PDA

View Full Version : A little something I'm working on right now..



PacMan
January 22nd, 2009, 01:16 PM
I was having trouble finding a serial board for my Altair at a reasonable price so I decided to build my own board. Here's a picture of it in its current state.

http://www.x-game.se/images/s-100-lab-mounting.JPG

The 40-pin dipper is a SCN2681A DUART, which is a great device to work with. It's still missing the transceivers but that is coming along.

When I started to make the schematics I also decided to put in a socket for an EPROM (or as in this case a flash memory). I also opted for a two PLD (16V8 ) decoding and bus control solution. The EPROM is selected by the LS85 and the DIP switches to the left in the board.

Later I plan to add a floppy controller to the empty upper right section of the board (notice the conveniently placed 34-pin connector), but that is for another rainy day.

The 2681 has a 8-bit output port and a 7-bit input port that will be used to (among other things) control the /PHANTOM signal to be able to disable the on board EPROM if I should ever want to run CP/M on with this board.

/Pac

gerrydoire
January 22nd, 2009, 04:24 PM
I never had a problem with the Altair, till I started to use it...:cool4:

NobodyIsHere
January 22nd, 2009, 06:27 PM
Hi! Nice! Where did you get the prototyping board? It looks new. Did you design it yourself?

Excellent job! Thanks for sharing!

Andrew Lynch

PacMan
January 22nd, 2009, 09:50 PM
Hi! Nice! Where did you get the prototyping board? It looks new. Did you design it yourself?

Yes, I designed it during the christmas holiday and got it here about a week ago. I will be making a small batch of them shortly, if you want one let me know.

They carry two 5V regulators and one +12 V and one -12 V regulator and a bunch of decoupling caps. Originally I wanted to place as much generic logic on the board as possible but then I found myself locked into an 8-bit card without the possibility to do a 16-bit card. Instead I made the lab board as simple as possible including only the voltage regulators.

/PacMan

Erik
January 23rd, 2009, 05:41 PM
Heya PacMan,

I've got a Tarbell 2022 and a CCS 2422 - both use the FD1793B.

You've got first dibs 'cause you asked, but they are generally available if you don't want 'em.

I can keep digging if you'd like. Those were from a more accessible box of S-100 cards. :)

NobodyIsHere
January 23rd, 2009, 06:12 PM
Hi Pacman! What did you use to design the prototype board PCB? KiCAD, Eagle, or ? I am big fan of KiCAD although there are lots of good EDA packages these days. You can't argue with KiCAD's price though!

Thanks and have a nice day!

Andrew Lynch

PacMan
January 24th, 2009, 03:29 AM
Andrew,

This board was actually done with KiCAD. It was my first board using this CAD package and I must say that I am impressed with it. I can mention (but i wont ;-)) several packages that costs thousands of dollars that is no better than KiCAD.

/Pac

NobodyIsHere
January 24th, 2009, 04:00 AM
Hi! Thanks! I thought I recognized the layouts. Are the schematics and PCB files available? You are free to post them on the N8VEM wiki if you'd like.

I am quite impressed with KiCAD as well. As a free tool, I kept my expectations in check but so far it can do everything I have wanted it to. The program continues to evolve and smooth out the rough edges too. Add in FreeRouting.net and it is a impressive combination!

One project that's been in my mind but never come to the forefront is a ECB to S-100 project. I have a circuit in a book that shows how to map the Z80 signals to S-100 boards and from there its a short leap to ECB. The book is called "S-100 and Other Micro Buses" and has the circuit in the back.

It seems to me a small ECB and S-100 motherboard with the necessary interface logic would allow an N8VEM SBC to control S-100 boards. However the complicating factor is not the logic interface because that's quite simple. Its the S-100's power supply requirements.

Here is a half thought out idea I've been bouncing around in my head recently to address the situation. Please let me know what you think of it...

I was sorting through some old PC scrap stuff and found some old Dell laptop power supply bricks which I thought might be useful. Instead of a straight linear S-100 power supply, how about using a couple scrap PC laptop bricks; one for the +20 VDC power rail (nominally +16v but I've seen this range quite a bit and 20 VDC is probably in the tolerable range) and run it through a 7808 or similar for the +8 rail? Use the other Dell laptop brick "flipped" to supply the negative rails in a similar configuration. I am not sure if the laptop power supplies would have the current capacity sufficient for more than a card or two. It might for a small S-100 motherboard system though. That would be perfect for home brew hobbyists.

Of course, replace the proprietary connector and replace with something more commercially available like something from MOLEX.

These are just speculations and I don't know how practical it would be. If there was an easy and cheap fix to the S-100 power supply situation that did not require a full blown linear or tweaking the internal circuitry of an ATX power supply I think small S-100 systems would be more practical for home brew hobbyists. I have various S-100 boxes like NorthStar Horizon, VG, and other home brew S-100 systems but they all are *large* units and require a lot of space to work on. ECB is *much* more compact and a lot more hobbyist friendly in many respects, IMO.

Thanks and have a nice day!

Andrew Lynch


PS, I re-read my post and I accidentally mangled the paragraph on the power supplies when writing it. I've revised it to say what I meant; one PS brick would supply the positive rail (+20v, and through a 7808, +8v). The other PS brick would be reversed to supply the negative rails (-20v, and through a 7908, -8v).

The neat thing about using the Dell laptop supplies is that they are very cheap on eBay. You can see used models of PA-6, PA-9, PA-10, PA-16, etc for $5 or less. Get any two of those, add a couple of 8 volt regulators, and you can form a simple low cost S-100 power supply. It won't have much capacity but enough for a couple boards.

PacMan
January 25th, 2009, 05:07 AM
Andrew,

It sounds like an interesting idea to build a ECB <-> S-100 interface. It would give you access to a lot of interesting S-100 functions. With that said I might add that I have found that S-100 equipment is getting more and more difficult and expensive over time. I have been trying for some time to find a floppy disk card for a reasonable price but most boards start at $100 and with shipping I often land upwards $130-$140 which is ridiculous.

Anyway, power supplies.... I always use good (yes expensive) switch mode power supplies when I work on equipment that need more than on or two different voltages. In the past I have spent to much time debugging hardware (and software) that was behaving incorrect due to faulty (or badly designed) power supplies. The power supply is as important (or maybe even more so) as the hardware that you are designing, building or working on.

I remember specially one episode where I designed a 8051 based system (only the MCU part) for a customer. The design worked according to the specification when it was handed over to the customer (using our internal lab power supplies). Months later when they were integrating our boards into their product nothing worked as expected. I was called in to "fix our faulty design" and ended up spending about a week trying to find the problem. In the end it turned out that their inexpensive power supply solution was causing all the electronics in their system to malfunction.

Lesson learned, don't skimp on the power supply ;-)

You should select something like this http://www.meanwell.com/search/rq-125/default.htm to work with. The +-15V is perfect for S-100 system and the +-5V can probably be jacked up to +-7V without problem which is also perfectly okay. I could not find a supply with those exact specs on ebay right now but I have bought several of them in the past so I'm sure there will be more. The last one I bought was $10 for a 150W +-12V and +-5V which which I turned up to +-15,5 and +-6,5V. Quite sufficient for a few boards.

/Pac

NobodyIsHere
January 25th, 2009, 07:55 AM
Hi! I like the dedicated power supply approach definitely! It really would simplify the PCB and probably reduce one side by 4 cm or so since the LM317/LM337's wouldn't be needed.

The N8VEM ECB backplane fits on a Eurocard 160x100 PCB and has 6 DIN 41612 connectors. The circuit in the book would probably take (I am guessing) about 12 20 pin DIP chips to implement. That would require probably about 4 x 16 cm area to place and route in a decent manner.

Assuming 3cm centers on the S-100 boards, to add 3 (maybe 4) connectors would require about 9cm or so. I am imagining a PCB of about 16 x 23 = 368 cm2 area (6"x9") or 54 in square. Thats a big PCB but not huge.

Add mounting holes for stand offs and skip the case and that's a low cost S-100 hobby system ideal for prototyping. It'd probably need a switch and a PS connector but the N8VEM ECB backplane already has that in its design. Since it is multiple input voltage though it may be worthwhile to just make the PS switchable rather than on the PCB.

Please take a look at the circuit in the book on TRS-80 to S-100 bus. I think it could be readily adapted to ECB as the TRS-80 is a straight forward Z80 SBC. Its a short step from there to ECB.

No doubt such a design would have major limitations with legacy S-100 PCBs but that is not uncommon either. Many S-100 systems are effectively custom due to the numerous variances in S-100 implementations. Staying with a single vendor helps but making a system of mix and match boards gets rather sporty. Making new boards with S-100 prototype boards should work though.

