Raspberry Pi 4 vs Odroid N2

gsgs
Posts: 37
Joined: Thu Jul 11, 2019 5:14 pm
languages_spoken: english
ODROIDs: n2
Location: Germany
Has thanked: 5 times
Been thanked: 1 time
Contact:

Re: Raspberry Pi 4 vs Odroid N2

Post by gsgs »

now I figured out how to measure CPU-temperature
https://praxistipps.chip.de/raspberry-p ... ehts_46615
vcgencmd measure_temp

the pi4b running stockfish with 4 threads gets to 82C after ~10min without any heatsink
and kn/s go down to 920kn/s (continues to fall) after an initial 1005 kn/s.

With one of the 40x40x20 mm heatsinks temperature is 74C (ambient=24C) and kn/s is stable at ~1002 kn/s

With two of the 40x40x20 mm heatsinks temperature is 72C (ambient=24C) and kn/s is stable at ~1005 kn/s

With three of the 40x40x20 mm heatsinks temperature is 68C (ambient=24C) and kn/s is stable at ~1005 kn/s

With three of the 40x40x20 heatsinks plus a 100x100x20 heatsink on top of that
temperature is 69C and 1006 kn/s , so that doesn't help anymore

With just one of the small pi-heatsinks temperature is 82C but kn/s still at >1000

harry4516
Posts: 56
Joined: Sun Jan 25, 2015 10:35 am
languages_spoken: english, german
ODROIDs: all Odroid's, most RPIs, some Banana-Pis
Location: Bavaria, Germany
Has thanked: 0
Been thanked: 5 times
Contact:

Re: Raspberry Pi 4 vs Odroid N2

Post by harry4516 »

I have the RPI4 and the N2 running here and did some simple measurements:

Power (plaback same video): N2=4.5w .... PRI4=5.8w

Youtube-Video in Cromium, max. resolution for stutter free full screen video:
N2 ... impossible
RPI4 ... 720p

CPU temp during video playback:
N2= 40 degC , RPI4= 65 degC

Ethernet: both running at 933 MB/s max (maybe limited only by my iperf3 server)

compiling a C program (sysbench sources) compilled using 4 threads:
N2=3m01s, RPI4=3m22s
(for any reason I don't understand the N2 takes a bit longer if I run "make" with 6 threads)

memory throughput (sysbench):
N2=1263 MB/s, RPI4=524MB/s

Price (Germany, Aug 2019), 2GB version:
N2=80 Eur, RPI4=46 Eur

"Feeling" the speed while working on the desktop:
almost the same, no big difference.

my conclusion:
--------------------

RPI4 advantage: better video support, lower price
RPI4 disadvantage: horrible heat, needs cooler and fan

N2 advantage: lower operating temperature, does not need a fan

for me, a fan is an absolute nogo. This is why I love my N2 !

stmicro
Posts: 251
Joined: Tue Apr 28, 2015 4:23 pm
languages_spoken: english, chinese
ODROIDs: Many Odroids and Rpis.
Location: shenzhen china
Has thanked: 0
Been thanked: 1 time
Contact:

Re: Raspberry Pi 4 vs Odroid N2

Post by stmicro »

Firefox could play 1080p Youtube video nicely on my N2. But Chromium can't do it at all. I don't know why.

IgaBiva
Posts: 115
Joined: Tue May 07, 2019 4:00 pm
languages_spoken: english, deutsch, srpski
ODROIDs: N2
Has thanked: 7 times
Been thanked: 17 times
Contact:

Re: Raspberry Pi 4 vs Odroid N2

Post by IgaBiva »

Pi 4 still does not support HDR and this is no go for me.

What use case you have for YouTube? Desktop?

Have you tried CoreElec? 1080p YouTube works perfect on it.

Also, CoreElec on N2 is only combination which plays absolutelly all of my videos, from very old SD camera videos to newest UHD HDR rips and all of this with excellent video quality.

I have used Kodi on multiple devices (Android TV, Pi3, Xbox X) and all of them had some issues. N2 is by far best video performer of all.

stmicro
Posts: 251
Joined: Tue Apr 28, 2015 4:23 pm
languages_spoken: english, chinese
ODROIDs: Many Odroids and Rpis.
Location: shenzhen china
Has thanked: 0
Been thanked: 1 time
Contact:

Re: Raspberry Pi 4 vs Odroid N2

Post by stmicro »

I saw that CoreELEC can play 4K Youtube videos via VP9 HW acceleration this morning. viewtopic.php?f=179&t=33956&start=150#p265507
I will try the same add-on on Android Kodi a few days later to see how it works with 4K Youtube + HDR.

User avatar
rooted
Posts: 8034
Joined: Fri Dec 19, 2014 9:12 am
languages_spoken: english
Location: Gulf of Mexico, US
Has thanked: 724 times
Been thanked: 252 times
Contact:

Re: Raspberry Pi 4 vs Odroid N2

Post by rooted »

stmicro wrote:I saw that CoreELEC can play 4K Youtube videos via VP9 HW acceleration this morning. viewtopic.php?f=179&t=33956&start=150#p265507
I will try the same add-on on Android Kodi a few days later to see how it works with 4K Youtube + HDR.
HDR isn't working on YouTube yet, they are working on a fix.

stmicro
Posts: 251
Joined: Tue Apr 28, 2015 4:23 pm
languages_spoken: english, chinese
ODROIDs: Many Odroids and Rpis.
Location: shenzhen china
Has thanked: 0
Been thanked: 1 time
Contact:

Re: Raspberry Pi 4 vs Odroid N2

Post by stmicro »

rooted wrote:
Mon Aug 12, 2019 11:50 pm
stmicro wrote:I saw that CoreELEC can play 4K Youtube videos via VP9 HW acceleration this morning. viewtopic.php?f=179&t=33956&start=150#p265507
I will try the same add-on on Android Kodi a few days later to see how it works with 4K Youtube + HDR.
HDR isn't working on YouTube yet, they are working on a fix.
I don't think Android Kodi/addon have any relation to the CE.

User avatar
rooted
Posts: 8034
Joined: Fri Dec 19, 2014 9:12 am
languages_spoken: english
Location: Gulf of Mexico, US
Has thanked: 724 times
Been thanked: 252 times
Contact:

Re: Raspberry Pi 4 vs Odroid N2

Post by rooted »


stmicro wrote:
rooted wrote:
Mon Aug 12, 2019 11:50 pm
stmicro wrote:I saw that CoreELEC can play 4K Youtube videos via VP9 HW acceleration this morning. viewtopic.php?f=179&t=33956&start=150#p265507
I will try the same add-on on Android Kodi a few days later to see how it works with 4K Youtube + HDR.
HDR isn't working on YouTube yet, they are working on a fix.
I don't think Android Kodi/addon have any relation to the CE.
Yeah I missed the Android part, I thought you were saying 4K YouTube + HDR in CE.

