M500 PSU cable transmits high RF noise

Moderators: RokuDouglas, RokuKen, RokuRyan, RokuJamesL

M500 PSU cable transmits high RF noise

Postby Nandy » Thu Jan 19, 2006 3:57 pm

Hi for all,

There was a topic about grounding, etc.
http://www.rokulabs.com/forums/viewtopic.php?t=2475
I also found some other topics about hiss, background noise in external DAC-s and other mystic problems and I think my problem is also related.

Problem:
Roku M500 (EU/Asian) is connected to a Panasonic SA-XR55 (EU) via coax. Receiving acceptable FM signal with the receiver in tuner mode is not possible, the RFI is so high. There is a modulated noise over the whole FM range. Modulated noise pattern closely related to display activity.

Tests:
0. Situation as above.
1. Coax disconnected both end, M500 operates in Internet Radio mode - no change, noise still exists
2. Roku PSU disconnect - perfect FM radio signal
3. PSU re-connect, M500 set to standby - noise changing from modulated to static (no display "activity" just time and date)
4. Web interface config - standby screen set to off - noise static, a bit lower but still exists.

I made further tests to judge how the RF noise goes into the receiver, trough the antenna wire and/or trough the 230V also. I disconnected my receiver from the wall plug and made run it from an UPS. No connection between the Roku and Pana units. RF noise is the same.

Test setup for reproducing the problem:
Roku M500 switching PSU is at floor level. M500 is on a high shelf, so the 9V supply wire is full lenght vertical. SA-XR55 with factory FM single wire antenna (not shielded coax for aerial) connected. FM antenna and supply wire are within 0,5 feet. No connection between units. Receiver is in tuner mode.

I am interested if anyone can reproduce the problem maybe with the SA-XR55, maybe with other receivers also.

IMHO source of problem:
The Roku M500 digital circuits and especially the display drive generates current spikes to the 9V supply. The current spike runs trough the long and unshielded supply wire and changes that wire to a FM transmitter antenna.

Possible workarounds:
1. place far or use shielded cable for aerial. But I dont want aerial, I live close to the FM transmitter.
2. no standby - when not used, disconnected. Lost standby functionality (bye, nice clock).
3. Using different receiver - I tried an old Yamaha RX-V520 and behaved a lot better. Maybe its FM tuner has better AM rejection.
These are NOT solving the problem, the RF noise is still emitted from Roku supply wire.

IMHO re-engineering solutions (needs testing, but I dont want to loose garantee):
1. In the M500 a low-ESR 1000u/16V electrolytic puffer cap paralelled an FM-shunt ceramic capacitor connected directly to the display supply connectors. This way the puffer cap locally feeds the transient current need, so the spike runs out to the PSU wiring is reduced - transmitting power also reduced.
2. A small size, low series R, at least 2A inductor connected series to the caps so the digital circuits get the power trough it.
Analogue output stage needs a separate stab IC what get its feed from the other end of the inductor (directly from the PSU).
2a. If display drive has a specific base freq or the RF emission shows a freq peek than with one more inductor there can be a T-filter tuned to that freq.
3. There can be an RF shunt cap in the PSU also and a ferrite donut on the cable.
4. Several small digital units (like modems) use small PSU with long cabling but those PSU output is small voltage AC. The RF-shunt, rectifier and the puffer cap is in the modem housing. It might not be the good solution because of 50Hz pollution, but worth a try.

Best regards,
Nandy
Nandy
 
Posts: 12
Joined: Thu Jan 19, 2006 3:49 am

Conducted/Unintentional Radiation

Postby davwebb » Thu Jan 19, 2006 5:15 pm

The Roku could not have possibly passed EU compliance testing (or FCC testing) with this problem. I can test the M1000 unit on a LISN/EMC test rig, but my Roku has a US 110VAC/60 Hz adapter, so the results wouldn't really be valid for your different adapter. Does the unit exhibit the same problem with either RCA (analog) or optical hookups?
davwebb
 
Posts: 352
Joined: Wed Aug 10, 2005 6:47 am

Postby Nandy » Thu Jan 19, 2006 11:57 pm

> Does the unit exhibit the same problem
> with either RCA (analog) or optical hookups?

Yes, and even without any hookup. In the test setup I suggested the effect can be reproduced without ANY hookup.
Nandy
 
Posts: 12
Joined: Thu Jan 19, 2006 3:49 am

Postby Nandy » Fri Jan 20, 2006 1:52 am

> but my Roku has a US 110VAC/60 Hz adapter,
> so the results wouldn't really be valid for your different adapter.

Well, and also the M1000 and the M500 has different adapters and different displays but the measurement might be very interesting anyway.
On the weekend I try to implement a small, pluggable RF filter network to the M500 power connector right under the plastic cup.
Also try to dig up some other 9V PSU for further M500 testing.

