Nexus 6P will receive at 20MHz, but will transmit at 40MHz or 80MHz, and should be on the 5GHz radio, not 2.4GHz.
What are you talking about here?
"I think i need to start the maintenance of my own mwlwifi fork, without this FCC-related limitations for BW and power."
Which antennas are you using ? I've just ordered Linksys High gain antennas (used ones for a decent price) but im not sure if that was the right pick...
Doesn't the WRT AC Series come with high-gain antennas? The high-gain antennas sold by Linksys appear to be the exact same antennas sold with the WRT AC Series, just with a slightly different aesthetic appearance.
Awhile back I did a side-by-side comparison of the stock and Linksys hi-gain antennas. While there was barely any noticeable improvement they do allow me to reach a bit farther. My opinion is they are worth it, but not at full price.
The stock WRT antennas are 2dBi and the hi-gain are 'up to' 7dBi on 5GHz band.
I bought them to get a range/coverage boost at the same floor (horizontally), hopefully @2,4+5ghz...
All my Clients are at the same floor so i'm looking forward to get at least a little bit better coverage @2,4ghz and maybe a lil performance boost as well.
I'm using a DIR-860L B1 as simple AP to get a decent wifi signal at my garden and it works pretty well with a decent performance.
My WRT3200acm is only responsible for the living room, bed room and kitchen but i struggle abit to get a decent signal at my bed room for netflix and amazon prime...
Read what I said. The 2.4GHz radio wouldn't come up properly if I set it to HT40 on mwlwifi-20180330. It would appear fine in the logs and in LuCi but wasn't visible to any clients. HT20 was my only option.
Originally, when I discovered that^ I presumed it was a regulatory issue and didn't bother looking into it.
I deliberately connected the 6P to the 2.4GHz to test with as it's one of the newer devices I own. 5GHz works fine at 80MHz. Android automatically selects the 5GHz for obvious reasons.
According to LuCI the 6P was only doing 72Mbit/s on 2.4GHz. I'm not sure the 6P supports 144Mbit/s, a couple of my devices do but I've never bothered working out which ones. What I do know is they're all showing 72Mbit/s with mwlwifi-20180614.
What would be helpful is if you could also test the lastest mwlwifi commit and see if any of your devices connect at 144Mbit/s.
@davidc502 question related to using the latest kaloz mwlwifi commit when building my own image.
I updated the PKG_SOURCE_VERSION and PKG_SOURCE_DATE in the package/kernel/mwlwifi/Makefile file to the latest values, but I'm not sure how to find the PKG_MIRROR_HASH to update that. Is this the right procedure and if so where would I find the PKG_MIRROR_HASH value?
Update PKG_SOURCE_DATE and PKG_SOURCE_VERSION to reflect the latest commit. Leave PKG_MIRROR_HASH blank for now.
Then type 'make package/kernel/mwlwifi/{clean,download} V=s'. When it's done type 'sha256sum dl/mwlwifi-2018-06-14-f2acbc75.tar.xz' and enter the result as the PKG_MIRROR_HASH in the Makefile.
Perhaps you should re-read what you said, as you stated this is no longer an issue:
The throughput of 144mbit/s is the max speed with a 20MHz broadcast width [the broadcast width determines the throughput speeds available, as it's not about a device supporting a specific throughput speed], however in order to garnish whether you're actually experiencing an issue would be to perform an iperf test between the client(s) and the the router.
You were telling me that I was halving the bandwidth by setting it to HT20. HT20 was my only option (previously).
Now, set to HT40 (at which it is now visible) I only see 72Mbit/s on all devices. I was seeing 144/Mbit/s at HT20 before so something is at the very least different. Yes?
iperf, ok I'll test with that. LuCI isn't a good indicator I take it.
Edit: Seems I need client and server with iperf? I'll do another build tomorrow after work.
Since this isn't the official davidc502 thread, I'd like to go slightly off topic. I'm in the market for second router and want the best.
How's the WRT3200ACM / WRT32X right now? I've heard the CPU is locked at max clock speed without scaling and RF output is not adjustable with OpenWrt. Neither of these are good for efficiency. How about the USB 3.0 port? I read on OEM firmware it's good for an insane 100MB/s read/write, basically NAS-like performance. Is it this fast on OpenWrt?
Otherwise I might go for the R7800 which seems to be more polished right now, albeit with slower USB 3.0 performance (I've heard closer to 50MB/s rw). Thanks.
Got myself a new router so I cleaned up my work and put it up on github. It's now organized so that it shouldn't conflict with the base tree and I can easily keep up with the latest master. I only build for ipq806x devices as that's what I have, but my profiles are easily converted to other devices.
Could u tell us which device you are using right now ?
And why u gave up on the WRT AC series ? 
I'm not entirely sure, but I don't think CPU overclocking is supported on any of the WRT routers.
I have a WRT32X running extremely well on OpenWrt SNAPSHOT r7177.
Samba transfer speeds are 90-110 Mbps.
I love the dual-partition. If you're like me and live in a house where family members will physically assault you if the WiFi is down for more than a few moments - it's a great feature. 
There are a few people on the DD-WRT Forum who left the WRT3200 for the R7800 and rave about the performance on Merlin.
Hope this helps!
The CPU scaling I'm referring to is the CPU clocking down, not up. Modern CPUs will scale down from the max as needed depending on load to reduce power/heat. This is something the WRT line does from OEM firmware, but I read still isn't supported with OpenWrt. The R7800 has it along with RF output scaling too. Just one of the things I was curious about. Thanks for the help though good to know USB is so fast on this router (assuming you meant 90-110 MBps which reviews have said OEM fw does).
CPU scaling is supported, the radios are locked at maximum power on V2 flavours, rango and venom(based on channel selection).
Does CPU scaling work with davidc's latest build ?
And the maximum power lockdown is a pretty lame setting imho ! Not everyone need to have their radios running at max power output !
Silly silly Marvel... Silly³ ! Hopefully that will change in the future... It's a real deal killer imho !
Does OEM firmware really do that?
Care to point the relevant bits from OEM GPL sources?
As far as I know, the CPU frequency scaling had not been merged into upstream Linux and I doubt it has been integrated into OEM sources.
Would be really interesting if it has been, as that would enable us integrating it also to Openwrt.
I got a TP-Link Archer C2600 which I'm very happy with. My mods aren't tied to a specific device, as mentioned you can just adapt one of my configs and compile my tree for your own device.
Never had a WRT AC so haven't given up on it 
From what I've read there's tons of issues with them? Also they look like crap IMO. Anyways, I think we derailed from the topic a bit here, so I'll leave it at that.
