Hello

I am facing a problem with the E310  that I do not encounter when using the
E100

Here is the relevant configuration to the device:

set_time_next_pps(time_spec_t{});  // Set the time at 0 at the next pps tick

#if device == E31X

set_time_source("gpsdo")

#endif

In both devices, the gps reports to be locked.

We then receive a signal which is externally synchronized to the GPS PPS.

When receiving this signal, the E100 correctly reports the reception time at
a few us after the PPS while the E310 reports this time at a large offset
after the  PPS. In addition, for the E310, this offset changes each time the
program is run.

It appears  as if on the E310 calling set_time_next_pps(time_spec_t{}) does
not set the time at the internal GPS 1 PPS tick? Did I miss a configuration
step?

I am using UHD 003.008.005

Thank you

Thierry

Thierry:

This question has come up so many times and there isn't a proper solution for it yet. To me it's the largest outstanding bug in the E310 and a deal-killer for several applications.

You can use my dance for synchronizing the E310 GPS previously posted here:

http://lists.ettus.com/pipermail/usrp-users_lists.ettus.com/2016-January/017483.html

But the best you can do is GPS time +- 2000ns. Even if you use an external signal source for time synchronization, the onboard pps trigger is never more than 20ns accurate to GPS time. As a quick summary, some of the other posts talking about this problem:

http://lists.ettus.com/pipermail/usrp-users_lists.ettus.com/2015-November/016928.html

http://lists.ettus.com/pipermail/usrp-users_lists.ettus.com/2015-November/016928.htmlhttp://permalink.gmane.org/gmane.comp.hardware.usrp.e100/18253

The issue seems to arise from an FPGA pps trigger issue, but it's still unresolved in UHD 003.010.x.

As a side note, you need to be using the Release 4 version of the SD image, previous releases used the GPSDO quite differently.

Good Luck.

Kyle Logue

From: USRP-users usrp-users-bounces@lists.ettus.com on behalf of Marcus D. Leech via USRP-users usrp-users@lists.ettus.com
Sent: Thursday, April 21, 2016 07:12
To: usrp-users@lists.ettus.com
Subject: Re: [USRP-users] E310 Clock Synchronization with internal GPS

On 04/21/2016 09:04 AM, Thierry Guichon via USRP-users wrote:

Hello

I am facing a problem with the E310  that I do not encounter when using the E100

Here is the relevant configuration to the device:

set_time_next_pps(time_spec_t{});  // Set the time at 0 at the next pps tick

#if device == E31X

set_time_source("gpsdo")

#endif

In both devices, the gps reports to be locked.

We then receive a signal which is externally synchronized to the GPS PPS.

When receiving this signal, the E100 correctly reports the reception time at a few us after the PPS while the E310 reports this time at a large offset after the  PPS. In addition, for the E310, this offset changes each time the program is run.

It appears  as if on the E310 calling set_time_next_pps(time_spec_t{}) does not set the time at the internal GPS 1 PPS tick? Did I miss a configuration step?

I am using UHD 003.008.005

If you can, I strongly recommend upgrading your E310 to the latest system image, which has a much-more-recent UHD.

Use the image from here:

http://files.ettus.com/e3xx_images/e3xx-release-4/ettus-e3xx-sg1/sdimage-gnuradio-dev.direct.xz

Using the instructions here:

http://files.ettus.com/manual/page_usrp_e3x0.html#e3x0_upgrade_sd_card

Also, we're probably going to have to see more of your setup code to make a determination about what might be wrong.

Thank you

Thierry

USRP-users mailing list
USRP-users@lists.ettus.commailto:USRP-users@lists.ettus.com
http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com

Thank you Marcus and Kyle,

It looks like anyhow I need to update my image :-)

The offset of 2 microseconds that you are experiencing will not be a problem
for my application. The offsets that I am currently seeing are more in the
order of several hundreds milliseconds so there is something fundamentally
wrong with either my setup or the UHD version that I am using.

I will see what I get after the upgrade.

