• Please review our updated Terms and Rules here

SCSI-1 to/from IDE drive converter

lynchaj said:
Due to IDE's inherent 16 bit nature, the 8 bit to IDE interfaces are going to be clunky no matter what although the PPI IDE is about the smoothest I've seen. Performance is OK but not great.

Microcontrollers struggle with IDE too since it requires ~24 GPIO pins. SCSI requires ~20 GPIO pins too (plus buffers) which means a single chip microcontroller SCSI to IDE solution requires at least 44 GPIO plus whatever programming interface, status, etc. I can see the GPIO pin count going to 55-60 GPIO quickly. Not many hobbyist friendly microcontrollers have that many GPIO pins available.
Click to expand...
Life will be easier with LSI parts. Keep in mind SCSI target mode code is not the same as initiator code. It will be simpler if you start off supporting just the basics and ignoring less common commands like reserve/release and use canned data for mode pages. But code will get seriously hairy if you plan on supporting those.
 
Hi! I've seen several 8 bit IDE implementations and the 8255 PPI + inverter implementation is the smallest part count. All of the 8 bit IDE implementations tend to be about the same speed wise as they are pure polled IO and clunky. I used an LSI 8 bit IDE implementation on the N8VEM DiskIO and a different but similar design for XT-IDE. Both use several chips and a lot more PCB real estate than PPI IDE and don't bring any appreciable improvement. You really need a 16 bit data path (CPU/uC) to get the speed advantage of IDE like the PC/AT and later can.

My plan is just for basic SCSI-1 software support. With a Z80 and 32KB ROM there is an upper limit to how much is practical to support the SCSI standard. However the degree of software support will be up to whom ever is the software lead. So far, that's no one. Are you interested in volunteering?

Thanks and have a nice day!

Andrew Lynch
 
Would you be interested in doing a single device bridge or multi device bridge? If going multi, I would strongly urge that DMA be taken into consideration. Otherwise a single (low performance) cpu servicing multiple devices may render the the I/O performance unbearable.
 
Last edited:
hexsane said:
Multi preferred but single would do for starters. If you can start the code I would certainly join in.
Click to expand...

Hi! "better is the enemy of good enough" A single SCSI to IDE bridge would be great for the bulk of most vintage computer hobbyists which is the intended audience. If you can't get a SCSI drive and the machine won't work *any* speed drive would be an improvement. Assuming they are using CF or SD, the Z80 to Z53C80 in pseudo DMA mode will be plenty fast enough. The XT-IDE board works fine with CF and fast compared to real MFM drives it is replacing on XT class systems and that is pure PIO mode. I would expect similar performance.

With a CPU dedicated 100% to the role of data transfer how is the DMA going to speed things up? On a Z80, DMA takes the CPU off the bus and idles it so the DMA can access the bus. The Z80 DMA cannot go dramatically (orders of magnitude) faster than a Z80 CPU in pure dedicated data transfer. For most vintage computer hobbyists, adding DMA to this bridge board adds a lot of parts and complexity but little in terms of actual enhancement. The pseudo-DMA mode eliminates the polling IO delays through use of /DACK and DRQ via the interrupt system. Keeping part count down is the key to keeping costs low.

On a heavily loaded general purpose computer with the CPU busy doing other things, the DMA offers tremendous improvement in data transfer. In this case, not so much.

I appreciate your comments and support. Let's make this project work!

Thanks and have a nice day!

Andrew Lynch

PS, I updated the files on the N8VEM wiki so interested builders can take a look at what I am proposing. Basically it is a Eurocard embedded controller with SCSI on one end and IDE and SD on the other. It is inserted between the SCSI controller and the IDE drive and power with a drive cable. You connect a serial terminal to set the parameters and off you go.
 
Last edited:
Hi! I am trying to decide what to do with this project. I think it would be a good and helpful thing to have but recognize it may be too specific or overcome by events. If you are interested in buying a prototype board please contact me at

mailto:LYNCHAJ@YAHOO.COM?subject=SCSI2IDE

If there are 5 people interested I can get a batch of five PCBs made for $150 or $30 each. If there are 10 I think I can get the prototype board cost reduced to around $20 each.

