Pine64 SOQuartz Blade Review

Pine64 SOQuartz Blade Review
Pine64 SOQuartz Blade Review

I’ve previously reviewed the Pine64 SOQuartz here on the site and found that if you used Armbian it was a feasible alternative to the CM4. The compatibility with other IO boards varies wildly though. We’ve definitely seen mixed experiences using the CM4 Blade when using the SOQuartz for example.

I was curious if I got a hold of one of the official Pine64 SOQuartz Blade boards if the experience would be better. Today we will review the Pine64 SOQuartz Blade IO board and see how well things are working at time of writing.

Let’s begin!

Hardware Used

Pine64 SOQuartz - Bottom View
Pine64 SOQuartz

The Pine64 SOQuartz has a 1.8GHz ARM Cortex-A55 processor as well as a Mali-G52 Bifrost GPU. Onboard ram is LPDDR4 and has variants of 2GB, 4GB and 8GB available.


Pine64 SOQuartz Blade
Pine64 SOQuartz Blade

The Pine64 SOQuartz Blade is an official Pine64 expansion board meant to provide a 1U compatible Blade-size I/O board (very useful for data centers as well as clusters).


Geekworm Copper Heat Sink Set
Geekworm Copper Heat Sink Set

The Geekworm copper heat sink set is designed to fit many different single board computers. It uses thermal conductive adhesive which many “cheap” heat sink kits for SBCs don’t have. Eliminates hot spots and reduces throttling. Can be further enhanced by powered cooling over the heat sinks.



These are from the Pine64 wiki page here.

  • SOQuartz BLADE Baseboard Dimensions: 200mm x 40mm x 15mm
  • Input Power:
    • DC 12V @ 3A 3.4OD/1.3ID (IEC 60130-10 Type E) Barrel DC Jack connector
    • PoE
  • Storage
  • microSD – bootable, support SDHC and SDXC, storage up to 2TB
  • USB – USB2.0 Host port (with header for setting OTG ID pin)
  • Expansion Ports
  • 2×20 pins “Pi2” GPIO Header
  • M.2 slot
  • PWM fan header
  • Jumpers
  • OTG ID jumper
  • GPIO voltage, select 3.3V or 1.8V
  • PoE Enable

Build Quality

Let’s start with the packaging. The blade ships in a clear hard shell like this:

Pine64 SOQuartz Blade Packaging
Pine64 SOQuartz Blade Packaging

Everything is wrapped in an anti-static bag. I’ve mentioned it before in my other reviews but I’m really happy to see all of the manufacturers moving away from shipping these in little boxes like Raspberry Pis used to/still are. They always got crushed. This will make it to you intact.

Now let’s take a look at the blade itself:

Pine64 SOQuartz Blade - Top View
Pine64 SOQuartz Blade – Top View

Here you can see where we mount the M.2 NVMe drive. We can also see a jumper labeled USB_OTG_ID.

There’s a fan connector on the right hand side of the blade as well as a HDMI port. SD Card slot, UART and a USB 2.0 port are on the left as well as the PoE Ethernet port.

Now let’s look at the bottom of the board:

Pine64 SOQuartz Blade - Bottom View
Pine64 SOQuartz Blade – Bottom View

As you can see there isn’t much to see on the bottom here. It’s well built and sturdy like all Pine64 devices I’ve used.

Testing the Blade

I first prepared a SD card on it with the Armbian image available from the Pine64 wiki here. Next I took the SD card and put it in the Pine64 Blade. I connected the HDMI as well as a keyboard and then plugged in a network cable connected to a PoE port to see of it would power itself purely from the Ethernet connection.

Here was the result:

Pine64 SOQuartz Blade - First Boot
Pine64 SOQuartz Blade – First Boot

Excellent, no power cords! Everything worked the first try. The WiFi is working but you must have the external antenna attached like I do in the picture. It won’t work well without it.

I went through the initial Armbian configuration and finally it’s time to test some NVMe performance with the SOQuartz!