Regards,
Nandy
Nandy
 
Posts: 12
Joined: Thu Jan 19, 2006 3:49 am

Postby stretch » Fri Jan 20, 2006 4:07 am

Have you tried using a generic linear power supply instead of the PSU supplied by Roku?

The standard PSU has a wide operating voltage which suggests that its a Switch Mode Power Supply (SMPS) and as I understand it they can be noisy (RF).

Trying a 9V linear PSU with the same current rating would confirm if the problem is in the SMPS or SB M500.

I've got an SB M1000 and could probably take it to work and get the electronics technicians to measure the emmitted RF noise.

On another thread regarding noise i think someone tried placing a ferrite inductor around the lead. Not sure if it helped though.
stretch
 
Posts: 623
Joined: Thu Dec 15, 2005 11:29 pm
Location: Australia

Postby SuzieQ » Fri Jan 20, 2006 4:18 am

If I were in your shoes I would try to:
- plug the M500 power supply in a different outlet and see what happens.
- shield the M500 power supply cable, for example, by wrapping it in aluminium foil.
SuzieQ
 
Posts: 420
Joined: Tue Nov 30, 2004 4:27 pm
Location: Canada

Postby Nandy » Fri Jan 20, 2006 6:51 am

Thanks for all,

> stretch: Have you tried using a generic linear power supply
I will get one on next week.

> stretch: someone tried placing a ferrite inductor around the lead
I went to the local electronic parts shop and bought some ferrite rings and the stuff I listed above (low ESR puffer caps, RF shunt caps, etc). I will build a pluggable filter and see how it works. And of course I will write my results.

> SuzieQ: plug the M500 power supply in a different outlet and see what happens
I did it and nothing changed. As I wrote in my initial letter I not only plugged to a different outlet but I made my receiver run from an UPS (battery). It means there was no coupling possible through the mains since the UPS input was in the air.

My main problem is that I can not state to myself that:
- is it rather a Roku or an SA-XR55 tuner problem,
- is it a general problem to all M500 unit/PSU or I have a faulty one,
- is it a general problem to all SA-XR55 tuners or I have a faulty one,
- are other Roku products (like M1000) behaves the same way or not,
- maybe the problem does not exist in the US... ...only in 230V areas.

Thanks again,
Nandy
Nandy
 
Posts: 12
Joined: Thu Jan 19, 2006 3:49 am

So far, I just can't see

Postby Nandy » Fri Jan 20, 2006 11:19 am

I built the pluggable RF shunt. I plugged series with Roku 9V cable and guess what???
No change! The ferrite inductor also did not help. It means it is not back-reflected impulse from the M500 unit.

This way the only possible source of the RF noise is the PSU itself.
So what this PSU does with an artificial load? I got a 12V/15W halogen bulb, connected to the 9V output and...
...yes...
...it is the source.
The bulb glows, the RF noise cames, the FM reception gone.
For testing purposes I will change the "Made in China" PSU what is the factory equipment of my 150$ unit to a non-switching one.

Needs to be tested.
Stay tuned.

The point goes to stretch.

Thanks,
Nandy
Nandy
 
Posts: 12
Joined: Thu Jan 19, 2006 3:49 am

Solution

Postby Nandy » Mon Jan 23, 2006 1:14 pm

I bought a new PSU, it is a 9V/3000mA unit, part number is MWS9125GS. I made the bulb test and this PSU does not disturbs the tuner.

I bought 4 metre of 75 ohm cable with good shielding for digital (Neotech KHS112C) and soldered high grade RCA connectors on it.

This new PSU solved my problem.

Regards,
Nandy
Nandy
 
Posts: 12
Joined: Thu Jan 19, 2006 3:49 am

Postby gcleveland » Fri Feb 03, 2006 12:54 pm

I have had RFI problems for the year and a half I've owned an SB. It picks up airborne RFI bursts from various sources (e.g., flash cameras, the spark ignition on my furnace) and transmits them downstream (tripping relays, causing loss of the phase lock loop in my DAC). I've never been able to solve it, and basically have given up. It's most definitely the SB, as I've done standard troubleshooting to isolate the offending device.

Here's my post about it from a while back, which died a quiet death, unsolved.

http://www.rokulabs.com/forums/viewtopi ... highlight=
gcleveland
 
Posts: 47
Joined: Wed Sep 08, 2004 7:26 am
Location: Ottawa, ON

Postby stretch » Fri Feb 03, 2006 4:36 pm

gcleveland,

in your last post you basically stated that it's a design / pcb layout fault. No wonder it died a quiet death.

If that were the case then why aren't more people seeing the same problem?

I suspect that it's installation specific or caused by a manufacturing fault in your SB (i.e. faulty component, dry joint, etc...).

Have you tried a linear PSU instead of the SMPS unit supplied with the SB?
stretch
 
Posts: 623
Joined: Thu Dec 15, 2005 11:29 pm
Location: Australia


Return to SoundBridge General Discussion

Who is online

Users browsing this forum: No registered users and 1 guest