What are the failure modes of PROMs?

On Sat, 7 May 2016, Mattis Lind wrote: I had a PROM fail on the 11/45 CPU board. This PROM is responsible for the Conditional Codes handling. There was one output that had failed. I think that the output drivers may fail, because I was able to temporarily fix the problem by adding a pull-_down_ resistor to the (open-collector) output. That fix was unstable of course, and I ended up in replacing the PROM with a GAL (ugly fix because the pinout doesn't match well, and there is no space for a nice adapter solution)

Christian

On Mon, May 9, 2016 at 5:13 AM, Mattis Lind <mattislind at gmail.com> wrote: Not impossible, but it seems more likely that one of the address input pad buffers failed.

Having just programmed some 24SA10's a couple of weeks back on a Data I/O 29B, I can report that you should be fine...

I am working on the first of my two 11/05s. Interestingly, it has the early version M7261E Control Logic & Microprogram board and the later version M7260 Data Paths board (with circular baud rate selector switch) as described in: http://gunkies.org/wiki/KD11-B_CPU From the description there, it seems like an older/newer combination, but maybe that was common. I would not have guessed that the four possible combinations would all work together, but maybe they do? I have a couple different drawing sets for the 11/05 and while some have the matching M7260 schematic, only the GT40 drawings (I found on bitsavers) has the M7261E schematic: http://www.bitsavers.org/pdf/dec/graphics/VT11/GT40_Graphic_Terminal_Engine… The GT40 drawings has the PROM listings and related, so I am hoping that they match what is in the two boards. Presently, the machine sometimes runs relatively well and other times it does not. It does have bit 1 stuck ON in memory, but that should be a relatively simple task to diagnose as it is not intermittent. When the machine is "working" I am able to deposit 0777 at 0100 and run. When running this simple program, I've experimented with flexing the boards and such, so it doesn't seem like an obvious poor connection, but that remains to be seen. The machine is a configuration #2 model (as described in the "gunkies" site) and my initial messing with KM11 boards, reveals that I can step the microcode with a KM11 in either the #1 or #2 position, but when two KM11s are installed at the same time, they do not function properly together. Is this expected or do I have an issue there too? Thanks much to those who have provided details and documents on the web, they have already been of great value and will most certainly continue to be a resource in the future. More updates in the future... --tom On 5/6/16 5:32 PM, Noel Chiappa wrote: chip albeit Does things PROMs. the apparently in form - prints the 'soon'. them are

I am working on the first of my two 11/05s. Interestingly, it has the early version M7261E Control Logic & Microprogram board and the later version M7260 Data Paths board (with circular baud rate selector switch) as described in: http://gunkies.org/wiki/KD11-B_CPU From the description there, it seems like an older/newer combination, but maybe that was common. I would not have guessed that the four possible combinations would all work together, but maybe they do? I have a couple different drawing sets for the 11/05 and while some have the matching M7260 schematic, only the GT40 drawings (I found on bitsavers) has the M7261E schematic: http://www.bitsavers.org/pdf/dec/graphics/VT11/GT40_Graphic_Terminal_Engine… The GT40 drawings has the PROM listings and related, so I am hoping that they match what is in the two boards. Presently, the machine sometimes runs relatively well and other times it does not. It does have bit 1 stuck ON in memory, but that should be a relatively simple task to diagnose as it is not intermittent. When the machine is "working" I am able to deposit 0777 at 0100 and run. When running this simple program, I've experimented with flexing the boards and such, so it doesn't seem like an obvious poor connection, but that remains to be seen. The machine is a configuration #2 model (as described in the "gunkies" site) and my initial messing with KM11 boards, reveals that I can step the microcode with a KM11 in either the #1 or #2 position, but when two KM11s are installed at the same time, they do not function properly together. Is this expected or do I have an issue there too? Thanks much to those who have provided details and documents on the web, they have already been of great value and will most certainly continue to be a resource in the future. More updates in the future... --tom On 5/6/16 5:32 PM, Noel Chiappa wrote: > > From: Mattis Lind > > > Thanks Noel for sorting this out. > > Eh, de nada. But thank you. > > > >> I wonder if the ucode in the two versions is identical? The uROM chip > >> numbers should give it, (if they are the same on both versions, albeit > >> in different locations on the board), but I have yet to check. Does > >> anyone happen to know? > > OK, so the situation here is pretty complicated. To start with / make things > worse, that CPU uses lots of PROMs. Lots and lots and lots and lots of PROMs. > > For the data paths board (M7260), both major versions appear to contain the > same PROMs (going by the DEC part numbers), but the chip location (Exx) > numbers are all different. > > For the control board (M7261), the C, E ('early' version) and F ('late' > version) etch revisions each contain mostly the same PROMs, but apparently > with slight differences between the sets of PROMs in each (as reflected in > different DEC part numbers). For details see: > > http://gunkies.org/wiki/PDP-11/05#Control_PROMs > > to which I have just added all the gory details. > > As to getting the contents of all of them dumped in machine-readable form - > oi vey! > > > >> on the earlier version (prints for that version are in the GT40 prints > >> online > > It turns out that I have hard-copy prints for the "C" etch revision of the > M7261, which do not yet appear to be online; the GT40 prints have the "E" > etch revision. > > I will scan the pages for that revision of the board, and put them up 'soon'. > (I'm not doing the whole print set, it's about 1" thick, and most of them are > for other things anyway, like MM11-L memory, etc.) > > Noel >