Minou666
Posts: 149
Joined: Sat Sep 14, 2013 9:40 pm
languages_spoken: French and English
ODROIDs: Odroid C1,Odroid C2, Odroid HC1, Odroid HC2 and XU3LITE
Location: White Pigeon Michigan
Has thanked: 0
Been thanked: 0
Contact:

Re: Raspberry Pi 4 vs Odroid N2

Post by Minou666 »

I find the flame war amusing, not quite as good as the flame wars our OS/2 groups had with the windows groups back in the 90s.

Joke aside, the best boards are the beagle bone and beagle board for my industry, odroid and RPI have no CAN and beagle boards have open source hardware which is not the case with Odroid and RPI.

For video I haven't seen anything that beat the Odroid and TI boards in the RPI world.

Most of my work involves boards using TI devices but I still love Odroid. I use RPI for small designs that do not require to run fast. The $10 version is a good for those.

Michel

powerful owl
Posts: 170
Joined: Thu Mar 28, 2019 8:57 pm
languages_spoken: english
ODROIDs: 2 x HC1, 3 x H2, 1 x H2+
Has thanked: 41 times
Been thanked: 18 times
Contact:

Re: Raspberry Pi 4 vs Odroid N2

Post by powerful owl »

Minou666 wrote:
Tue Aug 20, 2019 8:25 am
I find the flame war amusing, not quite as good as the flame wars our OS/2 groups had with the windows groups back in the 90s.
Hah, I bet that was something. I did run OS/2 for a short while, I think it was on a state of the art 33MHz CPU, monstrous!

Anyway, RPi4 cooling solution -

Image

Yeah I know it has a fan, but if you watch the video it's still very effective without it. I think I might get/try one.

https://www.seeedstudio.com/ICE-Tower-C ... -4097.html

boolean
Posts: 15
Joined: Thu Aug 31, 2017 8:27 am
languages_spoken: Dutch, English
ODROIDs: C2 & N2 ; RPIs: 3B & 4B
Has thanked: 0
Been thanked: 0
Contact:

Re: Raspberry Pi 4 vs Odroid N2

Post by boolean »

If you want adequate (and silent) cooling for your RPi3/4 or C2, just use this enclosure:

Image
Image

I use these on both my Raspberry's (3B & 4B) and C2. Unfortunately this design isn't available for the N2, so you just have to deal with the impractical heatsink. I bought this case to give it a less cheap-ass look.

Image

Hardkernel f*cked-up their C2 due to the lack of kernel support, but I still appreciate their effort so I bought the N2 as well. Let's see where it goes from here...

User avatar
rooted
Posts: 8034
Joined: Fri Dec 19, 2014 9:12 am
languages_spoken: english
Location: Gulf of Mexico, US
Has thanked: 724 times
Been thanked: 252 times
Contact:

Re: Raspberry Pi 4 vs Odroid N2

Post by rooted »

The N2 case is fine, and the heatsink isn't impractical at all. You are the first person I've seen complain about it actually, everyone else likes the cooling solution because it works so well. No need to spend additional money to properly cool the device, all other manufacturers should actually follow their lead.

I do like that metal case design for the N2.
These users thanked the author rooted for the post (total 2):
elatllat (Sun Aug 25, 2019 9:39 pm) • beos (Mon Jul 27, 2020 5:59 pm)

powerful owl
Posts: 170
Joined: Thu Mar 28, 2019 8:57 pm
languages_spoken: english
ODROIDs: 2 x HC1, 3 x H2, 1 x H2+
Has thanked: 41 times
Been thanked: 18 times
Contact:

Re: Raspberry Pi 4 vs Odroid N2

Post by powerful owl »

Well, in reality the heatsink on the N2 is not well designed, a large mass simply slows down temperature increase. Airflow is needed to dissipate heat (*) - having the fins horizontally on the bottom against a surface (as it will usually be) is the worst possible place for them. I'm not saying it doesn't work, just that it's far from an optimal solution in engineering terms. I use a fan on my HC1s for this reason. But yes, at least it has one.
(*) Added: or emission, again the bottom on surface is the worst place.

User avatar
rooted
Posts: 8034
Joined: Fri Dec 19, 2014 9:12 am
languages_spoken: english
Location: Gulf of Mexico, US
Has thanked: 724 times
Been thanked: 252 times
Contact:

Re: Raspberry Pi 4 vs Odroid N2

Post by rooted »

That's what everyone thought when the N2 was released (including myself), but in practice the heatsink works impressively well.
These users thanked the author rooted for the post:
elatllat (Sun Aug 25, 2019 9:41 pm)

Tenaja
Posts: 61
Joined: Mon Jun 24, 2019 2:05 am
languages_spoken: english
Has thanked: 1 time
Been thanked: 6 times
Contact:

Re: Raspberry Pi 4 vs Odroid N2

Post by Tenaja »

I don't have an n2 yet, but the cooling surface area looks great. As long as you don't set it on a conforming insulator (blanket or towel), that huge block with cooling fins on the bottom is likely to be far more efficient than a 1cm cube on top.

What confuses me, though, is why you wouldn't just place it upside down, if you really wanted cooling fins on the top. It's not like your cables will break... Just because a photo shows the interesting bits in top doesn't mean you can't hide them when you are having overheating issues. Add some standoffs if you don't want a case top.

powerful owl
Posts: 170
Joined: Thu Mar 28, 2019 8:57 pm
languages_spoken: english
ODROIDs: 2 x HC1, 3 x H2, 1 x H2+
Has thanked: 41 times
Been thanked: 18 times
Contact:

Re: Raspberry Pi 4 vs Odroid N2

Post by powerful owl »

Oh, I was responding to rooted "all other manufacturers should actually follow their lead" - it's not a good basic design, I hope not all manufacturers follow this lead. The specifics of the particular board aren't what I was commenting on. JMO.... :)

boolean
Posts: 15
Joined: Thu Aug 31, 2017 8:27 am
languages_spoken: Dutch, English
ODROIDs: C2 & N2 ; RPIs: 3B & 4B
Has thanked: 0
Been thanked: 0
Contact:

Re: Raspberry Pi 4 vs Odroid N2

Post by boolean »

rooted wrote:
Sun Aug 25, 2019 9:59 am
The N2 case is fine, and the heatsink isn't impractical at all. You are the first person I've seen complain about it actually, everyone else likes the cooling solution because it works so well. No need to spend additional money to properly cool the device, all other manufacturers should actually follow their lead.
It is impractical and the design could be more efficient. However, for the heat it produces it doesn't actually matters that much.
rooted wrote:
Sun Aug 25, 2019 9:59 am
I do like that metal case design for the N2.
I do too, it's about the only N2 case on the market that doesn't actually hurt your eyes. I was hoping that companies like Wicked Aluminum would jump the N2 train, but somehow currently no one is giving it any love or attention. Maybe it's because of the heatsink they have to deal with or its tiny market footprint, who knows...

blu
Posts: 84
Joined: Wed Mar 08, 2017 11:30 pm
languages_spoken: english
ODROIDs: XU4, N2
Has thanked: 3 times
Been thanked: 22 times
Contact:

Re: Raspberry Pi 4 vs Odroid N2

Post by blu »

rooted wrote:
Sun Aug 25, 2019 9:59 am
The N2 case is fine, and the heatsink isn't impractical at all. You are the first person I've seen complain about it actually, everyone else likes the cooling solution because it works so well. No need to spend additional money to properly cool the device, all other manufacturers should actually follow their lead.

