• Please review our updated Terms and Rules here

68008 Wire-Wrap Prototype Board for VME

H

Harry.Speer

Member
Joined
Dec 14, 2022
Messages
32
I've decided to use VME bus - or rather a subset of VME - as presented on the Tom Storey GitHub here: https://github.com/tomstorey/COMET
He's using only the signals needed for a 68000 16-bit Data Bus and Addresses from A1 to A23 - IIRC. Not intended for multiple CPU's but allowing for use of DMA on the bus for a Slave Card like Floppy Controller or Hard Drive Controller. He's put up an 8-slots VME Backplane and also both a 100mm x 160mm VME Prototype Card and also a 100mm x 220mm VME Prototype Card. So I'm thinking use a PCB House to run a few of each and then get to work.

His Backplane-8 - 8 VME Slots here: https://github.com/tomstorey/COMET/tree/master/VME_Backplane_8ActPas

His Prototype 160 Card here: https://github.com/tomstorey/COMET/tree/master/VME_Protoboard_160

His Prototype 220 Card here: https://github.com/tomstorey/COMET/tree/master/VME_Protoboard_220

Does anyone here have experience either with HomeBrew 68k SBC projects and / or Someone with experience working with the VME Bus ??

This is my first HomeBrew Computer and I've chosen the "smallest" of the 68000-Family of CPU's. Includes the PLCC 68008 via a PLCC to DIP package converter card, a UV EPROM or a EEPROM or a 512k x 8 Flash Chip with a Read-Only Jumper in place to prevent accidental writes.
Then a Single Static RAM 512k x 8 - One Chip !!! and then the usual 68681 or equivalent DUART, with MAX232-Family Chips for RS232 levels shifting. Possibly a 68230 PIT, a Battery-Backed RTC, a "ticker" clock, and Lastly or Firstly a Power Supervisor Chip like ADM705 and a Traditional 555 Timer for the Pair o f /RESET and /HALT through Open-Collector Buffer or Inverter with each having a Pull-Up Resistor and a Buffered LED for Each. A few other well-placed LED's like Address Line A23 and A22 and so on. A Debug Display can be very useful.

So anyone with experience, please chime in !! Harry S. Speer
 
There's a 'EC' variant of the 68000 that's worth looking into, Harry. I've built an entirely homebrew SBC around it. It is very fast. There's a thread about it in this sub-category. I run mine at about 28 MHz!.

I think that I remember seeing some 68k software quite a while ago written by Harry Speer. Is that you?

Roger
 
Thank You Roger and Dave. Good info from both of you.

But for my first foray into hand-building my first computer,
a type of handcraft I have never done before, I want this to
be "Dead Simple". Simple and Small, the 68008 has less
data bits, 8 rather than 16, and less Address Lines. Instead
of DIP I want to use the PLCC chip of the 68008 to get a
couple more Address Lines. 68008 in 48-Pins DIP has A0
to A19 1024k Bytes or 1 MB megabyte. The 52-pins PLCC
has A0 through A21 = 4 MB mega-bytes. Use 74LS138 to
select one of 8 "pages" that are 512k x 8 each. ROM starts
at Zero on Power-Up with /HALT and /RESET together, then
it Swaps out and RAM swaps in. Both the Static RAM and the
Flash as ROM max-out as DIP packages through-hole at
512k x 8 each. And Pages 0 through 3 Might be RAM, and
Flash normally on Page 6. Page 7 Selects Memory-mapped
IO devices up at the Top-Most Page.

My Idea is to keep this First one absolutely Dead Simple.
And Use as Few Chips as possible, also use as few Address
and data lines to wire-wrap - Less being better.

I have never done any of this before, so I am "Jumping Back"
into my own personal history to my college days, my early 20's
Like 1980 when I was 21 Y.O. In College. I'm now Retired,
mostly, and I want to try to do "Electronics From Scratch".

When I left College, I took Computer Repair, Upgrades, Installs,
etc, etc. Those kinds of jobs. All of the parts were "FRU's" -
Swapping whole boards, whole power supplies, etc.

