Empathy List Archives

DO

Daniel O'Connor

Sun, May 18, 2025 12:43 AM

Hi,
I’m trying to use an M8F to make a 10MHz GPSDO - I have designed a PCB with it, a VCO (AOCJY1-10.000MHZ-E) to EXTINT0 controlled by a Ti DAC8571 and divided by 10e6 (using a Renesas GreenPak) going to EXTINT1. ie the “external (coherent) PPS” design from the documentation.

I tried a naive configuration (using the gain value given by a VCOCAL) but that seems to cause the control loop problems, so I tried a much larger value and it stabilised, however it has done so at an offset from 0 - it is very peculiar.

I wrote up what I have done on the uBlox forums https://portal.u-blox.com/s/question/0D5Oj00000sxppHKAQ/m8f-smgr-not-steering-correctly?t=1747019984559 but I was wondering if anyone on this list has had experience and could shed some light on what I might be doing wrong.

Thanks.

PS I checked the list archives and found a few mentions but nothing concrete.

--
Daniel O'Connor
"The nice thing about standards is that there
are so many of them to choose from."
-- Andrew Tanenbaum

Hi, I’m trying to use an M8F to make a 10MHz GPSDO - I have designed a PCB with it, a VCO (AOCJY1-10.000MHZ-E) to EXTINT0 controlled by a Ti DAC8571 and divided by 10e6 (using a Renesas GreenPak) going to EXTINT1. ie the “external (coherent) PPS” design from the documentation. I tried a naive configuration (using the gain value given by a VCOCAL) but that seems to cause the control loop problems, so I tried a much larger value and it stabilised, however it has done so at an offset from 0 - it is very peculiar. I wrote up what I have done on the uBlox forums https://portal.u-blox.com/s/question/0D5Oj00000sxppHKAQ/m8f-smgr-not-steering-correctly?t=1747019984559 but I was wondering if anyone on this list has had experience and could shed some light on what I might be doing wrong. Thanks. PS I checked the list archives and found a few mentions but nothing concrete. -- Daniel O'Connor "The nice thing about standards is that there are so many of them to choose from." -- Andrew Tanenbaum

JL

Jim Lux

Tue, May 20, 2025 6:22 PM

Offset from zero for the VCO control voltage?
Or fixed frequency offset between GPS ticks and 1pps oscillator ticks?
Or fixed time offset between GPS ticks and 1pps oscillator ticks?

On Sun, 18 May 2025 10:13:19 +0930, "Daniel O'Connor via time-nuts" time-nuts@lists.febo.com wrote:

Hi,
I’m trying to use an M8F to make a 10MHz GPSDO - I have designed a PCB with it, a VCO (AOCJY1-10.000MHZ-E) to EXTINT0 controlled by a Ti DAC8571 and divided by 10e6 (using a Renesas GreenPak) going to EXTINT1. ie the “external (coherent) PPS” design from the documentation.

I tried a naive configuration (using the gain value given by a VCOCAL) but that seems to cause the control loop problems, so I tried a much larger value and it stabilised, however it has done so at an offset from 0 - it is very peculiar.

I wrote up what I have done on the uBlox forums https://portal.u-blox.com/s/question/0D5Oj00000sxppHKAQ/m8f-smgr-not-steering-correctly?t=1747019984559 but I was wondering if anyone on this list has had experience and could shed some light on what I might be doing wrong.

Thanks.

PS I checked the list archives and found a few mentions but nothing concrete.

--
Daniel O'Connor
"The nice thing about standards is that there
are so many of them to choose from."
-- Andrew Tanenbaum

time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe send an email to time-nuts-leave@lists.febo.com

Offset from zero for the VCO control voltage? Or fixed frequency offset between GPS ticks and 1pps oscillator ticks? Or fixed time offset between GPS ticks and 1pps oscillator ticks? On Sun, 18 May 2025 10:13:19 +0930, "Daniel O'Connor via time-nuts" <time-nuts@lists.febo.com> wrote: Hi, I’m trying to use an M8F to make a 10MHz GPSDO - I have designed a PCB with it, a VCO (AOCJY1-10.000MHZ-E) to EXTINT0 controlled by a Ti DAC8571 and divided by 10e6 (using a Renesas GreenPak) going to EXTINT1. ie the “external (coherent) PPS” design from the documentation. I tried a naive configuration (using the gain value given by a VCOCAL) but that seems to cause the control loop problems, so I tried a much larger value and it stabilised, however it has done so at an offset from 0 - it is very peculiar. I wrote up what I have done on the uBlox forums https://portal.u-blox.com/s/question/0D5Oj00000sxppHKAQ/m8f-smgr-not-steering-correctly?t=1747019984559 but I was wondering if anyone on this list has had experience and could shed some light on what I might be doing wrong. Thanks. PS I checked the list archives and found a few mentions but nothing concrete. -- Daniel O'Connor "The nice thing about standards is that there are so many of them to choose from." -- Andrew Tanenbaum _______________________________________________ time-nuts mailing list -- time-nuts@lists.febo.com To unsubscribe send an email to time-nuts-leave@lists.febo.com

EM

Ed Marciniak

Wed, May 21, 2025 6:08 PM

Does your source type need to be feedback ext oscillator instead of frequency source?

