ANAN-7000DLE MkII Ethernet failure - again

Can't talk to your radio? This is the place to ask!
User avatar
neoklis
Posts: 34
Joined: Thu Jan 17, 2019 11:51 am
Location: Souni, Limassol, Cyprus
Contact:

ANAN-7000DLE MkII Ethernet failure - again

Postby neoklis » Tue Dec 08, 2020 2:01 pm

Hi All

I am again experiencing failures with the Ethernet connection of my ANAN-7kDLE
MkII radio. In the morning when the rig is cold, an Ethernet connection with
the computer cannot be established. This I test using the "ethtool" in Linux
to display the properties of the Etherent port. The relevant part of the output
from ethtool is this:

Speed: Unknown!
Duplex: Unknown! (255)
Auto-negotiation: on

After about 2 min of warming up, I get this report:

Speed: 100Mb/s
Duplex: Full
Auto-negotiation: on

And then after a few more minutes this:

Speed: 1000Mb/s
Duplex: Full
Auto-negotiation: on

Only then can I use the radio from the client software.

This is the second time the Ethernet transceiver chip is failing. The first
time it was a complete failure and I had to dismantle the radio and send the
PCB to the States for repair under warranty. This time it seems to be a low
temperature related failure.

I would like to hear from other owners of the ANAN-7kDLE about their exerience
with this product, especially any failures of the Ethernet transceiver chip and of
any solutions to this problem, if any.

My thanks in advance.
dj1yr
Posts: 52
Joined: Mon Apr 10, 2017 2:32 am

Re: ANAN-7000DLE MkII Ethernet failure - again

Postby dj1yr » Tue Dec 08, 2020 3:33 pm

Hi,

have you ever used a high quality LAN cable?
I have had radio failures before, this was due to a simple LAN cable.

On a radio (7000DLE) I also had to replace the NET TRX chip, which had temperature problems like yours.
The cause was probably a non-soldered thermal pad under the chip.

Meanwhile I put on all my Orion MK2 boards heatsinks on the chips with thermal adhesive.

I can't say if this is a general problem, but it seems to happen.

best regards René
User avatar
w-u-2-o
Posts: 5539
Joined: Fri Mar 10, 2017 1:47 pm

Re: ANAN-7000DLE MkII Ethernet failure - again

Postby w-u-2-o » Tue Dec 08, 2020 4:42 pm

When you are running protocol 2 firmware (which it seems you are since you report GigE speeds) and a problem is solved by letting the hardware warm up, this is almost always indicative of the P2 firmware version you are running not making timing closure on your specific hardware (serial number).

Two things to try:

#1) go back to protocol 1. If P1 firmware works then now the problem is, with 99.9% certainty, a P2 firmware issue.

#2) try different versions of the the P2 firmware, all located at this link: viewtopic.php?f=32&t=3192

It is entirely possible you may have to go back as far as version 1.9 to find P2 firmware that works properly on your hardware.

Bottom line: it is way too early to be blaming the hardware for this problem.
User avatar
neoklis
Posts: 34
Joined: Thu Jan 17, 2019 11:51 am
Location: Souni, Limassol, Cyprus
Contact:

Re: ANAN-7000DLE MkII Ethernet failure - again

Postby neoklis » Tue Dec 08, 2020 4:48 pm

dj1yr wrote:Hi,

have you ever used a high quality LAN cable?

Well, I bought the best I could find and I have already tried a different one but no improvement.

dj1yr wrote:On a radio (7000DLE) I also had to replace the NET TRX chip, which had temperature problems like yours.

So it looks like a possible problem of the quality or suitability of the particular Ethernet chip.

dj1yr wrote:Meanwhile I put on all my Orion MK2 boards heat sinks on the chips with thermal adhesive.

My problem seems to be cold temperatures so its not to do with the heat sinks.

dj1yr wrote:I can't say if this is a general problem, but it seems to happen.

Well, I live in a country where its difficult to find parts and workshops to replace SMD chips, so I guess
I should hope that the situation doesn't get worse, I can wait for the warm up.

Thanks for your reply.
User avatar
neoklis
Posts: 34
Joined: Thu Jan 17, 2019 11:51 am
Location: Souni, Limassol, Cyprus
Contact:

Re: ANAN-7000DLE MkII Ethernet failure - again

Postby neoklis » Tue Dec 08, 2020 5:01 pm

w-u-2-o wrote:Bottom line: it is way too early to be blaming the hardware for this problem.

Thank you Scott, that sounds good to me! I will try other firmware once I am done with some health trouble.
User avatar
neoklis
Posts: 34
Joined: Thu Jan 17, 2019 11:51 am
Location: Souni, Limassol, Cyprus
Contact:

Re: ANAN-7000DLE MkII Ethernet failure - again

Postby neoklis » Wed Dec 09, 2020 2:19 am

w-u-2-o wrote:It is entirely possible you may have to go back as far as version 1.9 to find P2 firmware that works properly on your hardware.

I went back to v1.9 and the behavior of the rig changed, there is an alternation every 2-3 secs between Speed: 1000Mb/s and
Speed: Unknown! but after about 30 secs speed settles to 1000 Mb/s and my client works OK. So there is a definite improvement.

