Another pointless board. And  cpu is positioned stupidly, so no heatsink either. 

Edited March 23, 20179 yr by tkaiser
Moved from 'landing thread' https://forum.armbian.com/index.php?/topic/3576-nanopi-neo-2/ to here

16 minutes ago, hojnikb said:

cpu is positioned stupidly, so no heatsink either

 

I really love your valuable comments. All the time. They're so great. Awesome. You should really try to get a job as CEO at one of those stupid Chinese companies. They need you.

 

For all those people not thinking they're the only smart person on this planet: Check NEO heatsink here: http://www.friendlyarm.com/index.php?route=product/category&path=82

1 hour ago, tkaiser said:

 

I really love your valuable comments. All the time. They're so great. Awesome. You should really try to get a job as CEO at one of those stupid Chinese companies. They need you.

 

For all those people not thinking they're the only smart person on this planet: Check NEO heatsink here: http://www.friendlyarm.com/index.php?route=product/category&path=82

and as always you manage to be rude in every possible way. Good way to keep the community going

 

And back in real world, this heatsink wont do jack sh**, because you cant mount it directly to the soc (as the soc is on the back of the board). So your options are either metal case, that touches the soc or no case at all and a heatsink on the soc, which will in  turn make an ugly monstrosity. Just because guys at friendly couldn't slap the soc at the top of the board.

The thermal pad which you can put in between the SoC and heatsink will take care of the heat transfer. You can always opt to put a copper shim instead. Then pointing the heatsink upwards with the pin header soldered on that side too you have access to everything you need... 

Therefore in my opinion it is not a pointless board, and I do think the design is actually a well thought design (mechanically)...

1 hour ago, hojnikb said:

And back in real world, this heatsink wont do jack sh**, because you cant mount it directly to the soc (as the soc is on the back of the board). So your options are either metal case, that touches the soc or no case at all and a heatsink on the soc, which will in  turn make an ugly monstrosity. Just because guys at friendly couldn't slap the soc at the top of the board.

Please look at the photos: http://www.friendlyarm.com/index.php?route=product/product&path=82&product_id=134

Heatsink is attached on the bottom of the board, so it touches the SoC almost directly (thermal pad should be efficient enough). May not work the best if the board is laying on its bottom side with the heatsink, but who says that it can be used only horizontally and in this orientation?

 

@tkaiser

It would be better to argue or discuss stuff without referencing your opinion on the person who posted a message you don't agree with.

 

@hojnikb

It would be better to phrase your original message as "I don't see any use cases with this board for myself compared to other already existing boards". Some people will still find a use case for it or will choose this board insted of other ones.

10 hours ago, pzw said:

the design is actually a well thought design (mechanically)...

Exactly! 

 

We went through this design with original NEO already, I made series of tests concerned about heat transfer capabilities of the thermal pad and position of heatsink (works even on the bottom but you end up with 2°-3°C more compared to upright or top position.

 

Biggest problem with NEO so far was due to position of USB receptacle and Ethernet jack the thermal pad had to be rather thick, can be seen easily on the Wiki (and that's the only reason I put this picture online, so that people can inform themselve before making any assumptions): http://linux-sunxi.org/FriendlyARM_NanoPi_NEO_%26_AIR#NanoPi_NEO

 

With this Plus 2 USB receptacles and Ethernet jacks moved outside the mounting hole area so I would assume with this design thermal pad can be replaced with the usual 15x15mm copper shims and then combined with any 40x40mm heatsink or even a simple copper plate as enclosure bottom combined with a 3D printed top instead (or an ABS plastic case like this with a 16x16 hole on the bottom to connect copper shim with external heatsink). Or just a thermal pad or simple copper shim combined with an aluminium box as 'heatsink enclosure'.

 

Maybe FA will also provide a 2nd heatsink variant with lower distance and thinner thermal pad providing better heat dissipation? We'll see... at least my 1st thought when seeing NEO Plus 2 yesterday was: 'I would choose this over NEO2 for sure' due to 1 more USB receptacle and the possibility to improve thermal design if necessary (on Plus 2 we can switch to 1.3V VDD_CPUX voltage so better heat dissipation might be needed with constant full load).

