Searching \ for 'PIC architecture and fast ADC' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: piclist.com/techref/microchip/ios.htm?key=adc
Search entire site for: 'PIC architecture and fast ADC'.

Truncated match.
PICList Thread
'PIC architecture and fast ADC'
2005\08\01@210246 by Vasile Surducan

face picon face
On 8/2/05, Chen Xiao Fan <spam_OUTxiaofanTakeThisOuTspamsg.pepperl-fuchs.com> wrote:
> He actually told me that it is not a process issue even though
> he did mention that things like EMC/EMI get tricky when going to
> fine process like 0.25um or lower with lower Vdd and lower
> latch-up threshold voltage. He specifically said that it is
> an architecture issue and they would have to come out
> something like PIC20 to incorporate this. Of course this
> is not a highly sought features in the low end so they
> will not come out a PIC20 architecture just to make faster
> ADC possible. However I still do not quite understand this.


 Mybe things would be much clear after you'll take a look to the
fastest Maxim's AD converters arhitecture like MAX1448 or
MAX1181/1186. Maybe the example I've chosen is not the best, but
definitely will reveal the EMI problems. It's almost the same
capacitor charging arhitecture as used in PIC at very slow aquisition
time, but with a 1.5 bit chained conversion up to 10 bit. 500Ksps ADC
into an ordinary PIC will need a FIFO buffer, another clock and
everything can be much expensive. Or as a producer we need small
troubles and a lot of money from something we already have isn't it ?

cheers,
Vasile



{Quote hidden}

> {Original Message removed}

2005\08\01@220020 by William Chops Westfield

face picon face
>> He actually told me that it is not a process issue even though
>> he did mention that things like EMC/EMI get tricky when going to
>> fine process like 0.25um or lower with lower Vdd and lower
>> latch-up threshold voltage. He specifically said that it is
>> an architecture issue and they would have to come out
>> something like PIC20 to incorporate this.

Perhaps he means there simply isn't room in the SFR address memory
map for yet another "standardized" peripheral...  It looks like the
SFRs occur ("must" occur?) in the first 0x20 bytes of bank0/1 and
the first 0x18 bytes of bank2/3, and if you look at something like
the pic16f87x register file map, there aren't a lot of unused locations
there to add additional peripherals (and it's not as if the 87x has ALL
the possible peripherals, either...)

BillW

More... (looser matching)
- Last day of these posts
- In 2005 , 2006 only
- Today
- New search...