'Criticise this processor choice - Z8 versus PIC'
I'm looking for some "devil's advocates" who have used Z8's
I have an application where the new Z8E seems the ideal choice.
I haven't heard too many comments about Z8 except for one here last
October - he criticised them thoroughly - I won' mention his name but
I'll send him a copy of this query - may get some useful input. One
other user mentioned it without positive or negative comment.
All values need accuracies of at least 6 bits - the more the better.
1 x 1 KHz PWM - read mark/space.
1 x variable 50 - 3000 Hz square wave - read frequency.
1 x analog voltage (0 - 0.1 ish volts) - read to about 6 bits plus
1 x 10 KHz PWM. (average resolution about 6 bits)
1 x variable frequency 1 - 500 Hz (6 bits plus ?)
Output PWM is complexish function of all input signals.
Seriously, it's gotta be cheap.
Volume is 50,000 per year or so.
Considered 16C71 or new variants. About $US2 - maybe somewhat less.
0.1 volt signal will need to be amplified so one packet of op amps
needed (LM324 or similar OK)
If a VCO is used this analog signal could become a frequency and
remove the need for the A2D.
So - a PIC 16C505 could perhaps be used.
The single timer would make life hard.
The output PWM is challenging - a lower resolution with dithering
would allow say 4 bit (or less?) resolution.
The Z8E00110 has
3 timers (2 x 8, 1 x 16 bit)
2 x 8 bit will provide a 16 bit (or less) set-and-forget PWM output
Use VCO for analog in.
AND the volume price (say 10,000 up) looks at this stage to be a bit
under $1US. Slightly more than 16C505 but much easier to meet spec.
Going up to 16C71 and friends still leaves me with 1 timer and no
Its hard to escape the Z8 in this application.
Tell me all the bad things that you know or have heard (or that the
opposition rep has insinuated :-)) about the Z8.
Sure. I'll use PIC elsewhere still and I'd use it here if it works
but what would I use instead of the Z8
The Z8 looks like it will be slower than a PIC (better addressing and
general instruction set and architecture, poorer MIPs)and less code
efficient. Certainly MChips comparisons say so - the new core is
faster but not vastly so.
Some Z8E background
This has a 10 MHz max clock which is divided by 10 to give a 1uS
constant instruction time at 10 MHz.
It has 2 x 8 bit and 1 x 16 bit timer. The 2 x 8 bit timers can be
combined to provide a 16 bit PWM output.
It also has an analog comparator.
64 bytes RAM
1KB ROM (OTP)
No UART, no EEROM,
They have a real albeit simple ICE for $99 ($US99, $NZ99, probably
$AUS99 too :-))
Genuine program, assemble, download and run in target circuit !
For me the $NZ1.70 in true volume ($2.10ish in tube quantities) is
Last week I asked a question about what processor would be good
out of the Z8, AVR, SX, and SST's "FlashFlex 51" family. I received
about 4 emails straight to me. They all had used ZILOGs chips for sometime.
The thing that caught my attention was that they all basically had the
same complaint, ZILOG seems to have this habit of changing their die
and not telling people. A few people did say that overall ZILOG is a
good chip for selection, price, and are great number crunchers.
Hope this helps, if you decide to go that way let me know how it
Phoenix Arizona USA
I used Zilog Z86E30 and got it masked thru their Distri in Hong Kong. I
placed order for 10k pcs and paid 3000 USD for masking charges over and
above the cost of masked parts.
However distri got 10k pcs manufactured after my producted clikced in the
marker and sold it in the trade market. When I complained this to Zilog's
Hong Kong office, they simply said it is beyond their scope. They get
order from distri and do not have anything to do with end customer.
Now I use Motorola parts even if they art more expensive. No Zilog - Cant
On Wed, 9 Jun 1999, Steven Kosmerchock wrote:
Let me add a few more things about Zilog chips.
at work we use to use zilog chip more then 3 years
I found very strange problems with it.
1. any input you should have at list 150 mv that means
if you have a switch connected to the input directly
you are in trouble
put 1 mohm resister on crystal pins
2. Factory test mode accidentally may enable if osc1 input
gets 9 volts noise for 3 ms. we use our product in noisy environment
we had mystery problems with Zilog.
put clamping diode one side to the VDD and the other side to
the osc1 pin.
any way we do not use Zilog chips any more I use my favorite
Microchip product I always been happy with this company
excellent product and excellent support.
I think Zilog is good for toys or something if it didn't work properly
kids are not going to notice them they may thing that is how it
made any way Good lock with Zilog users.
Vinay Kumar wrote:
More... (looser matching)
- Last day of these posts
- In 1999
, 2000 only
- New search...