©ASUSTEK COMPUTER INC. ALL RIGHTS RESERVED.

D_S_O_ · ‎03-20-2015

I have suspected this for quite a while, and today it came the confirmation.

R5E is extremely (coincidently with the word in its name) sensitive to electromagnetic interference. I had the voltage fluctuations in my mains overnight, as my UPS system reported, and I could not start the computer in the morning.

Please note that the power to the board was not lost. UPS response time to power outages is very short and I had no problems with other computers. In my opinion all the problems we are getting here can be associated with on board EMC issues.

The above specifically applies to the random problems. Maybe the rushed design did not consider filtering conducted interference and adequate shielding for the radiated one. Please note that signal filtering (i.e. ferrites, passive filters) will reduce performance. It is the natural consequence of its application.

I know for example that using Platinum certified Seasoninc PSUs give me far less problems than the Gold one, where tolerances are relaxed. Also, R5E computers I have put together with no wireless and Bluetooth enabled are performing O.K.

Have you people thought about it, or it may be just me?

Raja · ‎03-20-2015

There isn't enough info in your post to determine what the cause of the non-POST is. Would need some objective testing with the correct test equipment.

The boards are designed to cope perfectly with PSUs that operate within ATX pull-up and ripple spec. In fact they will tolerate out-of-spec for a while as well (depending on how severe the PSU fault is).

The VRM solution is very robust and the noise is very low. Just a quick shot here dealing with a circa 300W load transient from software (default switching freq as well):

Based upon the content of your post - do you have any measurement results on your side showing issues? It seems to be all based on conjecture..

It's possible your UPS and PSU don't play well together (causing rail pull-up issues around POST). The board is the least likely cause in this (assuming the system is 100% stable).

D_S_O_ · ‎03-20-2015

Raja, of course I do not have the measurements results to back up my claims. That is why there is the word "possible" in the title of my thread.

I know that the board was immunity tested to EN 55024:2010. However, passing the test means that within specific conditions R5E works as desired. Considering the variety of computer components available today and generally unpredictable electromagnetic environments, it is very much probable that the board responds with random errors in circumstances outside of its test environment(s). That is actually why there is a list of compatible components guaranteed to work with the board. I only have doubts that the EMC immunity testing (not just functionality) was conducted for each combination. It would be simply too expensive.

My first post actually clearly described what happened. To give you more "prove" I have experimented with switching the power to the UPS briefly off/on. Then I had shut down my system and started again. Guess what... every time I had a problem booting up. My system had to adjust to the "new" conditions and after a few tries returned to normal. This basically means that the unstable power causes the issue. No doubt about it.

O.K. - to satisfy you and others with measurements results I will measure the mains disturbances before the power supply and after. I have access to necessary tools. The problem is to find the time. I do not really know when I will be able to conduct the tests. Waiting for this I am available for the discussion.

Praz · ‎03-20-2015

D.S.O. wrote:
O.K. - to satisfy you and others with measurements results I will measure the mains disturbances before the power supply and after. I have access to necessary tools. The problem is to find the time. I do not really know when I will be able to conduct the tests. Waiting for this I am available for the discussion.

Hello

That you have access to the proper and necessary tools is good. I would like to see the results as well. Looking forward to your spectrum analyzer results. I will be particularly interested in the harmonic range and levels this interference is occurring at.

D_S_O_ · ‎03-20-2015

Praz wrote:
Hello

That you have access to the proper and necessary tools is good. I would like to see the results as well. Looking forward to your spectrum analyzer results. I will be particularly interested in the harmonic range and levels this interference is occurring at.

The only problem I have is to get suitable current probe as the one we normally use is to measure emissions on large power cabling. I will need to fiddle a bit with the board's supply cabling. This is the only thing which pushed me away from conducting the tests to date.

Praz · ‎03-20-2015

D.S.O. wrote:
The only problem I have is to get suitable current probe as the one we normally use is to measure emissions on large power cabling. I will need to fiddle a bit with the board's supply cabling. This is the only thing which pushed me away from conducting the tests to date.

Hello

There must be some confusion here and maybe it is the way I am reading your post. A current probe is not used with a spectrum analyzer so is not needed for proper troubleshooting of EMI emissions/interference. I'm not sure what testing you would be performing with a current probe but it will be irrelevant to the problem creating the issue you are claiming.

D_S_O_ · ‎03-22-2015

Praz wrote:
Hello

A current probe is not used with a spectrum analyzer so is not needed for proper troubleshooting of EMI emissions/interference. I'm not sure what testing you would be performing with a current probe but it will be irrelevant to the problem creating the issue you are claiming.

Hi Praz, I think that you are a bit misinformed here. Could you please research the subject a bit more. I am using current probes for diagnosing of conducted interference in cables for ages. And.. with spectrum analyser.

Raja · ‎03-20-2015

You need to test what the UPS and PSU combo do to pull up or what happens at the PSUs output rails. The fault lies in that combo (according to the OP).

If the pull-up on any rail ends up being affected, the board may not POST. The power on/off condition also requires seeing whats happening to the standby power output of the PSU with a scope. If any settings applied constitute overclocking, then removing power (even if you cannot see it happening) may be causing a cold power up which can be tougher to pass with overclocked settings.

Like I said, not enough info in any of these posts and too much assumption. The opening post does not mention overclocking, or which POST code is seen on the board Hex code reader. The rest is simply conjecture of EMI -which is wrong. The only thing that gives a hint to me that it is that another PSU does not have an issue (implied by the initial post). The bulk capacitance of PSUs can be different so a momentary power supply glitch may not manifest with the same "knee" on its output. As you can't test that, there is not much one can pin it on per se (and that is technically not board "EMI susceptibility" either).

What you can do : either use the PSU that works, or see if the system is unstable from cold AC power up at stock versus overclocked. If it is, it will need more tuning of associated parameters. The other PSU can also be tried directly to the wall and if it does not happen from cold AC power up there, then the issue is with the UPS and PSU combo.

D_S_O_ · ‎03-20-2015

Raja, It should not matter if I am overclocking or not. The issue is that the system works fine for months and booting becomes problematic only when there is a power outage. If this is not enough to associate the problem with the power outage, I do not know what is?

F.Y.I: - I am not overclocking
- Q-codes are: randomly bd or 87, and occasionally b4

About b4: - except keyboard and mouse on rear USB2s I have water cooling controlled from internal USB2 port.
- on USB3s are; active UPS/monitor hub/card reader/camera, and inactive HDD dock/printer/scanner

I agree that other components could play a role here, and more testing is required, but it is the main board which controls all.
I doubt that APC will replace my 3 months old UPS worth $2k just because the R5E based computer (one of 3 connected) is failing to boot for short period after power failure.

Anyway, the testing I am planning to conduct is to measure conducted emissions on the input and output of the PSU at the time of power failure.

Raja · ‎03-20-2015

Me too Praz. This isn't EMI..

Possible electyromagnetic interference