User avatar
liz
Raspberry Pi Foundation Employee & Forum Moderator
Raspberry Pi Foundation Employee & Forum Moderator
Posts: 5212
Joined: Thu Jul 28, 2011 7:22 pm
Contact: Website

Re: Pete Lomas's take on the compliance issues

Thu Mar 29, 2012 6:51 pm

Pete wrote a blog post for element14 today. (I'm a bit peeved; I've not been successful in getting him to write one for us)

It's all about his take on the compliance stuff that we're dealing with at the moment. Pete, as most of you know, is our CTO and did the hardware design work for the production models of the Models A and B. The post's well worth a read - and if you want to discuss it here, we'd really like to hear what you have to say.
Director of Communications, Raspberry Pi

User avatar
Jongoleur
Posts: 1179
Joined: Thu Aug 11, 2011 12:47 pm
Location: O'erlooking the sea, and all those effin windfarms...

Re: Pete Lomas's take on the compliance issues

Thu Mar 29, 2012 7:05 pm

Good to read something from someone who actually knows what is going on.  Of course, they will do magnificent work to get the Pi into a compliant state, but Easter is in the way so I wouldn't realistically expect to hear of the final triumph of good over adversity until after we've scoffed all our Creme Eggs!

Good luck and crack on!!!

(btw Liz - have a look at the latest /. on the Pi, some confusion about "CE"  saynomore...)
I'm just a bouncer, splatterers do it with more force.....

User avatar
johnbeetem
Posts: 945
Joined: Mon Oct 17, 2011 11:18 pm
Location: The Mountains
Contact: Website

Re: Pete Lomas's take on the compliance issues

Thu Mar 29, 2012 7:17 pm

It's a great post.  It sounds like it's mostly HDMI that's the problem, and it can probably be fixed in software (I assume GPU software) by reconfiguring the HDMI drivers in the BCM2835.  This is the advantage and disadvantage of having so much programmable -- you can fix the problem when it occurs, but if you forget to configure one of the myriad parameters you get surprises.

Having a dodgy HDMI cable sounds familiar.  I spent a lot of time tracking down an HDMI cable issue which turned out to be because the cheap 1m Chinese cables I was using didn't bother to implement the individual ground shield lines for each HDMI differential pair: only the outer shell was grounded.  A high-quality cable that did implement the lines worked fine.

I don't know why Pete wrote it at element14 instead of here -- maybe he likes their trolls better? 

User avatar
grumpyoldgit
Posts: 1452
Joined: Thu Jan 05, 2012 12:20 pm

Re: Pete Lomas's take on the compliance issues

Thu Mar 29, 2012 7:23 pm

Amusingly, I think he is in discussion with the trolls who went there after being banned here! They have been invited to tell him what is wrong with his design. It could be interesting.

User avatar
liz
Raspberry Pi Foundation Employee & Forum Moderator
Raspberry Pi Foundation Employee & Forum Moderator
Posts: 5212
Joined: Thu Jul 28, 2011 7:22 pm
Contact: Website

Re: Pete Lomas's take on the compliance issues

Thu Mar 29, 2012 7:29 pm

About HDMI problems being familiar: we mentioned it to Steve Furber, Chris Curry and Sophie Wilson at the BBC Micro event on Sunday, and at the mention of the word "HDMI", all three of them rolled their eyeballs and said rude things about HDMI and their individual experiences of its tendency to spew EM. We feel we're in good company.
Director of Communications, Raspberry Pi

User avatar
Davespice
Forum Moderator
Forum Moderator
Posts: 1665
Joined: Fri Oct 14, 2011 8:06 pm
Location: The Netherlands
Contact: Twitter

Re: Pete Lomas's take on the compliance issues

Thu Mar 29, 2012 8:05 pm

I bought a HDMI cable for £9.99 the other day.  How would one know if the cable was dodgy? Would it just not work - or would it work but chuck out lots of EM that I would never know about? Sorry if this sounds like a stupid question.