The only issue I can think of right now is that the CPU clock on the N8VEM SBC is 4MHz where the phi signal on a S-100 bus is typically 2MHz. I wonder if that'd matter? The S-100 bus would be short enough to not require termination but long enough to still be useful.

Its an interesting thought experiment if nothing else. This idea has been floating about in the back of my mind for months and every once in a while it resurfaces.

Thanks and have a nice day!

Andrew Lynch

PacMan
January 25th, 2009, 02:04 PM
I finally had some spare time to devote to my little board today.

Added the serial buffers (top of the board).
http://www.x-game.se/images/s100/090125-1.JPG

And wired the input address buffers to the S-100 connector.
http://www.x-game.se/images/s100/090125-2.JPG

It actually went much quicker than I had expected. I have done quite a lot of wire wrap prototypes in my days but that involved flipping the board a gazillion times to figure out pinning. Not to mention wiring all VCC and GND signals. With my new prototype board power is always available close to the chips on the backside of the board making it a snap to solder power directly to the chip.

I also figured out that I should add X-Y markers on the board like the hand written ones. Makes it very easy to locate a chip on the board from the schematic (provided you have entered markers on the schematic that is :mrgreen:)

I don't know if you have seen it but the holes on this first board are much to small for the pin headers to fit. I probably had a big hole in my head when I selected the hole size. I probably just figured that the IC sockets needed to fit.

Anyway, that's all for today.

--> Andrew, I'll look at you post more tomorrow when I'm lucid again.

/Pac

PacMan
January 26th, 2009, 01:50 AM
Hi! I like the dedicated power supply approach definitely! It really would simplify the PCB and probably reduce one side by 4 cm or so since the LM317/LM337's wouldn't be needed.

The N8VEM ECB backplane fits on a Eurocard 160x100 PCB and has 6 DIN 41612 connectors. The circuit in the book would probably take (I am guessing) about 12 20 pin DIP chips to implement. That would require probably about 4 x 16 cm area to place and route in a decent manner.

Assuming 3cm centers on the S-100 boards, to add 3 (maybe 4) connectors would require about 9cm or so. I am imagining a PCB of about 16 x 23 = 368 cm2 area (6"x9") or 54 in square. Thats a big PCB but not huge.

Add mounting holes for stand offs and skip the case and that's a low cost S-100 hobby system ideal for prototyping. It'd probably need a switch and a PS connector but the N8VEM ECB backplane already has that in its design. Since it is multiple input voltage though it may be worthwhile to just make the PS switchable rather than on the PCB.

Please take a look at the circuit in the book on TRS-80 to S-100 bus. I think it could be readily adapted to ECB as the TRS-80 is a straight forward Z80 SBC. Its a short step from there to ECB.

No doubt such a design would have major limitations with legacy S-100 PCBs but that is not uncommon either. Many S-100 systems are effectively custom due to the numerous variances in S-100 implementations. Staying with a single vendor helps but making a system of mix and match boards gets rather sporty. Making new boards with S-100 prototype boards should work though.

The only issue I can think of right now is that the CPU clock on the N8VEM SBC is 4MHz where the phi signal on a S-100 bus is typically 2MHz. I wonder if that'd matter? The S-100 bus would be short enough to not require termination but long enough to still be useful.

Its an interesting thought experiment if nothing else. This idea has been floating about in the back of my mind for months and every once in a while it resurfaces.

Thanks and have a nice day!

Andrew Lynch

Andrew,

Looks to me as if it would be smooth sailing for you to make this happen. You could check Grant Stockly's small S-100 backplanes. A tad expensive but very high quality. Together with that kit your ECB card should be able to plug right into your adapter board and voila, instant cofee ;-) Or are you maybe aiming to have the S-100 connectors already on your adapter board ?! In that case you could choose a less expensive connector than the one used on Grants boards.

I don't think the hardware to adapt the ECB bus to S-100 should be to hard to create, depending a bit on how much you are aiming to support on your adapter card. I'd go for supporting the prioritized interrupt lines (VI0* - VI7*), maybe by using a I8259 or even simpler by using a 74LS148 and a 74LS373 or something similar. DMA could be nice but then you are adding a lot of complexity to the adapter card that could be implemented on the ECB bus or on the S-100 bus. I guess some consideration regarding DMA is needed to make the ECB bus talk DMA fluently on the S-100 bus and vice versa.

