26.black.dmt.ntelos.net) |
| 22:49.46 | *** join/#edev chouimat (n=dieu@kde/developer/chouinard) |
| 22:55.35 | *** join/#edev linac (n=lin@121.41.52.184) |