Upgrade, Killer NIC and panadapter flyback lines

[Neil]

Over the last 6 months I have replaced the micro atx motherboards in an AMD 8150 and an Intel i7 4770 with "better" ATX motherboards with
improved memory and system management along with the necessary new cases to fit the larger boards. Both new boards have KILLER nic installed which work, I know some users have problems getting these to connect properly to radios. As I have breathing problems I now work purely in datamodes. With both of these new systems on xmit there is a terrible mess on the panadapter, with and without PS, the mess is reminiscent of the flyback lines we could get on the old CRT tubes, right across the full width of the panadapter. Having checked with another sdr receiver this is also transmitted. I have been building and using computers since before Microsoft windows so I thought no problem, easy fix, Wrong. Checked for stray RF, checked for good grounding for both radio and PC, all ok. Tried best cables, very slight improvent, disabling Killer nic and installing bog standard gigabit card, better but nowhere near right. Eliminate all un-necessary connections to pc and radio, no real difference. Even went as far as using a long ethernet cable and setting radio up in another room away from the pc, no difference. Went back to rebuilding my old system and everything is good again. I also have a Lenovo z70 laptop, the laptop gives absolutely the best xmit signal you could ever wish to see/hear even without PS. Considering that the shielding on well made laptops is usually at least good this could be a clue.
Now having spent money and time on the new systems I wasn't going to not use them, although they weren't usable in their present state. The problem was identical on both new systems. Having eliminated the causes that I could think of, inadequate shielding on the new cases or the Killer nic adapters was seeming to be the likely cause. First thing wrap the pc with foil (temporarily) so as to cut down any emissions. No real difference. As I had already disabled the killer nic and tried a pci nic to no avail I was starting to resign myself to using the laptop or rebuilding the old systems. I then had my final idea, a last ditch attempt if you will, electrically isolate the radio from the pc and nic as much as possible. To do this I purchased a £10 usb3.0 hub with a built in gigabit nic. Works absolutely fine, no flybacks and a lovely xmit signal. Sometimes the solution is that simple.
I hope my experience and "cure" may be of value.
Best regards, Neil

[Tony EI7BMB]

Thanks for that info Neil, will come in handy for my next PC upgrade.

[w-u-2-o]

A couple of us have just completed some pretty high-end new PC builds (see here), both of them with Killer NICs on the motherboard. Neither of us experienced the same problems you have, Neil. You can see all the details of those builds in that thread. (Although, yes, the Killer NICs don't seem to talk to the radio, but the Intel NICs in the motherboards work fine. I plan to investigate this when I get some free time.)

I'd suggest it is something unique to the motherboard you are using, and not anything to do with the Killer NIC chipset.

What motherboard and case did you use specifically? So that we can all avoid buying them in the future!

It would also be interesting to see a screen shot of the display so that we can all understand what it looks like when this problem is occurring.

You might also check the latency performance of your new builds, as discussed in this topic.

73,

Scott

[Neil]

Hi Scott,
The post was simply info that may be of value to OM`s with similar connectivity problems. As you say that you have problems getting radios to work with the Killer nics that at least partly confirms my comments. However I was able to get the radio working with the Killer nics, just not cleanly. Bog standard nics were better but not what I would consider acceptable.
The Motherboards are MSI 990 FXA Gaming (AMD).
Asrock H97 Killer Series Gaming (Intel)
Cases are Corsair carbide 400c and Zalman Three Hundred Three. Only the corsair is new.
I am pretty confident that the cases are not the problem, both motherboards have been in both cases and the issue persisted, however when putting the old motherboards in the cases the problem was gone. I did mix and match on every conceivable permutation of mobo and case, old mobo`s fine, new mobo`s not fine. The only common factor in the poor xmit is the Killer nics unless there is something I have completely overlooked.
From working around computer hardware for around 40years one of the main things I have observed is that experiences are often dissimilar in similar circumstances. the usb nic works for me, that doesn't mean it will work for others. The only thing that has me scratching my head is why didn't the issue improve when the Killer nics were disabled and I used standard pci cards?.
The latency is not and never was a problem. latency figures compare favourably with other good figures that are posted on the forum.

Best regards, Neil.

[Neil]

Hi Scott,
My apologies for not fully replying to your last post regarding a screen shot, sorry I do not have one and I don't really want to strip down my pc again in order to reproduce the issue. At present I am using the i7 setup but next week I probably am going to rebuild the AMD into the corsair case, as the issue is not board specific I should then be able to give a screenshot. In the meantime a description would be that from the centre of the panadapter a vivid red, narrow line goes fully left and right of the entire panadapter. the lines appear to occur at each phase change of a datamode transmission.

73, Neil. M0WNF

[w-u-2-o]

No worries, Neil. I'll be interested to see one if you have the opportunity. Don't be rebuilding a PC just to do it!