Regards
/Pontus

NobodyIsHere
January 27th, 2009, 04:07 AM
Hi! I tried to post a reply on this yesterday but somehow it got tossed into the bit bucket. Oh well! This is the TRS-80 to S-100 bridge circuit I was referring to

http://www.hartetechnologies.com/manuals/Unclassified/S-100_and_Other_Micro_Buses.PDF

It starts on page 115, chapter 14. I have the schematics for the TRS-80 model 1, so I am thinking since it is a basic Z80 design itself, is just implement the minimally required TRS-80 circuitry required by the interface along with the bridge circuit. There really is nothing all that exotic there although I am fairly sure the compatibility of the bridge circuit with S-100 is probably quite limited.

http://incolor.inebraska.com/bill_r/trs80_schematics.htm

The main issue I see with a S-100 design of any sort is the procurement of the female card edge connectors. Certainly they are still available but they are quite expensive. Availability of connectors was one of the deciding factors, along with PCB cost, in going with ECB on the N8VEM design since the connectors (DIN 41612) are inexpensive and commonly available.

Please take a look at the bridge circuit and let me know what you think of the proposed approach. Does it sound reasonable? Do you think there would be interest in such an odd hybrid? For home brew computing, ECB is/was popular in Europe and S-100 is/was popular in the US. The design has the potential to leverage what remains of both hobbyist communities and resources.

Thanks and have a nice day!

Andrew Lynch

PacMan
January 27th, 2009, 05:27 AM
Andrew,
I glanced through the schematics and they look okey. What I would do is to check out the schematics of a few Z80 S-100 processor boards to see how they have solved the S-100 compatability issues. And as i mentioned earlier you might want to support the interrupt scheme available on the S-100 bus which would make many more S-100 boards run properly in your system.

The connectors are available from Digikey, they have a few different version in stock but the one from EDC (http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=EDC346102-ND) seems like a good choice. It's around $7 if you buy more than 10 of them. I recently bought a set of Sullins connectors for an original Altair 18-slot motherboard but those buggers cost me a little over 16 bucks each, which is a little steep even for me.

I actually have no idea if this kind of board would be interesting. Me, I was raised with S-100 computers s I have a strong link to them and love to play around with them but I don't know how many people that share that sentiment.

It would certainly be fun to do it and I could be willing to help out with testing and some funds if necessery.

/Pacman

NobodyIsHere
January 27th, 2009, 11:46 AM
Thanks! That is exactly the information I was looking for. Right now the how ECB to S-100 bus bridge project is just a thought experiment. My next step is to write up a schematic in KiCAD and post that for some comments.

On the N8VEM project, I am working on Zilog Peripherals board at the moment which is taking up all my time. The CTC and dual PIOs seem to work fine but the DART is giving me fits. I am not familiar with the DART/SIO way of doing things so I don't know if I am initializing it properly although I am able to get the DART to respond to simple commands like setting the RTS pin. Still can't get it to send any data out of the TX pins though.

Its great to see other hobbyists willing to make their own computers. I like S-100 and have several systems but it is not very hobbyist friendly IMO. Still, it is a lot of fun.

Thanks and have a nice day!

Andrew Lynch

PacMan
January 27th, 2009, 12:48 PM
Just let me know if you need any assitance, it really sounds interesting. Good luck with the DART, i have never worked with that chip so I can't help your there.

/Pac

NobodyIsHere
January 28th, 2009, 05:24 PM
Hi! Where did you get the 50/100 pin card edge connector for KiCAD?

Thanks and have a nice day!

Andrew Lynch

PacMan
January 29th, 2009, 10:22 AM
Andrew,
I created the entire board myself, which includes the connector (as J1). So when I do another S-100 design (o yes, you know they are coming) I will have a predefined board with everything (J1 connector and board outline) in place. The lab-board was a sort of a first step in this process.

I will try to make sure that you get the KiCad design data for this but at the moment I am really (and I mean really) busy with another commercial (this is how I make my money and feed my kids) project.

To keep my sanity though I will continue to make postings about this board so don't hold your breath waiting for it.

What do you know, I may be the next S-100 board king (OK that is kidding, you got to know that) *hehe*

/Pontus

Terry Yager
January 29th, 2009, 12:03 PM
What do you know, I may be the next S-100 board king (OK that is kidding, you got to know that) *hehe*

/Pontus

You might be surprised, there's prob'ly enough S-100 enthusiasts around that are starving for new toys. Mebbe not enough to feed the kids, but take them out to a nice restaurant occasionally. That prot board looks real nice, perhaps better than any commercial board I've ever seen. I'd buy one, for the right price. Perhaps you could add a logic probe to version 2.0?

--T

PacMan
January 29th, 2009, 12:37 PM
Perhaps you could add a logic probe to version 2.0?

--T

Ok, that sounds interesting, do you have anything specific in mind ?

Right now i'm thinking that I could add support to the board for both 16-bit as well as 8-bit boards by having a CPLD on board as well as all the necessary buffers to make that happen. In that case you would could make a board that would meet the IEE696 standard when it comes to memory accesses. And if that were the case you could easily (if you know VHDL) implement pretty much the entire standard.............

Whoops, what happened there, to much oxygen to my brain :mrgreen:

Sorry 'bout that, i'm back to reality :cool: Please... i'm really interested in what you mean. it's just my brain that is acting out on me (as it always does when I'm listening to Jennifer Rush (OMG am I really that old ??)).

The current board is a simple prototyping board (which has proven to be very useful) but as mentioned it could easily be improved to be more useful to other uses.

Regards
/Pontus

NobodyIsHere
January 30th, 2009, 04:36 AM
You might be surprised, there's prob'ly enough S-100 enthusiasts around that are starving for new toys. Mebbe not enough to feed the kids, but take them out to a nice restaurant occasionally. That prot board looks real nice, perhaps better than any commercial board I've ever seen. I'd buy one, for the right price. Perhaps you could add a logic probe to version 2.0?

--T

Hi Terry! I have an S-100 bus extender board that comes with a logic probe. It was broken and I managed to repair it. It actually works pretty well and the circuit is simple.

I've been mulling over this ECB/S-100 bridge design some more and started gathering my notes and entering them into a schematic. I can see already a some issues arrising.

First, the N8VEM SBC uses ECB for IO only and not a memory bus. All the memory and internal IO are kept off the ECB by the bus interface logic. There is nothing preventing ECB from being a general purpose memory and IO bus but it was a simplifying assumption on the N8VEM SBC to keep all the memory local. If I ever redesign the SBC, I intend to add jumpers to allow memory accesses over the ECB *and* the disabling of the lower 32KB page so it can be accessed by the CPU. This is relevant to the bridge circuit since it will have to be configurable so that (/MREQ OR /RD) are not interpreted as requesting the databus to send data to the N8VEM SBC. Fortunately, I think this can be tweaked with a 74LS32 ensure /RD is not part of a memory access before allowing the S-100 bus data buffers to send data to the ECB. This can be optional too.

Second, as Pacman discovered there is not much information available regarding S-100 male and female edge connectors for KiCAD. That is fixable as I have some unused connectors in the basement so making a custom connnector in KiCAD is part of the "to do" list now.

Third, the bridge circuit I am considering does not address the vectored interrupts. I will check with another book (Interfacing computers with the real world) to see if their discussion involves the vectored interrupts. I haven't looked into it yet but I think that would be a solvable problem regardless.

I am finding that there probably going to be more ICs needed than I originally thought. Of course I am just laying in the schematic to capture the data with no thoughts past that. Once its in and organized I can work out those details. There are some weird glue logic too like a 5 input OR which are going to require either group of 74LS32s or another clever solution which doesn't require unobtainium chips.

Probably given the complexity of this device I'll need to do a prototype on the bench. That adds some time of course. However the real question is how to build the prototype. If I recall correctly, the pitch spacing on an S-100 connector is 0.125" and not the typical 0.100" that most prototype boards use. Probably this would have to be a frame type prototype board with a regular 0.100" spaced board for the ECB connector and bridge components. Then mount the prototype board and the S-100 connector on a wooden frame and connect the two with wire wrap.

The good news is that my Zilog Peripherals prototype board is testing out just fine and everything seems to be working AFAIK. I'll be moving to its PCB layout and trace optimization soon as the XT-IDE PCB finally clears out of my basement PC I use for trace routing tasks.

Thanks and have a nice day!

Andrew Lynch

PS, oh yeah, now I remember... I've been studying the TRS-80 schematic for how it derives its signals for its expansion bus and they are really quite basic stuff. /TEST is just Z80 /BUSRQ pulled high. /IN is (/RD OR /IORQ) and so on and so forth. There does not appear to be anything very complex that's for sure. Once I get the schematic keyed in, I'll post a copy and hopefully get some constructive inputs.