Thanks for the tip!
dj1yr
Posts: 52
Joined: Mon Apr 10, 2017 2:32 am

Re: ANAN-7000DLE MkII Ethernet failure - again

Postby dj1yr » Wed Dec 09, 2020 11:43 am

neoklis wrote:My problem seems to be cold temperatures so its not to do with the heat sinks.

I have done this to avoid new problems.


The change of the firmware brought a change of behaviour, but to speak of an improvement?

My radios are found quickly.
User avatar
w-u-2-o
Posts: 5539
Joined: Fri Mar 10, 2017 1:47 pm

Re: ANAN-7000DLE MkII Ethernet failure - again

Postby w-u-2-o » Wed Dec 09, 2020 12:22 pm

Did you try any of the other versions of the Rev. 2.0 firmware? I'm surprised you had to go all the way back to 1.9.
User avatar
neoklis
Posts: 34
Joined: Thu Jan 17, 2019 11:51 am
Location: Souni, Limassol, Cyprus
Contact:

Re: ANAN-7000DLE MkII Ethernet failure - again

Postby neoklis » Wed Dec 09, 2020 2:03 pm

dj1yr wrote:The change of the firmware brought a change of behaviour, but to speak of an improvement?
My radios are found quickly.

Well, mine was too, but this low temperature issue cropped up, and since I can't experiment with a new Ethernet tcvr chip,
I have to accept the best I can get with work rounds. Waiting 30 secs for the warm up is no big deal, since after several minutes
of operation it seems to take 3-4 hours for the rig to cool off enough to trigger this issue.

Right now with this Crown virus I don't feel inclined to do much more. I also need a heart catheterization (angiogram) which
be done on the 23rd, so the prospect of Xmas in the hospital is an additional negative factor! ;)
User avatar
neoklis
Posts: 34
Joined: Thu Jan 17, 2019 11:51 am
Location: Souni, Limassol, Cyprus
Contact:

Re: ANAN-7000DLE MkII Ethernet failure - again

Postby neoklis » Wed Dec 09, 2020 2:08 pm

w-u-2-o wrote:Did you try any of the other versions of the Rev. 2.0 firmware? I'm surprised you had to go all the way back to 1.9.

Actually no. I thought I had a better chance going all the way back to 1.9 to do a quick test. I am in the middle of a difficult phase in
the development of my home brew HPSDR client for Linux, so will try different versions of firmware once I untangle myself from this
effort.
User avatar
neoklis
Posts: 34
Joined: Thu Jan 17, 2019 11:51 am
Location: Souni, Limassol, Cyprus
Contact:

Re: ANAN-7000DLE MkII Ethernet failure - again

Postby neoklis » Thu Dec 10, 2020 7:49 am

neoklis wrote:I went back to v1.9 and the behavior of the rig changed, there is an alternation every 2-3 secs between
Speed: 1000Mb/s and Speed: Unknown! but after about 30 secs speed settles to 1000 Mb/s and my client works OK.
So there is a definite improvement.

Well, looks like I spoke too soon. Today I had the same problems again with v1.9 installed. It took about 8 min of warm up before I could have the full 1000Mb/s speed. So it seems its not a matter of firmware version.
User avatar
w-u-2-o
Posts: 5539
Joined: Fri Mar 10, 2017 1:47 pm

Re: ANAN-7000DLE MkII Ethernet failure - again

Postby w-u-2-o » Thu Dec 10, 2020 12:09 pm

You can't make that assessment without trying at least 6 versions of the firmware. Not joking. The problem with all of the P2 firmware is that timing closure on the MAC to PHY interface is not guaranteed over all temperature ranges and serial numbers. You clearly have an outlier. The symptoms are classic. The only question is how bad an outlier. And it's actually the warming of the FPGA that is making the difference not the PHY chip. So you could replace that PHY chip again and it would make no difference.

You can easily prove this with a can of cold spray. Get it working, then hit the FPGA with the spray and watch it stop.
K9RX
Posts: 414
Joined: Fri Apr 14, 2017 3:47 pm

Re: ANAN-7000DLE MkII Ethernet failure - again

Postby K9RX » Thu Dec 10, 2020 1:37 pm

I have to say as a design engineer this sounds like a "myth" ... timing issues based on what would essentially be a small difference in temperature. Sounds more like there is something that is not right in the interface and its protocol or the chip is not designed to do what is being asked of it - it can - and does - but its not designed to and thus is appears as if it is variable. In that case that is not what the chip designers planned or envisioned for its use.

Seriously - electronics design is not suppose to be black magic.

