Re: gEDA-user: series terminators
After all that, I forgot to mention the line length. Rule of thumb: if the first reflection gets back to the driver before the driver has finished transitioning, then you probably don't have a problem, and you don't need termination. In your case, that's 12, or about 1ns. So, if your drivers have edge rates of ~1ns or faster, you may have a problem. Another rule of thumb: modern CMOS circuits which are designed to go 'fast-ish' (quick clock-out, so high slew rate drivers) probably have a slew rate in the range of 1.5 - 2.5ns. I don't know what chips you're using, but you probably don't have a problem. Evan ___ geda-user mailing list geda-user@moria.seul.org http://www.seul.org/cgi-bin/mailman/listinfo/geda-user
Re: gEDA-user: series terminators
I don't know what chips you're using, For completeness, M32C/83 and CY7C1049B ___ geda-user mailing list geda-user@moria.seul.org http://www.seul.org/cgi-bin/mailman/listinfo/geda-user
Re: gEDA-user: series terminators
Ahh, I'm using the CY7C1049 in some of my stuff. Did you need terminators? ___ geda-user mailing list geda-user@moria.seul.org http://www.seul.org/cgi-bin/mailman/listinfo/geda-user
Re: gEDA-user: series terminators
DJ - On Thu, Sep 21, 2006 at 11:47:38AM -0400, DJ Delorie wrote: I'm thinking I should design in some series resistors on the address and data lines for my RAM expansion board (30MHz, about 6 of 8 mil trace on DS FR4, 5v). 5 Volts? How quaint. You didn't give the thickness of the FR4 between the trace and the ground plane. Series resistors for a bidirectional trace go near the middle of their length. For unidirectional traces, place the resistor anywhere from the middle to the source end. How does one go about calculating an ideal resistance? The mcu spec doesn't have a min rise time spec, the ram chips are 20ns (50MHz) and imply they want a 3ns rise time. Think of it this way: you want to charge a parasitic capacitance (guess 10 pF) with a rise time of 3 ns. That means you need a series resistance less than 3.0 ns / 10 pF = 300 Ohms. I figure I can put in zero-ohm resistors for now, and experiment with one of the lines, but I'd rather just come close enough on the first try ;-) Any resistor in the range of 25 to 150 Ohms will damp the ringing. It takes a good FET probe and a fast 'scope to characterize the system well enough to optimize on the bench. Whenever I have this concern, I just put in 100 Ohms, observe that it works, and go on to the next chore. - Larry ___ geda-user mailing list geda-user@moria.seul.org http://www.seul.org/cgi-bin/mailman/listinfo/geda-user
Re: gEDA-user: series terminators
5 Volts? How quaint. Yeah, well, that's what the eval board uses. My furnace controller uses 5v too, because that's what the LCD panels need. It throws in a bunch of issues about level converting, though. You didn't give the thickness of the FR4 between the trace and the ground plane. Double sided, 0.062. Series resistors for a bidirectional trace go near the middle of their length. For unidirectional traces, place the resistor anywhere from the middle to the source end. I figured that; they're going on the SRAM side of the DIN connector, about in the middle of the run. It takes a good FET probe and a fast 'scope to characterize the system well enough to optimize on the bench. I have a 32 Ms/s scope and a 500MHz logic analyzer, neither of which is great for this kind of test, but might be close enough. Any resistor in the range of 25 to 150 Ohms will damp the ringing. Whenever I have this concern, I just put in 100 Ohms, observe that it works, and go on to the next chore. Ok, sounds reasonable. Thanks! ___ geda-user mailing list geda-user@moria.seul.org http://www.seul.org/cgi-bin/mailman/listinfo/geda-user
Re: gEDA-user: series terminators
DJ - On Thu, Sep 21, 2006 at 12:18:46PM -0400, DJ Delorie wrote: Double sided, 0.062. How quaint. ;-) The board I'm in the middle of bringing up is 6-layer, 0.045. I've probably linked to it before: http://recycle.lbl.gov/llrf4/ So the distance between a trace and its ground plane is an order of magnitude smaller than what you have. And yes, it's laid out with Gnu/PCB. I figured that; they're going on the SRAM side of the DIN connector, about in the middle of the run. On the board above, I have some 100-ohm series resistors [*] on digital traces. Signalling rate could be as high as 105 MS/s, and the longest traces are about 4.5 cm. I have a 32 Ms/s scope and a 500MHz logic analyzer, neither of which is great for this kind of test, but might be close enough. Only if you work with slow parts. Edge speeds faster than 1ns are fairly common these days. - Larry [*] On the board photo, look in the center of the board, to the left of the big FPGA. There is a vertical line of eight resistor packs, in series with the data pins from the four high speed ADCs just to their left. ___ geda-user mailing list geda-user@moria.seul.org http://www.seul.org/cgi-bin/mailman/listinfo/geda-user
Re: gEDA-user: series terminators
Double sided, 0.062. How quaint. ;-) Yeah, well, it's just going to sit on my desk anyway. Plus, it's panelized with the challenge boards, so cost is an issue. I was extra careful about adding planes to the bottom side, and bypassing all over the place. The proto-pcb industry is starting to catch up with what hobbiests want, but they're still way behind what high-tech boards demand. The board I'm in the middle of bringing up is 6-layer, 0.045. The furnace board will be 4 layer 0.062 (it's cheapest). Oh, for checking part footprints, I've found the paste layers to be most useful, at least for SMD parts. Only if you work with slow parts. Edge speeds faster than 1ns are fairly common these days. Well, I want slower edges ;-) (And I can't afford faster tools) ___ geda-user mailing list geda-user@moria.seul.org http://www.seul.org/cgi-bin/mailman/listinfo/geda-user
Re: gEDA-user: series terminators
If you can afford a 4 layer, you can definitely afford a faster scope. It's not just the cost, it's the justification and size. I just got a 500MHz logic analyzer that's the size of my hand (Intronix Logicport) and I have a 32Ms/s parallel-port scope (er, about 10-15MHz) and a Tek 561A (er, about 1MHz these days) with some plugins. The 4 layer board will go into the furnace, which will easily repay its cost, and it's slow enough that my existing tools are sufficient. The existing furnace controller has already paid for itself. Now, if I start doing more high speed stuff (which I hope happens :), I'll probably start shopping for a USB-based high speed scope then. I don't want to use up a lot of space on a big scope, I have too much stuff in my office already. ___ geda-user mailing list geda-user@moria.seul.org http://www.seul.org/cgi-bin/mailman/listinfo/geda-user