Investing the LTE issue


Askarus

Forum Addict!
Joined
Sep 28, 2011
Messages
4,989
Location
Germany
Here comes the LTE thread.
The aim is at first to write down everything that I know and to give others a proper place to collect information.
Thanks to @johnicboom we finally can collect some data on Letux OS and compare it with the Pyra OS output, so it finally makes sense to start this discussion here.

One of the first things I investigated once I picked up my Pyra again was LTE, so I'll start with my observations

Vanilla Pyra behavior

To get over this quick, let's do some bullet points:
  • The connection is generally bad
  • In my house I do not get any connection at all
  • In the garden, a connection can be established but it it nit stable at all
  • If a connection can be established, it is in general very slow
  • Others have reported that they can get a good if an LET station is very close and in direct sight range


Testing with external antenna

Apparently the Pyra has a connector for an external antenna. I am not sure if that is true or if it is just a debug port. I got different inputs on that, however I believe that it's an external antenna port.
At least an external antenna does fit and as such I tested it all with third party antennas.
2 antennas were tested. I chose them because they were the only ones that I could easily source via Amazon.

Those two were tested:
Code:
1: https://www.amazon.de/-/en/dp/B081JNMB5J
2: https://www.amazon.de/-/en/dp/B010GO59UA

Both antennas gave me some improvement. I was able to connect and even browse the web with them. However,
  • the connection in my garden happened almost instantly after booting
  • connecting was not reliable, so even at the same spot, I could not always connect properly or lost connection from time to time
  • connection chances still varied heavily where I stood in my garden
  • no in house connection still
How I tested:
  1. First antenna
    • Inside the housing, sticking on the SD cards, no SD card was inserted
    • Outside the housing, the connection wire sticking out of the SD card slot, sticking at the bottom of the battery lid, no SD card was inserted
    • Same as above but the antenna was dangling loosely
  2. Second antenna
    • Outside the housing, dangling loosely (the antenna is inside a PCB and does not fit inside the housing
There was no real difference between both.
I got the best results with antenna 1 sticking to the housing.
I recall for sure that I used that setup somewhere outside to browse the Internet with the Pyra. However the experience was not reliable enough that I wanted to keep the antenna.


Investigating frequencies

I raised the issue to @hns via E-mail. His response was like that:
  • This makes it seem like the modem is not properly compatible with the frequencies available
  • It might be that some part (for example MMC-interface) causes disturbances

We then had some conversation regarding frequencies.

I am using O2 and those are the frequencies they use:
Code:
 Deutschland     O2/Telefonica     20 (800 MHz)     Juli 2011    
 Deutschland     O2/Telefonica     3 (1800 MHz)     2016     neue Frequenzlizenzen aus der Auktion 2015
 Deutschland     O2/Telefonica     7 (2600 MHz)     2013    
 Deutschland     O2/Telefonica     1 (2100 MHz)     Sep. 2018    
 Deutschland     O2/Telefonica     28 (700 MHz)     Apr. 2019     Telefónica will LTE 700 künftig punktuell dort einsetzen, wo eine bestehende Basisstation auf 800 MHz überlastet ist.[13] LTE 700 wird besonders im ländlichen Raum, z. B. in Niederbayern, aktiv genutzt.
 Deutschland     O2/Telefonica     8 (900 MHz)     ab 2022     vereinzelt bereits dauerhaft aktiv, z. B. am Königssee (Oberbayern) oder bei Hausen/Niederbayern
(source: https://de.wikipedia.org/wiki/Liste_von_LTE-Netzwerken)

In comparison, here is the data sheet of the EU-modem: https://www.gs-m2m.de/fileadmin/Bilder/GSM_Module/Module/PHS8/gemalto_datasheet_PHS8_web.pdf
We use PHS8-E.

The more recent one would be PLS83W: https://www.gs-m2m.de/fileadmin/Bil.../pls83-W/Thales_Cinterion_PLS83_Datasheet.pdf
It supports more bands. Directly quoting hns here:

FDD-LTE: bands 1, 2, 3, 4, 5, 7, 8, 12, 13, 18, 19, 20, 26, 28, 66
TD-LTE: bands 38, 40, 41
translated with https://en.wikipedia.org/wiki/LTE_frequency_bands (each MHz)

700
800
850
900
1700
1800
1900
2100
2300
2500
2600

So apparently the PHS8-E does not support 2600 and 700MHz bands and hns raised that this might cause the issue if those were the most prevalent bands in my area.

I kindly disagree and I can support my stance quite well.

First, I have an older phone here. It is from 2015 and connects perfectly well. The phone supports 800/850, 900, 1700, 1900, 2100 MHz, so no 700 and 2600 MHz (source: https://forums.crackberry.com/black...ry-priv-models-supported-frequencies-1043968/)
Second, my current phone that already has a 5G modem and most likely supports all recent frequencies also connects to those older bands. I was able to investigate that because the phone is rooted and I was able to run an app that displays the connection strength and the frequencies I connect to.
Here are some screenshots:
photo_2024-03-04_17-08-20.jpgphoto_2024-03-04_17-08-26.jpgphoto_2024-03-04_17-08-30.jpg

The conclusion I draw is this:
Of course I can not ruse out that the more recent modem would make a difference, however I can not see why and I would not at all rely on a new modem to fix the LTE issue.