User avatar
rurwin
Forum Moderator
Forum Moderator
Posts: 4257
Joined: Mon Jan 09, 2012 3:16 pm
Contact: Website

Re: Pete Lomas's take on the compliance issues

Thu Mar 29, 2012 8:24 pm

It would chuck out loads of EMI that you wont know about. But you wont need to know about it either -- it's much less than your router or your mobile phone and probably less than your microwave oven. The only time you need to be bothered is when it causes interference on your TV or your radio or your HiFi speaker cables, or when the Radio Ham next door complains.

Regor
Posts: 2
Joined: Fri Jan 13, 2012 3:56 pm

Re: Pete Lomas's take on the compliance issues

Thu Mar 29, 2012 8:35 pm

Quoting Pete:-


Taking the longer view, this needed to be done before we get these into schools so ultimately its all good and in the right direction.


This might have been missed by some, and if the supply situation had been less fraught would be worth a front page post all by itself.

User avatar
Tass
Posts: 535
Joined: Sat Jan 21, 2012 11:15 am

Re: Pete Lomas's take on the compliance issues

Thu Mar 29, 2012 9:32 pm

Agreed – lets remember what this is all about.

EDIT: Hmmmm YouTube link doesn't work...

rpiguy
Posts: 25
Joined: Mon Mar 26, 2012 10:01 pm

Re: Pete Lomas's take on the compliance issues

Thu Mar 29, 2012 10:18 pm

I have a ton of experience with software, I've designed lots of custom algorithms and data structures, I know quite a few programming languages, I've written nontrivial assembly language programs, and even though I've never done kernel hacking I could probably write a device driver if I had clear specs of the CPU, the device in question, and other system peripherals like IRQ and DMA.

I know Linux sysadmin stuff pretty well; I've set up an LVM-snapshot capable system, built Linux with buildroot and Openembedded, done compiling and cross-compiling and kernel compiling and kernel cross-compiling and writing /sbin/init for an initramfs and compiling in ARM chroots on x86-64 and compiling with VirtualBox and compiling on an ARM embedded board.

I think "developer" describes me a lot better than "end-user."

But, my knowledge of hardware is mostly theoretical.  I wouldn't begin to know how to fix EMI issues.  I don't even own a soldering iron.

All the discussions on these forums of GPIO pin headers, bypassing power control circuits, and the merits and drawbacks of surface-mount technology are interesting but I don't really feel equipped to contribute to those discussions or practically apply the information contributed by others.

I also have no idea what legal or financial risks I would take on by operating a non-FCC-approved device here in the USA. [1]

I think it's better for users like me that the Foundation, manufacturers and distributors iron out the regulatory issues, since I have neither the experience nor the resources to deal with them.

That being said, it's my understanding from the blog post, front page posts, and forum that the device itself is technically in good shape.  If I somehow got my hands on one of the 2000 uncertified Pi boards and plugged it in, treating it the same way I (as a software guy) would treat any other computer, I wouldn't destroy the nearest mains transformer, set my house on fire or kill everyone with a pacemaker in a 1000m radius.  I probably wouldn't even cause measurable static on analog (now digital) TV or FM radio unless the receiver was really close to the Pi.

So for this particular device, at this point in time, it seems the approvals are more of a "rubber stamp" that's only necessary from a social/legal perspective rather than a technical/physical one.

I'm not saying that the FCC isn't a necessary evil or that the government should do away with certifications; the next board designer who sees the success of the Pi and wants to duplicate it might be more ethically challenged than the RPi folks, and might need that regulatory hurdle to keep them from cutting corners on safety and interference to increase margins on a product with rock-bottom prices.

And everyone who's feeling the pain of the delivery slippage certainly has a right to feel disappointed and upset.  But pointing fingers at the Foundation, manufacturers or distributors is certainly unproductive.  I would say it's also unfair, since they took legal advice on the necessity of certification early on, and were told by experts it wouldn't be necessary for the initial release, and relied on that advice in their announcements, planning and scheduling.