I had to get work I could do and that was my "career" for about
45 years to now. 2025 - 1980 = 45. Now Age 66.
So I am Deliberately making this first one as "dead simple" as
possible. The "K.I.S.S." principle. 68k does NOT get smaller
than 68008, but PLCC because I want those "Extra" Address
pins that DIP doesn't give me. I may use a Lot of PLCC to
DIP Socket adapters. Some of the Best Prices favors the
PLCC chips. Also PLCC package me be harder to "Fake"
Over in China. I should think that DIPs are easier to "Black
Top" than a PLCC package.

Also I want to get as many parts as I can, especially the Chips,
from Reputable Vendors like Mouser and DigiKey. Simple
switches - Both On/Off and Momentary I might risk Ebay or
Jameco. Also I discovered that UtSource.Net has a LOT
of Obsolete Chips that Mouser nor Digikey is likely to sell.
I've also been alerted to check out UnicornElectronics.Com

Can either of you suggest other online stores for these kinds
of Parts ?? Websites ??

For my Second project board, it probably will be a full 16-bit
data bus 68000 or "HC" or "EC" or whatever. And that will be
the main CPU on the bus. Either VME Bus or the Bus that
Steve has over at Mega-Micros.co.uk - some very nice work.
I'm struggling with my decision.

I am struggling between which bus to go with. So neither simple
nor easy to decide. Each has it's merits. VME was historically
started by Motorola for the 68000 family, or so I hear. Makes
sense, to me, to use that. I looked at the RC2014 and RCBus,
also in the S100 Google Group and S100 on Facebook. But there
is no sense of "This is a good fit" between 68k and 68030 and S-100.
S100 tops out at 16-bits data. I want a bus which I can properly
wire up for 32-data bits and 32-address bits. VME can do that.
But S100 obviously cannot.

I will keep "Lurking Without Posting" on the S100 Google and
Facebook, RC2014 on Google and Facebook, RCBus and
such in Google Retro-Comp, which is not restricted by Brands
of CPU's. But I'll continue "Watching, Listening, and Hopefully
Learning" on all of these. But me trying to only post on a
few Limited Numbers of Groups, Groups where the 68K Motorola
Family of CPU's and the VME Bus, or similar, are Welcoming to
me.

I have thought a long time on this. I really believe that my Very
First One should be Dead Simple. The "K.I.S.S." principle.

Thanks Again,
Harry
P.S. Sorry for the long post. HSS
 
A barebone 68008 can be as simple as 68008, RAM, 22V10 PLD, and oscillator. 22V10 serves as both the bootstrap ROM as well as a bitbang serial port.
Bill
 

Attachments

  • BB68008.jpg
    BB68008.jpg
    1.2 MB · Views: 17
If you are really after K.I.S.S., I'd avoid the S-100 68k card that is available through the S100computers group. I've built a LOT of 68k SBCs (68000, 68010, 68020), and I was unable to get that one to work for me!

As far as the availability of chips goes, you've named the best vendors. I've always had good luck with Unicorn, and of course, Jameco, Mouser, etc. are pretty good too, even if their prices aren't the best. Be very, very careful of ePay. If you need to go offshore, UtSource is very good.

PLCC is a very good idea!

Roger
 
Plasmo said:
A barebone 68008 can be as simple as 68008, RAM, 22V10 PLD, and oscillator. 22V10 serves as both the bootstrap ROM as well as a bitbang serial port.
Bill
Click to expand...
Yes, Bill. Thank you. I've always enjoyed your 68k chip designs, BTW. However, at this time, I do not have a programmer for programmable logic chips - PALs, GALs, CPLDs, etc. And No Experience with them either. I don't even have an UV EPROM nor EEPROM Programmer either. (Nor a UV Lamp eraser with a timer). Not yet, on both. I do look forward to it. So this board will be with 74LSxxx Family chips or 74ACTxx or HCT or "F" or "S". Your basic TTL logic chips. And at the moment, I have KiCad installed but haven't yet created anything with it. So I bought a pad of graph paper for doing schematics by hand with pencils. Very "Old School". :-) At age 66, whaddya expect ?? ;-) I have a LOT of learning ahead.
 
norwestrzh said:
If you are really after K.I.S.S., I'd avoid the S-100 68k card that is available through the S100computers group. I've built a LOT of 68k SBCs (68000, 68010, 68020), and I was unable to get that one to work for me!