I do like that metal case design for the N2.
Hear, hear! IIRC, N2 is the first high-performing SBC that comes with *adequate* passive thermal solution out of the factory. Other manufacturers have what to take note of.

ps: N2 case allows for effortless flipping of the device whichever side one prefers.
These users thanked the author blu for the post:
rooted (Tue Aug 27, 2019 3:53 am)

boolean
Posts: 15
Joined: Thu Aug 31, 2017 8:27 am
languages_spoken: Dutch, English
ODROIDs: C2 & N2 ; RPIs: 3B & 4B
Has thanked: 0
Been thanked: 0
Contact:

Re: Raspberry Pi 4 vs Odroid N2

Post by boolean »

blu wrote:
Mon Aug 26, 2019 10:38 pm
Hear, hear! IIRC, N2 is the first high-performing SBC that comes with *adequate* passive thermal solution out of the factory. Other manufacturers have what to take note of.
No, it actually needs the heatsink in order to operate on the advertised speed. I've tried to run it without and it fails on moderate workloads before it can adjust/lower its clock. Don't get me wrong, I do like the N2 (otherwise I wouldn't have purchased it) but its heatslink is nothing worth mentioning. It's just a $1 peace of anodized aluminum that you could buy yourself at RadioShack. It isn't even copper!
blu wrote:
Mon Aug 26, 2019 10:38 pm
ps: N2 case allows for effortless flipping of the device whichever side one prefers.
What kind of non-argument is this? Are you trying to put it up your rectum? It's a rectangular case, period.

User avatar
rooted
Posts: 8034
Joined: Fri Dec 19, 2014 9:12 am
languages_spoken: english
Location: Gulf of Mexico, US
Has thanked: 724 times
Been thanked: 252 times
Contact:

Re: Raspberry Pi 4 vs Odroid N2

Post by rooted »


boolean wrote: No, it actually needs the heatsink in order to operate on the advertised speed. I've tried to run it without and it fails on moderate workloads before it can adjust/lower its clock. Don't get me wrong, I do like the N2 (otherwise I wouldn't have purchased it) but its heatslink is nothing worth mentioning. It's just a $1 peace of anodized aluminum that you could buy yourself at RadioShack. It isn't even copper!

What kind of non-argument is this? Are you trying to put it up your rectum? It's a rectangular case, period.
Of course the N2 needs the heatsink, that's why it has it. What kind of argument is this?

No SBC will operate at advertised speeds without a heatsink

You can certainly not buy a heatsink like the N2 has for $1, heatsinks are not that inexpensive.

blu
Posts: 84
Joined: Wed Mar 08, 2017 11:30 pm
languages_spoken: english
ODROIDs: XU4, N2
Has thanked: 3 times
Been thanked: 22 times
Contact:

Re: Raspberry Pi 4 vs Odroid N2

Post by blu »

boolean wrote:
Tue Aug 27, 2019 2:54 am
No, it actually needs the heatsink in order to operate on the advertised speed. I've tried to run it without and it fails on moderate workloads before it can adjust/lower its clock. Don't get me wrong, I do like the N2 (otherwise I wouldn't have purchased it) but its heatslink is nothing worth mentioning. It's just a $1 peace of anodized aluminum that you could buy yourself at RadioShack. It isn't even copper!
You lost me here. It's as if we're talking about different things. Did I say the device can operate as advertised without a heatsink? Did I say it was a valuable-alloy heatsink? /scratches head
boolean wrote:
Tue Aug 27, 2019 2:54 am
What kind of non-argument is this? Are you trying to put it up your rectum? It's a rectangular case, period.
Somebody mentioned earlier in this thread the heatsink was underneath the device. I'm just saying it's whichever side you choose it to be -- it's a rectangular box (with no moving parts inside), as you so eloquently mentioned.

boolean
Posts: 15
Joined: Thu Aug 31, 2017 8:27 am
languages_spoken: Dutch, English
ODROIDs: C2 & N2 ; RPIs: 3B & 4B
Has thanked: 0
Been thanked: 0
Contact:

Re: Raspberry Pi 4 vs Odroid N2

Post by boolean »

Let met elaborate on this. You made it sound like Hardkernel's design is something epic that more manufacturers should incorporate. When a SBC doesn't has a heatsinkit, it generally doesn't need one to operate. The N2 needs one, that's why it's there, not because it's some nice afterthought. Yes, the N2's SOC is able to throttle its frequency, but without the heatsink in place that mechanism fails because the temperature is rising too quick. Most other SOC's don't have this issue and therefore don't necessarily need a heatsink.

Another peculiar thing I've noticed is that the RPi4 (A72) has a better per clock performance than the N2 (A73). I've clocked both devices to 2GHz (passively cooled), ran a few benchmarks (SciMark, Smallpt and PyBench) and the Raspberry came out on top on all runs. Has someone else also tried this? I'm curious for other results...

mad_ady
Posts: 8568
Joined: Wed Jul 15, 2015 5:00 pm
languages_spoken: english
ODROIDs: XU4, C1+, C2, C4, N1, N2, H2, Go, Go Advance
Location: Bucharest, Romania
Has thanked: 578 times
Been thanked: 483 times
Contact:

Re: Raspberry Pi 4 vs Odroid N2

Post by mad_ady »

I think the point is - HK could have gone with a small heatsink + fan design - like xu4. Many users are happier to have a larger, passive heatsink instead.
These users thanked the author mad_ady for the post:
elatllat (Thu Aug 29, 2019 7:23 am)

blu
Posts: 84
Joined: Wed Mar 08, 2017 11:30 pm
languages_spoken: english
ODROIDs: XU4, N2
Has thanked: 3 times
Been thanked: 22 times
Contact:

Re: Raspberry Pi 4 vs Odroid N2

Post by blu »

boolean wrote:
Thu Aug 29, 2019 3:09 am
Let met elaborate on this. You made it sound like Hardkernel's design is something epic that more manufacturers should incorporate.
N2 is one of the few passively-cooled SBCs on the market that does not suffer from chronic thermal issues, and it's like that out of the box, without the need for aftermarket parts and users' trial and error. If that was the norm everywhere, then we wouldn't be having this discussion. Since it's not -- yes, other vendors better take note.
When a SBC doesn't has a heatsinkit, it generally doesn't need one to operate. The N2 needs one, that's why it's there, not because it's some nice afterthought.
RPi4 also needs one, but it ships with none -- total absence of thought.
Yes, the N2's SOC is able to throttle its frequency, but without the heatsink in place that mechanism fails because the temperature is rising too quick. Most other SOC's don't have this issue and therefore don't necessarily need a heatsink.
Huh? Are you saying that temperature rises too quickly for the thermal sensors to detect it? Let me explain how thermal trip points work -- trip points are programmable events, driven by thermal sensors within the SoC, which events are set to trigger throttling and/or active thermal solutions at temps deemed by the manufacturer as 'about-to-get-hot', ''hot' and 'critically hot'. Without having participated in the selection of the trip points in the current odroid images, I'd venture to guess those trip points were chosen with this particular heatsink in mind. So if you remove the factory heatsink and expect the same trip points to maintain reliable operation that may not happen -- they were not calculated that way -- shutdown temps will be reached much quicker. That does not mean the device cannot have a proper setup of trip-points that allow its operation without a heatsink -- that would be absurd.
Another peculiar thing I've noticed is that the RPi4 (A72) has a better per clock performance than the N2 (A73). I've clocked both devices to 2GHz (passively cooled), ran a few benchmarks (SciMark, Smallpt and PyBench) and the Raspberry came out on top on all runs. Has someone else also tried this? I'm curious for other results...
Strictly uarch speaking, CA72 generally has better per-clock performance on fpu workloads compared to CA73. On integer and mixed workloads things largely depend on the specifics of the actual workload.

That said, even on some heavily-fp workloads factory-clocked N2 can still beat an overclocked, DIY actively-cooled 2GHz RPi4. Behold: https://youtu.be/u2rDYDv-BwA?t=666
These users thanked the author blu for the post (total 2):
elatllat (Thu Aug 29, 2019 7:31 am) • rooted (Thu Aug 29, 2019 2:48 pm)

boolean
Posts: 15
Joined: Thu Aug 31, 2017 8:27 am
languages_spoken: Dutch, English
ODROIDs: C2 & N2 ; RPIs: 3B & 4B
Has thanked: 0
Been thanked: 0
Contact:

Re: Raspberry Pi 4 vs Odroid N2

Post by boolean »

You clearly don't understand the difference between "need" and "nice to have". The RPi4 is designed to be able to run without a big ass impractical heatsink. If you think your application requires more cooling capacity, you could always add an appropriately sized heatsink yourself. The N2 has a stock heatsink which it actually can't run without (in it's current design) when dealing with heavy workloads. You don't want to enforce a limit either, because then you're loosing efficiency. For example, it can process small/moderate workloads at stock frequency without that plump heatsink attached, so you should keep it that way. However, when the workload seems te be bigger than expected, it should throttle its frequency down a bit so it won't overheat and crash. This mechanism works well with the heatsink attached, so apparently it isn't designed to work without and therefor it needs one.

I'm familiar with temperature trip points (TTRIP) so that's why I mentioned the temperature issue. If the events are programmed correctly, it should trigger throttling at a given temperature. I don't see why this should be heatsink depended. Even if they anticipated that it should be able to sustain a certain clock frequency within a specific temperature range, once the threshold has reached it should start throttling instead of shutting down cores or crashing.

User avatar
rooted
Posts: 8034
Joined: Fri Dec 19, 2014 9:12 am
languages_spoken: english
Location: Gulf of Mexico, US
Has thanked: 724 times
Been thanked: 252 times
Contact:

Re: Raspberry Pi 4 vs Odroid N2

Post by rooted »

Why not adjust the trip points and see if blu is correct?

If you don't want to use the big ass impractical heatsink you don't have to, but to expect it to function properly without it is ridiculous. The trip points are set to work in conjunction with the factory installed heatsink if you want to remove it you are changing the predefined layout of the N2 and all bets are off.

If you adjust the trip points and your N2 still hard locks due to heat then one of several things could be wrong.

1) You damaged the SoC somehow while removing the heatsink but added pressure of the heatsink alleviates it.

2) Something is wrong with your hardware from the manufacturer.

3) There is a bug in the kernel thermal code.

If you remove the heatsink from a PC how long will it function under load? (yes this is the same thing, the N2 was designed to be used with the installed thermal solution)

SlappyMcPhee
Posts: 105
Joined: Fri Aug 18, 2017 2:09 pm
languages_spoken: english
ODROIDs: XU4 (3 of them)
Has thanked: 6 times
Been thanked: 5 times
Contact:

Re: Raspberry Pi 4 vs Odroid N2

Post by SlappyMcPhee »

After reading through this thread it just seems like it is all about throwing shade at the N2 which just seems uncouth to do on the manufacturer's forum. If someone is so enamoured with the Pi 4 then why not just move along with their day using it?

I will give my two cents on the state of the Pi 4 and cooling. It is straight asinine that the Pi Foundation released the board without at least passive cooling especially with how closely the cores are fabbed together on the die. I have a couple already myself and the ENTIRE board becomes so hot without one it is just plain stupid. All in the name of keeping their form factor it would seem. They knew what they were doing, but they did it anyway. It is just like with the USB-C debacle. They knew what they were doing and released to the public anyway. Then the icing on the cake is that they knew how they were going to have thermal issues with the design because they downclocked the cores to 1.5. What I find laughable is how so many out in the wild call upclocking to 1.75 or 2.0 an OC....lmao.

So from the standpoint that the cores are supposed to run at 2.0 natively and they can barely run at the 1.5 without thermal issues for the entire PCB this whole attempt to knock the N2 on cooling from NIB seems utterly pointless.