Gary Myers
www.ServoCAT.com
dj1yr
Posts: 52
Joined: Mon Apr 10, 2017 2:32 am

Re: ANAN-7000DLE MkII Ethernet failure - again

Postby dj1yr » Thu Dec 10, 2020 1:54 pm

Hi,

what WU2O writes I have experienced with my Angelia-board.
It was always claimed that it is due to the firmware, I have found out during my experiments that the power supply of the Anan 100D for Angelia +P2 is too weak, Angelia needs more current, the external voltage board brings too little current on the 3.3V rail, I have installed a Pico PSU and the problems were gone.
The FPGA needs 100-200mA more current for P2.
User avatar
neoklis
Posts: 34
Joined: Thu Jan 17, 2019 11:51 am
Location: Souni, Limassol, Cyprus
Contact:

Re: ANAN-7000DLE MkII Ethernet failure - again

Postby neoklis » Thu Dec 10, 2020 3:45 pm

OK Scott

Now that you mention the possibility of the FPGA playing up, it could explain a couple of of things that I have been wondering about:
In the very beginning of this behavior, I could hear relays click even though the client app could not connect to the rig. I should have
realized this cannot be caused by peculiar behavior of the Ethernet chip. This behavior though has only happened on 2-3 cold starts
so I forgot about it.

Another thing that bothered me was that once warmed up and used for a while, it would take at 2-3 hours after switch off for the problem
to appear again. I thought its rather strange that the Ethernet chip would take so long to cool. But the FPGA is much larger so the delay
is more likely.

Anyhow, I will try various P2 versions eventually to see what happens.
User avatar
w-u-2-o
Posts: 5539
Joined: Fri Mar 10, 2017 1:47 pm

Re: ANAN-7000DLE MkII Ethernet failure - again

Postby w-u-2-o » Thu Dec 10, 2020 9:33 pm

dj1yr wrote:Hi,

what WU2O writes I have experienced with my Angelia-board.
It was always claimed that it is due to the firmware, I have found out during my experiments that the power supply of the Anan 100D for Angelia +P2 is too weak, Angelia needs more current, the external voltage board brings too little current on the 3.3V rail, I have installed a Pico PSU and the problems were gone.
The FPGA needs 100-200mA more current for P2.
That is a Hermes and Angelia problem. It does not occur on the Orion or Orion MKII boards.
User avatar
w-u-2-o
Posts: 5539
Joined: Fri Mar 10, 2017 1:47 pm

Re: ANAN-7000DLE MkII Ethernet failure - again

Postby w-u-2-o » Thu Dec 10, 2020 9:37 pm

K9RX wrote:I have to say as a design engineer this sounds like a "myth" ... timing issues based on what would essentially be a small difference in temperature. Sounds more like there is something that is not right in the interface and its protocol or the chip is not designed to do what is being asked of it - it can - and does - but its not designed to and thus is appears as if it is variable. In that case that is not what the chip designers planned or envisioned for its use.

Seriously - electronics design is not suppose to be black magic.

Gary Myers
http://www.ServoCAT.com

Gary,

You are welcome to take a look at both the schematic and the firmware source code.

I can tell you that when I suggested that Rick look at changing the thermal model in the design environment things improved a great deal across the board. Not perfect, but a lot better.

The reality is someone needs to get a high speed logic analyzer and a thermal chamber and profile 10 or so boards over temperature. Then we'd know for sure. Unfortunately nobody is in a position to do that.

73,

Scott
dj1yr
Posts: 52
Joined: Mon Apr 10, 2017 2:32 am

Re: ANAN-7000DLE MkII Ethernet failure - again

Postby dj1yr » Fri Dec 11, 2020 4:17 am

w-u-2-o wrote:That is a Hermes and Angelia problem. It does not occur on the Orion or Orion MKII boards.


Hi,

I just wanted to say in relation to "Mystic", and I meant with Angelia not the exchange of the fuse, but the installed power supply in the 100D, which is too weak for P2 and the 3.3V rail is already at P1 to 99% capacity.

But back to the topic...
User avatar
neoklis
Posts: 34
Joined: Thu Jan 17, 2019 11:51 am
Location: Souni, Limassol, Cyprus
Contact:

Re: ANAN-7000DLE MkII Ethernet failure - again

Postby neoklis » Fri Dec 11, 2020 4:43 am

Hi

This morning all worked well right away. So more observation is needed for a final conclusion.
User avatar
neoklis
Posts: 34
Joined: Thu Jan 17, 2019 11:51 am
Location: Souni, Limassol, Cyprus
Contact:

Re: ANAN-7000DLE MkII Ethernet failure - again

Postby neoklis » Thu Dec 17, 2020 4:45 pm

w-u-2-o wrote:You can't make that assessment without trying at least 6 versions of the firmware. Not joking. The problem with all of the P2 firmware is that timing closure on the MAC to PHY interface is not guaranteed over all temperature ranges and serial numbers.

Well, I have gone through all the six available v2.0 firmware versions as well as v1.9 but I always have the same problem.

At cold switch on, ethtool lists a speed of 1000Mb/s but after 3-4 sec it lists the speed as Unknown! After 2-3 min, it lists the speed as
100Mb/s and after 8-10 min I get raise the speed to 100-0Mb/s using

Code: Select all

sudo ethtool -s enp4s0 speed 1000 duplex full.

Then all works well and it takes 2-3+ hours off power for the rig to cool off enough for the fault to reappear.

I guess I will just have to be patient t switch on!

Return to “Network Connections & Network Hardware”