There’s a very long TLDR post on how I got here I’ll not bore you with. Instead let me state ‘the problem’

Setup is a SIM based external antenna with 2 NetGear Orbi Wireless/Mesh/Router inside. One is in the loft connected to the internet router (which is connected to the antenna) via Ethernet. The other is in the middle of the house. WiFi coverage is good in all rooms.

The Orbi has ‘speedtest’ embedded in its OS. I ran the test on the supplied app and speeds are consistently 60 down/20 up.

If I run the SAME application on a wireless device (PC/MAC/Phone/Etc) I get the 20 up but between 25/15 down. The speedtest starts at about 40 and falls away over the period of the test. Tried from all rooms including standing next to router in the loft.

Much (oh so much) testing has been done with different mesh solutions (started with Tenda MW3 but it is quite limited in speed as it doesn’t have separate backhaul-basically same problem as wifi repeaters), different devices, home plugs to try and bypass the wireless etc.

Nothing changes the measurement. So I wondered if it was a measurement issue. To test I ran 6 HD youtube streams across three devices pulling 45+meg through the router (which I can monitor using the router interface) with no buffering at all!

So my conclusion is there is something hooky in the venn of SIM based internet/Mesh Wireless/Speedtest app. But I’ve no idea what that might be. Especially as this only started happening a week AFTER we’d put the antenna up.

Anyway it’s not a problem as I’m not ‘losing’ bandwidth. But it is blooming annoying as I’ve spent hours chasing my tail trying to fix it. Much of which has been precariously walking on floor joists in the loft trying not to go through the ceiling!

Any ideas gratefully received.

Any ideas gratefully received.

Forget about it until actual capacity becomes an issue.

Yeah well I could. But I want something back for all the time I put into it. And there’s clearly a lightbulb moment I’m missing….

I have a USB-C to Ethernet adaptor somewhere. I might take my laptop into the loft, bypass the wireless completely and see if I can replicate the speed test from the Orbi. At least then I’d KNOW it’s wireless (and not specifically Orbi as I had it with Tenda as well). Then that might allow me to move on 😉

Yeah well I could. But I want something back for all the time I put into it.
As an analyst, I know you’ll know all about sunk costs and that you really need to ignore them 😉

Seriously though I hope you get it sorted (or at least some closure). I read your previous thread and I’m considering a similar setup once my current Virgin Media deal ends.

That’s a fair point. In fact if I’d used my ‘professional’ skills, I’d have identified it as a measurement problem far earlier.

In terms of how well it’s practically working tho, no issues at all. The Orbi was double the price of the Tenda Mesh but it’s way better. Lots more security built in and stronger wireless especially with my NAS plugged into the back of the Satellite. Now get 15-20 meg/sec from Mac doing transfers. Before was getting less than 1 most of the time.

802.11g will probably top out at around 20Mbps.

I started there. All the devices I’m testing with though are 802.11n. The mesh does transmit a 2.4ghz freq and a few older devices (like our TV) connect to that. But the Macs/phones are on the 5ghz part of the spectrum and if I believe – which based on my current feelings about measurement tools is a big if – the Orbi link speed ratings, they are close to the full 850meg.

And for additional super-weird point, when we first put the antenna up I was not getting this problem. Three have been working on the mast tho. I’m just going to blame them to stop me spending any more time one it 😉

Remember 5GHz has less range, so distance can become an issue before it does with 2.4GHz. Have no experience of Orbi, but how do they interconnect? Does the Orbi app show a max strength connection between all the modules? Ignore if it’s Ethernet backhaul. Worth raising a support ticket with Orbi? Perhaps all the security and extra features have an overhead?

https://community.netgear.com/t5/Orbi-App/Netgear-Orbi-app-speed-test-inaccurate-vs-speedtest-net-for/m-p/1728642#M153

The answer here is a laptop/whatever wired to your network running iperf then use iperf on client devices on wifi to measure actual bandwidth to the server.

E.g. for ios you can get a client for free: https://itunes.apple.com/us/app/iperf-3-wifi-speed-test/id1462260546

And for various OS https://iperf.fr/iperf-download.php

This is how you get good measurements, point to point, up or down, or both etc.

@jamze – that’s useful info. The Orbi uses a separate radio for backhaul. One of the reasons I switched from a 2 radio system. There are currently only 2 (that’s was the max when they came out, but now you can add more) and we’re definitely getting 2.4 band in places due to distance from unit(s). But my testing has been from lots of places and I make sure I’m on the 5ghz band.

The mgt software just tells you if the link is strong. I can see a lot of packets flying up and down it from the orbi console with no collisions or loss but that doesn’t mean much. I think I need to do what @IA suggests and plug a ethernet connected device directly into the internet router and bypass the Orbis completely.

@IA – thanks will download those and give them a try.