/Steps off soapbox

Sent from my Pixel 3a using Tapatalk

Owner The Retro Arena and Odroid Retro Arena

powerful owl
Posts: 170
Joined: Thu Mar 28, 2019 8:57 pm
languages_spoken: english
ODROIDs: 2 x HC1, 3 x H2, 1 x H2+
Has thanked: 41 times
Been thanked: 18 times
Contact:

Re: Raspberry Pi 4 vs Odroid N2

Post by powerful owl »

Um, no offence but I think you may have missed the point.

boolean
Posts: 15
Joined: Thu Aug 31, 2017 8:27 am
languages_spoken: Dutch, English
ODROIDs: C2 & N2 ; RPIs: 3B & 4B
Has thanked: 0
Been thanked: 0
Contact:

Re: Raspberry Pi 4 vs Odroid N2

Post by boolean »

rooted wrote:
Thu Aug 29, 2019 3:05 pm
If you don't want to use the big ass impractical heatsink you don't have to, but to expect it to function properly without it is ridiculous. The trip points are set to work in conjunction with the factory installed heatsink if you want to remove it you are changing the predefined layout of the N2 and all bets are off.

If you remove the heatsink from a PC how long will it function under load? (yes this is the same thing, the N2 was designed to be used with the installed thermal solution)
I'm glad we agree on something, because that's indeed EXACTLY the same thing! If you remove the heatsink of a (modern) PC/Server CPU, it just slows down without locking, crashing or burning itself. Only AMD had thermal issues in the past (they probably still have), but that's not relevant here.

Can you do me a favor by running Thomas Kaiser's script on your N2 and posting the results over here? Maybe I've just received a faulty unit indeed.

Code: Select all

wget https://raw.githubusercontent.com/ThomasKaiser/sbc-bench/master/sbc-bench.sh
SlappyMcPhee wrote:
Sun Sep 01, 2019 3:24 pm
I will give my two cents on the state of the Pi 4 and cooling. It is straight asinine that the Pi Foundation released the board without at least passive cooling especially with how closely the cores are fabbed together on the die. I have a couple already myself and the ENTIRE board becomes so hot without one it is just plain stupid. All in the name of keeping their form factor it would seem. They knew what they were doing, but they did it anyway. It is just like with the USB-C debacle. They knew what they were doing and released to the public anyway. Then the icing on the cake is that they knew how they were going to have thermal issues with the design because they downclocked the cores to 1.5. What I find laughable is how so many out in the wild call upclocking to 1.75 or 2.0 an OC....lmao.

So from the standpoint that the cores are supposed to run at 2.0 natively and they can barely run at the 1.5 without thermal issues for the entire PCB this whole attempt to knock the N2 on cooling from NIB seems utterly pointless.
Maybe you should follow a comprehensive reading course, because you're totally missing the point here. It would have been nice if the N2 was designed to be able to run without a heatsink, or at least had a more refined/practical one so its overal size wouldn't have been doubled. Removing its heatsink voids warranty, so you shouldn't temper with it (like I did).

I've got multiple SBC's, but I bought the N2 specifically for my bedroom to use with RetroPie (or a fork) and LibreELEC. Unfortunately, because of the impractical heatsink, it's too thick to fit behind the TV. Removing the heatsink or replacing it with a smaller one didn't provide stable operation, so I needed to look into other options as well.

I did a bit of research and decided to bet on a RPi4 and ordered one including a case with passive cooling. The Raspberry is about half the height of the N2 (both inside their case), so this fitted like a charm. Clocked @ 2GHz and at full load (sbc-bench) it won't get any hotter than 54.5°C, so that won't be an issue either. If you know how to read numbers, here are the results: http://ix.io/1TZA ;)

Having said that, performance wise I still prefer the N2. Its GPU is a lot beefier than the one packed into the RPi4. It's just a shame that it won't fit my use case and I can't make it fit without loosing warranty.

User avatar
rooted
Posts: 8034
Joined: Fri Dec 19, 2014 9:12 am
languages_spoken: english
Location: Gulf of Mexico, US
Has thanked: 724 times
Been thanked: 252 times
Contact:

Re: Raspberry Pi 4 vs Odroid N2

Post by rooted »


boolean wrote:
rooted wrote:
Thu Aug 29, 2019 3:05 pm
If you don't want to use the big ass impractical heatsink you don't have to, but to expect it to function properly without it is ridiculous. The trip points are set to work in conjunction with the factory installed heatsink if you want to remove it you are changing the predefined layout of the N2 and all bets are off.

If you remove the heatsink from a PC how long will it function under load? (yes this is the same thing, the N2 was designed to be used with the installed thermal solution)
I'm glad we agree on something, because that's indeed EXACTLY the same thing! If you remove the heatsink of a (modern) PC/Server CPU, it just slows down without locking, crashing or burning itself. Only AMD had thermal issues in the past (they probably still have), but that's not relevant here.

Can you do me a favor by running Thomas Kaiser's script on your N2 and posting the results over here? Maybe I've just received a faulty unit indeed.

Code: Select all

wget https://raw.githubusercontent.com/ThomasKaiser/sbc-bench/master/sbc-bench.sh
I tried but I'm running CoreELEC currently.

powerful owl
Posts: 170
Joined: Thu Mar 28, 2019 8:57 pm
languages_spoken: english
ODROIDs: 2 x HC1, 3 x H2, 1 x H2+
Has thanked: 41 times
Been thanked: 18 times
Contact:

Re: Raspberry Pi 4 vs Odroid N2

Post by powerful owl »

boolean wrote:
Mon Sep 02, 2019 5:20 am
Can you do me a favor by running Thomas Kaiser's script on your N2 and posting the results over here?
With heatsink on (or off)?

blu
Posts: 84
Joined: Wed Mar 08, 2017 11:30 pm
languages_spoken: english
ODROIDs: XU4, N2
Has thanked: 3 times
Been thanked: 22 times
Contact:

Re: Raspberry Pi 4 vs Odroid N2

Post by blu »

boolean wrote:
Mon Sep 02, 2019 5:20 am
I did a bit of research and decided to bet on a RPi4 and ordered one including a case with passive cooling. The Raspberry is about half the height of the N2 (both inside their case), so this fitted like a charm. Clocked @ 2GHz and at full load (sbc-bench) it won't get any hotter than 54.5°C, so that won't be an issue either. If you know how to read numbers, here are the results: http://ix.io/1TZA ;)
Is that with the Wicked Aluminum $70 case?

boolean
Posts: 15
Joined: Thu Aug 31, 2017 8:27 am
languages_spoken: Dutch, English
ODROIDs: C2 & N2 ; RPIs: 3B & 4B
Has thanked: 0
Been thanked: 0
Contact:

Re: Raspberry Pi 4 vs Odroid N2

Post by boolean »

powerful owl wrote:
Thu Sep 05, 2019 2:16 am
boolean wrote:
Mon Sep 02, 2019 5:20 am
Can you do me a favor by running Thomas Kaiser's script on your N2 and posting the results over here?
With heatsink on (or off)?
On and off, if possible. I don't want you to ruin your device.
blu wrote:
Thu Sep 05, 2019 7:31 pm
boolean wrote:
Mon Sep 02, 2019 5:20 am
I did a bit of research and decided to bet on a RPi4 and ordered one including a case with passive cooling. The Raspberry is about half the height of the N2 (both inside their case), so this fitted like a charm. Clocked @ 2GHz and at full load (sbc-bench) it won't get any hotter than 54.5°C, so that won't be an issue either. If you know how to read numbers, here are the results: http://ix.io/1TZA ;)
Is that with the Wicked Aluminum $70 case?
Yes, this was the 6th run in a row @ 2GHz with a room temperature of 26.5°C. My N2 starts throttling (with heatsink) after the 3rd run.