As far as the availability of chips goes, you've named the best vendors. I've always had good luck with Unicorn, and of course, Jameco, Mouser, etc. are pretty good too, even if their prices aren't the best. Be very, very careful of ePay. If you need to go offshore, UtSource is very good.

PLCC is a very good idea!

Roger
Click to expand...
Yes, Roger, I know EXACTLY why John Monahan's 68000 S-100 CPU card fails - He left out Either a 555 Timer for /RESET & /HALT on Power-Up - or - Use a Power On /RESET & /HALT from a Power Supervisor Chip. Examples ADM705 or a Texas Instruments TL7705ACP, etc. Etc. This OOPS !!! is why the card always fails. And if he left THAT out of the card, What else did he do wrong ?? For Example, his DMA that is NOT Compatible with the IEEE-696 Specifications ?? So Yes, I am Well Aware of John Monahan's 68K S-100 CPU card and his 68030 S-100 CPU Card. Which is why I decided to do it myself. Nuff Said ?? An S-100 68k SBC Card for "beginners" who want Motorola 68k on the S-100 Bus is my answer to that problem. My design starts with BOTH a Power-Supervisor Chip like an ADM705 AND a 555 timer. The physical momentary push-button /RESET switch input is debounced by the ADM705. The /RESET output of the ADM705 triggers the 555 Timer. Overkill ?? Yes, I know, but I am *Very Much* a "Belt And Suspenders" kinda guy. Make Damned Sure that the 68k CPU gets a proper Power-Up /RESET & /HALT !!! And I've abandoned my S-100 / S100 card idea. I'm looking at VME Eurocards for the VME Bus or a simplification of the VME Bus. VME can handle a 32-bit CPU Card and a 32-bit RAM card combo. Whereas John's design uses a Ribbon-Cable "Sky Bus" between the 68030 CPU Card and the 32-bit RAM Card. S-100 "native" bus can do 8-bit and 16-bit but NOT 32-bit. I've read through the IEEE-696 specs more than once. Which is why I will probably go with VME or possibly go with Steve's work of Mega-Micros.Co.UK. He has a bunch of cards designed and tested. 68000 and a Rev.1 and Rev.2 68030 motherboard with slots. Nice work. A difficult decision for me. If I go with VME, do I end up wire-wrapping every single card ?? That's not appealing. So I am not hurrying into a decision just yet. I have a LOT to think about.
 
Btw, note that the 68020 and 68030 has the ability to handle bus widths of 8, 16 and 32 bits, so at least in theory you should be able to tie the "bus width" inputs to the 8-bit selection, and just use your 8-bit memory and I/O as is, like on a 68008.

But also: I don't know how it's on the 68008, but specifically on the 68000 and 68010 there is the E clock and "E bus cycles" that makes it very easy to connect 68xx/65xx I/O chips. Thus you don't need the 68xxx UART and other chips if you don't want those, as I assume that the 6850 or 6551 are cheaper than the 68xxx UART. Going off on a tangent, it might be worth going through the effort of using Intel compatible peripherals as I would think that 16550 compatible UARTs are still cheap, and in particular ISA multi I/O cards are probably also still the cheapest way of getting various I/O for a DIY microcomputer project (minus the floppy part, that requires ISA/intel style DMA).
 

[B]MiaM[/B] - Thanks !! I had apparently forgotten dynamic bus sizing as I think it was called. So Yes, an '020 or a '030 could be "rigged" to do 8-bits only. A Temporary "testing" setting to "get up off our knees" so to speak. ​

And Yes, The 68000 Family, or at least some of them, can use the older 68B50 UART for a single serial port. However I think that would again be a "testing" step and once it works OK then step "up" to a DUART like the 68681 chip. And I would like to get a serial ports chip with deeper FIFO's - for example the MaxLinear Exar XR16C2852CJ-F - DUART - 128-Bytes FIFO but this chip is End-Of-Life. Instead maybe the Texas-Instruments TI TL16C754B 64-Bytes-FIFOs QUAD UART 68-Pins PLCC Chips. My reason for wanting deeper UART FIFO's is that the 68000 computer cannot receive data as fast as a 64-bit Desktop computer can send them. The deeper, the better. Anyhow that's kind of what I have in mind.

