Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
MikroTik LHG LTE18 inside and its m.2 EG18-EA modem
The MikroTik LHGG LTE18 has a similar construction to the earlier LHG LTE6, including a physically identical dish and pipe clamps, but with a few changes to the PCB.  Opening the top is much like the basic LHG models, where there are three plastic tabs holding the cover with the heat sink in place.  These easily pop out using a blunt butter knife or spudger.


The modem is an m.2 based Quectel EG18-EA in this EU model and has two U.FL antenna ports connected for 2x2 MIMO.  The rear of the PCB has the antenna elements physically attached and fed through to the two U.FL ports on the top of the PCB, just like the LHG models.  A DIY modification for 4x4 MIMO is possible by running U.FL to SMA connectors to the spare two ports on the modem, however, this would involve drilling additional holes for the SMA ports on the outer case, not to mention weatherproofing these ports. 

Its SIM slot takes a micro SIM and is easier to insert / eject than the earlier LHG models that took a full size SIM.  The SIM needs a blunt tip object to fully push in and to push again to eject.

The following shows the reverse side of the top cover with the thermal transfer pads.  As these make contact with a few components such as the black IC chip above, the unit should not be kept powered off while the cover is off. 


It has a Gigabit Ethernet port which can take a PoE voltage range of 12v to 52v and will also power from MikroTik routers that have a PoE output, such as the hAP ax³ WAN port I was testing it with.
I have now modified a LHG LTE18 with a damaged PCB to work as a passive MIMO antenna.  It turns out this is a reversible mod that does not require drilling holes, but take care not to damage the thermal pads.  This modification would be better off carried out on an LHG with a defective PCB where the dish and the internal antenna elements are still intact.  There is a small opening to the right of the SIM slot that is wide enough for thin coaxial pig tails to pass through with the outer cover in place.  

I used a pair of 10cm U.FL to female SMA pigtails.  To make this modification, remove the m.2 modem to access the U.FL connectors on the main PCB.  Carefully detach the two short coaxial leads and push on the U.FL connectors of the pigtails in its place.  Insert the two coaxial leads into the slot to the right of the SIM card slot.

Before fitting the outer cover, ensure the coaxial leads head right and that no coaxial lead goes over the square black chip.  There are thermal pads inside the outer cover that make direct contact with this chip, the m.2 modem as well as two small chips directly above the left button, which would likely cause damage if a coaxial lead gets caught in-between.  The damaged thermal pad on the left (compare to post above) above shows how easily these are damaged, although in this case the PCB was already damaged from another incident.


When attaching coaxial leads to the SMA ends, be sure to secure these in place such as with hot melt glue to prevent the leads tugging directly on the delicate pigtails, such as during windy weather.
I'd be curious to see how it performs signal wise vs your current setup Big Grin
With the dry calm weather I did a comparison of this LHG with 2m coaxial RG58 leads against my Poynting XPOL-2-5G that I use with a separate 4G router for our VoIP line.  I hooked up our Fritz Box to the main 4G connection temporarily so I could use my MikroTik Chateau LTE12 for this test.

I borrowed by Dad's cement mixer tripod and stuck a wooden pole into it.  For additional security against the wind toppling it, I placed a garden gnome on each corner. Smile

I started by mounting the XPOL-2-5G on the pole and spent at least 10 minutes moving it up/down and turning left/right to get the best signal readings I could get on Three and ran multiple speed tests.  I then removed the antenna and put on the modified LHG with the two coaxial leads, again spending some time fine tuning the height and direction to get the best signal readings and speed test.  I did not move the pole / tripod throughout the tests.

Carrier Aggregation readings XPOL-2-5G vs LHG with leads:


Band 3 is about 5dB better and Band 1 is about 10dB better.  Unsurprisingly, band 20 is weaker by several dB, particularly on the diversity:


As shown in the cell monitor, the main issue I have at the house is the band 20 cell ID 127 is clashing with cell 399 on another mast:  


My main 4G router set up is about 40m behind our house where a distant tree partially obstructs the mast operating band 20 cell ID 127.  The mast operating band 20 cell ID 399 performs much better at peak time than the other despite rarely operating band 1.  The band 1 cell ID 117 above is on a different sector which is why it is much weaker than band 3.  When band 1 cell ID 115 was operating in the past, it had a similar signal strength to band 3.  

Despite the big improvement the LHG offers on bands 1 and 3, I got the best speed test result with the XPOL-2-5G antenna, likely due to additional congestion by the time I swapped the antennas.  I also suspect the bottleneck is with the backhaul to the mast as I was expecting to see much higher download speeds particularly with huge SINR improvement on band 1.  I also wish I was within the range limit to connect to band 1 or 3 as the primary band to test the upload speed as the LHG's poor performance on band 20 cut the upload speed in half:


The following shows a size comparison between the LHG and XPOL-2-5G.  I had just one antenna mounted on the pole at a time during the above tests.


Pity there is no 5G band n78 within range as I'd love to see how the LHG performs on the 3.6GHz band, particularly with it capable of 19dBi on Three's n78 band highlighted below on MikroTik's LHG gain graph:

That's quite a nice signal improvement especially 16-28 SINR. The newer LHG looks a lot nicer than the old one Big Grin
Quite a nice difference on the higher frequency bands with the LHG . At least a gust of wind can't blow it down now Tongue
I remember the lower freq bands having terrible performance on my old LHG. I remember the time I managed to get B3 signal on Vodafone down the field with the 80m cable run it was maxing out the 100 mbps port on the dish. It also had unusable latency which was made me stop using it.

Nothing more frustrating than having more than enough signal but unable to actually connect because of however it works over distance.
I wonder would it be worth using an LHG as a passive antenna for N78 for me here. I'd say if I got the RSRP under -100 on N78 it might connect more reliably but I'm not sure.
19dbi seems like a lot. I think the EPNT-2 is 11? If I was able to just run long pigtails directly onto it would be even better.
If you get decide to get one, I suggest going for the basic LHG LTE to do the conversion as the dish and internal antenna elements of the LTE18 appear to be identical to the LHG LTE I bought a few years ago, which also had U.FL connections to its antenna elements.  They basically just changed the PCB and improved the cover with a heat sink.  From a quick check on Getic, they have used LHG LTE for €118 + delivery.  In the worst case scenario, you could undo the mod and can sell it off again.  At least it doesn't require drilling any holes and the front cover clips can be pried loose with a butter knife. Tongue

I also tested this mod on Eir Sunday.  Like Three I can only connect to the low bands 20 or 28 as the primary band, however, Eir is a right pain putting up a tantrum when it comes to using cell lock.  With the signal the LHG picks up on band 1 and 3, the network is determined to get me off band 20 and 28, causing the connection to cycle down/up.  If I let it try connecting on band 1 or 3, the connection just dies as it cannot synchronise (endlessly fails RACH).  The following screenshot shows a partial LTE log filtered by "detach" from a previous day where I was testing with Log Periodic antennas trying to cell lock on band 20:


If I cell lock to two cells across two bands 20 and band 28, it flip/flops back and forth between the two bands and after 10 to 15 minutes, it eventually settles down, usually staying band 28 and aggregating bands 1 and 3.  Cell locking on band 20 does not appear possible without disabling all the other bands as the network just persistently keeps trying to get the modem off band 20, even leaving it alone for 30+ minutes.  

When the connection finally stabilised on each antenna 28+1+3 CA and fine tuning each antenna as above, the following are the CA readings from the XPOL-2-5G vs LHG with leads:


When I checked the MIMO readings, the diversity ended up with the better RSRP, so I had to swap the leads and wait another 15 minutes with the network putting up another cell lock tantrum. 


Cell monitor:


Like Three, I got the slighter better download speed with the XPOL-2-5G.  However, while waiting for the network to quit fighting with the cell lock, it's likely congestion picked over the 30 minute period:


Like Three, I suspect the congestion is with the network backhaul on that distant mast.  The LHG only gets about 1/3rd the upload speed due to the gain being at its weakest on band 28.
I bought a USB3 to M.2 on AliExpress module (link) to test whether the Quectel EG18-EA M.2 modem that was inside the LHG would work externally with another MikroTik router, in this case a MikroTik hAP ax³.  One issue I ran into is that the kit came with four U.FL to IPEX4 pigtails.  Both the adapter board and the modem have U.FL connectors, so I just had the two that were inside the LHG.  For anyone considering get this or similar adapter, be sure to get a few 5cm U.FL female to U.FL female pigtails as these do not appear to be available on Amazon.

To my surprise the RouterOS firmware on the hAP ax³ has LTE modem support, so does not require any add-on module to use a USB modem.  Once I attached the modem, it appeared as an LTE interface just like on any other LTE enabled MikroTik router.  I tested on 3G for now as I mainly wanted to check that the USB module and modem would work, especially as the EG18-EA modem does not appear to be sold separately anywhere.  The only minor configuration I had to do was add an APN and set the WAN interface to LTE.  The main status page shows "No Signal" while in 3G mode, but operated fine as shown in the speed test:


In 3G mode, the LTE status page does shows the connection status.  The EG18-EA modem as well as the EG12-EA in the Chateau LTE12 only show the RSSI in 3G mode.


The "BetterRoaming" operator I was using here is an eSIM only roaming data provider.  I curious to see whether I could get it to work with a router. Big Grin
My brother who got the LHG LTE18 installed had an issue with very poor upload speeds.  Like me he needs to use band 20 or 28 as the primary band due to also picking up his signal from across Donegal bay.

While his RSRP reading was weak at about -103dBm, the upload speed seemed unusually low at just over 1Mbps.  After having little luck trying to improve this by moving his dish, I checked the MIMO signal reading for curiosity and sure enough the main polarisation had an RSRP of about -110dBm and the diversity cross-polarisation about -103dBm.  As the modem transmits on the main polarisation only, it meant physically swapping the two coaxial cables around on the modem. 

Unlike a router with external antenna leads, swapping the polarisation involves removing the front cover off the LHG.  The LHG must be powered off and ideally taken down to perform this procedure.  As the heat sink makes physical contact with a few chips inside, removing the front cover with the LHG powered on can lead to these overheating, not to mention damage to the modem if the coaxial cables are detached while powered.

Unlike the older LTE4 / LTE6 models, the short coaxial cable is long enough to reach the opposite connector, so it's just a matter of swapping the two and running the longer thin coaxial lead around the other side of the modem.  After the change on my brother's LHG, the two readings swapped around and his upload speed improved to around 6Mbps.

Note: Ensure the long coaxial lead does not come in contact with the black chip or grey heat transfer pad as shown below to the right of the coaxial lead.  Otherwise it could get trapped between the heat sink behind the outer cover and either cause damage or overheating. 


To check the MIMO main and diversity signal readings, run the following command:

/interface lte at-chat lte1 input="at+qrsrp"

The first two RSRP readings after "+QRSRP:" correspond with the "MAIN" and "DIV1" modem connections, respectively, as shown in the following example:


Forum Jump:

Users browsing this thread: 1 Guest(s)