Sincerely

Thierry

From: USRP-users [mailto:usrp-users-bounces@lists.ettus.com] On Behalf Of
Kyle A Logue via USRP-users
Sent: Thursday, April 21, 2016 1:09 PM
To: Ettus Mail List
Subject: Re: [USRP-users] E310 Clock Synchronization with internal GPS

Thierry:

This question has come up so many times and there isn't a proper solution
for it yet. To me it's the largest outstanding bug in the E310 and a
deal-killer for several applications.

You can use my dance for synchronizing the E310 GPS previously posted here:

http://lists.ettus.com/pipermail/usrp-users_lists.ettus.com/2016-January/017
483.html

But the best you can do is GPS time +- 2000ns. Even if you use an external
signal source for time synchronization, the onboard pps trigger is never
more than 20ns accurate to GPS time. As a quick summary, some of the other
posts talking about this problem:

http://lists.ettus.com/pipermail/usrp-users_lists.ettus.com/2015-November/01
6928.html

<http://lists.ettus.com/pipermail/usrp-users_lists.ettus.com/2015-November/0
16928.html> http://permalink.gmane.org/gmane.comp.hardware.usrp.e100/18253

The issue seems to arise from an FPGA pps trigger issue, but it's still
unresolved in UHD 003.010.x.

As a side note, you need to be using the Release 4 version of the SD image,
previous releases used the GPSDO quite differently.

Good Luck.

Kyle Logue

From: USRP-users usrp-users-bounces@lists.ettus.com on behalf of Marcus D.
Leech via USRP-users usrp-users@lists.ettus.com
Sent: Thursday, April 21, 2016 07:12
To: usrp-users@lists.ettus.com
Subject: Re: [USRP-users] E310 Clock Synchronization with internal GPS

On 04/21/2016 09:04 AM, Thierry Guichon via USRP-users wrote:

Hello

I am facing a problem with the E310  that I do not encounter when using the
E100

Here is the relevant configuration to the device:

set_time_next_pps(time_spec_t{});  // Set the time at 0 at the next pps tick

#if device == E31X

set_time_source("gpsdo")

#endif

In both devices, the gps reports to be locked.

We then receive a signal which is externally synchronized to the GPS PPS.

When receiving this signal, the E100 correctly reports the reception time at
a few us after the PPS while the E310 reports this time at a large offset
after the  PPS. In addition, for the E310, this offset changes each time the
program is run.

It appears  as if on the E310 calling set_time_next_pps(time_spec_t{}) does
not set the time at the internal GPS 1 PPS tick? Did I miss a configuration
step?

I am using UHD 003.008.005

If you can, I strongly recommend upgrading your E310 to the latest system
image, which has a much-more-recent UHD.

Use the image from here:

http://files.ettus.com/e3xx_images/e3xx-release-4/ettus-e3xx-sg1/sdimage-gnu
radio-dev.direct.xz

Using the instructions here:

http://files.ettus.com/manual/page_usrp_e3x0.html#e3x0_upgrade_sd_card

Also, we're probably going to have to see more of your setup code to make a
determination about what might be wrong.

Thank you

Thierry

USRP-users mailing list
USRP-users@lists.ettus.com
http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com

A quick correction:

I said this problem was unresolved in 3.10.x but that is not strictly true. It is unresolved in the E310 Release 4 image which is using 3.9.2.

Paul Garver posted to the mailing list today this tidbit:

For others who may be searching this same topic, I’ll provide an update. We had an error in our metadata recording code which was causing the mean time delay estimation between multiple B200s to be on the order of microseconds (~25uS) at 25 MSPS. Now, we have all three B200s sync’d within a sample on UHD 3.8.5 using a timing reference and 10 MHz/1PPS. The sample mean is around 40nS and the sample variance is extremely low (reads zero). To do better than this, I presume we will need to compensate for the random phase by calibrating on every run / retune.

We use our custom recording tool in gr-analysis to record the times, which looks like it time-tags accurately as long as the correct segment size is given (default is OK).
https://github.com/garverp/gr-analysis

Does this seem reasonable? Does anyone else have time alignment results for the B200 they would care to share?