This could be a fun vintage computer community project if there were sufficiently interested hobbyists. The PCB could be an ECB backplane, a small S-100 backplane, or an ECB/S-100 bridge bus. I need to find a decent power supply though. That has always been a major weakness of S-100, IMO.

ECB is much more hobbyist friendly; small standard sized PCBs, common PS voltages, cheap and commonly available connectors, etc. Its major problem is lack of PCB space though. The S-100 prototype boards are 5"x10" for 50 square inches of real estate and doesn't include the connector. Even when shared with regulators etc, its twice a 160x100mm Eurocard which does include the DIN 41612. I was just barely able to jam in an IDE and FDC controller into the Disk IO board using a 2 layer PCB. The Zilog Peripherals prototype is packed with ICs as well and that is *without* the connectors! I am doing its layout and it was another one that requires a lot of manual routing just to get it all to fit on a 2 layer Eurocard. Its all a trade off; big PCBs == big price.

PacMan
January 30th, 2009, 12:49 PM
Andy,
Holy crap, what were you thinking of :shock: Only I/O on the ECB bus :crazy:
*hehe* Sorry 'bout that. Well that should give access to most of the useful S-100 boards anyway since you already have the memory onboard.

Anyway, I've been going through your posts regarding the N8VEM system and I really like you CP/M implementation. I will have to get back to you on that when I start working on my S-100 super I/O board.

/PacMan

NobodyIsHere
January 30th, 2009, 02:27 PM
Andy,
Holy crap, what were you thinking of :shock: Only I/O on the ECB bus :crazy:
*hehe* Sorry 'bout that. Well that should give access to most of the useful S-100 boards anyway since you already have the memory onboard.

Anyway, I've been going through your posts regarding the N8VEM system and I really like you CP/M implementation. I will have to get back to you on that when I start working on my S-100 super I/O board.

/PacMan

Hi! Howard Harte has a Super IO for the S-100 controller. Its not cheap but its really nice!

http://www.cpm80.com/superio.html

The N8VEM SBC is dual mode; it can operate standalone or as an ECB bus controller. The decision to limit the ECB to IO only was to simplify the design by keeping all memory local to the CPU. ECB really shines as an IO bus but is not so great as a memory bus. The IO cycle has an additional wait state in it so it is a LOT more reliable over the bus. Also, M1 instruction fetches are the most aggressive timing cycle and really push the CPU <-> memory interface.

Putting memory stuff on a bus just adds too much complexity early on in the project so it made sense at the time. Considering S-100's history of troubled memory boards, especially with DRAM, this makes sense. Now that the rest of the system is stabilizing though opening the SBC up to bus memory starts to be more appealing. The changes necessary to modify the SBC are pretty minimal but important.

Tonight I will keep working on the schematic and post an early version soon. If you would, take a look at some vector interrupt handling circuits and let me know what you think. I will check my other book to see if they address it or not.

Thanks and have a nice day!

Andrew Lynch

PacMan
January 30th, 2009, 02:59 PM
Andrew,
Yepp, I've read about mr. Harte's Super I/O board but there is absolutely no way in h..l that I am paying $350 for something that I can do myself and learn something in the process of doing it. As a hobby I need to minimize the cost as far as possible, but when I see that there is something to be learned I can spend a few more bucks to do it.

I'm looking forward to going through your schematics.

Regarding the prototype board, I don't think I will add any probing feature. I assume you mean addings means to visually look at signals on the bus and maybe be able to single step. There is already such a board available on eBay (I have it and it's great) and I feel that it would increase the amount of logic to much that it would invade on your own logic.

/PacMan

NobodyIsHere
January 30th, 2009, 06:44 PM
Hi! Here is a helpful online reference for the S-100 pin definitions.

http://wiretap.area.com/Gopher/Library/Techdoc/Hardware/s100bus.txt

I jotted down as much of the schematic as I could tonight. Obviously this is barely a first draft and is extremely crude. The intent is just to get the ideas down on the schematic as a starting point and refine it later.

Unfortunately, I have many open questions about S-100 signals because their intention was not clear in the book. The Z80 and ECB uses a lot of active low control signals but it appears the S-100 uses many of those similar signals as active high. That will have to be straightened out. I can see now that a prototype is a required to make sure all the signals are correct.

