Abrupt change in freq & correction factor changes

A place for Apache HARDWARE discussions, from bare boards to the latest hardware. SOFTWARE discussions belong in the SOFTWARE sub-forums.
User avatar
KA9UVY
Posts: 41
Joined: Thu Mar 02, 2023 5:58 am

Abrupt change in freq & correction factor changes

Postby KA9UVY » Tue May 23, 2023 8:30 pm

I received my 7000 DLE MK 2/3 on Good Friday this year and have seen it jump frequency maybe 10 times or so since I have owned it. This happens randomly and maybe on band change? keeps catching me off guard.
I can't tell if this is a software or hardware issue but a quick frequency calibration routine fixes it but yields a different correction factor in Thetis software. The change is from about 0.99997676 to 1.00000586.
I guess , thinking about it it always happens on a band change, I change bands and find everyone is off a bit to even KC's then go and run the calibration, it resets to ,999 or vice versa?
Has anyone seen this issue? Any known cause? I haven't turned the rig off over 5-10 times since I got it but usually it is right on when cold... This is not a subtle change, just change bands and realize it has happened again.

Thanks, 73
Robert
KA9UVY
Anan 7000DLE- MKIII
Thetis 2.10.3.5 u2
FW 2.2.2A
Protocol 2 v 3.9

73, Robert
KA9UVY
User avatar
w-u-2-o
Posts: 5539
Joined: Fri Mar 10, 2017 1:47 pm

Re: Abrupt change in freq & correction factor changes

Postby w-u-2-o » Wed May 24, 2023 2:59 am

I don't think you're the first person to report this. It would be good if others who are experiencing the same problem on Cyclone V based systems would chime in so that some estimate can be made of how widespread the problem is.

If there really is a problem it's almost certainly either in the hardware or the firmware.

It's worth noting there is newer firmware that includes the QSK fix. It would be important to know if this firmware fixes your problem. See:

viewtopic.php?f=32&t=4588
User avatar
KA9UVY
Posts: 41
Joined: Thu Mar 02, 2023 5:58 am

Re: Abrupt change in freq & correction factor changes

Postby KA9UVY » Wed May 24, 2023 6:36 am

Thanks Scott, I was on the thread you referenced about the firmware and am using the beta firmware with the QSK fix... Before I installed it my radio was really acting up with audio glitches and sequence errors galore.. Didn't know anyone else was seeing the freq shift so it's good to know.
It can go a good while before it happens but have seen it do it more than once in an evening of band hopping.
73, Robert
KA9UVY
Anan 7000DLE- MKIII
Thetis 2.10.3.5 u2
FW 2.2.2A
Protocol 2 v 3.9

73, Robert
KA9UVY
rdwing
Posts: 92
Joined: Sat May 21, 2022 2:05 am

Re: Abrupt change in freq & correction factor changes

Postby rdwing » Mon May 29, 2023 10:01 am

Robert,
I'm one of the other users that has reported this issue and I'm sorry to hear it's affecting you. Sadly it continues to be a problem. I have one of the in-between 7000's from last year with the EP4CE Cyclone IV FPGA's.

The way this presents is; I'll make a transmission, and when I unkey I'll discover that sometime during the TX the radio abruptly shifted frequency by about ~241 Hz. If I go check WWV I'll notice that the carrier isn't centered where it should be, and the required correction factor is an order of magnitude larger.

It has happened to me less often on 40 meters, but occurs quite regularly on 20m and up, and nearly instantly on 6m, even into a dummy load. If go to 10m and do a 10w tuning carrier for 15 seconds, by the time I unkey the frequency will be off.

Here's a quick video from November last year when I was trying to solve this issue, and a few threads where it was discussed without conclusion.
viewtopic.php?f=9&t=4182&p=22311#p21184
viewtopic.php?f=9&t=4380&p=22328#p22288

Apache labs has gaslit me badly and played a lot of games with trying to get this resolved, and the issue is bad enough it causes me to not really use the anan that much and use my Flex instead, which has been rock solid.

