Microchip PIC CTMU
I ran a search through the time-nuts archives and found only one
mention of this a couple years ago asking the same question which is,
has anybody explored the capabilities of the Microchip PIC CTMU for
use as a time to digital converter or programmable delay?
The Microchip datasheets lack detailed performance specifications but
my conservative estimate is that resolution down to 500ps over 200ns
using the built in 10 bit ADC should be possible without complex
calibration. (The PIC series ADCs with more than 10 bits are pretty
horrible.)
See What You Can Do with the CTMU:
http://ww1.microchip.com/downloads/en/AppNotes/CTMU%2001375a.pdf
Overview of Charge Time Measurement Unit (CTMU):
http://www.microchip.com/stellent/groups/SiteComm_sg/documents/DeviceDoc/en542792.pdf
PIC18(L)F2X/4XK22 w/10 Bit ADC:
http://ww1.microchip.com/downloads/en/DeviceDoc/39977f.pdf
PIC18F66K80 w/12 Bit ADC:
http://ww1.microchip.com/downloads/en/DeviceDoc/39977f.pdf
I prefer more discrete implementations but this might be useful for
its higher integration, lower cost, and simplicity in less demanding
applications.
Hi
I have not dug into them for several years. Back when I did, you were pushing
their limits as a couple of nanoseconds without getting into things like lots of
averaging.
Bob
On May 10, 2016, at 8:29 PM, David davidwhess@gmail.com wrote:
I ran a search through the time-nuts archives and found only one
mention of this a couple years ago asking the same question which is,
has anybody explored the capabilities of the Microchip PIC CTMU for
use as a time to digital converter or programmable delay?
The Microchip datasheets lack detailed performance specifications but
my conservative estimate is that resolution down to 500ps over 200ns
using the built in 10 bit ADC should be possible without complex
calibration. (The PIC series ADCs with more than 10 bits are pretty
horrible.)
See What You Can Do with the CTMU:
http://ww1.microchip.com/downloads/en/AppNotes/CTMU%2001375a.pdf
Overview of Charge Time Measurement Unit (CTMU):
http://www.microchip.com/stellent/groups/SiteComm_sg/documents/DeviceDoc/en542792.pdf
PIC18(L)F2X/4XK22 w/10 Bit ADC:
http://ww1.microchip.com/downloads/en/DeviceDoc/39977f.pdf
PIC18F66K80 w/12 Bit ADC:
http://ww1.microchip.com/downloads/en/DeviceDoc/39977f.pdf
I prefer more discrete implementations but this might be useful for
its higher integration, lower cost, and simplicity in less demanding
applications.
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Was that a couple of nanoseconds of resolution or minimum time
interval?
Based on the simplified schematic which shows the current source being
controlled by a single XOR gate, it looks to me like the minimum time
interval is much longer than the achievable single shot resolution.
On Tue, 10 May 2016 21:37:08 -0400, you wrote:
Hi
I have not dug into them for several years. Back when I did, you were pushing
their limits as a couple of nanoseconds without getting into things like lots of
averaging.
Bob
On May 10, 2016, at 8:29 PM, David davidwhess@gmail.com wrote:
I ran a search through the time-nuts archives and found only one
mention of this a couple years ago asking the same question which is,
has anybody explored the capabilities of the Microchip PIC CTMU for
use as a time to digital converter or programmable delay?
The Microchip datasheets lack detailed performance specifications but
my conservative estimate is that resolution down to 500ps over 200ns
using the built in 10 bit ADC should be possible without complex
calibration. (The PIC series ADCs with more than 10 bits are pretty
horrible.)
See What You Can Do with the CTMU:
http://ww1.microchip.com/downloads/en/AppNotes/CTMU%2001375a.pdf
Overview of Charge Time Measurement Unit (CTMU):
http://www.microchip.com/stellent/groups/SiteComm_sg/documents/DeviceDoc/en542792.pdf
PIC18(L)F2X/4XK22 w/10 Bit ADC:
http://ww1.microchip.com/downloads/en/DeviceDoc/39977f.pdf
PIC18F66K80 w/12 Bit ADC:
http://ww1.microchip.com/downloads/en/DeviceDoc/39977f.pdf
I prefer more discrete implementations but this might be useful for
its higher integration, lower cost, and simplicity in less demanding
applications.
Hi
The same process issues that make for a “less than perfect” ADC, also
add more than a little noise into the circuit. Exactly which component(s)
are the biggest culprits … no idea. About all you can do is take data and
look at the results.
Bob
On May 11, 2016, at 4:55 AM, David davidwhess@gmail.com wrote:
Was that a couple of nanoseconds of resolution or minimum time
interval?
Based on the simplified schematic which shows the current source being
controlled by a single XOR gate, it looks to me like the minimum time
interval is much longer than the achievable single shot resolution.
On Tue, 10 May 2016 21:37:08 -0400, you wrote:
Hi
I have not dug into them for several years. Back when I did, you were pushing
their limits as a couple of nanoseconds without getting into things like lots of
averaging.
Bob
On May 10, 2016, at 8:29 PM, David davidwhess@gmail.com wrote:
I ran a search through the time-nuts archives and found only one
mention of this a couple years ago asking the same question which is,
has anybody explored the capabilities of the Microchip PIC CTMU for
use as a time to digital converter or programmable delay?
The Microchip datasheets lack detailed performance specifications but
my conservative estimate is that resolution down to 500ps over 200ns
using the built in 10 bit ADC should be possible without complex
calibration. (The PIC series ADCs with more than 10 bits are pretty
horrible.)