Performance Benchmarking

You can verify the performance of your SSD on Pi Benchmarks using the following command:

sudo curl | sudo bash

Here are the results:

     Category                  Test                      Result     
HDParm                    Disk Read                 377.76 MB/s              
HDParm                    Cached Disk Read          371.45 MB/s              
DD                        Disk Write                144 MB/s                 
FIO                       4k random read            33684 IOPS (134736 KB/s) 
FIO                       4k random write           7064 IOPS (28258 KB/s)   
IOZone                    4k read                   83772 KB/s               
IOZone                    4k write                  39826 KB/s               
IOZone                    4k random read            43358 KB/s               
IOZone                    4k random write           80605 KB/s               

                          Score: 14,559

The full Pine64 SOQuartz Blade benchmark can be viewed here on Pi Benchmarks.

This is PCIe 2.0 performance we are seeing out of the drive. This gives excellent speeds and is using true NVMe! It’s not PCIe 3.0 performance but I definitely would not expect that at the price point of the SOQuartz (which the CM4 it’s competing against doesn’t do PCIe 3.0 either).

One of the few ways I know of to get PCIe 3.0 performance in the single board computer world right now is to jump up to a Rock 5B.

Pros / Cons


  • Only costs $29
  • Supports PoE (power over ethernet)
  • Supports NVMe and has a dedicated M.2 slot with mounting sizes 2230-2280


  • Image support continues to be lacking while Pine64 works toward mainline kernel support (use Armbian in the mean time)


It’s not clear to me how much the kickstarter Blade is even going to cost or when it’s going to be more generally available. If this is already available to buy somewhere I apologize but I legitimately couldn’t find where to get one (or anyone selling one on eBay or anything like that).

EDIT 2/11/2022: This has since been made clear to me. It will be about $64 for the basic version at time of writing. The full kickstarter page can be viewed here for more details. There are different packages available for the kickstarter project intended for the CM4.

This product however is very real. You can buy it today for $29.99 from Pine64. I definitely recommend pairing it with the SOQuartz if you get one. I haven’t tested it with the CM4 yet to see if everything works properly but I will do that in a future article. I’d be surprised if the kickstarter CM4 board doesn’t cost more than the Pine64 SOQuartz + Blade combined.

I really like that I can use this without having to use up any of my power adapters or AC plug spots or anything else. I have plenty of PoE-enabled network ports around so it’s very likely I’ll continue to use the Pine64 Blade in a setup like this as a server. I may buy a larger NVMe drive to put in there like a Crucial P3 4TB drive* and make it a storage server.

If you are investigating wanting to use one for a server or multiple of these for a cluster I’d definitely recommend picking up a Blade + SOQuartz and seeing if it may serve your use case. I’ve had great experiences with Pine64 with my Pinecil as well as this blade and they make a lot of other cool stuff / are a great company.

The software/image support is lacking but they themselves will tell you that (and they do if you read through their wiki pages) as they are working toward mainline kernel support. That will be a glorious day for the SOQuartz and other boards they are working toward this goal for once it’s ready (and from the wiki tracker they’ve made significant progress but you can see some things still need porting such as the fan driver). Therefore I do not recommend this for absolute beginners at this time as the image support is limited.

In other words it’s still a bit early to buy in so I recommend it for developers or experienced users mostly. A lot of you out there that might be waiting or trying to get your hands on that other blade are probably more than qualified to make this work as well as I have here. If you have enough experience to be working with the Compute Module 4 though realistically you’ll probably be fine using Armbian with it.

Armbian still seems to be the best choice to use with the SOQuartz (especially for people just starting out with the board) but there are other options like the Manjaro (requires USB to TTY serial console). I’d definitely recommend trying it for anyone who is comfortable with compute modules (and using a USB to TTY serial console) and has a use case for a blade setup like this!

Other Resources

Don’t miss my Pine64 SOQuartz review here

I’ve also reviewed the Pinecil soldering iron from Pine64 here


Inline Feedbacks
View all comments

Darren Williams
Darren Williams
21 days ago

I have had 6 of these blades running now for a few months with production workloads, they work great for what I want. I want a linux server with nvme and ethernet, that will do me just fine. They are good for pimox and k3s and that will fo for me. At the price they are, you can’t go wrong.

Darren Williams
Darren Williams
20 days ago

Here’s a pic of them in my dirty cab with the 19 inch rackmount I 3D printed for them I’m currently working on building a 10inch rackmount, 2U high with 80mm fans for cooling:

David T.
David T.
1 month ago

As an early backer on several Pine64 projects I can say I will never buy one again. They regularly release cheap “clones” of things they see that are popular and never fully back it with software support. I backed the original Pine64 and the Pine Stamp which has since been rebranded to the Padi probably to escape all the criticism they received. Their hardware and software has always been second rate.

You’ve criticized Ivan for having multiple revisions over the course of some time. Do you have any clue how hardware development works? Of course it has multiple revisions. He is just doing it in the public unlike most companies. I’ve received the blade personally and am the US distributor for the beta blades ( and can tell you that the product is real and it works. JetBrains is already using it in their production environment to serve users on a daily basis. Can the same be said for the soquartz-blade?

There are so many uninformed and just bad takes in this blog that I cant imagine you are not being paid to shill Pine products. Maybe you should take the fact that no one is talking about it as an indicator that no one is interested in buying it because the reputation of the Pine group is not a good one. Best of luck in your endeavors tinkering to get the soquartz-blade to work. I flashed an os and have been using my compute blade with full software support for 2 weeks now without issues.

Jeff Geerling
Jeff Geerling
1 month ago
Reply to 

I think we should take a break on this…

Ivan has them running at JetBrains because that’s where he works. He builds things like these blades, custom Mac mini brackets for mounting and remote controlling huge numbers in racks, etc. as part of his work there.

He’s been running various versions of the Blade testing them in production at his day job to make the boards better, and wasn’t comfortable going to production, ordering thousands of units, because he likely doesn’t have hundreds of thousands of dollars to risk on production for a board that isn’t being built by a major corporation.

Even with companies like JLCPCB etc, you order a batch, then you realize manufacturing requires a lot of little tweaks for better reliability. You order another batch, then you realize a component doesn’t work because some traces are too close to something else. You order another batch, then realize the placement of something is just off when you try putting it in an enclosure… etc. etc.

He was doing that almost entirely on his own, and finally opened up a small batch to outside testers once he was comfortable getting more feedback before going to a kickstarter. Here’s the whole production history.

Ivan’s been incredibly open throughout the design process, detailing every step on Twitter, Instagram, his website, and now also on Discord as he’s gathering feedback from those who are participating in the Kickstarter.

Jeff Geerling
Jeff Geerling
1 month ago

You too! And hopefully he’ll be able to sell the Blades after the Kickstarter somehow, maybe through a partner like Seeed or DFRobot or something… I would love to see custom boards like it available in general sales just like Pine64 has been good about doing.

Ralph Hightower
Ralph Hightower
1 month ago
Reply to  Jeff Geerling

I hope to buy Ivan’s blade after Kickstarter. Right now, I’m putting together another Kickstarter projector, the TuringPi cluster. I’ve got two NVIDIA Jetsons and want to fully populate it. But there’s $$$ involved. The add-ons look more complex than choosing options in ordering a car.

Razor Burn
Razor Burn
1 month ago

Hi James,

Its great to see you pumping out the content and the Pine64 SOQuartz Blade expansion board looks to be an affordable alternative to the unobtanium Raspberry Pi CM4 with some welcome additions such as PoE and actual PCIe 2.0 NVMe speeds.

Again the lack of software options may put some users off (myself included due to inexperience working with say Manjaro) but Pine64 deserve praise for the efforts being made to get mainline kernel support for their devices and the build quality is top notch for what you can expect to pay but I wish they had a few more distributors as most of the time its a struggle finding stock of accessories or boards from the official site and the 30 warranty is disappointing unless you purchase from AmeriDroid or one of the EU sellers who offer extended warranty as per local consumer laws.

I look forward to seeing how it pairs with the RPi CM4 in a future post and until then keep up the outstanding work! Cheers from Australia…

Jeff Geerling
Jeff Geerling
1 month ago

I can confirm the Compute Blade is real, the production of the board is already arranged (Ivan is not a fly-by-night scammer AFAICT… he’s been ver communicative on Twitter, Instagram, and Discord), but of course as with any Kickstarter there is always the risk of it not being shipped.

The reason it has taken so long is the entire project was conceived of, developed, tested, and produced by one individual (Ivan). He produced a batch of 100 or so ‘RC2 boards’ and sent them to people who pre-ordered pre-production units a few months ago. Most of those I can confirm shipped and are in use (except for a few where there were weird shipping exceptions), and many of the beta testers have provided feedback via Discord leading to well-documented support for all board features already.