The only suggestion/temporary workaround I have found is to power cycle the radio, but you must leave it in the off position for at least 10 seconds otherwise the frequency cal will still be a little strange.

Couple of other things I have noticed is that:
#1 The crystal has aged in a bit and continues to drift (not related to the frequency jump problem). When the 7000 first arrived the required correction factor was 1.00000278 (7/28/22), then 1.00000313 (8/17/22) but as of today (5/29/23) is it 1.00000393. No idea if this amount of drift is to spec or not, since I haven't bothered yet to figure out the units on the radio, but the data sheet for the XO shows ±100 ppb for temp stability, and ±40ppb per day/±3ppm life aging.

#2 The TCXO seems to be getting more unstable and possibly more sensitive to temperature. If I transmit for a bit the required correction factor drifts upwards towards 1.00000405.

Strange business indeed. Please let me know if you discover anything.
User avatar
w-u-2-o
Posts: 5539
Joined: Fri Mar 10, 2017 1:47 pm

Re: Abrupt change in freq & correction factor changes

Postby w-u-2-o » Mon May 29, 2023 12:02 pm

@rdwing

Reviewing your excellent summary, and going back to the links you referenced, it seems that this has only been reported by three people on this forum:

rdwing (yourself)
Robert
N2QQF

As described, the problem is noticeable enough that if it were endemic it would seem that there'd be a lot more reports. And, given there are a very small number of reports, this does tend to point the finger at hardware, but firmware is certainly not off the hook entirely. And even if firmware was the culprit problem, then that suggests a timing problem only on a tiny percentage of the hardware out there.

rdwing & Robert: would you both please confirm what version name/number of firmware you are running?
User avatar
w9ac
Posts: 290
Joined: Sun Apr 09, 2017 4:01 pm

Re: Abrupt change in freq & correction factor changes

Postby w9ac » Mon May 29, 2023 4:20 pm

It may be worth attaching an ultra-stable 10 MHz frequency reference, then observe any changes in frequency behavior.

Paul, W9AC
User avatar
w-u-2-o
Posts: 5539
Joined: Fri Mar 10, 2017 1:47 pm

Re: Abrupt change in freq & correction factor changes

Postby w-u-2-o » Mon May 29, 2023 6:03 pm

Good point, Paul. That would likely identify if it's a hardware problem or a firmware problem. However, it would require someone who is having a problem to spend money they shouldn't have to to identify the problem, and maybe solve it if it's hardware.
rdwing
Posts: 92
Joined: Sat May 21, 2022 2:05 am

Re: Abrupt change in freq & correction factor changes

Postby rdwing » Mon May 29, 2023 6:58 pm

FWIW awhile back I tested locking the 7000 to a 10 MHz output from the TinySA Ultra, and I wasn't able to reproduce the problem. The current firmware is Orion_MkIII_Protocol_2_v2.1.27G.rbf, but I believe I was having this problem since the day the radio arrived with the factory protocol 1 firmware, and definitely all protocol 2 versions. I can test that more in depth though.

Part of the problem with using an external frequency reference is that is it not trivial to find such a reference that has the same or lower phase noise specification as the inbuilt TCXO, even a GPSDO might not, at least not without spending at least $500.
User avatar
w-u-2-o
Posts: 5539
Joined: Fri Mar 10, 2017 1:47 pm

Re: Abrupt change in freq & correction factor changes

Postby w-u-2-o » Tue May 30, 2023 3:28 am

rdwing wrote:FWIW awhile back I tested locking the 7000 to a 10 MHz output from the TinySA Ultra, and I wasn't able to reproduce the problem.

That strongly suggests the problem is hardware related, not software.

The current firmware is Orion_MkIII_Protocol_2_v2.1.27G.rbf, but I believe I was having this problem since the day the radio arrived with the factory protocol 1 firmware, and definitely all protocol 2 versions. I can test that more in depth though.

