circa77
Experienced Member
One more time, still not a good argument, as the 14411 was also CMOS. Just because HP did it once doesn't make it the right answer for all time.
- 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
KIM-1 Repairs continue
- Thread starter falter
- Start date
Dwight Elvey
Veteran Member
KiCad doesn't have a router that I know of. It does create a rats nest. The thing I don't like the most is that it breaks the multi gate parts into individual schematic gates.
When I was doing my 4004 project I had to create my own packages as complete packages and not gates as I wanted to better optimize the physical layout to minimize
the size of the finished board.
The problem I had doing the small boards was that the last small trace rat wire was hiding along the pins. I'd much rather a bowed rat
wire until it was actually placed. Straight wires tend to get lost on top of each other.
Dwight
When I was doing my 4004 project I had to create my own packages as complete packages and not gates as I wanted to better optimize the physical layout to minimize
the size of the finished board.
The problem I had doing the small boards was that the last small trace rat wire was hiding along the pins. I'd much rather a bowed rat
wire until it was actually placed. Straight wires tend to get lost on top of each other.
Dwight
circa77
Experienced Member
I don't understand this. At the end of the day (unless you're using those little teeny SMT single gates that Fairchild put out a bunch of years ago - I don't know if anyone else second-sourced them), all 7400 roads lead to the same 14p DIP (or SOIC - whatever). How does "creat[ing your] own packages" in the CAD system affect the final layout - unless the problem you experienced was that it wouldn't allow you to choose which gate to use for which connection?
In Mentor we have a "gate swap" function in layout. If you don't like the way the gates were assigned during packaging, layout lets you click on one gate, click on another gate, click "swap gates", and all the connections are (correctly) exchanged between the two. Then, when you do a back-annotate, that gate swap is propagated back to schematic capture ("Design Architect" in Mentor-speak) and the drawing updated to reflect the reassignment. Those swaps, of course, don't have to be within the same package - you can have two 7400s on opposite ends of the board and swap gates between them. Similarly, you can swap pins within a gate - if your layout will benefit from swapping pins 1 and 2 of that 7400, you can do that too, and it'll make its way back to the schematic as well.
As for the short-guide problem, I've never heard of any package that would bow a guide wire for the benefit of readability - maybe it's out there, but I've never seen it. When I'm getting close to the end of routing a board and want to see if there are any such that I've missed, there's a number on the screen that tells me how many unrouted nets (i.e. guide wires) are still there. Then, if I can't immediately spot it, I just turn off all the other layers - and there it is.
Last edited:
Hugo Holden
Veteran Member
Well, I don't know about this, the argument doesn't wash in my book.
The reason is that the MC14411 chip you cited, while saving a few IC's, only has a 2 bit input for Baud rate control. Meaning it can only provide four possible baud rate settings.
But, the HP system, used on the T-W unit has a 3 bit input, so it can provide 8 different baud rate settings.
On the T-W unit, you have options of:
110,150,300,600,1200,2400,4800 & 9600 BAUD.
I often try to remind people of this sort of thing.
Many LSI circuits were produced, purporting to replace arrays of logic IC's from the 70's era, claiming equal or better performance. But fell short.
The finest example I know of this was Atari's Arcade Pong, with 65 IC's and possible ball motion vector states of 42 (Just a coincidence with the answer to life the Universe and everything on Hitchhiker's Guide to the Galaxy). They squashed it all into one LSI IC for the home Pong consoles. The number of shortcuts reduced the ball vectors and the game was nowhere as good.
Dwight Elvey
Veteran Member
Yes, one can put the entire schematic together and then use the back-annotation method but it is clumsy. If the schematic shows the packages used one can visualize the physical layout at the schematic level. Waiting until you are at the layout level is like shouting yourself in the foot, if you know where things were going at an earlier stage. On my projects I've done with wire wrap I've always considered the package layout as part of the schematic, to better position the packages on the board. Why would I do it differently because I can do it different.
As for unrouted net reports, you have to consider I was a beginner at using the tool and didn't even look at the reports. It only showed the level of understanding of the tool I had at the time. I'm not real bright you know.
Dwight
circa77
Experienced Member
Hugo, you need to actually read the data sheet. Those two inputs select one of four baud rate sets, not one of four baud rates.