See What You Can Do with the CTMU:
http://ww1.microchip.com/downloads/en/AppNotes/CTMU%2001375a.pdf
Overview of Charge Time Measurement Unit (CTMU):
http://www.microchip.com/stellent/groups/SiteComm_sg/documents/DeviceDoc/en542792.pdf
PIC18(L)F2X/4XK22 w/10 Bit ADC:
http://ww1.microchip.com/downloads/en/DeviceDoc/39977f.pdf
PIC18F66K80 w/12 Bit ADC:
http://ww1.microchip.com/downloads/en/DeviceDoc/39977f.pdf
I prefer more discrete implementations but this might be useful for
its higher integration, lower cost, and simplicity in less demanding
applications.
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
It is always an extra challenge to get even datasheet performance out
of sampling converters which are integrated with a microcontroller and
of course Microchip says nothing about noise or complete AC
performance.
I suspect adding an external ADC to the Microchip CTMU while possible
would be a waste of effort. If you are going to do that, then you
might as well design in a real time to voltage converter.
On Wed, 11 May 2016 07:27:14 -0400, you wrote:
Hi
The same process issues that make for a less than perfect ADC, also
add more than a little noise into the circuit. Exactly which component(s)
are the biggest culprits
no idea. About all you can do is take data and
look at the results.
Bob
Hi
Given that a “real” TDC is a resistor and capacitor attached to a FPGA pin (plus the ADC)
the cost of doing it better is not all that much. You can get down to a few hundred ps without
a lot of crazy effort ( still using the MCU ADC).
Bob
On May 13, 2016, at 3:51 AM, David davidwhess@gmail.com wrote:
It is always an extra challenge to get even datasheet performance out
of sampling converters which are integrated with a microcontroller and
of course Microchip says nothing about noise or complete AC
performance.
I suspect adding an external ADC to the Microchip CTMU while possible
would be a waste of effort. If you are going to do that, then you
might as well design in a real time to voltage converter.
On Wed, 11 May 2016 07:27:14 -0400, you wrote:
Hi
The same process issues that make for a less than perfect ADC, also
add more than a little noise into the circuit. Exactly which component(s)
are the biggest culprits
no idea. About all you can do is take data and
look at the results.
Bob
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Based on my earlier experiences with pattern dependant jitter caused
by interaction between different logic circuits, a few hundred
picoseconds is about what I would expect if an FPGA (or
microcontroller) is controlling the charge cycle timing. I would be
surprised if the PIC CTMU cannot achieve that unaided with careful
design.
On Fri, 13 May 2016 07:36:10 -0400, you wrote:
Hi
Given that a real TDC is a resistor and capacitor attached to a FPGA pin (plus the ADC)
the cost of doing it better is not all that much. You can get down to a few hundred ps without
a lot of crazy effort ( still using the MCU ADC).
Bob
On Fri, 13 May 2016 07:36:10 -0400
Bob Camp kb8tq@n1k.org wrote:
Given that a “real” TDC is a resistor and capacitor attached to a FPGA pin (plus the ADC)
the cost of doing it better is not all that much. You can get down to a few hundred ps without
a lot of crazy effort ( still using the MCU ADC).
The PICTIC II did 250ps resolution with an effective resolution of 680ps[1].
The upgraded version III had IIRC a resolution of ~25ps with better than 100ps
achieved. Eventhough the PICTIC III design is a bit more involved than
then II, it's still relatively simple (it uses more stable current sources,
buffer opamps, external ADCs and an active offset compensation scheme).
Attila Kinali
PS: Does anyone know what happend to Richard McCorkle? I haven't heard
of him in ages.
[1] http://www.ke5fx.com/pictic.htm
--
Reading can seriously damage your ignorance.
-- unknown
Only found info about PICTIC 1 and 2:
http://www.ko4bb.com/doku2015/doku.php?id=precision_timing:pictic
Where´s info about PICTIC III?
Daniel
Em 13/05/2016 14:51, Attila Kinali escreveu:
On Fri, 13 May 2016 07:36:10 -0400
Bob Camp kb8tq@n1k.org wrote:
Given that a “real” TDC is a resistor and capacitor attached to a FPGA pin (plus the ADC)
the cost of doing it better is not all that much. You can get down to a few hundred ps without
a lot of crazy effort ( still using the MCU ADC).
The PICTIC II did 250ps resolution with an effective resolution of 680ps[1].
The upgraded version III had IIRC a resolution of ~25ps with better than 100ps
achieved. Eventhough the PICTIC III design is a bit more involved than
then II, it's still relatively simple (it uses more stable current sources,
buffer opamps, external ADCs and an active offset compensation scheme).
Attila Kinali
PS: Does anyone know what happend to Richard McCorkle? I haven't heard
of him in ages.
On Fri, May 13, 2016 at 07:51:07PM +0200, Attila Kinali wrote:
On Fri, 13 May 2016 07:36:10 -0400
Bob Camp kb8tq@n1k.org wrote:
Given that a “real” TDC is a resistor and capacitor
attached to a FPGA pin (plus the ADC) the cost of doing it
better is not all that much. You can get down to a few hundred
ps without a lot of crazy effort ( still using the MCU ADC).
The PICTIC II did 250ps resolution with an effective resolution
of 680ps[1].
The upgraded version III had IIRC a resolution of ~25ps with
better than 100ps achieved. Eventhough the PICTIC III design is
a bit more involved than then II, it's still relatively simple
(it uses more stable current sources, buffer opamps, external
ADCs and an active offset compensation scheme).
Hmm, sounds interesting ...
Is there a PICTIC III schematic somewhere?
Thanks in advance,
Herbert
Attila Kinali
PS: Does anyone know what happend to Richard McCorkle? I haven't heard
of him in ages.
--
Reading can seriously damage your ignorance.
-- unknown
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.