Banana Pi BPI-BE1900 Wifi7 Module design

The ASUS has this memory which s**** :wink:

Memory 256 MB Flash and 2 GB RAM

And if most people do not know what they buy (to be honest, they buy a board, a case, a card maybe a modem all as single product - which is not normal at “normal” routers ;)), they should not shout to much

1 Like

PLEASE remember: this is a thread about the BE19! The criticism of the BE14 likely led to the redesign that became the BE19 — I’m quite sure about that! Time will tell how well the board performs across the various compatible BPI boards :slightly_smiling_face:.

I’d also like to see a redesign of the BE14. But that’s not up to me…


“Most people who buy this board aren’t so-called “developers” — they’re regular consumers and tech enthusiasts drawn in by the supposedly powerful hardware specs.”

  • You may be right :slightly_smiling_face:

  • But that doesn’t change the fact that this is a development board. And it’s officially labeled as such:

  • It might become part of your daily setup — but that’s not the board’s purpose. YOU / WE have to make it suitable for that! Getting the BE14 to run (in the way you want) still requires a lot of work (may be it is impossible)!

  • What I mean is: it’s not garbage, because it does work — it just might be useless in your specific case (and for many others). Its Wi-Fi range is significantly shorter compared to common routers.

  • As for general performance: they’re still working on the software

  • You’re projecting your expectations onto the board — I totally understand that — but it’s not quite what it seems.

  • You bought a development board with no shielding on the main board or the Wi-Fi module.

  • And it was clearly declared as an iFEM:grafik

  • If we explore and improve this development board, what we gain is knowledge.

Your criticism is welcome! :+1:

Okay, one question about this one, did you tested it in within the BPI-R4 default case?

And are all antennas necessary? (of course for full mimo setup you need all i think)

Cause with the default case i have four 5g antennas too, so i would need at all 18 antennas :rofl: Looks like if i want to use the BE1900 card i also have to use(and wait for) the new case :roll_eyes: Another option could be that i will use some internal antennas for some stuff :thinking:

Can’t wait for it – and to be honest, the BE14 (signal) quality lacks really if using inside the case :stuck_out_tongue: and cause of dust and to protect the hardware i want to have all in a case :slight_smile:

You already know a kind of price or release date?

As of now, the official SDK update situation for the Banana Pi BPI-R4 with the BE1400 Wi-Fi module is disappointing. The SDK source code provided by BPI-SINOVOIP based on OpenWrt V21.02 has not been updated for over a year. At this pace, by the time the SDK is fully developed, the hardware will likely already be facing obsolescence in the market.

The lack of continuous maintenance on the official SDK results in serious lag in compatibility and feature optimization, leading to a subpar user experience. Although community-maintained projects are relatively active, they remain unofficial and cannot fully replace the timely updates and assurances provided by official support.

This situation reflects the insufficient attention from the official side to this hardware platform and imposes significant uncertainty and risk on users. If the official update pace remains this slow, the BPI-R4 and BE1400 combination will struggle to maintain a competitive edge in the fast-evolving market, making its future prospects bleak.

if you need a high-end router that is “end-user friendly”, dont go with a development board. if you complain so much about the bad wifi experience you have, then why tf dont u go and buy something which is made for end-users, like a good AP? i loved the idea behind a GNU/Linux mobile phone, so i supported the pinephone and bought one. But i KNEW this aint gonna come close to the performance and/or functionality a 50$ android phone delivers. if you buy “the wrong” stuff for your needs, dont shoot at the vendor. there are parts of critic i share, e.g. they do not use and support the official openwrt version and instead forked it. but please dont mix constructive critic with critic that is solely based on your (wrong) choices.

unfortunately, my experience with all chinese companies (and i mean really all) is they dont care about the software. no matter if its a SoC, a wifi chip or something else. And its a known problem, especially with MTK, that they do not release the software under GPL, even if they steal all IP from GPL-based software. And force the companies that use their stuff with NDA’s to not release the stuff as open source.

I appreciate your response and understand where you’re coming from, but I think there’s a bit of a misunderstanding. My decision to buy this board wasn’t because I needed a high-end router or something particularly “end-user friendly.” I’ve used Raspberry Pi from the original up to the Pi 4 and have always been satisfied. This was simply me trying to challenge myself, learn something new, and kill some spare time — a hobbyist experiment, not a production deployment.

That said, I admit I didn’t do enough research before buying, and I had no idea just how bad the Wi-Fi 7 (BE14) implementation would be. It genuinely feels like a product that shouldn’t have been released to the public in this state — not even for developers. Some parts are so poorly executed that it raises questions about whether such a board could ever lead to a polished end-user product. It’s less of a development platform and more of a tech demo with marketing polish.