And Thank you Bill for the reminder and the pointer to one of your many 68k SBC boards.

I have also been looking at the VME Bus and noticing that there is no Address Line A0 on the bus. So a 68008 cpu on a prototype board would basically be limited to what's on that one board. VME seems to be 16-bit data or 32-bit data. Makes sense with origin being the 68000.

Sometimes I get my focus on one tiny thing and I forget to "Pull Back" and take a look at the "Bigger Picture". Having adjusted my focus, it really makes the most sense to me to start with a 16-bit CPU or, better yet, a 32-bit CPU. Thus an '020 or '030 CPU.

And looking at how VME 32-bit was, historically, it looks like a 32-bit CPU would need to have both the P1 / J1 bus and the remaining address lines and data lines end up on the P2 / J2 part of the VME Bus. I am reading from a Vector PDF which is showing the pinouts. So I would want to look at buying the more expensive Monolithic Backplanes. Example: Schroff 7-Slots monolithic VME Backplane. On ebay.

Schroff VME Monolithic 7-Slots 23000-467 Backplane Board​

$125.11 - 2025-07-30 - ebay-texas - https://www.ebay.com/itm/167518574509

Another reason to "Go Big" is that Linux demands the MMU in the '030 to work. Without the MMU I could use something like uClinux but I would really like to (eventually) run the mc68k Linux like Debian, for example. 32-bit Linux. I would not, however, start with that OS. ;-) More like try to get CPM-68k working. Something a lot simpler, hopefully.

I am trying to get something that I "played with" and LOVED back in my youth in college - My Own Home Sage 2 or Sage 4 computer. It's sort of like my "first love" computer-wise. A little hard to put into words.

So Thanks for all the info and perspective. Very helpful to have another set of eyes.
 
Do you plan on using any existing VME stuff? Otherwise it might be a good idea to just go with your own bus. Or maybe a combination, say VME for lager/faster cards, and say PC ISA for cards that are readily available at cheap prices, like say a card with a 16550 UART or a network card. It's fairly easy to use an UART or "ISA" multi I/O chip as is in your own construction, it seems more complicated to DIY a network card (at least unless you scavenge parts from old network cards, and/or only provide an AUI interface).
 
MiaM -

I am using the VME Eurocards, 6U sizes, similar to Vector brand, 1st because they are a standard size, 2nd, The connectors have the number of address lines and data lines I need - A1 through A31 and D0 through D31 - for use with a 68030 and higher if I ever get that far, 3rd, Because there are "cases" or "chassis" or "crates" that match the 3U size and the 6U size, 4th, Backplanes that are the same. Although the "monolithic" Backplanes with both P1/J1 and P2/J2 are more pricey and a bit more rare.
I am looking at ebay for used at a discount.

I do NOT plan to ever actually Buy Off-The-Shelf REAL VME cards to stuff in there. The Current VME is 64-bit which far exceeds my needs. I'll be using Only a Hobbyist "Reduced Function" VME Bus, Just enough for a Master 68k CPU card and the Slave cards for Memory - Flash and RAM, Parallel and Serial IO Ports, CF and / or SD, also FDC and IDE / PATA interface. These last few are for use with CPM-68k and other Disk Operating Systems. No second CPU card. And no DMA, although I may regret that later. Interface with "user" through an RS-232 Serial Port wired DCE for Console: and a Terminal, and at least one other AUX: Second RS-232 serial port for bulk uploads and downloads To / From Host Computer. NOT over the Console: port. And maybe eventually some sort of Network interface card or board. Tom Storey has that as part of his Comet68k. Over on the S100 group, they are considering how to do networking too.