Interference

In Austria at the Level Up I talked to a visitor who told me that he works in the LTE business.
His estimate was that there is some interference causes, especially because the modem and antenna are so close to the charging electronics and the battery. He advised me to test LTE without a battery attached.
I did exactly that, booted the Pyra and removed the battery.
That however made zero difference and I got better reception with the battery in and an external antenna.


LeTux

To investigate whether the issue is hardware related or if we use an improper config, I wanted to test LTE on LeTux OS.
Unfortunately I was not able to make a boot SD and got stuck.
Finally thanks to the help of @johnicboom , we were finally able to investigate LTE under LeTux.

It seems like it made no real difference for him and LeTux leads to equally bad results.
His first tests can be seen here: https://pyra-handheld.com/boards/threads/letux-os.100034/post-1719870

the conclusion of this is that there is no software issue or that some config is wrong in both OS. Maybe a more recent versions of "something" causes the issue.


Other guesses

Wat follows is a collection of guesses that I remember, which have been raised on Discord or anywhere else.
  • We do not use the antenna at all
  • The antenna is not properly mounted (pretty much ruled out already)

And that is basically all I know in regards to LTE. May the discussion begin.
 
Last edited:
Didn't HNS also mention that the battery voltage was capped to a certain maximum voltage, else the modem wouldn't work?
I think I read about someone commenting that the battery is not charged to 100%. This is for purpose. Since 100% is equivalent to 4.2V but there is a 50mV voltage drop in the charging circuit and the way how the LTE modem is connected, the modem will get a too high voltage and turns its transmitter off. So an almost full battery and charging power turns off the modem. The only solution we found was to limit the charging voltage to 4.15V which is equivalent to ca. 85% as reported by the fuel gauge chip. What is missing is to multiply the fuel gauge % by 1.17 so that it is displayed as 100% and nobody would notice

Are there any firmware updates available for this chip. I vaguely remember from similar projects (I'm not familiar with the chip in the Pyra btw) in the past in some rare occasions had to perform those firmware updates to fix issues.
 
Are there any firmware updates available for this chip.

hns said this in another thread:
The modem comes with already flashed and approved firmware. It needs as much firmware as an stick type modem plugged into some USB slot. Of course it needs some kernel driver to turn in on and off and some user-space software to make use of the commands. But that is not firmware and does not depend on hardware.
So there is no need for a "newer" or "better" modem firmware to flash. Of course the modem manufacturer likely has different firmware but they can't give that away without violating certification requirements (remember that this is a RF device which connects to a public network and every hardware design and firmware version must go through CE and FCC to prove that it does not harm other users).

He says other interesting things related to this thread's topic, too, so the whole post is worth a read:
 
Didn't HNS also mention that the battery voltage was capped to a certain maximum voltage, else the modem wouldn't work?


Are there any firmware updates available for this chip. I vaguely remember from similar projects (I'm not familiar with the chip in the Pyra btw) in the past in some rare occasions had to perform those firmware updates to fix issues.
I do not think that is the case. Anyways, his response does not make sense to me.
When I charge the battery to 95%, the voltage is still 4.1V, so I can not see any way that we require the charge to be capped to 85%.
Those are the reports by the internal XFCE power management tools, no precise measurements.
However, I would not bet on that. The modem is on regardless of the charge, the reception is just bad. It was always visible in modem manager.
 
I wrote something about here...


... comparable results.
 
I do not think that is the case. Anyways, his response does not make sense to me.
When I charge the battery to 95%, the voltage is still 4.1V, so I can not see any way that we require the charge to be capped to 85%.
Those are the reports by the internal XFCE power management tools, no precise measurements.
However, I would not bet on that. The modem is on regardless of the charge, the reception is just bad. It was always visible in modem manager.
There are two ways to measure how much the battery is filled. Either by measuring the voltage and interpolate linearly. Or by a coulomb counter (bq27421) which takes into account that the battery voltage is not rising linearly. The /root/charger script reads out both and they differ.
 
While I was getting my bearings with AT commands, I did a lot of unscientific testing. I have not yet gone back and really taken systematic measurements that I can compare across OSs, but I get similar results no matter if I'm in LeTux or PyraOS, and I have switched back and forth running the same commands and getting the same results, so it may not be worth doing more carefully.

Connecting to the modem and using the AT command AT^SMONI, the best numbers I got were -106 (RSRP), -8 (RSRQ). This was in a building with big glass windows, pretty much right across the street from the cell tower.

@Raik, are you still using the same antenna setup from the post you linked? It's this one, right?
It looks like I can order it from here:
It's cheap enough that I can order one and give it a try, if you still think it helped.
 
Back
Top