(Are you effectively run open loop with the DAC not using the correct source to calculate feedback)?

From: Jim Lux via time-nuts time-nuts@lists.febo.com
Sent: Tuesday, May 20, 2025 1:22:47 PM
To: time-nuts@lists.febo.com time-nuts@lists.febo.com
Cc: Jim Lux jim@luxfamily.com
Subject: [time-nuts] Re: uBlox M8F as GPSDO

Offset from zero for the VCO control voltage?
Or fixed frequency offset between GPS ticks and 1pps oscillator ticks?
Or fixed time offset between GPS ticks and 1pps oscillator ticks?

On Sun, 18 May 2025 10:13:19 +0930, "Daniel O'Connor via time-nuts" time-nuts@lists.febo.com wrote:

Hi,
I’m trying to use an M8F to make a 10MHz GPSDO - I have designed a PCB with it, a VCO (AOCJY1-10.000MHZ-E) to EXTINT0 controlled by a Ti DAC8571 and divided by 10e6 (using a Renesas GreenPak) going to EXTINT1. ie the “external (coherent) PPS” design from the documentation.

I tried a naive configuration (using the gain value given by a VCOCAL) but that seems to cause the control loop problems, so I tried a much larger value and it stabilised, however it has done so at an offset from 0 - it is very peculiar.

I wrote up what I have done on the uBlox forums https://urldefense.proofpoint.com/v2/url?u=https-3A__portal.u-2Dblox.com_s_question_0D5Oj00000sxppHKAQ_m8f-2Dsmgr-2Dnot-2Dsteering-2Dcorrectly-3Ft-3D1747019984559&d=DwIGaQ&c=euGZstcaTDllvimEN8b7jXrwqOf-v5A_CdpgnVfiiMM&r=JsDsKeR7cZC8wbZhIlxxBQ&m=cBnmEKXUWqQhVLArpif2u4bmZvoaieEzHah45TqBA22M6Zbe3erEl72_x0V8EOs3&s=uIgvoJxd9cmKabbxiCtGcylC26Exv6PqQ8hUnyCBS5M&e= but I was wondering if anyone on this list has had experience and could shed some light on what I might be doing wrong.

Thanks.

PS I checked the list archives and found a few mentions but nothing concrete.

--
Daniel O'Connor
"The nice thing about standards is that there
are so many of them to choose from."
-- Andrew Tanenbaum

time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe send an email to time-nuts-leave@lists.febo.com

Does your source type need to be feedback ext oscillator instead of frequency source? (Are you effectively run open loop with the DAC not using the correct source to calculate feedback)? ________________________________ From: Jim Lux via time-nuts <time-nuts@lists.febo.com> Sent: Tuesday, May 20, 2025 1:22:47 PM To: time-nuts@lists.febo.com <time-nuts@lists.febo.com> Cc: Jim Lux <jim@luxfamily.com> Subject: [time-nuts] Re: uBlox M8F as GPSDO Offset from zero for the VCO control voltage? Or fixed frequency offset between GPS ticks and 1pps oscillator ticks? Or fixed time offset between GPS ticks and 1pps oscillator ticks? On Sun, 18 May 2025 10:13:19 +0930, "Daniel O'Connor via time-nuts" <time-nuts@lists.febo.com> wrote: Hi, I’m trying to use an M8F to make a 10MHz GPSDO - I have designed a PCB with it, a VCO (AOCJY1-10.000MHZ-E) to EXTINT0 controlled by a Ti DAC8571 and divided by 10e6 (using a Renesas GreenPak) going to EXTINT1. ie the “external (coherent) PPS” design from the documentation. I tried a naive configuration (using the gain value given by a VCOCAL) but that seems to cause the control loop problems, so I tried a much larger value and it stabilised, however it has done so at an offset from 0 - it is very peculiar. I wrote up what I have done on the uBlox forums https://urldefense.proofpoint.com/v2/url?u=https-3A__portal.u-2Dblox.com_s_question_0D5Oj00000sxppHKAQ_m8f-2Dsmgr-2Dnot-2Dsteering-2Dcorrectly-3Ft-3D1747019984559&d=DwIGaQ&c=euGZstcaTDllvimEN8b7jXrwqOf-v5A_CdpgnVfiiMM&r=JsDsKeR7cZC8wbZhIlxxBQ&m=cBnmEKXUWqQhVLArpif2u4bmZvoaieEzHah45TqBA22M6Zbe3erEl72_x0V8EOs3&s=uIgvoJxd9cmKabbxiCtGcylC26Exv6PqQ8hUnyCBS5M&e= but I was wondering if anyone on this list has had experience and could shed some light on what I might be doing wrong. Thanks. PS I checked the list archives and found a few mentions but nothing concrete. -- Daniel O'Connor "The nice thing about standards is that there are so many of them to choose from." -- Andrew Tanenbaum _______________________________________________ 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