EDIT: Since summer is on its way out, and my room temperature has gone down to 20°C, I've done another run which results have been posted here: http://ix.io/1UnL
Last edited by boolean on Fri Sep 06, 2019 7:03 am, edited 1 time in total.

elatllat
Posts: 1799
Joined: Tue Sep 01, 2015 8:54 am
languages_spoken: english
ODROIDs: XU4, N1, N2, C4, N2+, HC4
Has thanked: 51 times
Been thanked: 118 times
Contact:

Re: Raspberry Pi 4 vs Odroid N2

Post by elatllat »

boolean wrote:
Fri Sep 06, 2019 4:49 am
... results: http://ix.io/1TZA ...
would there be any change when switching from armhf to arm64 (like increased thermal load) on the A72?

boolean
Posts: 15
Joined: Thu Aug 31, 2017 8:27 am
languages_spoken: Dutch, English
ODROIDs: C2 & N2 ; RPIs: 3B & 4B
Has thanked: 0
Been thanked: 0
Contact:

Re: Raspberry Pi 4 vs Odroid N2

Post by boolean »

I don't have a clue to be honest. What does increases thermal load is de presence of a monitor. Both tests have been performed with a 4K screen attached to the micro HDMI interface. If you would run the test over SSH without a screen attached, it will show higher performance numbers and even lower temperatures.

powerful owl
Posts: 170
Joined: Thu Mar 28, 2019 8:57 pm
languages_spoken: english
ODROIDs: 2 x HC1, 3 x H2, 1 x H2+
Has thanked: 41 times
Been thanked: 18 times
Contact:

Re: Raspberry Pi 4 vs Odroid N2

Post by powerful owl »

I'm not able to run sbc-bench with a desktop (HK Ubuntu Mate), load average never drops to the required 0.1...

User avatar
rooted
Posts: 8034
Joined: Fri Dec 19, 2014 9:12 am
languages_spoken: english
Location: Gulf of Mexico, US
Has thanked: 724 times
Been thanked: 252 times
Contact:

Re: Raspberry Pi 4 vs Odroid N2

Post by rooted »

powerful owl wrote:I'm not able to run sbc-bench with a desktop (HK Ubuntu Mate), load average never drops to the required 0.1...
You can go into the script and change that function to run the test, I had to do this due to a bug that I thought was fixed on the N2.

boolean
Posts: 15
Joined: Thu Aug 31, 2017 8:27 am
languages_spoken: Dutch, English
ODROIDs: C2 & N2 ; RPIs: 3B & 4B
Has thanked: 0
Been thanked: 0
Contact:

Re: Raspberry Pi 4 vs Odroid N2

Post by boolean »

I'm not sure that's a bug. The <=0.1% CPU usage threshold is a bit too ambitious if you ask me. I also had to comment out that function before it would run on Raspbian Buster.

mad_ady
Posts: 8568
Joined: Wed Jul 15, 2015 5:00 pm
languages_spoken: english
ODROIDs: XU4, C1+, C2, C4, N1, N2, H2, Go, Go Advance
Location: Bucharest, Romania
Has thanked: 578 times
Been thanked: 483 times
Contact:

Re: Raspberry Pi 4 vs Odroid N2

Post by mad_ady »

I think rooted is referring to the video decoder modules that keep load above 1 even when not in use.
These users thanked the author mad_ady for the post (total 2):
Tenaja (Sat Sep 07, 2019 6:21 am) • rooted (Sun Sep 08, 2019 6:14 am)

SlappyMcPhee
Posts: 105
Joined: Fri Aug 18, 2017 2:09 pm
languages_spoken: english
ODROIDs: XU4 (3 of them)
Has thanked: 6 times
Been thanked: 5 times
Contact:

Re: Raspberry Pi 4 vs Odroid N2

Post by SlappyMcPhee »

boolean wrote:
Mon Sep 02, 2019 5:20 am
rooted wrote:
Thu Aug 29, 2019 3:05 pm
If you don't want to use the big ass impractical heatsink you don't have to, but to expect it to function properly without it is ridiculous. The trip points are set to work in conjunction with the factory installed heatsink if you want to remove it you are changing the predefined layout of the N2 and all bets are off.

If you remove the heatsink from a PC how long will it function under load? (yes this is the same thing, the N2 was designed to be used with the installed thermal solution)
I'm glad we agree on something, because that's indeed EXACTLY the same thing! If you remove the heatsink of a (modern) PC/Server CPU, it just slows down without locking, crashing or burning itself. Only AMD had thermal issues in the past (they probably still have), but that's not relevant here.

Can you do me a favor by running Thomas Kaiser's script on your N2 and posting the results over here? Maybe I've just received a faulty unit indeed.

Code: Select all

wget https://raw.githubusercontent.com/ThomasKaiser/sbc-bench/master/sbc-bench.sh
SlappyMcPhee wrote:
Sun Sep 01, 2019 3:24 pm
I will give my two cents on the state of the Pi 4 and cooling. It is straight asinine that the Pi Foundation released the board without at least passive cooling especially with how closely the cores are fabbed together on the die. I have a couple already myself and the ENTIRE board becomes so hot without one it is just plain stupid. All in the name of keeping their form factor it would seem. They knew what they were doing, but they did it anyway. It is just like with the USB-C debacle. They knew what they were doing and released to the public anyway. Then the icing on the cake is that they knew how they were going to have thermal issues with the design because they downclocked the cores to 1.5. What I find laughable is how so many out in the wild call upclocking to 1.75 or 2.0 an OC....lmao.

So from the standpoint that the cores are supposed to run at 2.0 natively and they can barely run at the 1.5 without thermal issues for the entire PCB this whole attempt to knock the N2 on cooling from NIB seems utterly pointless.
Maybe you should follow a comprehensive reading course, because you're totally missing the point here. It would have been nice if the N2 was designed to be able to run without a heatsink, or at least had a more refined/practical one so its overal size wouldn't have been doubled. Removing its heatsink voids warranty, so you shouldn't temper with it (like I did).