If anyone's to blame, it's those unnamed "experts."  But keep in mind that it may not even be fair to blame them; when their opinion was asked, the interest in RPi was much smaller, there was much less media attention, and the overall demographic of the prospective user base was more hardware-oriented.

I think they're doing the best they can and I think the new approach they're taking to certification is the right one, even though it's definitely resulting in delays and upsetting users.

[1]  The chances of getting into hot water due to operating an uncertified device are probably pretty small for me personally, since I'm in a semi-rural area...electromagnetic noise diminishes by 1/r^2, there aren't too many people within an r small enough that they'd notice the interference, and they're probably not technically clued-in enough to know they CAN call the FCC on me, even if they were inclined to, which they probably aren't.  Not to mention whether the FCC would have the resources/priorities to harass a random dude playing with a circuit board in his house...seems like they would have bigger fish to fry...

Phoenix RasPi
Posts: 19
Joined: Thu Mar 29, 2012 12:42 am

Re: Pete Lomas's take on the compliance issues

Thu Mar 29, 2012 10:24 pm

rurwin said:


It would chuck out loads of EMI that you wont know about. But you wont need to know about it either -- it's much less than your router or your mobile phone and probably less than your microwave oven. The only time you need to be bothered is when it causes interference on your TV or your radio or your HiFi speaker cables, or when the Radio Ham next door complains.


Ummm, even with the crappiest-built cable, the RasPi spurious emanations are still going to be waaaay lower in power than those from a router (a couple of watts of effective radiated power - ERP), a mobile phone (about a watt of ERP), and certainly at least 4 ~ 5 orders of magnitude less than those from a microwave oven (upward of 1,000 watts of ERP in a much higher frequency range of GHz, kept overwhelmingly inside the cooking chamber, we hope).  The HDMI frequencies don't really overlap ham bands (although we haven't seen the splatter diagrams that Eben says they will publish), and it's unlikely that even a junky 1960s TV would display any interference from the frequencies transmitted over HDMI, even in the harmonics (which I suspect are well above even the over-the-air broadcast TV frequencies).

The ~10 dB reduction needed is only about a one order-of-magnitude difference (9 dB would be 8:1), so, it's not surprising that it looks like they may be able to tweak the issue away in the HDMI driver configuration (as I understand Pete's post - it's not clear how much of the needed reduction can be accomplished with a better cable).  That's music to our ears ... or, in this case, eyes, I suppose

User avatar
Tass
Posts: 535
Joined: Sat Jan 21, 2012 11:15 am

Re: Pete Lomas's take on the compliance issues

Thu Mar 29, 2012 10:30 pm

I'd definitely accept an uncertified board!

P.S. Since I gave up getting embedded video to work and can't edit my original post, here's the link regarding what this project is all about from my post above

(no – not off-topic at all )



EDIT: Well of course all you have to do is directly paste the link!

rpiguy
Posts: 25
Joined: Mon Mar 26, 2012 10:01 pm

Re: Pete Lomas's take on the compliance issues

Thu Mar 29, 2012 10:34 pm

Actually I take back what I said about being clueless w.r.t. EMI...thinking back to the Faraday cage concept in physics class, I would deal with EMI issues by putting the device in some sort of metal box

But I probably wouldn't even think to test a new device I just got delivered for EMI unless I happened to turn on a radio, hear static on a normally clear station, and be bored enough to experiment with moving it around and switching on/off the ton of electronic gadgets I own to try to isolate it -- an unlikely chain of events.

