Dave there is a crying need to get some kind of general DRAM circuit that can get us GB or DRAM for these 32 bit CPUs. It's way past my pay scale! Any help you or others can give would be great. John John Monahan Ph.D e-mail: <mailto:monahan at vitasoft.org> monahan at vitasoft.org Text: <mailto:monahan at txt.att.net> monahan at txt.att.net From: n8vem-s100 at googlegroups.com [mailto:n8vem-s100 at googlegroups.com] On Behalf Of yoda Sent: Tuesday, August 13, 2013 10:22 AM To: n8vem-s100 at googlegroups.com Cc: cctalk at classiccmp.org Subject: Re: [N8VEM-S100:1811] S-100 68K CPU board V3 Hi Andrew!!! Yes I will be posting a lot of stuff in a couple of weeks (source code, pictures, etc). Yes I have not forgotten the other project and am using this as a stepping stone for monitor and other code. I have been thinking about it a lot and have come to the conclusion we may be better off with just a 68040 SBC then trying to marry the 68360 to one. The 68360 introduces a lot of complexity that is not really necessary. I think we are going to have to accept some type of small board with SMT memory on it to get to where we want to go. I think we are seeing the same thing happen on the 80386 board as well. I am looking at ways to make a small FPGA into a memory controller and still have room for some more functions in it like timers which we will need for linux or other unices. I am also going to look at the amount of work required to put uclinux on the s100 68K. It will be a swapping version but a lot of code should be reusable going forward so it is not a wasted effort. We should probably chat sometime on the direction we want to go on this project and come up with a plan. Dave PS I am glad we finally got here too - this is the architecture I like the most (not an Intel fan). On Tuesday, August 13, 2013 9:05:29 AM UTC-5, lynchaj wrote: Yay! Woo Hoo! Great news! We have S-100 68K CPU running! Once this project is working maybe we could take another look at the 68360 P1 and/or 68040 SBC? This is really great news! Please post pictures and notes on the wiki! Let's get the word out! Thanks and have a nice day! You've already made mine! w00t! Andrew Lynch PS, this project has been in work for 2 years+ so I am very happy to see it working. From: n8vem... at googlegroups.com <javascript:> [mailto:n8vem... at googlegroups.com <javascript:> ] On Behalf Of yoda Sent: Monday, August 12, 2013 10:40 PM To: n8vem... at googlegroups.com <javascript:> Cc: mon... at vitasoft.org <javascript:> Subject: Re: [N8VEM-S100:1808] S-100 68K CPU board V3 Hi John Thought I would give an update - found the problem with memory - it turns out pin 1 of U2 on the board had a cold solder joint - guess it finally gave out. Reflowed the solder - the ground plane is a pretty big heat sink around pin 1 and can easily get a cold solder joint. Any ways after repairing that and playing with jumpers and switch settings I have 1 Meg of memory working with the S100-68K-V3. I also have the oscillator at 16MHz so the CPU is running at 8MHz and is a buss master. I am using the short S100 mother board from the group so it may have better noise immunity than the 22 slot mother board like you are using. I will do some more testing and want to boost the memory to at least 2 Meg for some experiments. I will get started next on the new monitor and getting data to load from the IDE CF card. Looks like I am on my way to getting CP/M 68K ported. I have a little clean up to do in a couple of areas (upgrade to latest compiler code, etc). Dave -- You received this message because you are subscribed to the Google Groups "N8VEM-S100" group. To unsubscribe from this group and stop receiving emails from it, send an email to n8vem-s100+unsubscribe at googlegroups.com. For more options, visit https://groups.google.com/groups/opt_out.

Hi Ryan! ?Thanks! ?We need help with a dense, fast DRAM/SDRAM/DDRx board in the worst way. ?Phat stacks of DDR RAM in the GBs and that requires DDR encoder/decoder logic. Neither John nor I know anything about it other than it is quite complex. ?I've used the old school DRAMs (4164) but they are nothing in comparison to DDRx or whatever its called. ? Help! ?This is an urgent appeal. ?The S-100 80386 CPU board is basically boned without a decent RAM board to go along with it. ?We have an 8MB SRAM board that works and can design an 32MB SRAM board but that just gets you in the door. ? Sophisticated Linux/BSD OS's all require GBs of RAM and 32MB won't even get a boot screen. ?Maybe a dirty look from the barest of the bare bones Linux distributions. ?Not much else. Help! ?Thanks and have a nice day! Andrew Lynch ________________________________ From: Ryan K. Brooks <ryan at hack.net> To: On-Topic and Off-Topic Posts <cctalk at classiccmp.org> Sent: Tuesday, August 13, 2013 2:05 PM Subject: Re: [N8VEM-S100:1811] S-100 68K CPU board V3 On 8/13/13 12:45 PM, John Monahan wrote: Opencores has a number of ddr-ram controllers.? PCB routing is the toughest part IMHO. If you guys need help with the 68k stuff, let me know, I've done 060 to 008. -Ryan