PWG

Which may indicate that there is something in UHD 3.9.5+ that could address this problem, but i haven't seen any alpha images on files.ettus.com newer than release 4.

Something to keep an eye on,

Kyle

From: USRP-users usrp-users-bounces@lists.ettus.com on behalf of Thierry Guichon via USRP-users usrp-users@lists.ettus.com
Sent: Thursday, April 21, 2016 10:34
To: 'USRP Mailing List'
Subject: Re: [USRP-users] E310 Clock Synchronization with internal GPS

Thank you Marcus and Kyle,

It looks like anyhow I need to update my image :-)

The offset of 2 microseconds that you are experiencing will not be a problem for my application. The offsets that I am currently seeing are more in the order of several hundreds milliseconds so there is something fundamentally wrong with either my setup or the UHD version that I am using.

I will see what I get after the upgrade.

Sincerely

Thierry

From: USRP-users [mailto:usrp-users-bounces@lists.ettus.com] On Behalf Of Kyle A Logue via USRP-users
Sent: Thursday, April 21, 2016 1:09 PM
To: Ettus Mail List
Subject: Re: [USRP-users] E310 Clock Synchronization with internal GPS

Thierry:

This question has come up so many times and there isn't a proper solution for it yet. To me it's the largest outstanding bug in the E310 and a deal-killer for several applications.

You can use my dance for synchronizing the E310 GPS previously posted here:

http://lists.ettus.com/pipermail/usrp-users_lists.ettus.com/2016-January/017483.html

But the best you can do is GPS time +- 2000ns. Even if you use an external signal source for time synchronization, the onboard pps trigger is never more than 20ns accurate to GPS time. As a quick summary, some of the other posts talking about this problem:

http://lists.ettus.com/pipermail/usrp-users_lists.ettus.com/2015-November/016928.html

http://lists.ettus.com/pipermail/usrp-users_lists.ettus.com/2015-November/016928.htmlhttp://permalink.gmane.org/gmane.comp.hardware.usrp.e100/18253

The issue seems to arise from an FPGA pps trigger issue, but it's still unresolved in UHD 003.010.x.

As a side note, you need to be using the Release 4 version of the SD image, previous releases used the GPSDO quite differently.

Good Luck.

Kyle Logue

From: USRP-users usrp-users-bounces@lists.ettus.com on behalf of Marcus D. Leech via USRP-users usrp-users@lists.ettus.com
Sent: Thursday, April 21, 2016 07:12
To: usrp-users@lists.ettus.com
Subject: Re: [USRP-users] E310 Clock Synchronization with internal GPS

On 04/21/2016 09:04 AM, Thierry Guichon via USRP-users wrote:

Hello

I am facing a problem with the E310  that I do not encounter when using the E100

Here is the relevant configuration to the device:

set_time_next_pps(time_spec_t{});  // Set the time at 0 at the next pps tick

#if device == E31X

set_time_source("gpsdo")

#endif

In both devices, the gps reports to be locked.

We then receive a signal which is externally synchronized to the GPS PPS.

When receiving this signal, the E100 correctly reports the reception time at a few us after the PPS while the E310 reports this time at a large offset after the  PPS. In addition, for the E310, this offset changes each time the program is run.

It appears  as if on the E310 calling set_time_next_pps(time_spec_t{}) does not set the time at the internal GPS 1 PPS tick? Did I miss a configuration step?

I am using UHD 003.008.005

If you can, I strongly recommend upgrading your E310 to the latest system image, which has a much-more-recent UHD.

Use the image from here:

http://files.ettus.com/e3xx_images/e3xx-release-4/ettus-e3xx-sg1/sdimage-gnuradio-dev.direct.xz

Using the instructions here:

http://files.ettus.com/manual/page_usrp_e3x0.html#e3x0_upgrade_sd_card

Also, we're probably going to have to see more of your setup code to make a determination about what might be wrong.

Thank you

Thierry

USRP-users mailing list

USRP-users@lists.ettus.commailto:USRP-users@lists.ettus.com

http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com