I definitely wouldn't think of measuring physical characteristics of a device like heat/power/electromagnetic waves as a matter of SOP

User avatar
johnbeetem
Posts: 945
Joined: Mon Oct 17, 2011 11:18 pm
Location: The Mountains
Contact: Website

Re: Pete Lomas's take on the compliance issues

Thu Mar 29, 2012 10:36 pm

rpiguy said:

If anyone's to blame, it's those unnamed "experts."  But keep in mind that it may not even be fair to blame them; when their opinion was asked, the interest in RPi was much smaller, there was much less media attention, and the overall demographic of the prospective user base was more hardware-oriented.
IMO nobody is to blame.  The plan was (1) make a developer release with uncased boards.  Who else besides a developer would buy a board with no case, no power supply, no SD card, no software, and no certification?  (2) See how it goes with the developer release.  If there are any problems, fix them before general release.  (3) Get certification for the general release board.  Certification is expensive and time-consuming -- why do it twice?  (4) General release!

The problem was that the developer release became general release because the general public ignored the assumption in (1).  So RasPi has to do certs now, and with luck they won't be a need to do it again for a while.

Regarding FCC risk, I believe (JMO/IANAL) the worst that would happen if you had a non-compliant board is that if someone complained you could get a "cease and desist" order where you had to stop using your RasPi or bring it into compliance at your expense with a suitably grounded case.

shirro
Posts: 248
Joined: Tue Jan 24, 2012 4:54 am

Re: Pete Lomas's take on the compliance issues

Thu Mar 29, 2012 11:19 pm

Excellent post.

I haven't seen it mentioned anywhere but is compliance being looked at globally or just CE and FCC? For instance Australia and NZ have C-tick and from memory it isn't that hard, register yourselves for a number, keep a folder with a declaration of compliance and your test results etc. But the standards for all these different countries aren't necessarily the same as CE. I am just worried that no mention of this means the shipments outside EU/USA might be further delayed.

Enlightenment
Posts: 34
Joined: Wed Aug 24, 2011 10:16 pm

Re: Pete Lomas's take on the compliance issues

Fri Mar 30, 2012 1:10 am

#1 Make case out of special type of plastic.

#2 Nickel-plate the plastic.

http://www.azom.com/article.as.....icleID=525

#3 Ground RPi to case.

#4 Yes, I know this doesn't solve all the problems, but a bazillion people are trying to make cases, so someone just needs to "add" metal to create a shielded case.

Enlightenment
Posts: 34
Joined: Wed Aug 24, 2011 10:16 pm

Re: Pete Lomas's take on the compliance issues

Fri Mar 30, 2012 1:14 am

In USA, Monoprice is infamous of selling high-quality low-priced cables.  I'm sure a bunch of people from USA would back up this statement.   This is not meant to be spam.

http://www.monoprice.com/produ.....p_id=10231

User avatar
tzj
Posts: 302
Joined: Thu Jan 05, 2012 3:09 pm
Location: Barnsley, SY, UK

Re: Pete Lomas's take on the compliance issues

Fri Mar 30, 2012 1:33 am

on the show "how its made", they say the plating of plastic buttons was a trade secret... not any more
http://www.themagpi.com <---- Checkout the MagPi
(The MagPi - Co-Founder)

User avatar
Jessie
Posts: 1754
Joined: Fri Nov 04, 2011 7:40 pm
Location: C/S CO USA

Re: Pete Lomas's take on the compliance issues

Fri Mar 30, 2012 2:49 am

BigAxe said:


In USA, Monoprice is infamous of selling high-quality low-priced cables.  I'm sure a bunch of people from USA would back up this statement.   This is not meant to be spam.

http://www.monoprice.com/produ.....p_id=10231



I buy all my HDMI cables from Monoprice.  I always buy 24AWG and add ferrite chokes on them.  I have had trouble with cheap cables but never with any of the thick cables from Monoprice.  The only problem I have ever had with a Monoprice cable was that the wire was so thick that I had to order a thinner gauge wire to make a bending radius without damaging a connector.

User avatar
rurwin
Forum Moderator
Forum Moderator
Posts: 4257
Joined: Mon Jan 09, 2012 3:16 pm
Contact: Website

Re: Pete Lomas's take on the compliance issues

Fri Mar 30, 2012 5:38 am

Phoenix RasPi said:


rurwin said:


It would chuck out loads of EMI that you wont know about. But you wont need to know about it either — it's much less than your router or your mobile phone and probably less than your microwave oven. The only time you need to be bothered is when it causes interference on your TV or your radio or your HiFi speaker cables, or when the Radio Ham next door complains.