Please note if you buy a prototype board you don't have to actually build and test it. You can buy a board and have it sent to a volunteer builder for build and test.

If the SCSI2IDE project is to succeed we have to get past the prototype board phase and into the hands of builders for coding. Obviously with prototype boards there are no assurances and you can expect there will be problems. We have to find those and fix them before a manufactured PCB is even a possibility.

Thanks and have a nice day!

Andrew Lynch
 
Hi! I am updating the PCB layout based on some feedback I've received to better support IDE/CF and IDE/SATA adapters. The change is relatively minor and doesn't affect any functionality.

Now the dual row headers are all flipped so if builders want to use right angle connectors the will face outwards from the board. Also the IDE connector is pulled further into the board about an inch or so and the switches position with CR2032 battery socket. This gives more clearance for the battery socket but also allows IDE/CF or IDE/SATA adapters to set on top of the board without hanging over the edge.

If builders want both the IDE/CF adapter and the CR2032 battery socket they'll have to mount the battery socket on the copper side of the PCB and trim the leads. Better yet, use an external battery and just run wires. I am not a fan of batteries mounted on PCBs since they tend to leak and ruin the board.

So far a couple of builders have signed up for prototype boards. We'll give this a few more days and see how it goes. If we get to 2 or 3 more builders willing to go for the prototype boards then we'll proceed. If not, then there is probably not enough community interest to warrant this project at least for now. The work done so far won't be lost though, it will just stay where it is until if and when it picks up.

Thanks and have a nice day!

Andrew Lynch
 
You could count me in for 1 prototype board if it will help advance the cause here, I am more than comfortable assembling the card and flashing an EEPROM, and testing with several vintage SCSI based systems and several IDE drives and a handful of CF cards, however I would be of no help in the area of programming the thing.
 
Last edited:
RWallmow said:
You could count me in for 1 prototype board if it will help advance the cause here, I am more than comfortable assembling the card and flashing an EEPROM, and testing with several vintage SCSI based systems and several IDE drives and a handful of CF cards, however I would be of no help in the area of programming the thing.
Click to expand...

Hi! May I suggest, no disrespect intended, that maybe what you might want to consider buying a prototype board and send it to a volunteer builder for build and test and/or develop code. Actually if you can do the basic build to assemble the board that may be helpful too. There are builders involved in the XT-IDE project which have done this approach. You can contribute to the project without being a hardware or software guru by providing the raw materials to builders with the proper skill set but lack the resources to participate otherwise.

Yes, buying a prototype board, even if it just sits on a shelf and does *nothing* else helps because it pushes the total quantity up to make the minimum threshold for a prototype PCB buy. Then other builders will get a chance to work on it who otherwise wouldn't. So even just buying a prototype PCB helps a lot because it opens opportunity for the whole project to advance. I've seen this a lot on the S-100 boards. The real heros in the project are those who buy the prototypes and first round of boards since they are usually the most troublesome. However, if those don't move nothing else will either.

Thanks and have a nice day!

Andrew Lynch
 
hexsane said:
I'm not sure I'm included in the count here but I should be. ;)

-Matt
Click to expand...

I think you and the previous poster would be #3 and #4. Four prototype PCBs and me (I'll buy a PCB as the backstop in case everything falls through) makes five builders. Five prototype PCBs which is the minimum order from www.33each.com at $30 each. If we can get 5 more builders to get prototypes elsewhere which would push that down to into the low $20+ range. That's certainly a possibility.

Let's give this a few days to soak and see what happens. A very encouraging development! Thanks and have a nice day!

Andrew Lynch

PS, BTW, one of the N8VEM builders working on the SBC-188 posted photos of the PPI IDE in action on the N8VEM wiki. Apparently it works fairly well considering it has a fast CPU driving it. Bumping up the CPU speed tremendously helps the data rate up to a point. I am thinking we should decent performance with an 8 MHz Z80.
 
lynchaj said:
Hi! May I suggest, no disrespect intended, that maybe what you might want to consider buying a prototype board and send it to a volunteer builder for build and test and/or develop code. Actually if you can do the basic build to assemble the board that may be helpful too. There are builders involved in the XT-IDE project which have done this approach. You can contribute to the project without being a hardware or software guru by providing the raw materials to builders with the proper skill set but lack the resources to participate otherwise.

Yes, buying a prototype board, even if it just sits on a shelf and does *nothing* else helps because it pushes the total quantity up to make the minimum threshold for a prototype PCB buy. Then other builders will get a chance to work on it who otherwise wouldn't. So even just buying a prototype PCB helps a lot because it opens opportunity for the whole project to advance. I've seen this a lot on the S-100 boards. The real heros in the project are those who buy the prototypes and first round of boards since they are usually the most troublesome. However, if those don't move nothing else will either.

Thanks and have a nice day!

Andrew Lynch
Click to expand...

No disrespect taken, I am more than comfortable soldering, but I am also not opposed to forwarding it off to a builder who has the programming skillset to write and debug the software side of it and letting them build and test it. This is your project, so whatever I can do to help, let me know.

Somewhat off topic, but speaking of the XT-IDE, I cant wait to get my hands on one of those kits too, soldering is the most fun part, but for one of those I would settle for a prebuilt unit, would love to get my Compaq Portable and PS2 Model 25 booting off a hard drive ;-)
 
I just got to thinking, you could possibly add another whole "user base" to this project if a 2.5" laptop syle SCSI connector was added to the edge, this combined with a SD card could in theory replace the internal 2.5" SCSI hard drive in OLD Mac laptops (PB 1xx series through late PPC models), as long as the built card is less than about 15mm or so thick it should fit the HD bays. I know lots of Mac laptop users struggle to find usable SCSI 2.5" drives for their laptops, and the CF to 2.5" SCSI converters I have seen are rare and well over $100.

Maybe this isnt feasable, just something that just came to mind.
 
Hi! I am not familiar with the 2.5" laptop style SCSI connector. Do you or anyone have a part number? I presume it is one of those 2mm pitch dual row headers. We added one of the laptop IDE connectors to the S-100 IDE board. If the laptop SCSI connector is like the laptop IDE connector it has 4 extra pins for power and ground. Is that right? Also a pin out would be helpful. I will take a look at the PCB layout and see what can be done to add the extra connector. It is pretty tight on the right hand side though.

Thanks and have a nice day!

Andrew Lynch
 
I know it is a 2mm pitch connector, but I honestly am not sure of pin count or pinouts for sure, I was able to find a few sites, but I cannot verify the reliability. This is not an aspect of this device I would personally use, I just figured there are probably people trying to restore PowerBook 100, 140, 170, and 180 laptops to working condition who could really use a device like this combined with an SD card as a replacement for the almost un-obtainable 20,40, and 80mb 2.5" SCSI drives.

Regarding pinout I have found the following:
http://www.adtron.com/pdf/S25H-spec053004.pdf
http://www.scrapheap.net/conner_scsi_pinout.txt (this is pinout off a stock PowerBook drive, so it SHOULD be accurate)
 
RWallmow said:
I know it is a 2mm pitch connector, but I honestly am not sure of pin count or pinouts for sure, I was able to find a few sites, but I cannot verify the reliability. This is not an aspect of this device I would personally use, I just figured there are probably people trying to restore PowerBook 100, 140, 170, and 180 laptops to working condition who could really use a device like this combined with an SD card as a replacement for the almost un-obtainable 20,40, and 80mb 2.5" SCSI drives.

Regarding pinout I have found the following:
http://www.adtron.com/pdf/S25H-spec053004.pdf
http://www.scrapheap.net/conner_scsi_pinout.txt (this is pinout off a stock PowerBook drive, so it SHOULD be accurate)
Click to expand...

I have several PowerBooks in the 100x series, one of which I'm planning to remove the hard drive from to swap to another machine. If it is helpful I can take some photos and measurements of the connector and the inside of the system.
 
