Input stage for time interval counter
Hi all,
I am currently working on a 2 channel time interval counter. It should
measure frequencies from 5 to 20 MHz. And hopefully have a resolution
about 1E-13 at 0,1s to measure HP10811 OCXO stability.
For inputs should I use UNBAL Rf transformers 1:4 (1:16) or drive the
LTC9657 directly.
Thanks for your suggestions
hgl
I would go with the transformer input. It would remove any common mode
current flowing on the inputs that might 'modulate' the switching
thresholds.
-Mike-
On Wed, Nov 30, 2022 at 12:41 PM Hans-Georg Lehnard via time-nuts <
time-nuts@lists.febo.com> wrote:
Hi all,
I am currently working on a 2 channel time interval counter. It should
measure frequencies from 5 to 20 MHz. And hopefully have a resolution
about 1E-13 at 0,1s to measure HP10811 OCXO stability.
For inputs should I use UNBAL Rf transformers 1:4 (1:16) or drive the
LTC9657 directly.
Thanks for your suggestions
hgl
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe send an email to time-nuts-leave@lists.febo.com
The eval board for the LTC6957 uses capacitive coupling. However, the transformer coupled circuit shown in the datasheet allows operation over a wider input signal level range. Although intended for 10Mhz inputs it should work well at 5MHz as well.
Bruce
On 01/12/2022 09:52 NZDT Michael Fahmie via time-nuts time-nuts@lists.febo.com wrote:
I would go with the transformer input. It would remove any common mode
current flowing on the inputs that might 'modulate' the switching
thresholds.
-Mike-
On Wed, Nov 30, 2022 at 12:41 PM Hans-Georg Lehnard via time-nuts <
time-nuts@lists.febo.com> wrote:
Hi all,
I am currently working on a 2 channel time interval counter. It should
measure frequencies from 5 to 20 MHz. And hopefully have a resolution
about 1E-13 at 0,1s to measure HP10811 OCXO stability.For inputs should I use UNBAL Rf transformers 1:4 (1:16) or drive the
LTC9657 directly.Thanks for your suggestions
hgl
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe send an email to time-nuts-leave@lists.febo.com
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe send an email to time-nuts-leave@lists.febo.com
Hi,
On 2022-11-30 17:14, Hans-Georg Lehnard via time-nuts wrote:
Hi all,
I am currently working on a 2 channel time interval counter. It should
measure frequencies from 5 to 20 MHz. And hopefully have a resolution
about 1E-13 at 0,1s to measure HP10811 OCXO stability.
For inputs should I use UNBAL Rf transformers 1:4 (1:16) or drive the
LTC9657 directly.
Thanks for your suggestions
I think I would go without an transformer. The transformer only helps if
you have low-frequency or DC differences you want to isolate from. For
higher frequencies, it couples through capacitive effect straight
through. If you want to isolate from RF frequencies bring the signal and
"ground" through a ferrit choke to suppress the common mode, and be sure
to have capacitive coupling to ground on the "inside" of that.
For most designs I've only see ferrit choke for common mode current
reduction help. The coax shield should be bolted to the chassi and
ground, but only sensed on the inside as reference, but not distributed
further in on the board. RF transformer use is more relevant if you want
to impedance convert or voltage convert, but isolation wise rarely worth
the effort most of the times.
So, that is what my experience tell me. There are a couple of books from
Henry Ott I would recommend you to consult, if I where to do a narrow
pick from the ones in my bookshelf.
Good luck with your project!
Cheers,
Magnus
Hi all,
I am working on a time interval counter which basically consists of 4
components :
input (LTC9657), divider by 100 (XCR3064), time-to-digital converter
(AS6501; 10ps RMS) and a microcontroller STM32H743-Nucleo Board.
Because of the SPI speed max 100k samples/s are possible. The STM32H7
can store 32000 samples with 1ps resolution in RAM ~ 0.1s.
The AS6501 can also sample 10MHz directly via LVDS (DDR) outputs, but
this requires an FPGA. This will be my next project ;-)
One sample takes 5µs for acquisition so I have 5µs left for
calculations. STM32 runs at 480MHz.
Is there a way (on the fly) to combine e.g. 1000 samples to one value
(MDEV or TDEV) ?.
I have tested a HP10811 as input and LTC9657-4 (direct and inverted
output). Direct output goes to AS6501 REF input and inverted goes to the
divider. Raw data and ADEV,MDEV,TDEV from a test with HP10811-60111
attached.
Statistics for 32k samples
p1
MAX: 10000244 ps
MIN: 9999771 ps
MAX - MIN: 473 ps
AVG: 9999999 ps
p2
MAX: 10000206 ps
MIN: 9999797 ps
MAX - MIN: 409 ps
AVG: 9999999 ps
p3
MAX: 10000226 ps
MIN: 9999780 ps
MAX - MIN: 446 ps
AVG: 9999999 ps
p4
MAX: 10000229 ps
MIN: 9999775 ps
MAX - MIN: 454 ps
AVG: 10000000 ps
p5
MAX: 10000220 ps
MIN: 9999772 ps
MAX - MIN: 448 ps
AVG: 9999999 ps