I'm not to the point of connecting serial yet, but I did see this page which will help me turn the current loop to rs232 when the time comes: http://retrocmp.com/how-tos/interfacing-to-a-pdp-1105 --tom On 6/14/21 10:37 AM, Jay Logue via cctalk wrote: > I also have an 11/05 with the early CPU boards that exhibited stuck bits on arrival.? Turned out > to be bad transistors in the inhibit circuits on the G110.? Pretty easy fix once I tracked it > down. So far I've found the GT40 print set to be a fairly accurate, at least for the boards I have. > > I'll be curious to learn how your serial console works.? Mine had a manufacturing defect that had > to be corrected before input would work. > > --Jay > > On 6/13/2021 1:44 PM, Tom Uban via cctalk wrote: >> I am working on the first of my two 11/05s. Interestingly, it has the early version M7261E Control >> Logic & Microprogram board and the later version M7260 Data Paths board (with circular baud rate >> selector switch) as described in: >> >> http://gunkies.org/wiki/KD11-B_CPU >> >> ?From the description there, it seems like an older/newer combination, but maybe that was common. I >> would not have guessed that the four possible combinations would all work together, but maybe >> they do? >> >> I have a couple different drawing sets for the 11/05 and while some have the matching M7260 >> schematic, only the GT40 drawings (I found on bitsavers) has the M7261E schematic: >> >> http://www.bitsavers.org/pdf/dec/graphics/VT11/GT40_Graphic_Terminal_Engine… >> >> The GT40 drawings has the PROM listings and related, so I am hoping that they match what is in the >> two boards. >> >> Presently, the machine sometimes runs relatively well and other times it does not. It does have bit >> 1 stuck ON in memory, but that should be a relatively simple task to diagnose as it is not >> intermittent. When the machine is "working" I am able to deposit 0777 at 0100 and run. When running >> this simple program, I've experimented with flexing the boards and such, so it doesn't seem like an >> obvious poor connection, but that remains to be seen. >> >> The machine is a configuration #2 model (as described in the "gunkies" site) and my initial messing >> with KM11 boards, reveals that I can step the microcode with a KM11 in either the #1 or #2 position, >> but when two KM11s are installed at the same time, they do not function properly together. Is this >> expected or do I have an issue there too? >> >> Thanks much to those who have provided details and documents on the web, they have already been of >> great value and will most certainly continue to be a resource in the future. >> >> More updates in the future... >> >> --tom >> >> On 5/6/16 5:32 PM, Noel Chiappa wrote: >>> ???? > From: Mattis Lind >>> >>> ???? > Thanks Noel for sorting this out. >>> >>> Eh, de nada. But thank you. >>> >>> >>> ???? >> I wonder if the ucode in the two versions is identical? The uROM chip >>> ???? >> numbers should give it, (if they are the same on both versions, albeit >>> ???? >> in different locations on the board), but I have yet to check. Does >>> ???? >> anyone happen to know? >>> >>> OK, so the situation here is pretty complicated. To start with / make things >>> worse, that CPU uses lots of PROMs. Lots and lots and lots and lots of PROMs. >>> >>> For the data paths board (M7260), both major versions appear to contain the >>> same PROMs (going by the DEC part numbers), but the chip location (Exx) >>> numbers are all different. >>> >>> For the control board (M7261), the C, E ('early' version) and F ('late' >>> version) etch revisions each contain mostly the same PROMs, but apparently >>> with slight differences between the sets of PROMs in each (as reflected in >>> different DEC part numbers). For details see: >>> >>> ?? http://gunkies.org/wiki/PDP-11/05#Control_PROMs >>> >>> to which I have just added all the gory details. >>> >>> As to getting the contents of all of them dumped in machine-readable form - >>> oi vey! >>> >>> >>> ???? >> on the earlier version (prints for that version are in the GT40 prints >>> ???? >> online >>> >>> It turns out that I have hard-copy prints for the "C" etch revision of the >>> M7261, which do not yet appear to be online; the GT40 prints have the "E" >>> etch revision. >>> >>> I will scan the pages for that revision of the board, and put them up 'soon'. >>> (I'm not doing the whole print set, it's about 1" thick, and most of them are >>> for other things anyway, like MM11-L memory, etc.) >>> >>> ????Noel >>>

Tom said My original PDP-11/05 print set that came from the BHP steelworks is titled just 'PDP-11/05 Engineering Drawings' and the drawings are for machine revision 1105-0. It covers the early 11/05 with the cast metal grille with no slots (as seen in the 11/05 handbook) rather than the later 11/05 with the plastic slotted grille. The dates on the drawings are 1972 and some of the microcode listing dates are different to the 'PDP-11/05-S, 11/10-S systems engineering drawings' bitsavers scan and the GT-40 scan, and some of them are the same. I'm not sure where I look to see what revision of the M7261 it shows, as the Rev box in the bottom right seems to jump around, eg. Rev N, Rev K, Rev H or is that just the minor drawing revision for that page? Steve