RWallmow said:
I know it is a 2mm pitch connector, but I honestly am not sure of pin count or pinouts for sure, I was able to find a few sites, but I cannot verify the reliability. This is not an aspect of this device I would personally use, I just figured there are probably people trying to restore PowerBook 100, 140, 170, and 180 laptops to working condition who could really use a device like this combined with an SD card as a replacement for the almost un-obtainable 20,40, and 80mb 2.5" SCSI drives.

Regarding pinout I have found the following:
http://www.adtron.com/pdf/S25H-spec053004.pdf
http://www.scrapheap.net/conner_scsi_pinout.txt (this is pinout off a stock PowerBook drive, so it SHOULD be accurate)
Click to expand...

Hi! Thanks! The adtron datasheet is what I needed. I am adding the laptop SCSI connector to the board. Unfortunately it requires a major rearrangement of the PCB layout to fit the connector. Due to the pitch of the 2mm connector it does not allow for routing trace between its pins so it will take up a surprising amount of space. I am glad there is some way to help the Mac hobbyists and maybe others with the 2.5" SCSI drive. I had no idea they were stranded with hard to find drives.

I will update the schematic and PCB layout on the wiki. I will post when they are updated. One way the builders here can help is to review the schematic against the adtron datasheet, the Z53C80 datasheet, etc. Also just look over the schematic and ask questions, imagine the PCB and check the layout, etc. Feedback, questions, comments, suggestions, constructive criticisms, etc all welcome.

Thanks and have a nice day!

Andrew Lynch

PS, another builder just joined the project so we are up to 6 total. 4 more and we can get the larger number of boards and reduce the price.

PPS, I could use some help with the fit of the board into a Mac laptop. My plan is to use a standard PCB size (Eurocard) which is 160x100mm. Please take some photos and/or see with a cardboard cutout if a 160x100mm board will fit along side a CF or IDE hard drive.
 
Last edited:
Hi! I updated the schematic and PCB layout files on the N8VEM wiki. Trace routing has just started and needs several days to trim the board enough for a prototype board. In the meantime please review the schematic and PCB layout for any fixes.

Thanks and have a nice day!

Andrew Lynch
 
Hi! I noticed there were two unused input pins on the UART and a couple of unused pull-up resistors so I added a couple of option jumpers. These spare input pins can be used for configuration settings. There are configuration jumpers for ID1, ID2, and ID4, and parity. Those options are read by the Z80 CPU and programmed into the Z53C80. There are also spare jumpers for MISC5, MISC6, MISC7, MISC8, plus the two I just added. Hopefully we won't need all of those but it is good to have some flexibility. The ID1, ID2, ID4, PARITY, and MISC5-8, are jumpers connected to an 8 bit input port.

My plan is to use the NVSRAM and EEPROM to store most configuration data and update the software as needed. The builder would use the serial port to set the options and then store them. The theory being is this bridge is intended to help the oldest and least standardized of the SCSI platforms so we need to take into account things are going to require some unique support. I would expect that Mac SCSI would not necessarily be the exact same as VAX or Amiga or Atari, etc.

Since this is a free and open design as part of the N8VEM project if builders need to write modified software to support their specific platform, we could roll those changes into a main source tree and use the configuration jumpers and/or NVSRAM to distinguish their specific configurations. Of course this presumes the board works in the first place but I think we have a good shot at a usable board.

With the recent changes, the density of the board has climbed significantly and it now requires some serious time for trace routing and optimization. It won't require full optimization for the prototype board but it should at least complete on the first setting. It is more complex than my liking right now with 142 vias and ~504 inches overall trace length. That's a bit much so we need to let it soak in the trace routing optimizer for a while to grind that down to size. I think 100 vias for a Eurocard sized PCB is reasonable (~4 vias per square inch) plus the pads for devices means we are in the 800+ range for drill holes.

Thanks and have a nice day!

Andrew Lynch
 
You must log in or register to reply here.
Back
Top