The difference though was the additional 4MB daughter card. The i386 boards come up often enough - typically for a few hundred dollars. The 2MB and 4MB daughter cards though are rare, and folks that have i386 boards have been asking if someone can make them as a result.
- VCF South West - June 14 - 16, Davidson-Gundy Alumni Center at University of Texas at Dallas
- VCF West - Aug 2 - 3, Computer History Museum, Mountain View, CA
- VCF Midwest - Sept 7 - 8 2024, Schaumburg, IL
- VCF SoCal - Mid February 2025, Location TBD, Southern CA
- VCF East - April 2025, Infoage Museum, Wall NJ
-
Please review our updated Terms and Rules here
Seeking memory card for InBoard/386
- Thread starter nc_mike
- Start date
cb88
Experienced Member
So what happened here did you end up getting a 4MB card yet? Even if you couldn't' get your hands on one someone could certainly document the wring of their board if they didn't want to sell it. I have two Inboards waiting for a ram upgrade! One in a portable, and the other is going in a 5160. I've gotten about halfway through figuring out a TI 486 adapter for the Inboard 386... so now only 90% of the work left!
IBM Portable PC
Veteran Member
Has anyone made any progress cloning the 4MB expansion board, or perhaps creating an 8MB or greater version?
harrison.frazier
Member
- Joined
- May 7, 2012
- Messages
- 47
I have actually been working on a clone board that uses SIMM modules. Stynx was kind enough to share the discovery work he had done on the pinout and the schematics he could find that he had done back in 2014. Last month, I remade the layout on a 4-layer board and reviewed and created a prototype that just barely squeezes in. I just got 5 prototype boards in from JLCPCB today. The Inboard detects 2 of the 8MB that are installed but freezes within 10 minutes or so due to parity errors.
I have located the machined pin header needed on Digi-Key also in order to mate the board to the inboard. I also was able to make 'debug' headers using 90 degree IDC headers.
Is there anyone who is knowledgeable with the Inboard or memory and timings that is willing to help?
Nice work
Please try the following steps:
1. remove any other unneeded card from the computer.
2. use a minimal DOS to load the inboard TSR
3. check the voltages of the psu and on the inboard-memory header
4. try different memory modules
5. try tying the unneeded address lines of the memory module to ground or 5V
6. populate only one bank of memory to minimise load on the TTL-ICs
I just checked the PCB layout you have sent me a month ago.
You may have pin 19 and pin 24 of the memory modules not tied to anything on the inboard header.
You seemingly only tied the pins together with the matching pins on the other modules.
That would mean that the pins are floating...
If that does not help, try removing R10 and tie that address line to ground.
I just realised that you used 7404 TTL ICs... these may be too slow. Try 74ALS04 or 74F04.
-Jonas
Last edited:
I have checked the data sheets.
The 7404PC has a 22ns popagation delay from low to high and 15ns from high to low.
A 74LS, 74H and 74ALS part have a maximum of 10ns for HL and LH transition.
A74F, 74S and 74AS have typically around or lower than 5ns for HL and HL transition.
Since there are a lot of chips on the output of the 7404 and 7432, they may get a bit hot (please check).
TTL-ICs get slower when hot.. maybe your TTL ics are just about fast enough and are too slow once they get hot.
The 7404PC has a 22ns popagation delay from low to high and 15ns from high to low.
A 74LS, 74H and 74ALS part have a maximum of 10ns for HL and LH transition.
A74F, 74S and 74AS have typically around or lower than 5ns for HL and HL transition.
Since there are a lot of chips on the output of the 7404 and 7432, they may get a bit hot (please check).
TTL-ICs get slower when hot.. maybe your TTL ics are just about fast enough and are too slow once they get hot.
Last edited:
harrison.frazier
Member
- Joined
- May 7, 2012
- Messages
- 47
Thank you for the suggestions!
-19 and 24 are already tied to the ground plane, so we can cross that off the list. I have also tied the unused 7404 inputs to ground as suggested by the datasheet.
-I do have some faster logic chips on the way, so I will try those as well.
I'll try it with minimal DOS and no additional cards. I have also played with the arrangement, using only four simms and I have similar results.
As for the SIMM modules - I have 8 1MB simms with MN414400ASJ-07 Panasonic chips (and a 1x1M parity chip) that I'm using, actually pretty close to the MN41C1000-08 chips that are on your genuine piggyback.
-19 and 24 are already tied to the ground plane, so we can cross that off the list. I have also tied the unused 7404 inputs to ground as suggested by the datasheet.
-I do have some faster logic chips on the way, so I will try those as well.
I'll try it with minimal DOS and no additional cards. I have also played with the arrangement, using only four simms and I have similar results.
As for the SIMM modules - I have 8 1MB simms with MN414400ASJ-07 Panasonic chips (and a 1x1M parity chip) that I'm using, actually pretty close to the MN41C1000-08 chips that are on your genuine piggyback.
Compatibility with modules comes down to timing and other rather esoteric stuff
There is no real rule which module works or which doesn't since you are on unsupported terrain anyways.
The 4Mb piggyback is from a user here on the forum, not me. I have only the 2Mb version.
What exactly is the error you get, could you post a screenshot/photo?
When do you get the parity errors, how long does it take to get parity errors?
Does the behavior change when you cold-start the machine versus longer runtime?