I put a lot of faith into this thing (enough that I made a 2nd video on it once I was able to test the RC2 version of the board!), and I have no financial incentive in the production or sales of the blades—I actually ordered the 10-node cluster so if it goes belly-up I’m out $1000 🙂

The reason Ivan put the high goal is he could not produce *any* blades without meeting a 7.5k production quota, with the manufacturer he is working with. And he also secured a batch of CM4s that he is selling for a slight markup because after months of trying to get them from Raspberry Pi (and failing at that), he finally bit the bullet and paid for a batch from a reseller at cost (plus shipping).

But as I mentioned in my video a couple weeks ago, other boards like the SOQuartz do work with this blade. And the CM4 also worked in the SOQuartz Blade (which was a pleasant surprise).

Finally, as to the 100,000k overnight—following along in the Discord, there were a lot of orders where a business that is planning on using the blades + rack units had to put in an order through their typical bureaucratic process… and I know more than one hosting company who is planning on deploying at least 1U of the blades for CI or other hosting purposes.

The project had over 6,000 people signed up to get notified once the Kickstarter went live—seeing 729 backers at this point, that seems like a typical ratio of people who are interested to those who purchased boards.

Murray The Goz
Murray The Goz
1 month ago

Some more insights on KS campaign (from a beta tester)

The only blame I would cast on Ivan is the habit of having complete control on the whole process. And being a damn perfectionist. He made a custom PCB just for the fan units 🙂
“old school” you say, and I can somehow agree with you. Ivan is not backed by a big company as a Pine64, and he feels all the responsibility to deliver a quality product.

About custom ordered PCB: the first batch of CARDS (not PCBs) sent to beta testers was checked by him one-by-one (and the fab he used made some mistakes) since they are not just plain PCBs but also have quite a number of components on them.
A big-backed company can afford a “oh the card was bad, have a replacement and throw away the 30$ card we don’t care” since they work in far bigger numbers.
Having 7k with PCB, components and reliable soldering lead to that price. Maybe with 100k (or with a company like Pine who gives all kind of cards) I guess the number could be way more lower.

About the financing: 400k was done in the first 12hours from the campaign launch, and my best guess that quite a portion of this was raised by the enthusiast beta-testers who put their hands on this and had a chance to check this is quite real. 100k was added following the common logarithmic curve in this case.

A good point and a bad point on you:
First, thank you so much for showing a good Raspberry alternative. I’d like the idea of having a low power (even lower than CM4 due to A55 vs A72 cores) card to run a cluster.
Sofar my experience with RPI clones was limited to Orange PIs and they run quite well.
For the others the usual was “ok we made a good HW product, now let people get mad on SW”
First time I read of Soquartz the sw support was “not so brilliant “. Happy to see it evolved (the sw page is still a mess, but I trust you). And you almost convinced me to give this one a shot (as soon as I have funds)

bad point: I got scammed quite some times from IGG/KS, and I agree the campaigns should be chosen wisely nowadays. Starting insulting an unknown person “for a scam backed off by a famous youtuber” is a lack of background check which is, at least in this case, widely available. And this makes me a bit upset.
Besides, you had the nerve to apologize, and this is a good thing.

BTW, I again thank you for the good work on alternatives. Alternatives always show you there is a different way of doing thing, and sometimes the different way is overall better, or it fits better your needs.

thank you. take care!

[using my nick I use in cyberspace to comment… real name is in the mail 😉 ]

Razor Burn
Razor Burn
1 month ago

It appears that the Compute Blade has been getting some added attention with some posts on popular sites such as CNX and an interview with the guys on Tom’s Hardware Podcast for those interested where he explains the project.

Razor Burn
Razor Burn
1 month ago

You make some strong arguments and I’ll clarify this type of device isn’t really targeted at me so what do I know but judging by the funds raised on Kickstarter and the interest this blog has attracted there clearly is a market that largely remains untapped.

Then again I’m in two minds as my initial thought is that the Pine64 version/s exist yet rarely get mentioned and you can argue its down to the poor reputation for software support, although your reviews show its more than capable for its purpose. They obviously have done a poor job marketing both devices as I’ve only really seen yourself, Jeff and maybe Jean-Luc from CNX review the SOQuartz/Blade and the odd mention elsewhere that it was in the works but took little notice as its not something I see myself really using and judging by comments the consensus is that the idea is fantastic (largely from backers/users) or mixed (needs some work/why bother/expensive or well priced).