The main thing is that other buses Do NOT have these same standard Power, Address, Data, and Control signals. I Began looking at RC2014, then RCBus, then S-100, and now VME. I believe that VME is what I will stick with. VME Began with the 68000 CPU. PCI and ISA also do not have this combination of "needs" or "Features".

Another factor is the number of Square-Inches (USA) of board space. S-100 was Huge at 5-inches x 10 inches. 50 S.I. 6U size 223mm = 8.8 inches by 220mm = 8.7 inches = 76.6 S.I. So also Huge. The VME 3U 100mm x 160mm cards are Tiny by comparison. 24.8 square inches if I have my math right. To me, 3U makes me feel claustrophobic which I do actually have. I don't know how anyone can squeeze into 3U - especially Without Programmable Logic chips. I do not yet have that skill.

I just now placed an order for 3x 6U prototype boards from Newark.Com - similar to Vector Boards. Two that are 6U "Big" 223mm x 220mm and the third was 6U "Small" 223mm x 160m. I can paste in more info if anyone wants to go see them. Total with Shipping and Tax = USD $33.69 for 3x of 6U Prototype boards which is a real bargain. I am comparing to prices of Vector Boards at Digikey.Com and Mouser.Com have for comparison. I had some network problems with getting into Newark.Com but I managed to get around them. Using Chrome Helped. More details if anyone is interested.

I've posted an offer for my S-100 stuff to the Seattle, Washington, USA "Retro Computers" Group and I am hoping to have those gone soon. It stings a bit but S-100 was NEVER going to do 32-bits of Address and 32-bits of Data for a 68030 or higher. My bad. *heavy sigh*

I hope that all that info helps you to understand my decision. This is what "works" for me and my goals for this Hobby. As for other cards, Tom Storey has a CF Card and a Centronics-style Parallel Printer Port card Here:

https://github.com/tomstorey

and also another VME "retroverse" here:

https://github.com/transistorfet/retroverse/

I have been browsing around a LOT for 68k related stuff and now VME stuff for 68k specifically most recently. Other Hobbyist "VME-Like" computers. I think Tom Storey says it better than I can. His Comet68k looks very nice. His Comet68k is also on Hackaday here:

hackaday.io

Introducing COMET68k | Details | Hackaday.io

<h1>In The Beginning</h1> <p>The story of COMET68k really starts back in July 2021.</p> <p>I had introduced a friend to the Motorola 68000 CPU and he was in the process of designing and building his first project using one of these processors. During the preceeding couple of years I had been...
hackaday.io hackaday.io

So I feel like this is a good choice for me, seeing that others have been down this same road for Hobbyist mc68k VME-Like HomeBrew computers. I think that there is enough info out there that I might be able to download the "docs" needed to send a PC Board "design" to a custom board maker. Something I want to do but I am just not there yet. I'm a lot slower than other people are. I've always been this way.

Others in the S100 Google Group have talked about doing this a lot, custom boards, and while they are developing for S-100 Bus - Altair / IMSAI version of the Bus and also later IEEE-696 Bus compatible. I would be going through the same steps but for VME and 68k instead. They are doing mostly 8080, 8085, Z80, Z180, Z280 and so on. Some few 8088 & 8086 too. All Intel and Zilog chips. Not Motorola. I will be Starting with KiCad and ending up receiving custom made PC boards. I am looking forward to how far I can go down this "parallel" road myself. A Different Bus and Different CPU's but much the same steps, and much the same process. I have examples to guide me. I've always had to rely on others' work for using my own.

I am also daily monitoring the Google Groups for S100, Retro-Comp, RC2014, CP/M, etc. for ideas and the steps that they are taking. A different road but a "parallel" road so to speak. I also watch these same groups on Facebook but they are much less active than the Google Groups. I tend to not do much Facebook very much. A huge sink-hole to pour time into and get very little back. :-(

And my End-Goal is to own my own 68k "Sage" Home Personal Computer. A love that has lasted for decades. About 40 years now. :-) See Here:

en.wikipedia.org

Sage Computer Technology - Wikipedia

en.wikipedia.org en.wikipedia.org

So my Thanks to you MiaM and to you Bill for your inputs.
 
