More free stuff!

Post Reply
User avatar
Dave
Posts: 1280
Joined: 11 Aug 2012 18:16
Contact:

More free stuff!

Post by Dave »

Hi folks,

You may be aware of the fake CMOS Z80s that we have found recently. The chips were marked as 20MHz CMOS Z80s, but were anything but!

Lez sent me a batch of additional CPUs for testing to assess the quality that he was getting from a couple of Chinese suppliers, testing has shown that they are not what the labels say they are, but they are still functional at lower speeds.

The chips are :-
NMOS
1 runs at 2MHz
2 runs at 4MHz
3 runs at 8Mhz
4 runs at 4MHz
5 runs at 4MHz
6 runs at 4Mhz
7 runs at 2 Mhz
So, they look like a mixture of Z80, Z80A and possibly a Z80H or maybe it's just a "B" that can be overclocked a little.

CMOS
1 runs OK at 8MHz
2 runs OK at 8MHz
3 runs OK at 8MHz
4 OK at 16MHz
5 OK at 16MHz
6 runs OK at 8MHz
7 OK at 8 MHz
8 OK at 16MHz
9 OK at 8 MHz
10 OK at 8 MHz
So, they look like 10MHz CMOS, remarked as 20 MHz
 
If anyone is interested in having a couple of these for the cost of postage, perhaps as spares for slower systems, including MTXs, let me know.

Regards
Dave
lezanderson
Posts: 186
Joined: 14 Aug 2012 15:31

Re: More free stuff!

Post by lezanderson »

Hi.

If anyone has tested the PLCC44 Z80s then let us know how you got on. I can't see the PLCC versions being counterfeit or 'Re-labelled' so I'd guess they're the genuine article.
Martin A
Posts: 799
Joined: 09 Nov 2013 21:03

Re: More free stuff!

Post by Martin A »

I've build a simple 32k rom, 32k ram CPU board, using one of the 20Mhz PLCC Z80s, that will plug into the MTX+ backplane.

It's fitted with 45ns Flash and 35ns RAM, so more than capable of running at 20mhz. And should be capable of running at around 30Mhz.

The Initial CMOS/NMOS test rom confirmed this IS a CMOS processor as advertised.
Initial test 16mhz
Initial test 16mhz
PLCC Z80 at 16 Mhz.jpg (111.04 KiB) Viewed 5617 times
Th memory test rom image was run at 16 Mhz to make use of test software written for the MTX+

As the screenshot shows, the test rom was able to run. Confirm working RAM in the upper 32k.

The next test was to change the clock speed from 16 to 20, by changing the oscillator.

Everything worked as expected. Not something that any of the suspect PDIP chips managed.

The next test was to put the 32mhz oscillator back on board, and move the shunt to wire the oscillator output directly to the CPU clock input, Bypassing the clock divider circuit.
Running at 32 Mhz with the divider bypassed
Running at 32 Mhz with the divider bypassed
PLCC Z80 at 32 Mhz.jpg (100.74 KiB) Viewed 5617 times
The test runs as expected with the clock at full speed. This isn't conclusive evidence that the chip is a 20mhz part, as I've a 10mhz rated CMOS DIP Z80 that will run at the same speed on a similar board. BUT I don't think that matters, the chip is running at it's rated speed and beyond.

It's a good one!
Martin A
Posts: 799
Joined: 09 Nov 2013 21:03

Re: More free stuff!

Post by Martin A »

More testing, The test rom was re-written to increase the VDP access delay to a point where it could work at 40mhz.

At 40mhz, assuming the chip will go that fast, the M1 memory access which is 1.5 cycles is only 37ns.

Even before allowing for the path though the select logic, this is outside the spec of the flash chip though within the abilities of the ram chip.

Initial tests showed that "something" was happening, as the VDP was being programmed with the correct screen mode, so further experimentation was definitely worthwhile.

I swapped out the CMOS logic for faster F series chips to reduce the select times to a minimum.

After confirming the board still worked 20mhz, not guaranteed because of the different voltage spec of the 2 families of chips, I ran it again with the divider bypassed.

It ran the full test.
running at 40.jpg
running at 40.jpg (669.34 KiB) Viewed 5613 times
You can just about see the scope screen in the photo reading 40mhz. Interestingly, with the scope attached the 16k ram in block 2 is failing the test. With the probe removed the tests pass again, confirming just how close to the edge the board is running.

Being a skeptic, I hand timed the tests at the various jumper settings

5mhz - 5 sec
10mhz 2.6 sec
20mhz 1.2 sec
40mhz 0,8 sec

So given that I was using an old clockwork stopwatch with a 0.2 sec resolution I'm happy the chip IS running at 40mhz, and not some weird harmonic.

All in all a good result for a chip labeled as 20mhz.

(and that Master is STILL turned off)
Post Reply