So that means the problem seems limited to the EP4CE version boards. If the problem exists with both P1 and P2 firmware then, combined with your reported success with the external 10MHz reference, it's 99.99% certain it's a hardware problem, and you'd have a good case with Apache for replacing the Orion MK-whatever board.

Part of the problem with using an external frequency reference is that is it not trivial to find such a reference that has the same or lower phase noise specification as the inbuilt TCXO, even a GPSDO might not, at least not without spending at least $500.

Not true at all. It's completely trivial. And, even if it weren't, you'd never, ever notice a 5 or even 10dB reduction in phase noise performance in real world use, only on the test bench.

That said, you are worried about nothing. As long as you choose carefully you can easily, and inexpensively, obtain a GPSDO with phase noise performance that exceeds the onboard 10MHz TCXO.

The onboard TCXO is a Connor-Winfield M100F with the following advertised phase noise performance:

Code: Select all

SSB Phase Noise for Fo=12.8 MHz
 @ 10 Hz offset -90 dBc/Hz
 @ 100 Hz offset -120 dBc/Hz
 @ 1 KHz offset -140 dBc/Hz
 @ 10 KHz offset -150 dBc/Hz
 @ 100 KHz offset -150 dBc/Hz
 @ 1 MHz offset -152 dBc/Hz

If you shop carefully on eBay for a BG7TBL (real or derivative) model GPSDO, you'll find them with the OCXO prominently featured in a photo as part of the listing (and that photo is important because you really want to know what you are buying). There are plenty on eBay right now for less than $150 that sport the venerable CTS 196 series OCXO. That OCXO is an outstanding performer with phase noise performance that exceeds the Connor-Winfield part that is on the board. I've got a similar unit with a Russian Morion OCXO in it, again essentially identical spec's to the CTS OCXO, it works fabulously well, no problem with any mode, digi or analog.
User avatar
KA9UVY
Posts: 41
Joined: Thu Mar 02, 2023 5:58 am

Re: Abrupt change in freq & correction factor changes

Postby KA9UVY » Tue May 30, 2023 1:30 pm

Sorry Guy's not been on forum for a few days...
I've loaded Metis_Orion_MkIII__CV_P2_v2.1.12A.rbf into my radio.
Thetis reports it in (General/H/W Select) tab as:
Ver: 2.1.12
Protocol 2 (v3.9)
PLL Locked
I am so glad mine doesn't do it that often and really concerned to hear that some have acted much worse, (rdwing)
I had thought about an external 10 Mhz source also but wanted to try and see if it could be a software/firmware issue first before spending more money on my setup.

Thanks, 73
Robert
KA9UVY
Anan 7000DLE- MKIII
Thetis 2.10.3.5 u2
FW 2.2.2A
Protocol 2 v 3.9

73, Robert
KA9UVY
User avatar
w-u-2-o
Posts: 5539
Joined: Fri Mar 10, 2017 1:47 pm

Re: Abrupt change in freq & correction factor changes

Postby w-u-2-o » Tue May 30, 2023 1:40 pm

KA9UVY wrote:Sorry Guy's not been on forum for a few days...
I've loaded Metis_Orion_MkIII__CV_P2_v2.1.12A.rbf into my radio.

So that means you have a Cyclone V FPGA, while rdwing has an EP4CE FPGA.

Not sure what that means, just pointing it out.

Again, if a 10MHz external reference fixes the problem, then it sure seems like a hardware issue and not a firmware issue.
User avatar
KA9UVY
Posts: 41
Joined: Thu Mar 02, 2023 5:58 am

Re: Abrupt change in freq & correction factor changes

Postby KA9UVY » Tue May 30, 2023 2:01 pm

Good Morning Scott;
Yes I believe I have the newest Cyclone V FPGA in my radio, Haven't actually opened it up yet to look but specifically asked Apache support when it got here on Good Friday so I could decide on Firmware and that is what Tony told me was onboard.
I would like to employ an external 10 Mhz source but looking for one with two outputs so it can also lock a Transverter, (Eventually )hooked to my Anan at the same time. (Pun intended)
Looking at the China units you mentioned in this thread and they all seem to only have one 10 Mhz output so it still looks like the Leo Bodinar unit is where I might eventually have to go.
As stated before my Anan doesn't jump freq that often so I am not pressed for time to decide. Overall I am still pleased to have this radio, knew it was going to be a challenge for me going in, but felt like it had and has the potential to be my Dream Radio :)
73, Robert
KA9UVY
Anan 7000DLE- MKIII
Thetis 2.10.3.5 u2
FW 2.2.2A
Protocol 2 v 3.9