Each of those sets gives you 16 simultaneous outputs (including the 1.8432MHz clock), which you can then choose from with a DIP switch.
Last edited:
circa77
Experienced Member
Okay, I get it. It's a bit of an odd approach, but if it works for you, sure. Personally, I wouldn't bother with creating the already-packaged schematic symbols to be able to do it that way; generally speaking, schematic capture is for getting the logical/electrical function of the circuit right in a symbolic sense, exclusive of physical considerations. To me it makes ample sense to have in my head a rough intuitive idea of what the placement is likely to be, and once I'm packaged and in layout I'll usually let it do an autoplace and then stare at the resulting ratsnest for a while and push the components around toward their final locations. Then my last lines of defense are the gate and pin swaps I mentioned previously. In the case of my system there's nothing at all clumsy about the backannotation process - it's extremely tightly integrated and I don't think I'd have any trouble at all achieving the same result as yours.
The part I don't like about it is that you're forcing a physical view onto the schematic rather than a logical one, and while beginners probably won't have a problem with it, just about any pro is going to react with "Who the hell drew this? Show me the gates!"
Last edited:
circa77
Experienced Member
That's a pretty broad claim. I'm not going to dispute it, but I'm unlikely to accept it without citations.
Seems to me like the original design is just begging to be implemented in an FPGA - or maybe just a CPLD, 'cause 65 SSI/MSI packages ain't nuthin'.... Yeah, there we go. A guy at Columbia did just that about ten years ago.
hanso
Member
It has been done for the KIM-1. Look at http://retro.hansotten.nl/6502-sbc/kim-1-manuals-and-software/kim-1-6530-replacement/ All 6532s are functionally equal btw, no mask programming. The 6530s have mask programming. The logic in the replacement circuits take care to map the 6532 to the 6530-002 and -003 specific memory layout.I wonder if this would physically fit in a KIM-1 and work with the right ROM contents and the right 6532 to replace the 6530 ?
6530 RRIOT Replacement kit for CBM IEEE Drives
If you have diagnosed your CBM IEEE drive and discovered the 6530 is faulty, you can use this to replace it. I also carry other parts used in these drives including 2114 SRAM, and 6502 and 6504 CPUs. This does not come assembled . You must solder it together. The photos only show the completed...monotech.fwscart.com
(I know these are good for the drives, I bought one in the past and they are good value for money too)
I think this is the same basic design as the adapter board made to replace the 6530 in the SFD-1001 and some other Commodore drives.
But, in all cases, as far as I know you cannot use just any 6532, because of the mask programming of the chip select pins. So only a certain number of 6532's will be compatible for the specific application.
Hugo Holden
Veteran Member
Ok, thanks I'll have to check the data sheet. And then it totally out-does the one made from CMOS IC's. I guess there must have been some factor of economy or availability of the Motorola thip chat caused the T-W people not to use it at the time in the late '70's.
Last edited:
Hugo Holden
Veteran Member
I liked the look of the Monotech adapter because it is very compact.
It is good to know that any 6532 will work. I did try to get some, they turned out to be fakes, that did nothing and just generated heat, clearly they were some other IC.
circa77
Experienced Member
Shouldn't have been - I knew the part because it's what was used in my first system, the SWTPC 6800, so they'd been around for years and were common as muck. And as I said, once you factor in its replacing multiple parts, reducing real estate, and assembly time and cost, there shouldn't have been much reason not to go with it.
Or maybe their thinking was "looks good, but HP's never used it..."
circa77
Experienced Member
I can't be the only person who's really struggling with the very existence of all these goddamn counterfeits. I know who does that, but who does that? It doesn't seem much in keeping with either Confucian or Christian principles, but I guess "sell roundeyes' garbage back to him" is the order of the day.
Hugo Holden
Veteran Member
One thing that is really messing with my head lately on fakes is:
I try to always get new old stock vintage IC's, stored well. Used IC's bother me, as I don't know what past traumas they have been subjected to.
Recently I bought a Rockwell 6545 CRTC controller IC said to be New & unused. On inspection, the pins had been soldered to before. The top surface had been sprayed black and the white writing and logos re-applied.
It is clearly a "refurbished IC", pulled from some apparatus and made to look new-ish.