So while I accept the criticism for my own choice, I think it’s also fair to criticize vendors who release underbaked hardware. It’s not just a matter of wrong expectations — some of these so-called “development boards” seem more like dead ends rather than stepping stones. Just like how some countries might forever stay “developing,” this board seems stuck in that phase too.

3 Likes

you dont want to compare a rpi against the r4. the first prototype of the rpi was in 2006, just so u know. also, it had no wifi and only 100mbit ethernet. comparing the r4 with a rpi is like comparing the pinephone with a samsung galaxy.

when i understood you correctly, all except wifi works as expected. so why dont u get an AP then?

then NEVER buy ANY chinese SoCs, as this is - at least from my experience - always the case. I planned to replace my odroid n2 with a rk chip (e.g. orange pi4 or even pi5 nowadays), but i did not. now guess why. because i checked the mainline support. they work great if you know where the limits are, but if you simply buy them because “oh the spec is awesome!”, then you are doomed.

so you think the rpi1 was perfect? i can tell you, it was not. and only the developers made (rpi in general) that great.

i don’t want to stand up for sinovoip here, but if you buy something, you should check whether the part is suitable for you. and with chinese hardware you always have the problem with the software, no matter if mainline support or updates. i think it’s good that the r4 was released. and now i just hope that sinovoip takes the feedback and creates a new hw version with which everything will be better. → thats why everyone should write an email, if disappointed, and tell them what exactly is wrong.

Yeah, I’ve had the same experience as you. It’s not just BPi — I also bought another “OP” “WiFi 7” board, specifically ChangWang’s Q3600 (IPQ5312), and honestly that thing is a huge trap too. It has one 2.5G port and three gigabit ports, but the most important one — the 2.5G port — is defective. When you connect a 2.5G device, it looks like it’s linked, but in reality there’s no connection at all (I even provided video proof & wireshark to the manufacturer showing that basic ARP doesn’t work). This issue happens, conservatively, about 5 out of 10 times.

When it does happen, if you’ve set the 2.5G port as your WAN, you have no internet — DHCP won’t even work. If you’ve set it as LAN, your client devices lose all connectivity. There’s no software workaround on the router side (restarting the port, rebooting the system — nothing helps). Oh, and sometimes this happens right at boot, leaving you with zero options except physically unplugging and re-plugging the cable.

The manufacturer has never fixed it. They claim you can solve it by flashing certain firmware, but I tried four different versions (including official iKuai 3.18 through 3.20 and third-party OP builds) — none worked. After I sent them testing videos and packet captures for each version, they just went silent. I’m pretty confident this is yet another hardware design flaw. (*update: After I returned the unit, the manufacturer claimed that the English version of firmware 3.20 fixed the problem, but the Chinese version 3.20 had a bug. That honestly sounds pretty hard to believe. Because someone actually went and downloaded the so-called “Chinese version,” only to find it was flashing the English system — they didn’t even get the firmware file names right. When someone in the group commented, “ChangWang’s router firmware just doesn’t have a perfect version,” the group’s tone — and the official response — was: > “There’s no such thing as a perfect product.” > “Look at Windows — after all these years it’s still getting endless updates and patches.” > “Bill Gates is so rich and has so many programmers, yet why hasn’t he made an operating system that’s close to perfect?” That’s just absurd.)

And honestly, even if the 2.5G link works, it’s pointless. iperf and OpenSpeedTest from a client to the device show that — yes, even though it negotiates at 2880 Mbps@160Mhz — the wireless performance is under 1 Gbps. (For comparison, other WiFi 7 gear@160Mhz can exceed 2 Gbps, and even WiFi 6 can do better than this.)

At first, I wondered why, in their official chat group, nobody had posted speed test screenshots for almost a month after release. Then I asked around and realized: everyone had fallen into the same trap and was waiting for the “fix.” As of now, I haven’t seen any sign of a real solution. In fact, they never bring up the issue in the group — instead, they regularly post screenshots showing off how their next product is being designed. If you ask about the problem, they just say it’s been fixed — but if you ask other group members, no one’s board has actually been fixed.

Now they’re saying that not many units were sold domestically (I assume they mean mainland China) and they plan to ship it overseas. I think people should be careful — this isn’t even labeled as a “development board,” it’s sold as a real router (white-label, so you can stick on any brand or logo you like).

