the science & technology of Commodore 8 bit computers and circuits

Disk drives, Monitors, SuperCPU etc.
User avatar
eslapion
Active Member
Active Member
Posts: 1215
Joined: Mon Jul 20, 2015 10:11 am
Location: Canada
Contact:

Re: the science & technology of Commodore 8 bit computers and circuits

Post by eslapion »

Here are a few things I discovered about Commodore PLAs that were useful in creating PLAnkton.

Sidenote: I consider I didn't really create PLAkate but rather discovered why it works perfectly well.

- Genuine Commodore PLAs signal at about 0-4 volts so claims by people at Tynemouth software that an adapter circuit is required for low voltage circuits to work well are false. We run the XC9536XL of PLAnkton at 3.6 volts - this circuits being CMOS, it signals rail-to-rail.

- The CASRAM latency on the 82S100 PLA used on early 326298 boards which makes it compatible with both early and late generations of C64 breadbin mainboards is near exactly 40ns. Use a longer delay and there will be DRAM problems with boards equipped with an RC filter (C204 and R42); use a shorter delay and there will be problems with boards 326298, KU14194 and early revisions of the 250407.

- The slew rate is very important. NMOS and TTL-LS technology chips have a relatively low slew rate compared to modern CMOS chips. What makes the difference betwenn 99% compatibility and 100% compatibility is probably this final detail - PLAnkton uses the slew rate limiting capabilities built into the XC9536XL. A high slew rate is what causes the SuperPLA V3 to be incompatible with FastLoad and a few other devices and it causes ringing on all logic transitions on the output because the C64 mainboards weren't designed to handle such fast switchings as those generated by the MACH210. The Atmel chip suggested by Tynemouth software also suffers this problem.

- Most 27C512 PROMs do generate multiple signal or glitches when there are transitions on input which should not result on changes on the output and these are called "static hazards". They are very easy to detect on the ROML and ROMH lines of the cartridge port when no cartridge is attached. PLAkate was created when I discovered the M27C512-90B6 does not suffer this problem specifically because it has a low slew rate. Using the cartridge game Super Zaxxon (which is very different from the disk version) is one of the easiest ways to detect these glitches because it uses a flip-flop tied to ROML to switch which section of the larger ROM is available to the C64.

- While the old 82S100 PLA consume a whopping 95mA (at 5 volts, this represents about half a watt of heat) to work, I was stunned to realize most PROMs (faulty or good) used as a PLA replacement consume less than 1/10th of that power. Being a programmable logic chip, the XC9536XL consumes about 20mA so this is more than a PROM but still about 1/5th the power of an original chip - nearly the same as a blue LED.

- Commodore 16, 64, Plus/4 and 1551 drive PLAs use only combinatorial logic; no registered logic is involved at all. For a given input, you will always obtain (after propagation delay) the same binary value on output. In the case of the C64, the trick is simply to ensure there is a slightly greater delay on the activation of the CASRAM line and no more than one output is low at any given time (except when the CPU is writing to color RAM).

- Genuine Commodore PLAs - just like all TTL-LS chips - consider a logic 0 all signals of less than 1.3 volts and all signals above that are considered a logic 1. As posted "elsewhere", I used a simple triangular wave of about 75KHz to determine the exact value. Anyone who claims this "threshold value" to be 2.5 volts is simply ignorant of standards in the logic circuit industry.


Wealth, like happiness, is never attained directly. It comes as a by-product of providing a useful service. -Harland D. Sanders
User avatar
eslapion
Active Member
Active Member
Posts: 1215
Joined: Mon Jul 20, 2015 10:11 am
Location: Canada
Contact:

Re: the science & technology of Commodore 8 bit computers and circuits

Post by eslapion »

For those who want the ugliest looking PLA replacement that's guaranteed to work as well as a genuine Commodore made PLA, sPLAtt is still available for 7.60 Euros or 8.50$US apiece.

Since sPlatt requires a logic inverter to select which one of the 2 PROMs is active, I used a 74HCT14 which has a few extra gates and used 2 of them to introduce adequate CASRAM latency. The consequence is this contraption works well on all C64 boards including the old 326298, KU-14194 and... the C64r.

Reminder: sPLAtt is based on ST's M27C256B-90B6 which has the same output characteristics as ST's M27C512-90B6 thereby eliminating any and all glitches or possible bus contention you would have with the Atmel AT27C512-45PU.

It is covered by the same guarantee as PLAnkton; 90 money back and 3 years afterwards.

Posted last year on Denial:
I noted the M27C256B-90B6 from ST too also has the special "lazy output" that characterizes the M27C512-90B6 - it too has a low slew rate that prevents any and all bus contention from ever being manifested on a C64's bus when used a a PLA substitute. Then the question is: How do you put 65536 possible combinations in a chip that has only 32768 possible entries? Answer: By using two chips and using what was the 16th input as a means of selecting the right chip.

A pair of stacked M27C256B-90B6 can become a single PLA sub and I call this ugly looking but perfectly working contraption sPLAtt.

When tested with a logic analyser, sPLAtt generates EXACTLY the same type of signal as PLAkate but being built around 2 PROMs, it doesn't have the energy saving benefit of PLAkate. I do have 50 of them in stock, giving the potential to make another 25 PROM based PLAs that operate to perfection.

Since I need a logic inverter to select which PROM signals, i used the left over gates from the inverter to increase the CASRAM latency thereby giving sPLAtt the same level of compatibility with older boards as PLAnkton.
Wealth, like happiness, is never attained directly. It comes as a by-product of providing a useful service. -Harland D. Sanders
Post Reply Previous topicNext topic

Who is online

Users browsing this forum: No registered users and 14 guests