I barely expected it would work, most refurbished parts like this can be any kind of IC, but it is actually a 6545 and (to my horror) it does work.
Now I'm conflicted, because it is a used part, with an unknown history, has had multiple steps in the handling and refurb process and could well have latent ESD damage.
I cannot really complain to the Chinese seller either, they may have thought it was new, and it does work. So I put it in my box of spares with a note on it "Works but used & not trustworthy"
Oddly, I would have rather this refurbished IC was a fake re-labelled part and didn't work, then I would have gleefully thrown it in the bin. But because it works, I'm stuck with the uncertainty of it that it could fail any time, and it drives me crazy.
I don't know what to call this problem, perhaps "The curse of the working fake".
(if you saw the way IC's are removed from pcb's in the refurb plants you would be shocked, blowtorch like flames heat the boards and they are shaken by a machine and most fall off into a container. Then there are many other steps applied, pin treatments, surface coatings etc etc, it is frightening)
Last edited:
circa77
Experienced Member
I feel you, but what you did with it probably wouldn't be my approach.
If it works but I'm "concerned" about it, there's nothing better than to press it into service, and if it fails, it fails. The only things that are going to unwrap that mystery are volts and time; it'll be suspect until operation proves it otherwise... or kills it. Either way, you'll know.
On the other hand, the last thing I want on the shelf is an unknown part, or one I'm nervous about, because I'm probably going to reach for it when I most need something that works.
If it works but I'm "concerned" about it, there's nothing better than to press it into service, and if it fails, it fails. The only things that are going to unwrap that mystery are volts and time; it'll be suspect until operation proves it otherwise... or kills it. Either way, you'll know.
On the other hand, the last thing I want on the shelf is an unknown part, or one I'm nervous about, because I'm probably going to reach for it when I most need something that works.
Hugo Holden
Veteran Member
I don't know if that would help me, because each day it kept working I'd be thinking, hmmm......I wonder if it will fail tomorrow and be back to square 1, worrying about its integrity.
Perhaps I should keep a journal and document the exact date it fails, and think of the Journal like the Almanac on Back to the Future. Then if I could borrow the DeLorean, go to the future, I could look in the journal and see what date it failed. Then come back to present time and remove the IC and bin it the day before it died, but still document the date it failed in the journal, or it would create a tautological problem. But I'm thinking, it might be easier to bin it now.
circa77
Experienced Member
You are mos def overthinking this.
(And since my last play on words bombed here, that was another one. Don't disappoint me twice.)
(And since my last play on words bombed here, that was another one. Don't disappoint me twice.)
hanso
Member
Looks like a Rev D, quite common. Sticker on the brand name, which is Commodore C-MOS. f the 003 is in a socket, maybe it it has replaced a damaged 003. 6530-002 and -003 were for sale in those days.
Dwight Elvey
Veteran Member
I see no reason it couldn't be used. You have to use the correct EPROM or what ever and preprogram it.
You might need to modify the I/O logic as that is also custom for each 6530. That board was setup for the disk controller.
It is smaller than the Corsham board.
Dwight
Hugo Holden
Veteran Member
I hope the KIM-1 repair is going well.
The talk of gate packages and the 7400 reminded me of something interesting.
A while back I was working on the logic boards of two Tek 2465B scopes, mainly 74HCT ICs' in SOIC packages. I just happened to notice that there were two different logic IC's in the exact same position on each board. Seemed unusual. However when I inspected how the gates in the IC's were used it became obvious, they were functionally equivalent in the application. (see attached)
But the fact that somebody at the Tek factory figured this out, and were able to use another type of IC, presumably when they were running low of one type, was interesting, nothing appears to get past them.
The talk of gate packages and the 7400 reminded me of something interesting.
A while back I was working on the logic boards of two Tek 2465B scopes, mainly 74HCT ICs' in SOIC packages. I just happened to notice that there were two different logic IC's in the exact same position on each board. Seemed unusual. However when I inspected how the gates in the IC's were used it became obvious, they were functionally equivalent in the application. (see attached)
But the fact that somebody at the Tek factory figured this out, and were able to use another type of IC, presumably when they were running low of one type, was interesting, nothing appears to get past them.