Wat happens when you populate only 3 of 4 Simms on one bank?
Im pretty sure that the slow TTL logic chips have to be replaced with either 74F or 74AS series ICs. Maybe the errors will go away with the swap.
It could still be another problem though...
-Jonas
There is no real rule which module works or which doesn't since you are on unsupported terrain anyways.
The 4Mb piggyback is from a user here on the forum, not me. I have only the 2Mb version.
What exactly is the error you get, could you post a screenshot/photo?
When do you get the parity errors, how long does it take to get parity errors?
Does the behavior change when you cold-start the machine versus longer runtime?
Wat happens when you populate only 3 of 4 Simms on one bank?
Im pretty sure that the slow TTL logic chips have to be replaced with either 74F or 74AS series ICs. Maybe the errors will go away with the swap.
It could still be another problem though...
-Jonas
Cimonvg
Experienced Member
hello
dont know exact about your network, since the schematics not public - but you manage to get going with less 74* chips then Intel does (?)
i did order som pins (dont know if sold out?) , and have not tried them out jet - perhaps they do fit nicely
https://www.ebay.com/itm/181530259256
/cimonvg
dont know exact about your network, since the schematics not public - but you manage to get going with less 74* chips then Intel does (?)
i did order som pins (dont know if sold out?) , and have not tried them out jet - perhaps they do fit nicely
https://www.ebay.com/itm/181530259256
/cimonvg
harrison.frazier
Member
- Joined
- May 7, 2012
- Messages
- 47
I swapped the slower chips with ALS variants and the system will still either freeze, give a beep and freeze, or, occasionally display "parity error" in text mode on the top left of the screen. I believe I had MSD open when that happened. I can only assume that message is put there by the inboard driver. Also, the chips barely feel lukewarm after use.
It also appears that the inboard is detecting and testing the first 256K of each stick, as it should behave for an actual piggyback. (If I put 4 256K sticks in I only get a report of the parity chips being faulty, as all I have are non-parity sticks of that capacity.) So there will be some additional wiring (or re-wiring) needed.
If I try to boot with connectors 1-3 populated, it will report that the fourth row of memory chips is bad, excluding the parity chip, as the layout appears on the genuine 1/2Mb boards.
Would using simms with slower ram chips make a difference? (Sorry, this is a bit of a new territory for me!)
I can also see if I can order some 74F chips and try those.
@Cimonvg - those connectors look to be the ones! The machined pin diameter is .5mm. It uses 2 74* chips for the Row and Column strobe and two hex inverters for the address lines. That does beg the question- what is the 5th 74* chip doing on the original design? I don't have one, this is the only pic I could find online. It looks to be another hex inverter.
It also appears that the inboard is detecting and testing the first 256K of each stick, as it should behave for an actual piggyback. (If I put 4 256K sticks in I only get a report of the parity chips being faulty, as all I have are non-parity sticks of that capacity.) So there will be some additional wiring (or re-wiring) needed.
If I try to boot with connectors 1-3 populated, it will report that the fourth row of memory chips is bad, excluding the parity chip, as the layout appears on the genuine 1/2Mb boards.
Would using simms with slower ram chips make a difference? (Sorry, this is a bit of a new territory for me!)
I can also see if I can order some 74F chips and try those.
@Cimonvg - those connectors look to be the ones! The machined pin diameter is .5mm. It uses 2 74* chips for the Row and Column strobe and two hex inverters for the address lines. That does beg the question- what is the 5th 74* chip doing on the original design? I don't have one, this is the only pic I could find online. It looks to be another hex inverter.
harrison.frazier
Member
- Joined
- May 7, 2012
- Messages
- 47
Update: I have it working stably with eight 80ns 256K Simms and 74F logic chips. I'll post more details soon.
I think that the additional chip was for the address lines as well. Im not sure anymore but the address lines were split into 2 "busses" for the banks. Since there are originally 9 address line (18bit=256k) the 3 hex-inverters would very nicely split the lines to have individual address lines per bank. Its pretty cool to see that my original layout was not wasted and you have brought it back to life ... you must be very close to getting the module to work. Maybe there are timing issues since you get the parity errors later on?
- Could your inboard have a problem with parity? I own 5 Inboard 386 of which only 2 work anymore and i had to replace a parity chip on one of them.
- Voltages may drop or be unstable. Do you use the original PSU or a new one?
- Do you have another inboard to check the memory card?
- Do you have another memory card?
Maybe you can give some PCBs to other users to cross-check their results?
I will try to order a bare PCB based on your provided layout.
I will try some PCB manufacturers tomorrow, since PCBway has problems with the layout.
-Jonas
Last edited:
Al Kossow
Documentation Wizard
Harrison has sent me the following as PM:
The expansion board is completely stable with 74F logic chips and eight 256K 80ns simm modules! Once I swap to 70ns modules I get parity errors again.
Very nice
sounds like setup/hold time problems with faster ram
Maybe. But it could have to do with the way the ram chips react as well. There are a lot of things that can be different between ram chips... we have to test several chip types before we can have answers. Harrison will send me 2 of his 5 prototype PCBs for testing. We will try to get some more results since we now know that the design is mostly working.
Maybe we should open a new thread for the further research
I have opened a new thread to have a central point to post updates:
Inboard-386-PC-2mb-expansion-CLONE
-Jonas
Last edited: