So Disappointed But I Rolled it Back. 7000 MK III To Protocol 1

[KD1GA]

Owning my 7000 MKIII for a few months now I really had no choice but roll it back to protocol 1. Radio worked moderately well but was plagued by intermittent issues, I.e., buffer overruns that would cause receive audio issues, sequence errors
that also caused actual transmit audio issues, infrequent but frustrating Thetis crashes and intermittent comments on my bad transmit audio. These issue occurred with Mic and speakers direct to radio or utilizing Vac with numerous top of the line external sound cards. I built an brand new machine in anticipation of my new SDR, i9, Gigabyte MB, M.2 drives, second dedicated nic, 128 gigs of ram clean Windows install etc. The machine is fast, runs cool and rock solid running applications much more demanding than Thetis. Added external cooling all over the 7000, I tried every trick in the book from all the posts here and great help from everyone but still could not get this thing to the point that I trusted it. So with a tear in my eye I rolled it back to protocol 1 and it is rock sold. No issues buffer set at 64 and it just works flawlessly. I do notice the lower sampling that P1 is restricted too and bit slower response from the radio but that’s it and now I at least trust it will work.

I am 100 percent convinced that the radio is fine and all the issues I experienced are firmware coding related. I check here everyday now in anticipation of newer Protocol 2 version being released, My fingers and toes are crossed that the fact this radio is discontinued won’t deter interest from the gracious volunteers providing such great free work getting the timings corrected, but I also understand that this being open source it may never happen now. A sad day at the shack.

[w-u-2-o]

It seems like it would be worth asking Apache if they have an old/used/repaired Orion MK-whatever board kicking around that they could use to replace your lemon MKIII board with.

[KD1GA]

So you think it could actually be a hardware problem? I guess I was simply convinced it was a timing issue, but I guess slight differences in chip run could mean some are less tolerant than others. I'll ping Apache and see if they will help. Thanks as always for the great support you provide to the community, its really is truly appreciated.

[KD1GA]

Okay a plea to Apache Labs has been sent. Thanks again Vin KD1GA

[w-u-2-o]

So you think it could actually be a hardware problem? I guess I was simply convinced it was a timing issue, but I guess slight differences in chip run could mean some are less tolerant than others.

You are exactly right, it is a timing issue. For a long list of reasons that have been discussed on this forum many times, the timing solution, primarily in the GigE MAC logic, is not as perfect or robust as we all would prefer. Thus there is the occasional serial number hardware that simply won't close timing.

All honor still goes to Rick, N1GP, for getting it as good as it is. Early versions of P2 would not run properly on probably half the serial numbers out there!

[w3ub]

You should look into sending it to Doug for a warranty repair as needed. Apache should make sure the occasional P2 timing problem gets resolved. Most of us have only minor P2 issues, mainly at wide BWs, but not enough to really hinder operating.

Doug

[kc2rgw]

Curious if you tried protocol 2 but at the lower spectrum width? If there are issues in networking, the added data throughput exaggerates them for sure as it’s a lot more data on the line to process.

I had a lot of headaches a while back due to an antivirus that had a shim on the IP stack (I disabled it) and then I made a bunch of network optimizations like forcing speed and duplex and other driver setting changes to optimize for gigabit throughput. It’s been very stable for a long while now.

It can be a frustrating puzzle.

[KD1GA]

I would love to hear what NIC tweaks you applied that actually help, please share.. I have purchase three different NIC's messed with the settings until I was blue in the face and still had a very sketchy connection. Anyway its all a moot point now, Apache labs has got back to me and explained very clearly that Protocol 2 is beta and unsupported.

"Hello Vincent,

Thank you for your email,

If doug has already assisted you then he has covered all bases.

Protocol II firmware is beta and is NOT guaranteed to work, there is ongoing work happening to improve reliability, however, at this point I would suggest using the protocol1 firmware.

The updated firmware versions will be released once available.

Regards,

Tony
Apache Labs Support"

Guess I'm shit out of luck 3K later!

[n1gp]

Hi Vin,

Sorry to hear of your troubles getting P2 to work on your ANAN.

I take it you have a MkIII which is using the Cyclone V FPGA, 5CEFA

I wonder if you could try the below gateware, it is generally all of the
Protocol 2 with the exception of Gigabit Ethernet support. It runs the
same speed as P1, 100 Mb.

It would be interesting to see how it runs on your ANAN. It runs fine on my
Cyclone V based Orion2 board:

https://www.dropbox.com/scl/fi/xghgifzswfwio4zkf9j0g/Orion_MkIII_P2_CV_21.9_100T.rbf?rlkey=xmpm7v1ndjmslwxezjg01370a&dl=0

BTW, could you let me know which P2 gateware images you have tried,
there were initially (for 5CEFA) Orion_MkIII_CV_P2_v2.1.12A.rbf, Orion_MkIII_CV_P2_v2.1.12B.rbf, and
Orion_MkIII_CV_P2_v2.1.12C.rbf announced where Orion_MkIII_CV_P2_v2.1.12A.rbf seemed
to work for most people.

-Rick / N1GP

[kc2rgw]

I would love to hear what NIC tweaks you applied that actually help, please share..

I forced speed and duplex and enabled most of the settings in this tech article

https://learn.microsoft.com/en-us/windo ... uning-nics

I also set the NetworkThrottlingIndex registry key to ffffffff (needs a reboot)

Using regedit create a DWORD entry called NetworkThrottlingIndex under
Computer\HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows NT\CurrentVersion\Multimedia\SystemProfile

Set the hexadecimal value to ffffffff

[K1LSB]

If you invoke Thetis in Administrator mode, Thetis is capable of setting the NetworkThrottlingIndex tweak automatically. See Settings>General>H/W Select>Network Settings>Network Throttle Index Tweak.

Mark

[KD1GA]

Hi Vin,

Sorry to hear of your troubles getting P2 to work on your ANAN.

I take it you have a MkIII which is using the Cyclone V FPGA, 5CEFA

I wonder if you could try the below gateware, it is generally all of the
Protocol 2 with the exception of Gigabit Ethernet support. It runs the
same speed as P1, 100 Mb.

It would be interesting to see how it runs on your ANAN. It runs fine on my
Cyclone V based Orion2 board:

https://www.dropbox.com/scl/fi/xghgifzswfwio4zkf9j0g/Orion_MkIII_P2_CV_21.9_100T.rbf?rlkey=xmpm7v1ndjmslwxezjg01370a&dl=0

BTW, could you let me know which P2 gateware images you have tried,
there were initially (for 5CEFA) Orion_MkIII_CV_P2_v2.1.12A.rbf, Orion_MkIII_CV_P2_v2.1.12B.rbf, and
Orion_MkIII_CV_P2_v2.1.12C.rbf announced where Orion_MkIII_CV_P2_v2.1.12A.rbf seemed
to work for most people.

-Rick / N1GP

Hi Rick,

I'm out of town for the week but will certainly try this firmware when I return and advise, I will also try these earlier versions as well and will report back. I do believe I tried these earlier versions months ago, less the the firmware you provided with no luck, but ill try them again. I do have to say unless they are completely rock solid I'll be staying with Protocol 1at least until something new is introduced. 1 has been amazingly solid for the last week with no issues at all. Thanks everyone for the help, its truly appreciated.

[StefanoAlvaro]

Hi Vin,



It would be interesting to see how it runs on your ANAN. It runs fine on my
Cyclone V based Orion2 board:

https://www.dropbox.com/scl/fi/xghgifzswfwio4zkf9j0g/Orion_MkIII_P2_CV_21.9_100T.rbf?rlkey=xmpm7v1ndjmslwxezjg01370a&dl=0

BTW, could you let me know which P2 gateware images you have tried,
there were initially (for 5CEFA) Orion_MkIII_CV_P2_v2.1.12A.rbf, Orion_MkIII_CV_P2_v2.1.12B.rbf, and
Orion_MkIII_CV_P2_v2.1.12C.rbf announced where Orion_MkIII_CV_P2_v2.1.12A.rbf seemed
to work for most people.

-Rick / N1GP

Novelty? I also downloaded this "hybrid" fw and will try it calmly in the next few days. In the meantime I tried to further cool my 7000 (5CEFA) but the problem of glitches in TX (only in TX) remains the same. In rx, even after hours, no problem, so I would rule out it being something that depends on the heating.



Update: as promised I loaded the fw, creating a new setting from scratch to avoid compromising any functionality by loading a previously saved one.
-The anan is shifted in frequency, I solved it with calibration
-In SSB (I haven't tried other ways) there is no feedback from the puresignal (the little square with the correction numbers remains still)
-By keeping the PTT alone without speaking, the linear amplifier goes into high SWR protection. When transmitting tone/words it seems to work well and the signal has an acceptable cleanliness.

-Fixed the glitch problem, even when holding the tone for a long time and on all possible samples

Thanks RIK, I hope the feedback can be helpful

[n2gq]

When I was having a issue with my 7000dle mkIII and protocol II I went into the network card setting and reduced it down to 100mbps with a bandwidth of 768000. And I still run 768000 just to keep it tamed. I no longer have to run the nic card at limited 100mbps after the latest firmware. I hope something works out for you.

[StefanoAlvaro]

When I was having a issue with my 7000dle mkIII and protocol II I went into the network card setting and reduced it down to 100mbps with a bandwidth of 768000. And I still run 768000 just to keep it tamed. I no longer have to run the nic card at limited 100mbps after the latest firmware. I hope something works out for you.

Thanks, I didn't think it could be done. So you're saying to decrease the speed of the PC ethernet card from 1000/1000 to 100/100 I understand correctly?

[KD1GA]

When I was having a issue with my 7000dle mkIII and protocol II I went into the network card setting and reduced it down to 100mbps with a bandwidth of 768000. And I still run 768000 just to keep it tamed. I no longer have to run the nic card at limited 100mbps after the latest firmware. I hope something works out for you.

Hi Jay,

Thanks for the information. I'm curious what version are you referring to, as the latest? I'll give it a try.

[StefanoAlvaro]

I changed the speed of the ANAN dedicated ethernet card from “auto negotiation” to 100mbps.
With the original P2 fw it doesn't work as I imagined (mine is 2.1.12 a) but thanks to this I discovered something interesting.
Going back to “autonegotiation” or higher settings (2.5 and 10 GBps) the system becomes stable!
If I then turn it off and on again the seq errors return, but every time I make a change to the network settings, until I turn it off ,the system with the P2 is perfectly stable.
As I imagined it's not a cooling/hardwre problem, there's something in there that needs to be corrected but I'm not able to find out what it is because I'm not a computer expert, much less an electronics expert, but something tells me that the problem to be solved is nearby.

[w-u-2-o]

It might be time to try a different Ethernet card. They are cheap enough.

I had a NIC with a Qualcomm chipset that wouldn't work at all with our radios. I had another with an Intel chipset that might have been a counterfeit device, it worked but poorly.

[StefanoAlvaro]

It might be time to try a different Ethernet card. They are cheap enough.

I had a NIC with a Qualcomm chipset that wouldn't work at all with our radios. I had another with an Intel chipset that might have been a counterfeit device, it worked but poorly.

It doesn't depend on the network card, in the PC the motherboard (proart creator z790 asus) has two different eth. of 10 and 2.5 GBps, I also have an external one (also 2.5 GBps), the test has been done on all combinations. Even the PC is new, but the problem was also on the previous one (another card, and we would be at 4)
I'm sure that anyone who has the same problem as me and the same rtx as me (5cefa 2023) can repeat the same experience.
I feel the solution is close, but I lack the experience to find the right way, arrrrggghhhhhhhhhh

[KD1GA]

It might be time to try a different Ethernet card. They are cheap enough.

I had a NIC with a Qualcomm chipset that wouldn't work at all with our radios. I had another with an Intel chipset that might have been a counterfeit device, it worked but poorly.

Thanks for the suggestion. I have tried a few NIC's, built into the board an Asus ASUS XG-C100C and a TP Link TX-40, both fast high quality cards with no luck.

I did discover something that may have been my problem all along and hope to have time to test it this weekend. The SDR has been super rock sold for weeks running Protocol 1no issues what so ever, just perfect. I had just noticed a bit of RF getting back into the computer via a serial to USB cable I use to control my Palstar HF-Auto. Not enough to adversely effect Thetis running Protocol 1, but it was screwing with audio on other apps, intermittently causing some windows system sounds to play when transmitting over 800 watts. So the plan is to remove any connections from other radios and accessories that are connected to the computer, reflash with Protocol 2 and see what happens. My hope is my problem was RF messing with the delicate timings required for for solid 1 gig operation. fingers crossed, I'll share the outcome.

[KD1GA]

Okay folks so far so good. Reloaded Protocol 2, Metis_Orion_MkIII_P2_CV_21.7. Removed and installed latest Thetis and restored existing data base. So far rock solid, fingers crossed. I would like to apologize for being convinced it was a hardware problem and spreading information that so far appears to have been totally off base; as it seems to have been just a tiny bit of RF getting into the computer from a serial to USB cable used to control my HF-Auto.

Protocol 1 was not bothered by the RF, but the timings in the P2 firmware evidently didn't like it. If you have a 7000 III and are having issues running Protocol 2, try disconnecting anything that could introduce RF into the computer and reinstall everything, my guess is, it will help. Thanks Vin KD1GA

[w-u-2-o]

Great news

[KD1GA]

Nope all the issues have returned... Errr guess a G2 is in my future.

[KD1GA]

I found another source of RFI through and Mic switch that showed up on my audio even with Protocol 1. Removed the offending switch and ran XLR direct to mt sound card and all is fine with Protocol 1. I'll reflash back to 2 and advise, fingers crossed again.

I guess when the leaves come down I will have some antenna work to do as well .

[K1LSB]

Vin,

I don't think your problem has anything to do with your antenna, IMO it's more likely related to RF grounding in the shack.

My own antenna is a multiband dipole 40 feet directly above my shack, so all of my equipment is deep inside the near field of my antenna and is unavoidably being slammed with hundreds of watts of radiated power every time I transmit, but I have no problem at all with RF in the shack. Power being radiated from the antenna is not per se a problem, it's simply something that needs to be dealt with because it can't be eliminated or even reduced (at least not without just reducing the TX power, which IMO isn't an option). Inadequate RF grounding, bonding or choking are much more likely culprits.

Mark

[n2gq]

I had the same issue with rf. But my problem, the mic wiring was wrong. I use a Heil Gold line mic, and of course I reversed the mic + lead. But it made me re-do the shack ground a lot nicer. Finally fixed the mic issue.

[n2gq]

When I was having a issue with my 7000dle mkIII and protocol II I went into the network card setting and reduced it down to 100mbps with a bandwidth of 768000. And I still run 768000 just to keep it tamed. I no longer have to run the nic card at limited 100mbps after the latest firmware. I hope something works out for you.

Hi Jay,

Thanks for the information. I'm curious what version are you referring to, as the latest? I'll give it a try.

Thetis ver. 2.10.0 6/19/2023
Firmware 2.1.12
Protocol II (v3.9)

Sorry for the late reply.

[KD1GA]

I wanted to post a follow up to my Protocol 2 issues, so it wouldn't just be hanging out there. I believe I found the issue , so with my fingers crossed it appears to have been RF into the computer. I looks like some stray RF through an HDMI cable from a monitor close the amp output. I switched out the cable with a higher quality shielded HDMI cable and then reinstalled P2 firmware. So far so good, no issues and Thetis has been rock solid. With the older cable no issues with protocol 1 running 1kw. So to me it appears P2 is more sensitive to any stray RF than P1, I certainly don't understand why, but in my case it seems to be true. Vin KD1GA

[n2gq]

I wanted to post a follow up to my Protocol 2 issues, so it wouldn't just be hanging out there. I believe I found the issue , so with my fingers crossed it appears to have been RF into the computer. I looks like some stray RF through an HDMI cable from a monitor close the amp output. I switched out the cable with a higher quality shielded HDMI cable and then reinstalled P2 firmware. So far so good, no issues and Thetis has been rock solid. With the older cable no issues with protocol 1 running 1kw. So to me it appears P2 is more sensitive to any stray RF than P1, I certainly don't understand why, but in my case it seems to be true. Vin KD1GA

Cover those unused ports with termination caps as well.

[Chiron]

Hi Vin,



It would be interesting to see how it runs on your ANAN. It runs fine on my
Cyclone V based Orion2 board:

https://www.dropbox.com/scl/fi/xghgifzswfwio4zkf9j0g/Orion_MkIII_P2_CV_21.9_100T.rbf?rlkey=xmpm7v1ndjmslwxezjg01370a&dl=0

BTW, could you let me know which P2 gateware images you have tried,
there were initially (for 5CEFA) Orion_MkIII_CV_P2_v2.1.12A.rbf, Orion_MkIII_CV_P2_v2.1.12B.rbf, and
Orion_MkIII_CV_P2_v2.1.12C.rbf announced where Orion_MkIII_CV_P2_v2.1.12A.rbf seemed
to work for most people.

-Rick / N1GP

Novelty? I also downloaded this "hybrid" fw and will try it calmly in the next few days. In the meantime I tried to further cool my 7000 (5CEFA) but the problem of glitches in TX (only in TX) remains the same. In rx, even after hours, no problem, so I would rule out it being something that depends on the heating.



Update: as promised I loaded the fw, creating a new setting from scratch to avoid compromising any functionality by loading a previously saved one.
-The anan is shifted in frequency, I solved it with calibration
-In SSB (I haven't tried other ways) there is no feedback from the puresignal (the little square with the correction numbers remains still)
-By keeping the PTT alone without speaking, the linear amplifier goes into high SWR protection. When transmitting tone/words it seems to work well and the signal has an acceptable cleanliness.

-Fixed the glitch problem, even when holding the tone for a long time and on all possible samples

Thanks RIK, I hope the feedback can be helpful

I've noticed a frequency offset when using protocol 2 with a 7000DLE MK3. It is around -160Hz. When using protocol 1, the frequency is correct (within several Hz). I asume that with only changing between protocol 1 and 2 firmware, the frequency should be identical.
Do you also have the frequency offset only with protocol 2?