I've got multiple SBC's, but I bought the N2 specifically for my bedroom to use with RetroPie (or a fork) and LibreELEC. Unfortunately, because of the impractical heatsink, it's too thick to fit behind the TV. Removing the heatsink or replacing it with a smaller one didn't provide stable operation, so I needed to look into other options as well.

I did a bit of research and decided to bet on a RPi4 and ordered one including a case with passive cooling. The Raspberry is about half the height of the N2 (both inside their case), so this fitted like a charm. Clocked @ 2GHz and at full load (sbc-bench) it won't get any hotter than 54.5°C, so that won't be an issue either. If you know how to read numbers, here are the results: http://ix.io/1TZA ;)

Having said that, performance wise I still prefer the N2. Its GPU is a lot beefier than the one packed into the RPi4. It's just a shame that it won't fit my use case and I can't make it fit without loosing warranty.
I think that you are missing the point of the product. The fact that they are touting the ability to emulate on it and then also use it as a media centre machine is only a small part of their clientele. Hard Kernel's boards are used in primarily industrial, aerospace, and other applications. The fact that HK releases the board with the ability and some support for media centre and emulation is really a plus. hell, the raspberry pi wasn't released as a primarily emulation/media computer. It is just a blessed circumstance that someone made it work. this reaction just seems kind of entitled because it won't fit behind your TV. let's be straight up here. NONE of the SBC companies design their boards with any design sense for gaming especially. This is practically screamed at you with how as an example the Pi Foundation redesigned the HDMI output with the pair of mini HDMI ports for desktop 4k or duel screen functionality is proof of that. There has been a slew of retrogaming designed cases for the Raspberry Pi for YEARS and now their new design and especially it being shadow dropped is proof that even though they think retrogaming on it is cool and all that isn't their focus. Really even as standalone media players they aren't so appealing from a design standpoint. Port placement especially. True gamers don't use wired controllers because they are cheaper. They do it for input response and latency purposes. If someone to complain about the N2 about design choices it would be that there are no UBS ports on the front therefore using wired controllers is a bit of a cludge. The fact that you can't fit it behind your TV to be mounted is a small bit as that is more so a design situation on your part. I have over a half dozen TVs in my home. They are ALL wall-mounted and even with slim mounting kits there is still plenty of room. Hell, I could even fit a PS4 if I wanted to mount it. Yes, the N2 board design does make things a little challenging from certain perspectives, but my take is that your perception is that all SBCs should meet a form factor like the Raspberry Pi in which case once again their design is far from idea. At least with the N2 you can run it with the simple case and it really isn't so bad looking. You can easily do something to mod it to be more appealing. It just takes some creativity. For many that is part of the fun. At the end of the day the community has actually been spoiled to the point of some entitlement with the amount of hardware support the Raspberry Pi has received. Like I said the N2 does have its design flaws, but adapt and overcome. The design of it will not change at this point. Lastly your bench testing by no means actually replicates real life scenarios. We have been developing our retrogaming build since we received our kits for the N2 in the spring and run into nary an issue when it comes to throttling. The fact that it is silent is a huge plus.

Essentially the complaint about it being "too tall" and can't behind my TV" is a first world problem. It is the same as the fact that the Pi4 can't be used in the current cases from Retroflag, Kintaro, etc.
Owner The Retro Arena and Odroid Retro Arena

powerful owl
Posts: 170
Joined: Thu Mar 28, 2019 8:57 pm
languages_spoken: english
ODROIDs: 2 x HC1, 3 x H2, 1 x H2+
Has thanked: 41 times
Been thanked: 18 times
Contact:

Re: Raspberry Pi 4 vs Odroid N2

Post by powerful owl »

rooted wrote:
Fri Sep 06, 2019 5:55 pm
powerful owl wrote:I'm not able to run sbc-bench with a desktop (HK Ubuntu Mate), load average never drops to the required 0.1...
You can go into the script and change that function to run the test, I had to do this due to a bug that I thought was fixed on the N2.
Thanks, I wasn't even thinking, I commented out the call to CheckLoad, @boolean here is the result in ambient ~17, Armbian desktop running 1080p with no windows open - http://ix.io/1UtB

powerful owl
Posts: 170
Joined: Thu Mar 28, 2019 8:57 pm
languages_spoken: english
ODROIDs: 2 x HC1, 3 x H2, 1 x H2+
Has thanked: 41 times
Been thanked: 18 times
Contact:

Re: Raspberry Pi 4 vs Odroid N2

Post by powerful owl »

For comparison, HC1 with same type heatsink design - http://ix.io/1UtW

(Single unit unstacked, case, no drive, Armbian Buster)

powerful owl
Posts: 170
Joined: Thu Mar 28, 2019 8:57 pm
languages_spoken: english
ODROIDs: 2 x HC1, 3 x H2, 1 x H2+
Has thanked: 41 times
Been thanked: 18 times
Contact:

Re: Raspberry Pi 4 vs Odroid N2

Post by powerful owl »

SlappyMcPhee wrote:
Sat Sep 07, 2019 3:53 am
Hard Kernel's boards are used in primarily industrial, aerospace, and other applications.
Do you have any examples? It would be interesting to see how they are used.

boolean
Posts: 15
Joined: Thu Aug 31, 2017 8:27 am
languages_spoken: Dutch, English
ODROIDs: C2 & N2 ; RPIs: 3B & 4B
Has thanked: 0
Been thanked: 0
Contact:

Re: Raspberry Pi 4 vs Odroid N2

Post by boolean »

SlappyMcPhee wrote:
Sat Sep 07, 2019 3:53 am
I think that you are missing the point of the product. The fact that they are touting the ability to emulate on it and then also use it as a media centre machine is only a small part of their clientele. Hard Kernel's boards are used in primarily industrial, aerospace, and other applications. The fact that HK releases the board with the ability and some support for media centre and emulation is really a plus. hell, the raspberry pi wasn't released as a primarily emulation/media computer. It is just a blessed circumstance that someone made it work. this reaction just seems kind of entitled because it won't fit behind your TV. let's be straight up here. NONE of the SBC companies design their boards with any design sense for gaming especially. This is practically screamed at you with how as an example the Pi Foundation redesigned the HDMI output with the pair of mini HDMI ports for desktop 4k or duel screen functionality is proof of that. There has been a slew of retrogaming designed cases for the Raspberry Pi for YEARS and now their new design and especially it being shadow dropped is proof that even though they think retrogaming on it is cool and all that isn't their focus. Really even as standalone media players they aren't so appealing from a design standpoint. Port placement especially. True gamers don't use wired controllers because they are cheaper. They do it for input response and latency purposes. If someone to complain about the N2 about design choices it would be that there are no UBS ports on the front therefore using wired controllers is a bit of a cludge. The fact that you can't fit it behind your TV to be mounted is a small bit as that is more so a design situation on your part. I have over a half dozen TVs in my home. They are ALL wall-mounted and even with slim mounting kits there is still plenty of room. Hell, I could even fit a PS4 if I wanted to mount it. Yes, the N2 board design does make things a little challenging from certain perspectives, but my take is that your perception is that all SBCs should meet a form factor like the Raspberry Pi in which case once again their design is far from idea. At least with the N2 you can run it with the simple case and it really isn't so bad looking. You can easily do something to mod it to be more appealing. It just takes some creativity. For many that is part of the fun. At the end of the day the community has actually been spoiled to the point of some entitlement with the amount of hardware support the Raspberry Pi has received. Like I said the N2 does have its design flaws, but adapt and overcome. The design of it will not change at this point. Lastly your bench testing by no means actually replicates real life scenarios. We have been developing our retrogaming build since we received our kits for the N2 in the spring and run into nary an issue when it comes to throttling. The fact that it is silent is a huge plus.