If you send me a S-100 50 position/100 pin connector parts library I'll add that to the schematic. Otherwise, I can create my own using libedit.

Thanks and have a nice day!

Andrew Lynch

PS, the files are here: http://n8vem-sbc.pbwiki.com/browse/#view=ViewFolder&param=ECB%20to%20S-100%20Bus%20Bridge

NobodyIsHere
February 3rd, 2009, 07:13 PM
Hi! I did a bit more work on the ECB to S-100 bus bridge circuit. Since I was able to recreate the S-100 connectors PCB footprint, I can now start running the traces.

I used an old broken S-100 connector to verify the pads are placed accurately and am using 40 mil holes which should enough for the pins.

Luckily I was able to go to the surplus store tonight and bough 7 S-100 connectors for fairly inexpensive so now I have enough bench stock to do more repairs and/or build a prototype.

My plan is to keep working on the schematic and PCB layout but then build a partial prototype with only a single DIN 41612 and a S-100 connector. Then test and fix until I get an N8VEM SBC to control a simple S-100 IO card.

Take a look at the revised drawings and let me know what you think. Its far from done but further along than it was.

Thanks and have a nice day!

Andrew Lynch

MikeS
February 3rd, 2009, 08:03 PM
Andrew: While I'm here, you asked what I had in the way of Z80 chips.
In case you or someone still wants any:

Lots of Z80ACTCs - SGS & Zilog
Lots of Z80ADART - SGS
Some SGS Z80BCPU
A couple Mostek MK3880BN-6
A couple SGS Z80ASIO-0

Also lots of 6116 RAMs and small EPROMs

mike

NobodyIsHere
February 4th, 2009, 03:58 AM
Hi Mike! Thanks! Yes, I am interested in some Z80 family chips especially since Zilog Peripherals prototype board seems to be working. The PCB is in trace routing/optimization at the moment. I did some initial testing and everything seems to be working but the Zilog Peripherals have so many states and modes I have no idea how to test everything.

Once I got the PIOs to send out some arbitrary patterns on the parallel port, the CTC can generate various pulse trains and interrupts, and the DART can initialize and generate an arbitrary stream of ASCII "0"s out the serial port, I rather ran out of ideas on how to test this thing. Of course, then I got distracted by something else floating by and rather lost focus.

The ZP project is one of those odd ones where I know it needed to be done to flush out the problems in some of the dark corners of the N8VEM design. Going through the design and test process resolved a whole bunch of mysteries for me. For example, now I know why the ECB bus has the CPU clock on pin C29. :-/ (I detest putting system clock signals on busses and resist this whenever I can but it didn't work out in this case. Intel's asynchronous IO approach is the only way to go, IMO, and their requiring a system clock is a major shortcoming of the Zilog Peripherals)

It did that job quite well but I have no particular use for the board myself. Its just more serial and parallel ports with a system timer. Some of the builders are interested in it for using their own RTOS or maybe MP/M? I don't know. I do know that people have been asking about the Zilog Peripherals since before the SBC PCB was even designed so hopefully this will be of some help to the N8VEM builders. I guess we'll see. Regardless, it pretty much had to be done.

Any way, this is drifting off topic so to get it back relevant to Pacman's original thought about S-100 prototyping, here is a link from a TRS-80 clone in Australia called "System 80" by Dick Smith. Why this is relevant is that it includes a schematic of a interface to attach a S-100 bus to a Z80 based system called the X-4010. This is a useful check to compare with the ECB to S-100 circuit I am proposing to use. Maybe we can pick up a couple of more signals using it.

As the S-100 hobbyists know, there are several S-100 implementations and according to Herb Johnson's website, there are at least three distinct groups of S-100 implementations out there not including IEEE-696. When I made the custom S-100 connector for the ECB to S-100 bridge, I used the Altair/IMSAI pin outs. So the choice as to which S-100 bus to implement further adds to complexity of this project.

http://classic-computers.org.nz/system-80/manuals_technical.htm

Really what I'd like to know is what sort of interest is there in the vintage computer community for a small S-100 motherboard and prototyping kit is there? Pacman has made an excellent S-100 prototyping board. I have several S-100 prototyping boards as well but unless I use a legacy chassis (like the CompuPro wreck I just bought on eBay) there is nothing to plug them into.

