CUViper 0 Posted July 1, 2003 I was wondering - does anyone have the Windows XP SP1 DDK CD that they could rip and let me download? I ordered one from Microsoft for free (well, with $15 shipping), but my order status says they're backordered, and they aren't giving an ETA. I need this to write a driver for a new product at work, so the sooner I get this, the better. So here's hoping someone here can hook me up... Share this post Link to post
CUViper 0 Posted July 2, 2003 It would be nice if MS would let me download it, but... Quote: We will not be offering the Windows XP SP1 DDK as a download on the WHDC Web site. However, it is part of the standard MSDN® subscription download and is included with the standard MSDN subscription And I don't have an MSDN subscription. I emailed them to see if they would let me download it, since I did place an order, but I haven't heard back yet. Share this post Link to post
CUViper 0 Posted July 2, 2003 I just got a response back from the Microsoft folks, and they just quoted me what I said in my last post... Share this post Link to post
CUViper 0 Posted July 17, 2003 It's actually for work - I've got a summer internship at Sonora Medical Systems in Longmont, CO. They've got a product called FirstCall, which currently runs through the parallel port to communicate with the PC. Of course many computers these days do not have a parallel port (especially laptops), so I am upgrading the system to run through USB. Which in turn requires custom USB drivers... I actually did get the CD earlier this week, so I'm ready to go. If you ever need to write a driver, the DDK is awesome - it comes with a ton of examples. I was able to take one of the examples, modify a few of the files in it, and I already have my basic communication working! Share this post Link to post
CUViper 0 Posted July 17, 2003 Quote: That's an EXCELLENT internship & experience man... seriously! Consider yourself VERY lucky!!! You must be pretty sharp if your prof.'s recommended you for it by the way, they don't do that for just any student! Yeah, I'm psyched. There were really slim pickings for intern's this year, so I was pleased when this one pretty much fell in my lap. And it's not one of those worthless interships where you just sharpen pencils for engineers and whatnot. It's a fairly small company so I have the opportunity to get involved in a lot of different projects. Quote: Yup, I have looked it over in Windows 2000 DDK (which I have here along with MASM 6.x release etc.), & that is NICE about MS providing the templates they do! (This is why I always really disliked the fact that guys that write drivers for Win32 Os' @ least try to "snow" others in saying it is SO massively difficult, when I know that templates for most types of devices are already available in the DDK!) Without the templates I would be completely lost. I consider myself a pretty proficient programmer, but there's whole a lot of arcaen-looking details in the innards of a driver. But with the example, I can just leave the basics alone and tweak the parts to my needs... Quote: * Not saying it's not tough, it probably is, because you have to deal with timings in the traces on boards & idiosyncracies with diff. equipment's ASICS, but it's not like working with no idea or help either from scratch! The hardware is a lot harder to deal with than software issues. The 80-100 volt pulses into the ultrasound transducers tend to wreak havoc on digital signals, which has been my biggest headache this week... This job is a great learning experience for me though, because I'm getting a lot of practical knowledge (versus books and equations) in a lot of different areas... Quote: P.S.=> Congratulations on being awarded that internship #1, & secondly kick some butt! That job alone may be setting you up for a future writing device drivers, & who knows? One day, you may write drivers we all use here @ this forum for devices we use... apk Thanks! Share this post Link to post
CUViper 0 Posted July 17, 2003 Ok, a quick summary of how it works, is the system tests each element of a probe one-at-a-time by sending the big pulse down the element. Most transducers have between 128-256 elements. So there are a ton of relays to pick out the correct element during the test. And since you need some way to control the relays, the selection part is done with a digital circuit. So there's really no way to completely separate the digital from the analog. We do the standard stuff like separating ground planes and minimizing parallel traces in the layout, but there's still a whole lot of EMI from the pulse. Having said all that, it does still work really well, and the only time we have problems is when we're prototyping new adapters. Which is what I was doing the past few days was debugging a new board that was designed by a previous intern (and not well designed if you ask me ;( ) Now if you've got some idea for a different way to select the elements, please do tell! Hell, if it's good enough, I'll even give you the credit! BTW, the transducer is just a fancy name for the ultrasound probe. If you follow the link I gave, you'll see a small picture of the system -- the probe is on the cable coming off the top of the box. The length of the cable is fixed by the OEM manufacturers, but both the cable and the probe are well shielded. It's only where the pulse goes through the board that it has a chance to cause problems. Share this post Link to post
