C128 Schematic Blunder

Started by hydrophilic, February 23, 2007, 07:54 AM

Previous topic - Next topic

0 Members and 3 Guests are viewing this topic.

hydrophilic

Since the floor is open, I have two related questions for this trivia topic.

Which logic gate in the C128 Schematic (as it appears in the Commodore 128 Programmer's Reference Manual and even other 'revised' versions), is incorrect?

What is the error, or equally, what gate should it be?

I found several inconsistancies, but I am not refering to resistors, complete ICs, labels, or values.  Just the wrong GATE.  It is the only one I found wrong, but Commodore never paid me to do a throrough analysis...

hydrophilic

Nobody even wants to guess? :( Less than 48 hours remaining! :P

In case you don't have the C128 PRG (or don't want to download it, kinda big) a schematic is availbable here. :rolleyes:

Mangelore

Is it U57 pins 1 and 2? The symbol should be a buffer not inverter.

hydrophilic

#3
You got it :cool:


After X number of revisions, you think they would have fixed that!  No QS-9002 qualification for Commodore!

Mangelore

Woohoo :D

Well, as I don't have any trivia questions handy, I declare the floor open.

Blacklord

Quote from: MangeloreWoohoo :D

Well, as I don't have any trivia questions handy, I declare the floor open.
Remember, it doesn't have to be C128 related - just C- related.

Lance

BilHerd

Lol... I remember this actually, yes it got fixed in a core schematic, there would not have been anything that would have triggered an update to any other document, in fact the incentive would have been very low for a bubble.

When we looked at the schematic we would have overlooked the bubble without trying as the nearby signal names signifies function and inversion, I probably could have looked at it all day long and seen the function as not inverted without catching the error.

Okay here is a question, why the heck was a high voltage and transistor used in the clock circuit of the Z80?

Here is another one: Why was a Z80 put on the main board in the first place and not just supported in the standard Z80 cartridge?

Bil Herd
C128 Designer

Mangelore

Hi Bil,

Wow, thanks for dropping in and posting on this forum!

Quote from: BilHerdHere is another one: Why was a Z80 put on the main board in the first place and not just supported in the standard Z80 cartridge?
I've read that the Z80 was used as a workaround to get the Magic Voice cartridge working with the C128. Something about needing the C128 to start at memory address zero?

Cheers
Fotios

RobertB

Hi, Bil,

Nice of you to drop on by.   Glad to have met you personally on Saturday.

CommVEx v3 info at  http://www.portcommodore.com/commvex
Robert Bernardo
Fresno Commodore User Group
http://videocam.net.au/fcug

JamesC

Quote from: BilHerdHere is another one: Why was a Z80 put on the main board in the first place and not just supported in the standard Z80 cartridge?
If what I read on the Internet was written and attributed correctly, you yourself said that the Z80 was put on the 128 motherboard because the extra 0.5A needed to run it on a cartridge wasn't possible on the then-current design.

Plus, the C64 CP/M cartridge was flaky as it was, and even less reliable on the 128.

In addition, Commodore Australia threatened to rip the Z80s out if you shipped 128's with Z80s in them, and whatever made AU mad was probably a good idea to West Chester Engineering. ;)

As far as the transistor in the clock circuit... I have no idea.

Now... I have a question for you, and Haynie didn't know when I asked him a couple of weeks ago: on the Plus/4 motherboard, centered at the front of the board (under where the spacebar would be) there are 6 jumpers marked J1 to J6. Each jumper is set from the factory to the "A" position. What are the jumpers for?

James
www.cbm264.com

hydrophilic

Wow, the man himself!

Answer to question 1 is because the Z80 requires a clock signal better than standard TTL (a high very close to 5V and not say 4.1V or so...)

Answer to question 2 is, like others have said, to allow reliable use with C64 cartridges which may check the address lines on power up -- the 8502 first fetches from address $fffc which cartridges might watch for... so the Z80 starts from $0000 and if needed will switch the hardware for 64mode, once this happens, the 8502 starts up 'fresh' in 64mode at address $fffc.

Steve Gray

Quote from: JamesCNow... I have a question for you, and Haynie didn't know when I asked him a couple of weeks ago: on the Plus/4 motherboard, centered at the front of the board (under where the spacebar would be) there are 6 jumpers marked J1 to J6. Each jumper is set from the factory to the "A" position. What are the jumpers for?

James
www.cbm264.com
According to the schematics (310164, page 4) they are to select the ROM sizes for U25 and U26 which are the function ROMs (built-in software apps).

Steve

BilHerd

Wow, I am impressed, all of the answers were true about the Z80: I was trying to find a way clear of the power supply cost problem, fixed the timing issue I knew would come up, and in the end saved my ass by having a processor that would boot from a different reset vector.  Calling Au's threat was just icing on the cake . =)

When managment went home for the night the Magic Voice cartridge didn't work, when they came in, it did.  We had an inverter already in the reset line if memory serves, as Z80 and '02 were already reversed, so we just changed the default state.  Von Ertwine, and his wife Cathy (who knew Z80 opcodes) basically walked us through the codes as we stood at the eprom burner and hand burnt the Z80 boot rom, later we added the Commodore Key sense logic as a failsafe.

Management walked back in, they found out it worked, I went and caught a nap.

Jumpers would be for Rom size and maybe even type when it comes to some /CS logic.  When you can't outguess what Marketing has in store for you, add jumpers.


Answer to transistor on Z80 clock is yes the Z80 needs 4.90V in a very little time or some such voltage.  Dave DiOrio thought he could coax the VIC chip into doing it in spite of the fact that NMOS drops out of active state around VCC-1V  As it turns out it did lose too much juice at the high end.  The problem was that if I used a 5 volt anything it would have an RC tail (exponential curve) and have the same problem and TTL just didnt even make the voltage in real life so in a fit of insanity I used the higher voltage knowing it was still on it's way to a higher voltage at a higher rate when it slammed into the 5v rail.  I had to study whether I was really screwing the transistor by doing this as it relied on it acting like a diode at the end of the day, and some transistor geometries aren't really set up for forward biased BC junction, or at least that was the theory I worked under.  In the end we did not pop transistors in production as I feared we might.  The problem when you get silly like this is that if your wrong and it makes it past preprod you are wrong times a few hundred thousand if not a mil.


Looking at the circuit I saw at least two other silly things we did with open collector parts... I had completely forgotten about one of  them (two back to back gates)

Bil Herd