DO

Daniel O'Connor

Thu, May 22, 2025 11:20 PM

On 21 May 2025, at 03:52, Jim Lux via time-nuts time-nuts@lists.febo.com wrote:
Offset from zero for the VCO control voltage?

There shouldn't be any (the DAC output goes to the VCO input via an RC filter) but even if there were.. The feedback mechanism should cancel it out.

Or fixed frequency offset between GPS ticks and 1pps oscillator ticks?

To simplify things for now, I have disable EXTINT1 in SMGR so it should only be looking at the 10MHz frequency/phase.

Or fixed time offset between GPS ticks and 1pps oscillator ticks?

Well, there is some intrinsic delay due to logic gate propagation but again, feedback should cancel it out.

In any case a fixed offset would not be an issue for my application.

The current state of affairs is that the system appears stable, but it appears to be stable with a non-zero error reported..

Ie SMEAS shows:
UBX packet len 108 TIM:SMEAS: iTOW: 429581000 ms
Id 0 Internal freqOffset -0.0000012 / 0.000 ppb phaseOffset +0.0000005 / 2.000 ns
Id 1 GNSS freqOffset -0.0000006 / 0.000 ppb phaseOffset +0.9999981 / 2.000 ns
Id 2 Ext0 freqOffset +0.0227220 / 0.000 ppb phaseOffset +107487531.0000006 / 4.000 ns
Id 3 Ext1 freqOffset +42.9496601 / 0.000 ppb phaseOffset -460219108.0000004 / 6.000 ns

Ext0 has a 0.0227 PPB frequency offset which the M8F does not seem to care about.
There is still room in the adjustment range (current DAC value is 50070).

--
Daniel O'Connor
"The nice thing about standards is that there
are so many of them to choose from."
-- Andrew Tanenbaum

> On 21 May 2025, at 03:52, Jim Lux via time-nuts <time-nuts@lists.febo.com> wrote: > Offset from zero for the VCO control voltage? There shouldn't be any (the DAC output goes to the VCO input via an RC filter) but even if there were.. The feedback mechanism should cancel it out. > Or fixed frequency offset between GPS ticks and 1pps oscillator ticks? To simplify things for now, I have disable EXTINT1 in SMGR so it should only be looking at the 10MHz frequency/phase. > Or fixed time offset between GPS ticks and 1pps oscillator ticks? Well, there is some intrinsic delay due to logic gate propagation but again, feedback should cancel it out. In any case a fixed offset would not be an issue for my application. The current state of affairs is that the system appears stable, but it appears to be stable with a non-zero error reported.. Ie SMEAS shows: UBX packet len 108 TIM:SMEAS: iTOW: 429581000 ms Id 0 Internal freqOffset -0.0000012 / 0.000 ppb phaseOffset +0.0000005 / 2.000 ns Id 1 GNSS freqOffset -0.0000006 / 0.000 ppb phaseOffset +0.9999981 / 2.000 ns Id 2 Ext0 freqOffset +0.0227220 / 0.000 ppb phaseOffset +107487531.0000006 / 4.000 ns Id 3 Ext1 freqOffset +42.9496601 / 0.000 ppb phaseOffset -460219108.0000004 / 6.000 ns Ext0 has a 0.0227 PPB frequency offset which the M8F does not seem to care about. There is still room in the adjustment range (current DAC value is 50070). -- Daniel O'Connor "The nice thing about standards is that there are so many of them to choose from." -- Andrew Tanenbaum

DO

Daniel O'Connor

Fri, May 23, 2025 3:40 AM

On 22 May 2025, at 03:38, Ed Marciniak via time-nuts time-nuts@lists.febo.com wrote:
Does your source type need to be feedback ext oscillator instead of frequency source?