CUViper 0 Posted July 18, 2003 The components are very well shielded wherever possible, but as I tried to explain before, the very nature of the switching circuitry requires SOME interlinking between analog and digital. It's just that little part where they connect that manages to create interference. The thing that makes dealing with EMI of this nature so difficult is that you really can't model EMI into any calculations. It really ends up being just a guessing game of finding a solution that will solve the interference without creating other problems. I did get my problem fixed - in this case a ferrite bead on the clock line cleaned it right up. As one of the engineers was telling me, dealing with EMI really becomes a matter of experience - as you go on you develop your own cookboook of solutions to various types of problems. My boss, who was the lead designer of the system and is now VP of R&D, is very good about encouraging me to look for better solutions. The relays especially are my biggest gripe - they work well, but they also account for about 80% of the board realty. I have yet to find anything that would be any better though, or even the same. Which leads evidence that it really was thought out well the first time through. Share this post Link to post
CUViper 0 Posted July 18, 2003 Quote: Whew... that does not sound like a ton of fun! That's like coding without watch windows & step tracing on my end doing MIS/IS/IT info.systems type work (where I make my living). I do get a sort of watch window, by using oscilloscopes to probe various traces and characterize their behavior. Perhaps a more apt analogy of the problem would be that I don't have the source code. To tie this in to our other thread, I have to try to "disassemble" the analog (ie real-word) characteristics of the circuit and EM interactions, and then modify (hack?) it to meet my needs. Or perhaps it's just late and I'm seeing correlations where they don't exist... :x Quote: You're lucky, sounds like you have a GOOD boss who's actually been there, I have had many of those, but also many who did not know a thing about this stuff... yet ran IT/IS/MIS dept.'s, which made not alot of sense to me! I agree - the right leadership makes all the difference. Already we've established a sort of "mentoring" relationship, which is great for me to enhance this learning experience. Quote: I think you will do really well, I wish you luck... you get good at it, you can probably make a nice living (since most coders don't gravitate to that end of things, I did not for one), & maybe even write drivers for the boards we talk about here. Like Video cards... who knows? One day, You may be @ say, NVidia! Thanks man... I can only hope that I get into a big company like that. I may be crazy, but I want to go even deeper than drivers and be on the team that actually designs the boards and chips! So who knows... Share this post Link to post
AndyFair 0 Posted July 19, 2003 CUViper Just a thought, but is there a way that you could isolate the 2 parts of the circuit using, for example, optical links rather than electrical links. I started to get into RobotWars a couple of years ago, and this problem of EM interference happens in robots too (low power radio control circuits controlling high power motor control circuits) - one of the solutions mooted was whether you could use some kind of an opto-isolator to separate the 2 parts of the circuit? I don't remember the specifics, but it's a thought. And congrats on the internship, BTW - I know how much competition there is for those things! Hope you learn lots and have fun doing it! Rgds AndyF Share this post Link to post
CUViper 0 Posted July 19, 2003 Quote: Just a thought, but is there a way that you could isolate the 2 parts of the circuit using, for example, optical links rather than electrical links. I started to get into RobotWars a couple of years ago, and this problem of EM interference happens in robots too (low power radio control circuits controlling high power motor control circuits) - one of the solutions mooted was whether you could use some kind of an opto-isolator to separate the 2 parts of the circuit? I don't remember the specifics, but it's a thought. Hmm, interesting idea... That would certainly help with ground-bounce issues. It won't do anything for the radiated EMI, but we might be able to use more aggressive shielding with the isolation in place. I will look into it! Quote: And congrats on the internship, BTW - I know how much competition there is for those things! Hope you learn lots and have fun doing it! Thanks! Share this post Link to post