I agree that Radxa could easily jump in and “design” something similar and it would certainly help push sales for their own Compute Modules but I assume they’re in a similar boat to RPi and can’t get enough supplies to manufacture the volumes necessary to make the project viable and I see Pine64 showing stock on their shop so I just don’t see the mad rush to buy this type of device but the hype for the Compute Blade is real but I just don’t see how revolutionary it is compared to the Pine64 one that anybody can buy now…?

You’re correct that the time is now and surely somebody in marketing at Radxa has seen the attention Ivan has garnered and with reports that his units are scheduled to ship August 2023 (all going to plan) there is plenty of time for a clone to swoop in take advantage of the hype (real or manufactured) and I wouldn’t be surprised to see that happen soon as the concept is clearly achievable and wouldn’t cost too much to manufacture in bulk quantities for one of the bigger players so watch this space but what I do know is that the project is real and has been a success with pressures to get units out their and into backers hands.

You can be proud knowing that your blog was one of the only few covering this as many have since popped in to show faint interest and the comments have been spicy but you’re right to question things and I’ll happily admit I was wrong to be sceptical and wish Ivan well as the pressure is really on to get units ready and into peoples hands and there’s nobody to blame if we see newer clones released as his success is sure to attract interest from other vendors/manufactures looking for a piece of the pie…

Great work covering these obscure devices and hopefully somebody at Pine64 reaches out to you for advice as you’re one of a handful of reviewers willing to purchase their units and share honest reviews so surely they could listen as they’re ahead of the trends for some products but largely ignored so makes sense to seek the opinion of a respectable reviewer/user but I’m not holding my breath nor suggesting you go down that avenue as I appreciate the transparency and honesty that your blog brings to the tech community… Take care!

Razor Burn
Razor Burn
1 month ago

Funny you mention the viability of RK3588 as I recall seeing work teased by Radxa (CM5), Banana Pi (RK3588_CV1 SoM), Forlinx (FET3588-C SoM) back in 2022 so we’re nearly there but pricing will play a big factor as so far much of the devices are too expensive, with the exception of the Orange Pi 5 and some extent Rock 5B and as you’re well aware the mainline kernel support just isn’t there yet but its only a matter of time before we see replacements for the overpriced CM4 which is non existent!

You’re spot on about people needing to look at things objectively as this tribalism does nobody any good as Coke and Pepsi are equally bad for you (jokes aside) and RPi, Radxa, OPi could care less whether you own 10 of their boards or none as it all comes down to $$$ and RPI has demonstrated that it values its industrial clients more than the average consumer so from a personal perspective they can no longer rely on my business as talk is cheap plus the alternatives are out there be they ARM or X86 and for MCUs you have so many options such as ESP32 that means one doesn’t need to rely on a Pico for smaller projects but thankfully supply and pricing still makes the Pico worth it so the RPi Foundation get some praise from me on that aspect.

Its great to see the comments coming in at the OPi 5 Thread as people feel confident enough to share useful tips which is great to see and I’m delighted that somebody has worked out how to get Home Assistant working as its something I’d be keen to do so great work! You’ve built a great platform here so don’t let the detractors but you off as its better to be true to oneself than sellout and I’m not ashamed to say that I have learnt a lot these last few months having frequented the site on a more regular basis. Thanks for sharing the links and I’d be happy to see you do a future review if/when the “Blade” becomes available for general sale as it would make for a great read but in the mean time stay safe and take care!

Razor Burn
Razor Burn
1 month ago
Reply to  Jeff Geerling

Thanks for the detailed clarification Jeff and apologies to Ivan and yourself for any offense as my comment meant no harm and I truly wish Ivan and his backers success in seeing this project through where other startups have failed.

Razor Burn
Razor Burn
1 month ago

I recall seeing mention of the board on one of Jeff’s reviews and took little notice as its not really a product aimed at me but having read your comments and comparison with the Pine64 SOQuartz Blade left me scratching my head as one is a finished product costing $30 USD and is available for purchase whereas the other has seen 9 iterations since 2020 and has managed to raise over 500K USD on crowd funding in large part to the reviews of an influential Youtuber and at time of writing has no finished products and is openly marketing the device as “Powered by Raspberry Pi” and brandishing the RPi logo so questionable marketing not to mention that the individual is a self described “hacker” then what does that tell you about the post sale support or commitment to the longevity of the product…? I see links to similar projects on the social media sites and a few cluster/server projects but hardly original or innovating and again no finished products for sale?!?!

