How to increase response range of nRF52832 in Bluetooth Mesh?

[master.chipset]

Hi everyone.
I have 4 boards based on nRF52832 that I've designed by myself.
I'm testing Bluetooth Mesh capabilities by them and I'm using Mesh Light example for it and I see their behavior using indicator LEDs on the boards. I saw that the range of boards are low so I tested them by changing TX power and PHY.
I tested them indoor with straight line in corridor.
For advertising the maximum range is about 16m with RSSI -97dbm
For Bluetooth Mesh I placed boards serially in a straight line and you can see nodes features like as relay or proxy settings in below. It's the maximum range that I could achieve:

By changing PHY and TX values, the below result achieved between every two nodes:
PHY: 1Mbps   TX: 0dbm    > 1.6m
PHY: 1Mbps   TX: +3dbm  > 2.4m
PHY: 1Mbps   TX: +4dbm  > 4.4m
PHY: 2Mbps   TX: 0dbm    > 1.6m
PHY: 2Mbps   TX: +3dbm  > 2.4m
PHY: 2Mbps   TX: +4dbm  > 4.4m
As you can see, the result isn't differ for 1Mbps or 2Mbps PHY setting.

I'm using nRF Connect SDK 2.5.0 and I edited the configurations by nRF Kconfig GUI as below:
   

I know that the most efficiency is related to antenna design and tuning but because I haven't any experience about it, so I copied the antenna part of PCA10040-nRF52832 PCB board.
You can see my PCB in below:

I have some questions. I'll be very delight if you reply them:
1. Is my configuration true about choosing 1Mbps PHY by disabling 2Mbps PHY?
    If my configuration about 1Mbps PHY is true, so why the result is same for 1Mbps PHY and 2Mbps PHY as I showed above?
2. Is there any extra configuration available that I can set to improve the efficiency?
3. Why advertising range is about 16m but for Bluetooth Mesh, node to node range is below than 5m? (I checked advertising range by nRF Connect app)
4. Why Node 1 responds me in the range of 16.8m without losing the connection but the other Nodes can't communicate by themselves in this range and can't be more than 4.4m?
    Is it related to proxy feature of Node 1 that others lack it? In the other words what's the difference between proxy feature that my phone and Node 1 use it and relay feature that the other communicate by themselves and how it influence the response distance range? (It's very brilliant for me if other boards can communicate with each other in 16.8m, too).
5. How is the antenna tuning process?
    Does it need a special equipment that I must provide? or
    Is it experimental by changing passive components in antenna path? or
    Is it related to calculations that I must do before designing the PCB?

Thanks in advance
Mehdi Sadeghian

[Joakim Jakobsen]

Hi.

1) Why do you expect the range to differ som 1mbps and 2mbps?

2) Increasing the TX power would be the best bet to increase the range significantly. Having only 4m is very low, but that might be due to HW limitations.

3) This might be because the communication before provisioning is significantly "lighter" compared to after. And less prone to packet loss.

4) Not sure, but it might be related to the antenna or hw layout on the phone is allowing it to maintain longer range.

5)
There are more then one way to "tune" the antenna, but the best and fastest way is to connect it to a Networks analyzer and modify length and matching components. 

It is possible to simulate and calculate what size, width, length an so on would be best and do simulations with matching components to get the best possible result. This would have to be incorporated in to the design before production. 

Easiest option is to send two devices to us and we can do the tuning of the antenna as well as match the radio for best performance. 

Br,
Joakim

[master.chipset]

Thanks a lot Joakim for your quick and complete reply.
I'm here with new questions.
How soldering the exposed pad of nRF52832 can affect the antenna range?
    As I learned by Nordic documents, it's very important to provide a very good ground plane for the chip. But because I soldered the chip by hand, I couldn't solder the exposed pad.
Can selecting the internal LDO regulator affects the antenna range compared with internal DCDC regulator?
    My PCB configuration is based on internal DCDC regulator.
How new versions of chip can affect the antenna range?
    I saw that in some posts, Nordic engineers ask about the version code of chip. For my, the chip version is E0 and is manufactured in 23th week of 2018:
    N52832
    QFAAE0
    1823FC

Best regards
Mehdi

[Joakim Jakobsen]

Thanks.

Not 100% sure about this, I'll have to ask internally before I can get back to you with an answer on this.

Br,
Joakim

[Joakim Jakobsen]

Hi again. 

master.chipset said:

How soldering the exposed pad of nRF52832 can affect the antenna range?

The response I got was that this would probably be fine for prototyping, but could result in reduced performance. But with no indication about how much it would affect it. 

master.chipset said:

Can selecting the internal LDO regulator affects the antenna range compared with internal DCDC regulator?

It could cause 1-2 dB less sensitivity for the nRF52832 with DCDC compared to LDO. This could also be worse if the exposed pad isn't soldered. 

master.chipset said:

How new versions of chip can affect the antenna range?

As far as I know this shouldn't affect the range noticeably. 

Br, 
Joakim

[master.chipset]

Hi again Joakim
Thanks a lot because of your complete and exact response.

Have a good time
Mehdi Sadeghian