 

Links for DIY tinkerers:

https://aliexpress.com/wholesale?initiative_id=QRW_20170322235019&SearchText=copper+plate&productId=32771212441

https://aliexpress.com/wholesale?catId=708041&initiative_id=AS_20170322224725&SearchText=40*40+heatsink

https://aliexpress.com/wholesale?catId=0&initiative_id=AS_20170323005119&SearchText=aluminium+box+electronics+case

10 hours ago, zador.blood.stained said:

Please look at the photos: http://www.friendlyarm.com/index.php?route=product/product&path=82&product_id=134

Heatsink is attached on the bottom of the board, so it touches the SoC almost directly (thermal pad should be efficient enough). May not work the best if the board is laying on its bottom side with the heatsink, but who says that it can be used only horizontally and in this orientation?

 

@tkaiser

It would be better to argue or discuss stuff without referencing your opinion on the person who posted a message you don't agree with.

 

@hojnikb

It would be better to phrase your original message as "I don't see any use cases with this board for myself compared to other already existing boards". Some people will still find a use case for it or will choose this board insted of other ones.

Obviously someone might find usecase for this board, as it's the case for anything really. If there really wasn't ANY market for it, i doubt they would bother designing it. It still think it would be wise if those companies at least tried doing some software work. What good is a piece of hardware, if the software provided is well, crap. Good thing we have sunxi/armbian to remedy this.

 

As for the heatsink; it seems awfully small (as it should be to even fit down there) so it's purpose might be limited. It will obviously help with some of the peak loads, but thats about it i'd say.

14 hours ago, pzw said:

Then pointing the heatsink upwards with the pin header soldered on that side too you have access to everything you need

 

And even when soldering the pin headers 'the usual way' on the other PCB side still everything works perfectly especially when combined with any of the currently available or future HATs/Docks: http://www.friendlyarm.com/index.php?route=product/category&path=86

 

With 'NEO Plus 2' Ethernet jack and the 2 USB jacks will then just project a little over compared to how it will look like with NEO, Air or NEO2:

2 hours ago, hojnikb said:

It still think it would be wise if those companies at least tried doing some software work

 

You should try harder, please flood this forum (yes, ours and not theirs) with such wishes, repeat it over and over again. Maybe then something like this will happen: http://wiki.friendlyarm.com/wiki/index.php/BakeBit  (wait, what? It happened already?!)

 

FriendlyELEC is one of the few board makers out there who do a great job on software. But reality won't affect you, you'll still flood threads with your 'valuable opinions'  

 

For those kind of people interested in information/news: please be aware that BakeBit stuff might not work with Armbian and you would've to rely on FriendlyELEC's OS images for BakeBit code examples to work. It's not on Armbian's priority list to support FriendlyELEC's various eco systems around their boards since we focus on slightly different use cases with H3 boards (the Matrix series does also exist and code examples are here, Igor added a huge patch to our H3 legacy kernel some months ago to get most if not all Matrix modules supported). Both Matrix and BakeBit stuff might work since this is all just standard interfacing stuff but we can't provide any support.

In my opinion it is much better to have the CPU on the solder side of the board.

It allows efficient passive cooling in a metal box that can be necessary for EMC in the real world.

 

2 hours ago, tkaiser said:

 

You should try harder, please flood this forum (yes, ours and not theirs) with such wishes, repeat it over and over again. Maybe then something like this will happen: http://wiki.friendlyarm.com/wiki/index.php/BakeBit  (wait, what? It happened already?!)

 

FriendlyELEC is one of the few board makers out there who do a great job on software. But reality won't affect you, you'll still flood threads with your 'valuable opinions'  

 

For those kind of people interested in information/news: please be aware that BakeBit stuff might not work with Armbian and you would've to rely on FriendlyELEC's OS images for BakeBit code examples to work. It's not on Armbian's priority list to support FriendlyELEC's various eco systems around their boards since we focus on slightly different use cases with H3 boards (the Matrix series does also exist and code examples are here, Igor added a huge patch to our H3 legacy kernel some months ago to get most if not all Matrix modules supported). Both Matrix and BakeBit stuff might work since this is all just standard interfacing stuff but we can't provide any support.