Its not surprising to see very little mention of Pine64 devices as they tend to be targeted to more experienced users and I’d be surprised if Pine64 was sending out “free” devices to influential reviewers as was the case a while back with their first phone/laptop as the general consensus was that the ideas seemed great but unfinished and poor software so not really marketed to your average user but in this case we have a finished carrier board that spec’s wise looks close to the Compute Blade yet thanks to some clever marketing one gets the impression that the Kickstarter campaign is for a much needed device that will transform how you use your RPi CM4 (if you’re lucky to own one) despite other options that cost less and come with the same form factor but then again that’s due to Pine64 not marketing their version strong enough and hopeful your review/s wake people up to alternatives that may otherwise be unknown to a willing audience.

As for fundraising campaigns, I’m highly sceptical for a host of reasons and feel the risks out weight any potential savings and knowing that Kickstarter takes a cut if something fails should be enough to make you stay away and in this case 500K is more than enough to get production started for what is ultimately a carrier board and I’m sceptical they’ll manage to source the required number of RPi CM4 compute modules adding to any delays especially if RPi get wind that they’re using their logo openly but buyer beware and only time will tell if it actually sees the light of day or backers get swindled out of their funds…?

Of all the internet sites/channels out there I hold yours in high regard with the few that you mentioned as the content is educational, well researched and the fact you’re not obligated to any special interests fills me with confidence as I know I’m getting a genuine review/guide that isn’t afraid to critique so I hope the audience understands your message as its important as the last thing we need in the industry is to see another start up fail as we’re looking at significant amounts of money for what Pine64 have shown is achievable and a more than suitable replacement to the unobtanium RPi CM4 module or a substitute if you already have one and seek a carrier board for use in a server rack at an affordable price point.

You’re right to be upset yet as was the case with the recently launched Orange Pi 5 I’m hopeful you’ll be seeing more attention and recommend that readers go back and checkout your other work as you’ve tested plenty of cool gadgets and popular SBCs including some of the most comprehensive testing of the Raspberry Pi range on the internet so I look forward to seeing how the RPi CM4 pairs with the Blade and as we Australian’s like to say “Keep the bastards honest!” Take care!

Razor Burn
Razor Burn
1 month ago

Hi James,

Apologies to you too for possibly causing a storm with my comments as I never attended to offend Ivan nor Jeff as I vaguley recalled seeing his review video and thinking the board looked cool yet not really my cuppa tea and only reacted after you posted your review of Pine64s alternative and naturally assumed the worst as you hear so many bad stories with crowd funding going bad so I truly wish Ivan success although it seems he chose a hard way at getting the product manufactured…?

I appreciate Jeff taking the time to correct me and in future I’ll be mindful not to name-names as I’d hate to have my ill informed comments take away from the important work that you do as I’m not in the business to besmirch any content creators and whether somebody works closely with a vendor/maker plays little part in my judgement so I’m happy to be corrected whenever I make an incorrect statement and greatly respect the honesty you and guys like Jeff demonstrate so hopefully no hard feelings!?

I’m happy to learn that your OPi 800 and monitor have arrived and will keep an eye out for both reviews, especially the monitor as I’m in the market for a portable display and the suggestion it could be used as a HA monitoring device sounds right up my alley and as great as the Wio Terminal is its somewhat limited to my requirements so I wait to see your findings in a future post… Thanks as always and take care!

Jeff Geerling
Jeff Geerling
1 month ago

Kickstarter is definitely hit-or-miss, and even the best Kickstarters sometimes turn out rotten after years of people having trouble sourcing components, or working with manufacturers, or getting stuck with shipping/customs.

Hardware manufacturing + logistics for a new product is difficult even for companies that have been doing it for years—many people who do a Kickstarter campaign are doing it for the first time! So I often back a project (well, about one every year) mostly because I think it’s a great idea that I’d like to see come to fruition. Less in that I expect to get the product in a timely fashion!

I still have two other Kickstarters I backed last year that are a few months behind in shipping 🙁