Recently I saw another “A***i” (again, wouldn’t mention vender name before I tested it) WiFi 7 product (they claim it’s the world’s first MT7987), but after these experiences I’m a bit allergic now — I’ll wait for someone else’s test results first. That said, the vendor claims, “BPi’s products are never calibrated; ours are fully calibrated. We’re the only DIY router maker that does full wireless calibration.” Looks like BPi’s reputation has really gotten around.

Recently SinoVoip has also released a few new… “case” I’m not sure how that’s supposed to help.

If they could provide real, reliable test results, that would actually be useful.

Of course, if future devices turn out to be uncalibrated, I’ll have to think very carefully before buying.

This product might have been ahead of the curve two years ago, but now I’m not sure they’ll even get WiFi 7 working properly before WiFi 8 hits the market.

1 Like

Can someone explain the purpose of the “2G SM/5G DFS” and “6G/5G-H Filter” antennas on the schematic and can those be omitted without much loss of performance or functionality?

How are things with BE19? I liked that there are 3 separate antennas for each frequency range and a separate chip for each range. I really want to buy R4 Pro + BE19 + a large case with all the necessary cooling kit

We would like to inquire about the release schedule for the BPI-R4 Pro and BE19000. Before the official launch, would it be possible for you to share preliminary test data with us?

We have been waiting for these products for quite some time and are very eager to see them released. Regarding the BPI-R4 and BE14000, the community has already provided valuable feedback. We sincerely hope the upcoming products can be introduced soon while also incorporating improvements to address these concerns.

We look forward to your update and thank you in advance for your support.

I can’t tell you much about the BE19000.

→ It does look better than the BE14000.

But honestly, don’t put yourself through this right now!

→ If you’re not actively developing Wi-Fi 7 and just want to use it with OpenWrt, you should really wait!

This advice also goes out to @Marazm.

I think the BE19000 still needs a year or two to really mature :see_no_evil:.

We’ve had quite a few people here argue that Wi-Fi 7 isn’t new anymore and is already available on the market. And that’s true! But OpenWrt operates under very different circumstances:

https://forum.openwrt.org/t/wifi-6e-wifi-7-or-wait/238414/11

I don’t mean this in a bad way :slightly_smiling_face:, it’s really meant as a friendly recommendation :+1:!

1 Like

Looks better? You mean it’s still junk, but a little better?

No, compared with commercial routers, the design looks promising :+1:

https://forum.banana-pi.org/t/banana-pi-bpi-be1900-wifi7-module-design/23388/35

But I did not see any shielding or cooling. How much time will it take for OpenWrt to get full support … ?

I see tracks where the shields should be soldered. I really hope that there is no shielding in the photo to clearly show what is soldered on the board. Because if you take a photo of a board with shielding, you will see just a metal plate with several protruding elements, this is not very impressive.

I see two possible approaches for shielding and cooling:

However, the R4 Lite is already on the market, and it doesn’t include any shielding or cooling in the package. The board itself clearly has options for adding shielding, though.

What I really don’t understand is why so many people are so focused on the release date.

If I had to guess which product will become stable first, my bet would be on the R4 Pro rather than the BE19. But buying one within the next year will be for those who are willing to tackle both mechanical and software challenges. And even then, success isn’t guaranteed.

I liked the shielding as in the first photo (green board) more. I’m waiting for R4 pro and BE19. Why am I focused on the release date? Because I want to know how soon the R4 pro + BE19 bundle will be released, and if the term is too long for me, I will look for another manufacturer, if not, I will wait.

I really hope that R4 pro and be19 will have decent shielding, I emphasize - not some formal one, but a very good one. Otherwise, we will get another garbage product with a lot of noise, because of which it will be impossible to use the device. After all, the R4 pro + BE19 bundle is very serious, and errors like the lack of shielding are simply ridiculously disgusting.

This is exactly what I am not happy with. If a manufacturer produces a serious device, for a lot of money with capabilities for corporate-level tasks, then this device MUST meet this level. Personally, I can wait until the software for the device is finalized, but it is completely unacceptable if the device is physically faulty. And if the manufacturer, instead of a recall campaign or at least helping to reduce the problems, simply closes its eyes to the problems, then this is a VERY bad manufacturer. What is being sold is not a toy/constructor, but a device with clearly stated characteristics, and BE14 does not meet the characteristics at all. Therefore R4 pro + BE19 will be the last chance for banana pi, if it turns out to be as problematic as BE14, then I will permanently delete all chat history and log out of my account on this site and will never touch banana pi products again, and I will advise all my friends to avoid this company.

1 Like