(Are you effectively run open loop with the DAC not using the correct source to calculate feedback)?

I used CFG-ESRC to configure EXTINT0 as a frequency source and EXTINT1 as a feedback source.

I tried running it with EXTINT1 disabled in SMGR in case my divide by 10e6 circuit was broken but it still does not steer correctly.

I’ve opened a case with ublox but I was hoping someone on this list might have prior experience :)

Thanks.

--
Daniel O'Connor
"The nice thing about standards is that there
are so many of them to choose from."
-- Andrew Tanenbaum

> On 22 May 2025, at 03:38, Ed Marciniak via time-nuts <time-nuts@lists.febo.com> wrote: > Does your source type need to be feedback ext oscillator instead of frequency source? > > (Are you effectively run open loop with the DAC not using the correct source to calculate feedback)? I used CFG-ESRC to configure EXTINT0 as a frequency source and EXTINT1 as a feedback source. I tried running it with EXTINT1 disabled in SMGR in case my divide by 10e6 circuit was broken but it still does not steer correctly. I’ve opened a case with ublox but I was hoping someone on this list might have prior experience :) Thanks. -- Daniel O'Connor "The nice thing about standards is that there are so many of them to choose from." -- Andrew Tanenbaum

AB

Azelio Boriani

Fri, May 23, 2025 12:20 PM

Why running with EXTINT1 disabled? Disable EXTINT0 instead and run only
with the PPS feedback, this is shown in the manual (page 83 of the 2016 R15
version).

On Fri, May 23, 2025 at 1:59 PM Daniel O'Connor via time-nuts <
time-nuts@lists.febo.com> wrote:

On 22 May 2025, at 03:38, Ed Marciniak via time-nuts <

time-nuts@lists.febo.com> wrote:

Does your source type need to be feedback ext oscillator instead of

frequency source?

(Are you effectively run open loop with the DAC not using the correct

source to calculate feedback)?

I used CFG-ESRC to configure EXTINT0 as a frequency source and EXTINT1 as
a feedback source.

I tried running it with EXTINT1 disabled in SMGR in case my divide by 10e6
circuit was broken but it still does not steer correctly.

I’ve opened a case with ublox but I was hoping someone on this list might
have prior experience :)

Thanks.

--
Daniel O'Connor
"The nice thing about standards is that there
are so many of them to choose from."
-- Andrew Tanenbaum

time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe send an email to time-nuts-leave@lists.febo.com

Why running with EXTINT1 disabled? Disable EXTINT0 instead and run only with the PPS feedback, this is shown in the manual (page 83 of the 2016 R15 version). On Fri, May 23, 2025 at 1:59 PM Daniel O'Connor via time-nuts < time-nuts@lists.febo.com> wrote: > > > > On 22 May 2025, at 03:38, Ed Marciniak via time-nuts < > time-nuts@lists.febo.com> wrote: > > Does your source type need to be feedback ext oscillator instead of > frequency source? > > > > (Are you effectively run open loop with the DAC not using the correct > source to calculate feedback)? > > I used CFG-ESRC to configure EXTINT0 as a frequency source and EXTINT1 as > a feedback source. > > I tried running it with EXTINT1 disabled in SMGR in case my divide by 10e6 > circuit was broken but it still does not steer correctly. > > I’ve opened a case with ublox but I was hoping someone on this list might > have prior experience :) > > Thanks. > > -- > Daniel O'Connor > "The nice thing about standards is that there > are so many of them to choose from." > -- Andrew Tanenbaum > _______________________________________________ > time-nuts mailing list -- time-nuts@lists.febo.com > To unsubscribe send an email to time-nuts-leave@lists.febo.com

BC

Bob Camp

Fri, May 23, 2025 1:09 PM

Hi

The DAC / VCXO combination on the M8F can only just so much. The EFC range on the VCXO is large
and the DAC has limited steps. You will never get to “zero" frequency error with it. Indeed, due to lags, you
will never get to zero error with any GPSDO if you look closely enough.

Bob

On May 22, 2025, at 11:40 PM, Daniel O'Connor via time-nuts time-nuts@lists.febo.com wrote:

On 22 May 2025, at 03:38, Ed Marciniak via time-nuts time-nuts@lists.febo.com wrote:
Does your source type need to be feedback ext oscillator instead of frequency source?

(Are you effectively run open loop with the DAC not using the correct source to calculate feedback)?

I used CFG-ESRC to configure EXTINT0 as a frequency source and EXTINT1 as a feedback source.