So if I’m reading this correctly everything works but a test run on individual items doesn’t seem to support what the system is capable of and now you are disappointed because while everything does what it should do you are getting numbers lower that you were expecting.

All I can suggest for your peace of mind is that you never measure your penis as this could end with a similar result. It might do everything you ask of it but you’ll be forever fiddling with it trying for that extra millimeter 🙂

Tried iPerf. Gave me the same results – 20meg ish download. Reran Youtube test, still goes 40 meg+. Plugged laptop (via USB ethernet) into satellite, then main unit and received same speeds. So it isn’t the backhaul between (there was a bit of difference between primary and slave orbi’s but it’s 10% max) nor the wireless the devices are connected to. Then plugged directly into router to take wireless out and….

… same. So I’m down to

1) there’s a problem with the mast since the upgrade
2) there’s something strange about how speedtest works
3) the speedtest app on the Orbi is hooky and it’s telling me lies
4) something else

I spoke to three on chat and they think there is still a problem with the mast that will get fixed in the next few days… which means I can ignore it until they close my case. I don’t expect it’ll fix it but it does give me a reason to leave it alone.

Just out of interest did a similar speed check on my Deco mesh setup, slightly surprised I’m getting pretty much full speed at my phone.

Internet speed using the built-in Deco app – 69.6/18 Mbps
On my Android phone/WiFi using Ookla – 68/15.2 Mbps

As another sense-check, what 4G speeds do you get on a phone outside? Does it match the 60Mbps Orbi’s speed test is claiming?

@jamze – pretty much. If I walk 300 yards so I can SEE the mast, I get about 60 on the phone. Walk a bit further and I get 100+. When we had an indoor antenna got about 15-20. So this feels about right as we don’t have line of sight due to pesky wood in between us and mast.

In the house (thick walls, in a bit of a bowl) prob 10-15 max. We always knew we’d need an external antenna. Which makes me think it’s not the mast but… arrrgghhh

Anyway, I’ll leave it with this

Bit fuzzy, but the top pic is the router CLI. That’s real time recording over 5 mins when I was hammering it with multiple 1080p video streams for 4 devices. Right hand one is a speedtest I took immediately after on the Orbi. Bottom one is a Speed test off one of the devices I’d just been running 3 or 4 HD streams on.

Boggled. Completely boggled!

But thanks for all your help. I still think I’m being an idiot. But at least it’s not a totally obvious idiot. I’m wondering now if I should go back to the Tenda Mesh and send the Orbi’s back. Not sure they’re adding much (other than giving me super high speed to our NAS)

Point of note, iperf on Windows is pretty broken.

The mesh does transmit a 2.4ghz freq and a few older devices (like our TV) connect to that.

Try disconnecting them? In fact, disconnect everything bar the device you’re testing from and see what happens. It shouldn’t really be an issue on modern hardware I don’t think, but I’ve seen cases where Wi-Fi networks only run as fast as the slowest device.

(sorry, I missed your last few posts, I’ve had this window sat open for a while.)

@cougar – I have tried disconnecting everything apart from one client, one wireless router and the internet router. It does get a bit better but not consistently.

I’ve decided to go with the mantra ‘if at first you don’t succeed, define EXACTLY what you mean by success’ 😉 It all seems to work so I don’t have a problem to fix. It’s just annoying.

Agreed, don’t worry about if working. After some more checks here, I think I have a similar issue, mesh system with 5 units, and I always assumed the point was they gave you a decent, consistent speed everywhere via a dedicated backhaul. Doesn’t seem to be the case, speed depends on which unit I connect to?

If I’m in the house, i.e. no more than a unit away from the router, getting the 65/20Mbps speeds. Out in the garden/workshops, it’s more like 25/14Mps. The only difference is there’s another unit in the chain. Everything is showing max WiFi strength, both on the mesh units and phone 🤔

That was why I ‘upgraded’ to a 3 radio/dedicated backhaul. The Orbi stuff tho – so far – is rock solid in terms of providing a great signal everywhere. And regardless of what speedtest says, I can push 60+ meg from anywhere in range, so I’m going to just put it down to a measurement issue.

I know with the 2 radio mesh, every time there was a ‘repeater hop’ the BW definitely suffered. Should not be the case with a dedicated backhaul.

I’m still confused.

If you send the Orbi stuff back you may want to try Uni-Fi equipment. They do a MESH system or you can go old school with router and wifi access points cabled from the router. So many configuration options, on the latter option, I reckon you’ll be busy for months sorting it all out!!

Thanks. I think I’m done tho. Everything works and seems – at the moment – to be stable. I’m kind of loathe to start again. I’ve lost a few hours I should have been working (and billing) and I don’t think I won’t to lose anymore!

That Uni-Fi stuff does look interesting tho.