Yeah, of the standard buses VME is probably the only really suitable for 68020 or higher.
 
Hi! I'm the creator of the retroverse project. I've spent many years looking at different computer buses to use for this project and VME, despite some of its shortcomings, has always been on top. I had at one point thought about making a non-standard extension to VME to have all 32-bit signals available on a 4-row connector but abandoned that approach because a limited amount of interoperability is much more approachable (eg. using existing off the shelf backplanes, like you mentioned). If cost and space are not concerns, then the 6U card size is the way to go, but you can also start with a simpler 3U and use the same 6U backplane until you're ready for the full 32-bit address space. There is actually a less well documented mode for multiplexing a 32-bit address and data cycle onto a 3U/P1-only bus using the A40/MD32 cycle. It requires some extra transceivers and fairly complicated control logic, but if you're already planning to use a programmable logic device for the control logic, that might not be much of a barrier.

About the lack of an A0 signal, it's actually just broken out into the /DS0 and /DS1 signals, which determine whether the byte should appear on the upper or lower part of the data bus. This matches the 68000's bus interface, which has the /UDS and /LDS signals. For the 68020/68030, you can get the exact same behaviour by acknowledging all bus operations as a 16-bit data port, which makes interfacing to 3U VME dead simple. The only reason my card is as complicated as it is is because I plan on eventually supporting the A40/MD32 cycle.

It's entirely possible to have a 68030 with only a 24-bit address bus and 16-bit data bus on a 3U (or 6U) VME backplane, and still use the MMU with full translation support. No need for 32-bits unless you want more than 16MiB of addresssable memory. All virtual addresses would translate to a physical address below 16MiB, and all 32-bit memory accesses would be transparently translated into multiple 16-bit accesses, but it would work perfectly fine.

By the way, I'm also working on a simple unix-like OS that runs on my system, which could be a stepping stone for you to eventually getting Linux running: https://github.com/transistorfet/gloworm

If you have any questions, I'm happy to answer them. I also hang out in the Usagi Electric discord. Good luck on your project!
 
Re 32-bit CPU on a 24-bit address bus:
Also you can obviously have more than 24 bits local to the CPU card, I.E. ROM and in particular RAM above address 16M, and have the lower 16M generate bus accesses on the VME bus.

Re buses of choice:
I haven't had a look at how hard it would be ti use PCI, but a very brief look at it tells that it multiplexes the address and data lines on the same set of 32 signals (or 64 for 64 bit cards). The problem for a hobbyist is that most PCI cards are likely not easy to use. Like there were a few PCI cards that were simple 16550 compatible serial ports and such, but otherwise you'd find things like graphics cards, network card, disk controllers and sound cards as PCI cards. In theory it might not be that hard to tell a graphics card to expose it's entire memory as linear memory on the bus and thus use it as generic RAM, but otherwise PCI cards seem not particularly useful for generic hobbyist usage.

On the other hand, how much stuff do you need for a hobbyist 68k system? The main thing that needs a large address bus and where a wide data bus really increases performance is memory, and that can be local to the CPU card.
 
Follow up on that thought, 4 meg RAM is enough for 68000 and 16meg memory is enough for 68020/30; a SIMM72 can satisfy the RAM requirement cheaply and compactly; programmable logic is contemporary of 68020/30 which can greatly reduce the logic complexity; three rows of VME connector and buffers take significant board space. So these observations lead to small format CPU, local memory, programmable logic that fit in 100mmX100mm pc board with 8-bit data 16-bit addresses for I/O expansion. RC2014 style bus is really not a bad choice for hobbyists even for 32-bit processors.
Bill
 
I considered CompactPCI, and it has some interesting features, but to implement it somewhat to spec requires either a lot of logic (so realistically a PLD) or a custom PCI controller. I can't remember the details, but I think some of the features pretty much require the configuration address space to be implemented, and it can't just be a chunk of memory, to configure irqs among other things. It doesn't include bus arbitration for multiple controllers (not applicable for this post, but for my project that is essential), and it doesn't have a mechanism for adding wait states, so if you have a slower device, you have to buffer it somehow. It also multiplexes the command/byte signals as well as the address/data signals, so you end up with a lot of transceivers. VME might not be as fast or configurable by comparison, but it's very adaptable, even when staying within spec, which is not hard to do, but you can also step a little bit outside the spec and still make use of standard cards, depending on what you're disregarding.
 