 

There is no need to be rude, you're not achieving anything good with this kind of attitude.  An inexperienced user getting into the  cheap SBC world might even feel this whole comunnity is toxic, given some of your posts. This is not something you want in a community oriented project. It's not that hard to keep opinions directed to users to yourself. I'm sure mods will clean any flood and unnecessary postings as needed.

 

 

 

26 minutes ago, Ford Prefect said:

In my opinion it is much better to have the CPU on the solder side of the board.

It allows efficient passive cooling in a metal box that can be necessary for EMC in the real world.

 

Only assuming you're actually have metal case, that can contact the soc via thermal pad. In most other applications (bare board, plastic casing) this will be worse,thermal wise.

28 minutes ago, Ford Prefect said:

It allows efficient passive cooling in a metal box

 

Sure, obvious and confirmed an hour ago by a friend of mine. His thermal values sounded too good to be true so I decided to order his two recommendations and try myself (comparing again FriendlyELEC's NEO heatsink and my usual H3 one -- please keep in mind that back then this was NEO PCB Rev 1.0 that had a heat problem, fixed by FA later. But it's easy to realize that FriendlyELEC's heatsink is a huge improvement over the usual ones with thermal adhesive)

• Aluminium enclosure
• Copper shims for H3
• Copper shims for A20 (since I want to try this out with Banana Pi Pro too)

 

1 hour ago, hojnikb said:

I'm sure mods will clean any flood and unnecessary postings as needed.

 

[x] Done.

 

Feel free to continue here (so not polluting the 'landing thread' regarding two new boards we (might) support) explaining to your audience why FriendlyELEC is stupid, why a heatsink that works quite well does not work at all in your opinion or why FriendlyELEC's (non existing according to you) software efforts lead to 'crap' as you call it.

 

Try to do full quotes whenever possible even if not needed and try to use as less words as possible per post so that the 'contents to signature' ratio is as bad as possible. Thanks.

Quote

explaining to your audience why FriendlyELEC is stupid

Was never implying that.

 

Quote

why a heatsink that works quite well does not work at all in your opinion

 

I'm sure you have hard data to back up this claim.

 

Quote

your opinion or why FriendlyELEC's (non existing according to you) software efforts lead to 'crap' as you call it.

 

Well, if friendlyelec images are all fine and dandy, why the hell do we have armbian for h3 in the first place ? I mean, if their images are top notch, a modification of their images for other h3 products (like from xunlong) would be more in place.

 

 

Quote

 

Try to do full quotes whenever possible even if not needed and try to use as less words as possible per post so that the 'contents to signature' ratio is as bad as possible. Thanks.

 

 

 Like this

Again, rudeness isn't really fitting for a technical forum like this.

 

On 03/23/2017 at 2:09 PM, Ford Prefect said:

In my opinion it is much better to have the CPU on the solder side of the board.

It allows efficient passive cooling in a metal box that can be necessary for EMC in the real world.

 

I totally agree, I've been pissed off by many boards on which it was not possible to have correct heat dissipation due to too large the distance between the CPU and the top of the enclosure. With the NEO/NEO2 now you can simply press the PCB (hence the CPU) against a thermal pad touching the enclosure and you're done. It's in fact one of the very rare board capable of spreading the heat outside of the enclosure and not to slowly heat its own environment.

I wish other vendors would realize this as well. And yes, I know it's difficult and expensive to make dual-side BGA designs so that will probably limit the possible candidates here. I would have had a much easier design for my MiQis if the RK3288 had been on the other side!

2 hours ago, wtarreau said:

It's in fact one of the very rare board capable of spreading the heat outside of the enclosure and not to slowly heat its own environment.

 

Hehe, it took some time until I realized that myself ~2 years ago (speaking about Banana Pi/Pro here).

 

BTW: guess what the massive aluminium block in the middle of this enclosure is supposed to do? Transferring heat away from S805/S905 in so called 'Spreedbox' since ODROID-C1+/C2 also use the 'wrong' SoC placement.

 

Hi

Nice Box for sure.

Main problem with such a solution is that when you want access to the hardware on a bench you have to use

a temporary heat sink.

I killed a DVD writer this way.

Only later I noticed there was a thermal pad on the cover I removed