I tried running it with EXTINT1 disabled in SMGR in case my divide by 10e6 circuit was broken but it still does not steer correctly.

I’ve opened a case with ublox but I was hoping someone on this list might have prior experience :)

Thanks.

--
Daniel O'Connor
"The nice thing about standards is that there
are so many of them to choose from."
-- Andrew Tanenbaum

time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe send an email to time-nuts-leave@lists.febo.com

Hi The DAC / VCXO combination on the M8F can only just so much. The EFC range on the VCXO is large and the DAC has limited steps. You will never get to “zero" frequency error with it. Indeed, due to lags, you will never get to zero error with any GPSDO if you look closely enough. Bob > On May 22, 2025, at 11:40 PM, Daniel O'Connor via time-nuts <time-nuts@lists.febo.com> wrote: > > > >> On 22 May 2025, at 03:38, Ed Marciniak via time-nuts <time-nuts@lists.febo.com> wrote: >> Does your source type need to be feedback ext oscillator instead of frequency source? >> >> (Are you effectively run open loop with the DAC not using the correct source to calculate feedback)? > > I used CFG-ESRC to configure EXTINT0 as a frequency source and EXTINT1 as a feedback source. > > I tried running it with EXTINT1 disabled in SMGR in case my divide by 10e6 circuit was broken but it still does not steer correctly. > > I’ve opened a case with ublox but I was hoping someone on this list might have prior experience :) > > Thanks. > > -- > Daniel O'Connor > "The nice thing about standards is that there > are so many of them to choose from." > -- Andrew Tanenbaum > _______________________________________________ > time-nuts mailing list -- time-nuts@lists.febo.com > To unsubscribe send an email to time-nuts-leave@lists.febo.com

DO

Daniel O'Connor

Sun, May 25, 2025 8:45 AM

On 23 May 2025, at 21:50, Azelio Boriani via time-nuts time-nuts@lists.febo.com wrote:
Why running with EXTINT1 disabled? Disable EXTINT0 instead and run only
with the PPS feedback, this is shown in the manual (page 83 of the 2016 R15
version).

Mainly because the PPS is generated by a GreenPak which I have never used before so I was not confident I got it totally correct (eg wether I am dividing by 10e6 or 10e6 plus or minus one) so I went for the simple approach of just steering the 10MHz VCO to get started :)

--
Daniel O'Connor
"The nice thing about standards is that there
are so many of them to choose from."
-- Andrew Tanenbaum

> On 23 May 2025, at 21:50, Azelio Boriani via time-nuts <time-nuts@lists.febo.com> wrote: > Why running with EXTINT1 disabled? Disable EXTINT0 instead and run only > with the PPS feedback, this is shown in the manual (page 83 of the 2016 R15 > version). Mainly because the PPS is generated by a GreenPak which I have never used before so I was not confident I got it totally correct (eg wether I am dividing by 10e6 or 10e6 plus or minus one) so I went for the simple approach of just steering the 10MHz VCO to get started :) -- Daniel O'Connor "The nice thing about standards is that there are so many of them to choose from." -- Andrew Tanenbaum

DO

Daniel O'Connor

Sun, May 25, 2025 8:47 AM

On 23 May 2025, at 22:39, Bob Camp via time-nuts time-nuts@lists.febo.com wrote:
The DAC / VCXO combination on the M8F can only just so much. The EFC range on the VCXO is large
and the DAC has limited steps. You will never get to “zero" frequency error with it. Indeed, due to lags, you
will never get to zero error with any GPSDO if you look closely enough.

Well of course, but if I can manually trim it by sending HOC commands to zero it out then I think the firmware should be able to do it also.

It seems odd the device is not steering to an error of 0 but was stable at +44Hz - that suggests a misconfiguration to me, hence my original query asking if anyone had done such a thing before.

--
Daniel O'Connor
"The nice thing about standards is that there
are so many of them to choose from."
-- Andrew Tanenbaum

> On 23 May 2025, at 22:39, Bob Camp via time-nuts <time-nuts@lists.febo.com> wrote: > The DAC / VCXO combination on the M8F can only just so much. The EFC range on the VCXO is large > and the DAC has limited steps. You will never get to “zero" frequency error with it. Indeed, due to lags, you > will never get to zero error with any GPSDO if you look closely enough. Well of course, but if I can manually trim it by sending HOC commands to zero it out then I think the firmware should be able to do it also. It seems odd the device is not steering to an error of 0 but was stable at +44Hz - that suggests a misconfiguration to me, hence my original query asking if anyone had done such a thing before. -- Daniel O'Connor "The nice thing about standards is that there are so many of them to choose from." -- Andrew Tanenbaum