Going off on a tangent, PCI is kind of similar to Zorro III on the Amiga, with multiplexed address/data signals and a configuration phase. Zorro III is an add-on to Zorro II which is just a 68000 bus (more or less) with two pins for daisy chaining auto configuration signals (ensuring that one card at a time responds to a specific auto configuration address range). In particular the auto configuration is really easy to implement from the computers perspective, as you can just daisy chain the signal although you are supposed to have circuits that handles presence/absence of cards, and then you just read an address rage at boot, and if something usable is there, write to that area to tell where in memory the card should respond, rinse and repeat until all cards are processed.

Unfortunately any usable Zorro II/III cards would already be in use by Amiga users, so it's not a source of cheap cards. It could be used to do some special things, like you could use the PC emulator cards for your DIY project or so.
 
I'm wondering if there are any "hobbyist" Terminal Cards or Video Cards for VME ??
It's not hard to find "terminal" boards out there:

1. https://geoffg.net/terminal.html
2. https://www.tindie.com/products/denjhang/versaterm-a-versatile-diy-serial-terminal/
3. https://www.tindie.com/products/petrohi/geoffs-vt100-terminal-kit/
4. https://www.ebay.com/itm/225010233671
5. https://www.tindie.com/products/lilygo/lilygo-ttgo-vga-v14-controller/#shipping
transistor said:
Hi! I'm the creator of the retroverse project. I've spent many years looking at different computer buses to use for this project and VME, despite some of its shortcomings, has always been on top. I had at one point thought about making a non-standard extension to VME to have all 32-bit signals available on a 4-row connector but abandoned that approach because a limited amount of interoperability is much more approachable (eg. using existing off the shelf backplanes, like you mentioned). If cost and space are not concerns, then the 6U card size is the way to go, but you can also start with a simpler 3U and use the same 6U backplane until you're ready for the full 32-bit address space. There is actually a less well documented mode for multiplexing a 32-bit address and data cycle onto a 3U/P1-only bus using the A40/MD32 cycle. It requires some extra transceivers and fairly complicated control logic, but if you're already planning to use a programmable logic device for the control logic, that might not be much of a barrier.

About the lack of an A0 signal, it's actually just broken out into the /DS0 and /DS1 signals, which determine whether the byte should appear on the upper or lower part of the data bus. This matches the 68000's bus interface, which has the /UDS and /LDS signals. For the 68020/68030, you can get the exact same behaviour by acknowledging all bus operations as a 16-bit data port, which makes interfacing to 3U VME dead simple. The only reason my card is as complicated as it is is because I plan on eventually supporting the A40/MD32 cycle.

It's entirely possible to have a 68030 with only a 24-bit address bus and 16-bit data bus on a 3U (or 6U) VME backplane, and still use the MMU with full translation support. No need for 32-bits unless you want more than 16MiB of addresssable memory. All virtual addresses would translate to a physical address below 16MiB, and all 32-bit memory accesses would be transparently translated into multiple 16-bit accesses, but it would work perfectly fine.

By the way, I'm also working on a simple unix-like OS that runs on my system, which could be a stepping stone for you to eventually getting Linux running: https://github.com/transistorfet/gloworm

If you have any questions, I'm happy to answer them. I also hang out in the Usagi Electric discord. Good luck on your project!
Click to expand...
Transistorfet -
I was wondering about a Terminal or Video Card Board for VME ?? I know I can just buy a "terminal emulator" card for less than $100 but I was just wondering if such a thing already existed in the VME Bus - specifically on the J1 / P1 Bus NOT including the J2 / P2 part of the VME Bus. Or just buy and install an RS232 PCI or PCI-Express x1 card slot in an IBM PC "clone" ?? And use some sort of Terminal Emulator Software ??? Thoughts ??
 
You must log in or register to reply here.
Back
Top