Essentially the complaint about it being "too tall" and can't behind my TV" is a first world problem. It is the same as the fact that the Pi4 can't be used in the current cases from Retroflag, Kintaro, etc.
I take back my words, you shouldn't follow a comprehensive reading course as it would be a waste of money since they won't be able to get through to you. Are you having a dialogue with yourself? Non of your rattling is relevant.

The only point that I was trying to make is that it would have been nice if the N2 had been designed to able to run without a heatsink as well (without voiding warranty), so the end user could tailor it to its needs. My use case has nothing to do with it, that was merely an example.

The Raspberry has "design flaws" too, but that's also not relevant here.

boolean
Posts: 15
Joined: Thu Aug 31, 2017 8:27 am
languages_spoken: Dutch, English
ODROIDs: C2 & N2 ; RPIs: 3B & 4B
Has thanked: 0
Been thanked: 0
Contact:

Re: Raspberry Pi 4 vs Odroid N2

Post by boolean »

powerful owl wrote:
Sat Sep 07, 2019 2:37 pm
rooted wrote:
Fri Sep 06, 2019 5:55 pm
powerful owl wrote:I'm not able to run sbc-bench with a desktop (HK Ubuntu Mate), load average never drops to the required 0.1...
You can go into the script and change that function to run the test, I had to do this due to a bug that I thought was fixed on the N2.
Thanks, I wasn't even thinking, I commented out the call to CheckLoad, @boolean here is the result in ambient ~17, Armbian desktop running 1080p with no windows open - http://ix.io/1UtB
Thank you for taking the time and effort to run the bench! I'll compare the results to my N2 once I get home.

cjheath
Posts: 10
Joined: Sat Feb 22, 2020 4:03 pm
languages_spoken: english
ODROIDs: None yet. Want N2 for Software Defined Radio app
Has thanked: 0
Been thanked: 0
Contact:

Re: Raspberry Pi 4 vs Odroid N2

Post by cjheath »

Hello all, newbie here.

I'm considering an N2 (as one of other options) to run a remote (powered by PoE) Software Defined Radio. This means putting it into a waterproof diecast box with some other stuff, up on an antenna tower.

If the diecast box has a sufficiently large flat surface area, could I remove the ODroid's heatsink and sandwich the diecast box with the N2 inside and the heatsink it came with outside? That is, is the underside of the ODroid heatsink flat, or sculpted?

Otherwise are there any other options for after-market heatsinking for the N2?

The other stuff in the box will be PoE splitter, antenna relays, LNA, filters, and one or two SDR modules (RTL-SDR and HackRF). It's the HackRF that needs the compute power of the N2.

Clifford

skeetre
Posts: 71
Joined: Mon Apr 29, 2019 11:27 am
languages_spoken: english
ODROIDs: Odroid-N2
Location: FL, USA
Has thanked: 16 times
Been thanked: 11 times
Contact:

Re: Raspberry Pi 4 vs Odroid N2

Post by skeetre »

A little late, but I did mean to share this a long time ago:
https://openbenchmarking.org/result/190 ... -190712762
Friends don't let friends use Spectre/Meltdown/Zombieload mitigations.

o_lampe
Posts: 175
Joined: Wed Jan 01, 2020 12:35 am
languages_spoken: english german
ODROIDs: MC1 solo
Has thanked: 11 times
Been thanked: 16 times
Contact:

Re: Raspberry Pi 4 vs Odroid N2

Post by o_lampe »

cjheath wrote:
Sat Feb 22, 2020 4:14 pm
Hello all, newbie here.

I'm considering an N2 (as one of other options) to run a remote (powered by PoE) Software Defined Radio. This means putting it into a waterproof diecast box with some other stuff, up on an antenna tower.

If the diecast box has a sufficiently large flat surface area, could I remove the ODroid's heatsink and sandwich the diecast box with the N2 inside and the heatsink it came with outside? That is, is the underside of the ODroid heatsink flat, or sculpted?

Otherwise are there any other options for after-market heatsinking for the N2?

The other stuff in the box will be PoE splitter, antenna relays, LNA, filters, and one or two SDR modules (RTL-SDR and HackRF). It's the HackRF that needs the compute power of the N2.

Clifford
I'm sure, you would get more responses to your question in a separate thread.
My opinion: no, you can't connect the stock heatsink outside a diecast housing (without cutting off the edges.)

AFAIK, the N2 doesn't run with PoE?

cjheath
Posts: 10
Joined: Sat Feb 22, 2020 4:03 pm
languages_spoken: english
ODROIDs: None yet. Want N2 for Software Defined Radio app
Has thanked: 0
Been thanked: 0
Contact:

Re: Raspberry Pi 4 vs Odroid N2

Post by cjheath »

Thanks, I got my answer in a separate thread. I can remove but not re-use the N2 heatsink.
Initial experiments will be with a C2, since I could get one quickly and cheaply. That needs a 15mm thick (or more) aluminium block between the chip and the case.
I'm planning on splitting off the PoE and converting to clean 5V before feeding USB and power to the ODroid.

Pendragon11
Posts: 6
Joined: Sat Apr 11, 2020 1:22 pm
languages_spoken: english
ODROIDs: N2
Has thanked: 0
Been thanked: 0
Contact:

Re: Raspberry Pi 4 vs Odroid N2

Post by Pendragon11 »

Howdy, I'm a pi4 convert. On the pi, I've been running raspbian, openvpn, a deluge torrent client, and a Plex server. The pi was awful for decoding video. May someone point me in the direction for doing this on the n2 or in tandem with the pi4? Should I go Ubuntu? Do I go coreelec (but what about openvpn, plex and torrents)? Thank you!

Post Reply

Return to “General Topics”

Who is online

Users browsing this forum: No registered users and 2 guests