Ummm, even with the crappiest-built cable, the RasPi spurious emanations are still going to be waaaay lower in power than those from a router (a couple of watts of effective radiated power – ERP), a mobile phone (about a watt of ERP),


I believe that's what I said.


and certainly at least 4 ~ 5 orders of magnitude less than those from a microwave oven (upward of 1,000 watts of ERP in a much higher frequency range of GHz, kept overwhelmingly inside the cooking chamber, we hope).


But with the door closed, how much is allowed to escape? Most of us don't expose ourselves to the EM inside a microwave oven, but may be concerned with the EM that gets out of it. IIRC that is in the order of single figure microwatts.


The ~10 dB reduction needed is only about a one order-of-magnitude difference (9 dB would be 8:1),


10dB is exactly an order of magnitude by definition. 9dB is approximately 8:1. (7.9432…:1 to be exact.)

Andre_P
Posts: 241
Joined: Sun Aug 28, 2011 7:57 am

Re: Pete Lomas's take on the compliance issues

Fri Mar 30, 2012 6:06 am

It sounds like there are options which is fantastic news !

I can remember doing stuff for a Train 'wing mirror' monitor. Had 25 MHz harmonics all over the shop. Cleaned it by using a slower cpld .

It's a bit of a black art so if this if the options are SW controllable then that is fantastic news, no board/BOM changes .

Please send my congratulations to the Peter and Eben in doing the tests and finding the solutions so fast ! That is Brilliant .

Phil Spiegel
Posts: 210
Joined: Tue Jan 17, 2012 8:17 am
Contact: Website

Re: Pete Lomas's take on the compliance issues

Fri Mar 30, 2012 6:25 am

shirro said:


....(snipped)....

I haven't seen it mentioned anywhere but is compliance being looked at globally or just CE and FCC? For instance Australia and NZ have C-tick........... I am just worried that no mention of this means the shipments outside EU/USA might be further delayed.


The announcement from E14/RS states that they will do CE then FCC followed by C tick ands rest-of-the-world approvals, with full production arranged for once its approved. It also describes how they are both cooperating with a 3rd firm for the testing procedures to get them completed as soon as possible.  They also mentioned WEE and RHoS and....and... and ... (go to their site(s) for up to date information)

I'm glad I only produce boards for my own use on my model railway(s)

Our loft-conversion's insulation appears to create quite a Faraday Cage!

arm2
Posts: 253
Joined: Thu Dec 15, 2011 3:46 pm

Re: Pete Lomas's take on the compliance issues

Sat Mar 31, 2012 4:37 pm

BigAxe said:


#3 Ground RPi to case.


Any ideas how?

Were on the RPi would you connect?

mole125
Posts: 228
Joined: Tue Jan 10, 2012 2:01 pm

Re: Pete Lomas's take on the compliance issues

Sat Mar 31, 2012 6:04 pm

That's easy, the ground pin on the GPIO connector.

User avatar
johnbeetem
Posts: 945
Joined: Mon Oct 17, 2011 11:18 pm
Location: The Mountains
Contact: Website

Re: Pete Lomas's take on the compliance issues

Sat Mar 31, 2012 6:16 pm

arm2 said:


BigAxe said:


#3 Ground RPi to case.


Any ideas how?

Were on the RPi would you connect?


According to the power supply schematic (ses RasPi Wiki), you can use the outer shell of the power supply connector S1 provided that 0 Ohm resistor R51 is installed (as it is on the Beta board photo).  I hope the HDMI shell is also grounded or my cheap HDMI cables won't work.  The Ethernet and USB shells may also be grounded as well.  We'll know better when we have complete schematics and/or boards.

Return to “General discussion”