As usual theS S-100 power supply issues continue to plague me. I am looking for one of the MeanWell PS's that Pacman suggested. I am searching eBay but those specific PS's but they appear about as rare as hen's teeth. Pacman is right though, the PS is the foundation of the system and has to be good. Also using the laptop supplies and SCRs, although very cheap, are probably current limited and also would excessively drive the cost of the project up in terms of additional PCB area required. Its better to isolate the PS from the PCB but then we have the same old problem of how to get the proper voltages in the first place.

Thanks and have a nice day!

Andrew Lynch

Terry Yager
February 4th, 2009, 01:59 PM
then I got distracted by something else floating by and rather lost focus.

I thought I was the only one who had that pr...ooooh, lookit the pretty lights...

--T

NobodyIsHere
February 5th, 2009, 03:54 AM
Hi! I've been thinking this over some more and have decided to completely revamp my approach on the ECB to S-100 bus bridge project. Rather than an integrated unit, I think the more practical approach is to design this in three components; a small S-100 motherboard, an ECB to S-100 ECB bridge board, and reuse the existing N8VEM ECB backplane.

The new approach allows the components to be more generic and the project a bit more digestible. Tackling everything at once is probably not feasible.

The first part is already available (ECB backplane) I think the next step is to make a small generic S-100 backplane. A 4 slot unit should be relatively inexpensive especially since it will accept the unregulated voltages "as is". Basically, it would be 4 slots in parallel with terminals for the +16V, -16V, and +8V rails.

The power supply continues to be a problem area but I am thinking that this could be done with a pair of inexpensive SWPSUs or even surplus laptop PSs. The first one provides the +15V and -15V and the second one provides 9V. S-100 voltage rails are unregulated so they tend to be more of a range of voltages rather than specific. Since the boards do their own voltage regulation we have some wiggle room on the exact voltages.

Since the load is spread out and there are only 4 slots, the individual PSs don't really have to be large. If I recall correctly, the rule of thumb for S-100 power supplies is about 1A per board. The key to keeping the project low cost is to keep it simple. Reuse scrap components whenever possible.

There is a lot on my "to do" list right now so this project is getting pushed to the back burner for now. When I am ready, I'll start a new thread on the S-100 motherboard.

Thanks and have a nice day!

Andrew Lynch

PacMan
February 5th, 2009, 10:21 AM
Andrew, sorry about being so non responsive for the past few days. Lot of things to do right now.

I looked through your schematic and although I haven't gone through all the logix it looks like a very good start. One short comment though, I'd skip the Altair/Imsai compatibility and leave out things like PROT and UNPROT which really only makes sense for a system with a front panel. Plus the fact that these signals were rarely used anyway.

When you get to it we should probably have a discussion regarding what really needs to be supported.

This schematic however gives you an example on how the vectorized interrupts can be handled: http://classic-computers.org.nz/system-80/manuals_tm3_drawing%204.15_X-4024%20S-100%20interface%20schematic.jpg .

I have used that technique together with 8080 and Z80 computers before and it is a neat and simple solution. The 8085 requires a little different solution and there a i8259A chip is the solution. You often see those guys together with 8085 based boards.

Maybe you could open a S-100 <-> ECB bridge thread already so that one could jolt down ideas when one has them so they don't get lost in the binary abyss.

I haven't been able to do much on my prototype prototype :sigh: board since my last post but it is on my lab desk in my office so at least I get to see it every day :happy8:

/Paccy

PacMan
March 4th, 2009, 12:34 AM
Crap, crap, crap,
My PLD programmer blew out yesterday so now I am stuck until my new programmer arrives. I got myself one of these little guys
http://www.kanda.com/images/25-03-2008-1616vp-190.jpg

USD$77 on ebay was quite reasonable for a device that powerful.

Anyway, the circuit I am building on my lab board is finally working as it should. A dual UART and a 32KByte FLASH together with the required S-100 circuitry is in place. A floppy disk controller and maybe an IDE interface will eventually work its way on there as well. It was when I was preparing one of the PLD's for more chip select signals when my programmer blew out.

/P

jackrubin
April 13th, 2009, 05:44 PM
Heya PacMan,

I've got a Tarbell 2022 and a CCS 2422 - both use the FD1793B.

You've got first dibs 'cause you asked, but they are generally available if you don't want 'em.

I can keep digging if you'd like. Those were from a more accessible box of S-100 cards. :)
Erik,

Do you still have the Tarbell available? I've been looking for one of these for quite a while.

Thanks,
Jack