A (Long) Conversation Chuck Peddle, Bil Herd, Jeri Ellsworth

[BigEd]

Just published: a raw unedited 3 hours of time with Chuck Peddle, hosted by Bil Herd and Jeri Ellsworth.

There are a few amusing minutes at the beginning about Chuck's teletype, and then best to skip to 24mins in for quick bit about Bil's job interview, and then at 41mins a parenthetical story about home made rocketry injuries, then a few comments on NMI, but the interview proper starts at about 45mins, which is where the first segment as found on youtube picks up - an easier way to get started if you don't want that 'in the room while rebooting windows to get skype working' experience.

A Conversation with Chuck Peddle, Bil Herd, Jeri Ellsworth: part 1, part 2 (about 33min in, mentions needing to trim many PLA outputs, by condensing and putting in don't cares),part 3

It's in a flash player but also downloadable as mp4.

(Thanks to RobertB on comp.sys.cbm)

Edit: updated the timestamp where he starts talking about meeting the price point by controlling the chip area. Now resolves to this url.

[kc5tja]

"Mensch and Peddle together again." -- Bil Herd

OK, my interest has been piqued.

[BigEd]

Might as well add some other finds to this thread - these are all PDF:
- a few chapters of On the Edge (The spectacular rise and fall of Commodore) are online - chapter 1 is about Peddle, Mensch and co up to the 6502 launch. There seems to be a second edition published with extra material.
- many early issues of Byte including Nov 75 with 6502 pre-publicity article on p56, and Dec 75 with 2-page spread for the Jolt 6502 card on the inside cover. (Also the August 80 Forth issue)

[fachat]

BigEd wrote:

A Conversation with Chuck Peddle, Bil Herd, Jeri Ellsworth: part 1, part 2 (about 40min, mentions needing to trim many PLA outputs, by condensing and putting in don't cares),part 3

Watched it today. Great stuff! Thanks for posting!

André

[kc5tja]

Huh, interesting. It seems that the PET and I are the same age, having been "born" in the same year.

[BigDumbDinosaur]

kc5tja wrote:

Huh, interesting. It seems that the PET and I are the same age, having been "born" in the same year.

When I was born the abacus hadn't been invented.

[chitselb]

I tried to find these video interviews today, and sadly, Blip has taken them down. Is there another place they can be found?

[Tor]

Sigh. This just goes to show that if you see anything interesting on the internet then better save it quickly.
My 4TB disk has started to fill up after I learned that lesson. Didn't get those shows though. Going to buy more disks soon anyway.

[BigDumbDinosaur]

Any chance the way-back machine caught them?

[kc5tja]

Or, just ask Jeri to re-post the videos? I'm sure she has copies.

[BigEd]

The uncut videos are on youtube, I think: start at https://www.youtube.com/watch?v=7U64ZDJbu_Y

[BigEd]

It's tricky to find the particular points of interest in such a long conversation, but here are a couple of pointers:

About using Moto's customers to help define the subset of the 6800 instructions which would allow for a cheap microprocessor:
https://youtu.be/wiv9417eYoY?t=1h7m54s

About the Moto lawsuit:
https://www.youtube.com/watch?v=wiv9417 ... t=1h22m47s

6502 for control - not for big software projects - and also about the strict die size target and the Special Bus, and the PLA (ROM) being 20 lines too wide:
https://www.youtube.com/watch?v=wiv9417 ... t=1h35m35s
The story as told here says the datapath was already done and was not changed at that point.

However, there's a hint in the layout of the datapath drivers that there were changes to the layout at some point, which has been taken to be evidence that the datapath used to be wider and perhaps had extra function. Specifically, the floating rectangle of diffusion (north east of the ALU out here
http://www.visual6502.org/JSSim/expert. ... &zoom=12.0
) which looks like a point where one or more control line drivers were hacked out to reduce the X size of the chip.

In this thread
viewtopic.php?f=1&t=2015
there's a quote from Chuck Peddle to the effect that they knew the extra die size cost of pushing A on interrupts, and couldn't afford it. It seems plausible that they knew the cost because they'd implemented it, and then had to remove it.

Elsewhere, Segher has opined that the 6502 originally had two accumulators, and that there are indications for this in the instruction encoding (and decoding), but that they were dropped for the same reasons of chip size.

Quote:

My strong suspicion is that originally the 6500 had *two* accumulators, and the instructions with low bits 11 were like those that now have 01, but for reg B instead of reg A.

It looks like "B" was removed when making the die less wide; mostly it saves a whole lot of PLA lines (the register itself takes almost no space). That's also probably why LDX # etc. are jumbled around: it takes less PLA space to decode this way.

You can easily see where PLA lines were removed: originally there was a ground divider every twenty (iirc) lines. More to the left the blocks have more lines missing now, which tells you what features were removed (or altered, simplified, etc.)

Edit: see also the discussion over here.

[BigEd]

There are more Chuck Peddle quotes in Brian Bagnall's book:

Quote:

Near the end of the design process, disaster struck. The engineers realized their architecture would not fit within the allotted area of the microchip. “When we sat down to optimize the system, we discovered we were 10 mills too wide,” says Peddle. “The design was almost done. Mathis and I put a big piece of paper down on a table and sat there and optimized every line until we got rid of 10 mills.”

The engineers were on a tight deadline to have the product ready for the upcoming Wescon show in September. They obsessively searched for ways to recycle lines in the schematic, thus reducing the area. Peddle grimly recalls, “Mathis and I had to keep redoing the architecture to make sure they stayed within that area.”

[fastgear]

Interesting number of Urban legends being debunked in the video.

Talking about urban legends. A couple of years ago I have read Wikipedia's article on the 6502. I can recall that they mentioned when the person have drawn out the mask layers by hand for the 6502, not a single mistake was made.

Well, according to the video this was also an urban legend.

I have just looked at this article on Wikipedia again, and seem that it was changed accordingly.

[jac_goudsmit]

fastgear wrote:

Talking about urban legends. A couple of years ago I have read Wikipedia's article on the 6502. I can recall that they mentioned when the person have drawn out the mask layers by hand for the 6502, not a single mistake was made.

I remember reading a quote from Chuck Peddle where he admired Bill Mensch for being able to design chips and get them right the first time(*). Bill Mensch was on the team for the 6502 at MOS (he also came from Motorola) and he designed the I/O chips: PIA VIA (and possibly the RIOT and RRIOT but I could be mistaken).

===Jac

(*) I suspect that this was an exaggeration.