73, Robert
KA9UVY
K4IBC
Posts: 64
Joined: Sun Apr 09, 2017 7:31 pm

Re: Abrupt change in freq & correction factor changes

Postby K4IBC » Tue May 30, 2023 3:44 pm

Q5 has signal splitters. You can get it with or without built in reference.
Here is a simple splitter without the reference.
http://q5signal.com/index.php?route=pro ... duct_id=90
electronics-tech
Posts: 21
Joined: Sat Apr 08, 2023 11:02 pm

Re: Abrupt change in freq & correction factor changes

Postby electronics-tech » Tue May 30, 2023 8:02 pm

I was looking for a Thread like this because I have been puzzled for a bit now. My new Anan 7000 DLE is also doing the same. I received mine in late April 2023. I have not checked to see what FPGA is in it.

After calibrating to 10.0Mhz WWV. And watching the waterfall in WSJTX on 6M for about 15 mins..... It just wanders... up and down. Clearly visible.

Transmitting also really affects it. Tried to work a Kansas station on 6M FT-8. He copied me +8. After 1 more exchange I could even decode him he looked like a backslash in the waterfall.

I do have a reference 10.0 Mhz Oven Osc that is good to 12 decimal places so I am going to jack that into it and see if it still wanders.

Results in a bit.

Ronnie
VE3NLS
rdwing
Posts: 92
Joined: Sat May 21, 2022 2:05 am

Re: Abrupt change in freq & correction factor changes

Postby rdwing » Tue May 30, 2023 8:07 pm

electronics-tech wrote:I was looking for a Thread like this because I have been puzzled for a bit now. My new Anan 7000 DLE is also doing the same. I received mine in late April 2023. I have not checked to see what FPGA is in it.

After calibrating to 10.0Mhz WWV. And watching the waterfall in WSJTX on 6M for about 15 mins..... It just wanders... up and down. Clearly visible.

Transmitting also really affects it. Tried to work a Kansas station on 6M FT-8. He copied me +8. After 1 more exchange I could even decode him he looked like a backslash in the waterfall.

I do have a reference 10.0 Mhz Oven Osc that is good to 12 decimal places so I am going to jack that into it and see if it still wanders.

Results in a bit.

Ronnie
VE3NLS


Ronnie, I have a feeling your issue is with over/underflows in the VAC setup. Right click on the VAC1 button and see if the over/under counters are going up a lot.

This issue is with an abrupt frequency change independent of VAC not a slowly wandering one.
electronics-tech
Posts: 21
Joined: Sat Apr 08, 2023 11:02 pm

Re: Abrupt change in freq & correction factor changes

Postby electronics-tech » Tue May 30, 2023 8:18 pm

Correction. It is staying stable just receiving. VAC over and under are minimal (I think at 40 and 52.). Am I wrong there?

It has been receiving now for 30 mins undisturbed.

I transmitted. And once again, the waterfall has shifted. I went back to WWV and sure enough under calibration it changed the correction value again.

I am plugging in the reference to try it.

Ronnie
VE3NLS
rdwing
Posts: 92
Joined: Sat May 21, 2022 2:05 am

Re: Abrupt change in freq & correction factor changes

Postby rdwing » Tue May 30, 2023 8:21 pm

w-u-2-o wrote:If you shop carefully on eBay for a BG7TBL (real or derivative) model GPSDO, you'll find them with the OCXO prominently featured in a photo as part of the listing (and that photo is important because you really want to know what you are buying). There are plenty on eBay right now for less than $150 that sport the venerable CTS 196 series OCXO. That OCXO is an outstanding performer with phase noise performance that exceeds the Connor-Winfield part that is on the board. I've got a similar unit with a Russian Morion OCXO in it, again essentially identical spec's to the CTS OCXO, it works fabulously well, no problem with any mode, digi or analog.