On 2013-08-13 12:26, Andrew Lynch wrote: Did you have a look at PSRAMs? I know, they only exists in BGA, but you could put 4/8 on a small PCB an make a DIL out of it. The controller for it would be definitely easier ...

This is true only if you insist on using all the latest bloatware. NetBSD/i386 4.x is perfectly usable with 64M, provided the things you use are not themselves gobbling RAM as if megabytes were free. (I expect 5.x is too, but I run it on only two machines, neither of which has less than multiple gigs of RAM, so I can't be sure.) 32M works, but can't self-host (which for my purposes amounts to "broken") because the gcc it comes with is too much of a pig. I suspect 16M would install just fine and might even run OK provided you're careful what you do, though I haven't tried it. Older NetBSD works on much smaller machines; I even ran a 5M 1.4T machine, though that was tight. Of course, using the latest in flashy bloatware may be what you mean by "sophisticated", in which case what you say is true but tautological. Linux? Other BSDs? I don't know. My relationship with Linux is probably best described as mutual antagonism; my relationship with the other BSDs is approximately equal to my relationship with Botswanan stock-trading houses - ie, nonexistent. /~\ The ASCII Mouse \ / Ribbon Campaign X Against HTML mouse at rodents-montreal.org / \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B

Wow, this is old but got stuck in my drafts box. Thought I'd send it now; still willing to work on this sort of thing, but I have massive time constraints for the next few months. On Aug 13, 2013, at 14:26, Andrew Lynch <lynchaj at yahoo.com> wrote: I'm glad to help (I do this professionally), but I'm out of town for the next two weeks. Ping me after that, though; you should be able to at least do some standard 3.3v SDRAM (not DDR) with a small CPLD. You can do DDR 1/2/3 with small, low-cost FPGAs. All of this will require surface mount (neither RAM nor FPGAs come in anything else now), but you could at least use SODIMMs to avoid most work with BGAs. I can't do the pin math in my head to think if you could use QFP packages for the FPGA, but I suspect not; this may be something that would be worth doing at an assembly house, even if just for the SMD components. You're not going to get away without using surface-mount bypass caps for an FPGA, either. Long story short, if you can deal with the space provided by standard SDRAM, it's probably your best bet. Let me know how I can help. I'm glad to help with schematic and layout in my spare time. Don't even THINK of touching SDRAM layouts with an autorouter. It will not save you any time or effort, because you will spend more time cleaning up after it than it would have taken to hand route it to begin with. - Dave

On 12/13/13, 1:27 PM, Al Kossow wrote: 060. I've built a couple of boards with 060 connected to 3.3v SRAM. Also, there's 5v tolerant 3.3v cplds.

On 12/13/13 2:07 PM, Al Kossow wrote: Ah, gotcha. Well, then yes, no go.

On Tuesday, 13 August 2013 18:22:09 UTC+1, yoda wrote: Shiny :) I've been looking for an S-100 board which could potentially run NetBSD. Theoretically the baseline would be a 68020 plus 68851, though '040 would probably make more sense there as well :) (or Coldfire, but thats a different path). If the 040 SBC gets to the "pre-order but with no defined timescale" stage then I'd definitely be interested :)

On Tuesday, 13 August 2013 20:09:14 UTC+1, yoda wrote: 16MB should be quite usable for even a recent NetBSD. I wouldn't recommend compiling using a recent *gcc* in that amount of RAM, but otherwise the system should be fine [*]. 4MB would boot but probably not run anything useful in a timely fashion :). Certainly it runs on a variety of m68k machines where 16MB was a common capacity (mac68k, sun3, etc). I suspect the code for the motorola VME SBC could be a reasonable place to start ( phttp://www.netbsd.org/ports/mvme68k/ ) [I've run X under NetBSD on a Dreamcast in 16MB with root & swap over NFS, but that was more proving a point :]