Thanks for the info Scott, I will check into that. I hadn't yet taken too much time to really dive into the nitty gritty yet so I appreciate it.

What about transmit phase noise also? The EP4CE and later revisions include the addition of a
c. Inclusion of an Ultra low noise LDO for the master clock will considerably improve Rx/Tx phase noise


TX phase noise is a little more difficult to measure for me, but comparing the flex 6400m and the anan here, the flex was the winner in this department. At least in the amateur hour relative tests I can do here. This was a little surprising to me, because the ARRL tests show that the anan should be the winner. Wonder if that's possibly a clue.

electronics-tech wrote:Correction. It is staying stable just receiving. VAC over and under are minimal (I think at 40 and 52.). Am I wrong there?

It has been receiving now for 30 mins undisturbed.

I transmitted. And once again, the waterfall has shifted. I went back to WWV and sure enough under calibration it changed the correction value again.

I am plugging in the reference to try it.

Ronnie
VE3NLS


Hi Ronnie, ok, just wanted to check. That symptom lines up. Upon transmitting on certain bands, the frequency will jump so number of Hz. For me its ~240hz. If I power cycle the radio leaving it off for at least 10 seconds, it will go back to the original correct frequency.
electronics-tech
Posts: 21
Joined: Sat Apr 08, 2023 11:02 pm

Re: Abrupt change in freq & correction factor changes

Postby electronics-tech » Wed May 31, 2023 12:09 am

After using my 10 Mhz Reference Oscillator as the time reference for over an hour, I have come to the conclusion that I cannot reliably make this 2019 iMac Core i7 running Win10 Pro work stably at all with Thetis in BootCamp. It seems all over the place. The reference oscillator is up and down and never seems to settle. The VAC under and over are fine. Low Hits.

So, I retract my previous statements. I have a new Core i7 Lenovo box here. So I will start from scratch and at least eliminate this variable.

My advice so far is don't try to do this on an Apple Product in Bootcamp. I did it because of the 5K Retina Display which is really nice. Not so good so far actually getting it to work.

And I also apologize. Most of that is off topic.

Ronnie
VE3NLS
User avatar
w-u-2-o
Posts: 5539
Joined: Fri Mar 10, 2017 1:47 pm

Re: Abrupt change in freq & correction factor changes

Postby w-u-2-o » Wed May 31, 2023 11:56 am

rdwing wrote:
w-u-2-o wrote: What about transmit phase noise also? The EP4CE and later revisions include the addition of a
c. Inclusion of an Ultra low noise LDO for the master clock will considerably improve Rx/Tx phase noise

Because both receive and transmit share the same 122.88MHz sample clock, and because that sample clock is phase locked to the same 10MHz reference (old or new, internal or external), phase noise performance must, by definition, be fundamentally the same on both. However, timing jitter in the DAC will be different than in the ADC on receive, and perhaps within the FPGA DSP processing as well. Thus there will probably be some slight differences in phase noise performance on receive and transmit.

However, phase noise performance is already so many times better than is necessary for any conceivable use of the radio it's really a "don't care" specification, unless one is looking to "win" on the test bench.

Finally, improved phase noise performance merely by the addition of a low drop out (LDO) voltage regulator on the input to the on-board 10MHz TCXO is going to be extremely modest (just a few dB). These are not microwave radios, the highest frequency in the box is 122.88MHz. Here's a good paper on the subject: https://www.ti.com/lit/an/slwa066/slwa066.pdf?ts=1685533791798

If you haven't already read the GPSDO tutorial thread you might want to take a look: https://community.apache-labs.com/viewtopic.php?f=17&t=2499

Return to “Apache-specific Hardware Discussions”