EVDO upload

getut · EVDO Fledgling Joined: 16 Aug 2006 Posts: 11

Ok.. per this thread http://www.evdoforums.com/thread3109.html, I have now got my omni antenna up in the air on my rohn tower and am finally getting decent signal strength.

My S720 is showing -81 to -85 EVDO signal strength where I was registering no EVDO at all and only -99 to -103 of 1xRTT with just the card by itelf. The newfound level of connection strength is buying me about a 45-50 kBps download rate but my upload rate is horrible. Usually around 1 to 2 kBps... 1xRTT or 56k modem is better than that by far on the upload side.

If I force the card to 1xRTT mode I can get 10-12 kBps upload rate but when in EVDO mode the uploads are horrendous. Is this due to low uplink strength as compared to downlink? Would making the last improvement I could possibly make (a yagi) help this performance increase?

Complete rig setup is below:

KR1 router -> 2' pigtail -> amp -> 2' LMR400 -> lightning arrestor -> 30' LMR400 -> 9dbi gain omni 24" mast antenna 38' up on a rohn tower

If I replace the omni with a yagi, will it improve the send performance?

Please note.. I do NOT have direct line of site to the tower BUT my rohn tower is in an open field with a huge dairy farm between myself and the tower. I have over a mile of unobstructed terrain to aim in the direction of Sprints tower.. so it is definitely not like I have tree coverage directly around the location where the yagi would be. There is nothing at all for at least a mile.

kaefer · EVDO User Joined: 05 Dec 2005 Posts: 65

Wow, you sound very motivated to get this to work. I'm also located in a fringe reception area and have tried many things to get good signal strength and more importnatly good download/upload speeds.

http://www.evdoforums.com/thread2551.html

I've even tried running an amp at each end of the LMR-400 cable. Everywhere else in the wifi world (outside of evdo) suggests using the amp as close to the antenna as possible.

Right now I don't know what to think. I've heard the cell carriers put distance limitations on the towers. So even if you have a good 80dbm evdo signal if you're too far away from the tower you'll get crappy data throughput. This may be what's happening with you.

Look at Zerog's post in the thread below:
http://www.evdoforums.com/viewtopic.php?t=389&highlight=distance

Good luck with your setup. Hopefully you break through the barriers you're facing...
-Scott

n6gn · Posted: Tue Sep 19, 2006 2:12 pm Post subject:

Your signal strength is adequate to get full rate EV-DO download. If everything is right you should see many 100K bps. I don't think a few more dB of signal will change things.

If I were you, I'd take out the amplifier between the card and antenna and see what happens. The LMR400 is pretty low loss and unless you put the preamplifier right at the antenna, you won't improve the C/N at the card, unless the card is no darn good.

The danger with the amplifier is additional distortion. Also, I'm not clear on what the additional gain (n either direction) does to the power control protocol. Hopefully nothing but just to see where you really stand, try it with no amplifier in the line and report back on signal strength and performance.

n6gn

MaximumSignal · Posted: Tue Sep 19, 2006 2:18 pm Post subject:

MaximumSignal · Posted: Tue Sep 19, 2006 2:20 pm Post subject:

You may see improved performance with a Yagi but keep in mind they are directional and must be aimed directly at the tower so you would need line of site

n6gn · Posted: Tue Sep 19, 2006 2:48 pm Post subject:

The amp does not add distortion . The amp is improving his anywhere from 20 - 30 db so I think you will find removing it will end any EVDO signal he is receiving
Every amplifier adds distortion. the question is how much and whether the subsequent degradation to the C/N (carrier to noise ratio) is significant.

The syptoms he describes are exactly those one gets with poor C/N.

Consider this, an amplifier, transmit or receive, placed at the radios antenna connector can only serve to add signal gain and potentially output power capacity. If the card has a low noise figure, then no preamplifier at its input will improve things, it can only make matters the same or worse. It only improves things when its noise figure is less than that of the card alone and it helps matters by amplifying both signal and noise equally up to a level where the cards own noise level is insignificant. At that point, the C/N at the input to the preamp has been transferred to the card's radio.

On transmit, the amplifier adds gain. However, unless it has a greater maximum (distortion free) output rating than the card alone, there's no benefit. Again, all it can do is add distortion. In short, an external (pre)amplifier only helps matters if it can do a better job than the native amplification inside the card, otherwise why wasn't it already inside the card. The exception is when you can put an amplifier closer to the antenna than the card, since the line loss can be bypassed that way.

To the EV-DO or 1xRTT system, distortion generally looks like noise. Third order distortion products end up back in the same frequency range (more or less) as the fundamental and are probably the biggest concern.

If the card is truly OK and has the expected transmit power capability, the best thing he can do is add antenna gain, probably in the form of a WiFi dish or something similar with feed modified for 1.9 GHz and to keep the line loss as low as possible by using good line (which he has) and keeping it as short as possible. Doing this is a double win since it improves C/N ratio at the card on receive and it increases ERP (effective radiated power) on transmit.

n6gn

MaximumSignal · Posted: Tue Sep 19, 2006 3:20 pm Post subject:

n6gn · Posted: Tue Sep 19, 2006 4:40 pm Post subject:

The reason the amp is the only amplifier allowed on Alltel's network is because how low the distortion level is on the unit and because it takes zero additional bandwidth . You cannot lump the amp in with other amplifiers . It uses software driven amplication as compared to hardware amplification that other amplifiers use. It is the only amplifier that has the ability to communicate with the device and adds power as needed by the device .I am wondering about the lightning suppresser . I have never had a customer insert one mid line. Perhaps try everything without the lightning suppersser I have never had this problem with a amp but would be happy to exchange the amplifier out but I really do not feel this is the problem.
As for taking the amplifier out of line he already stated that without the amplifier he could not even get an EVDO signal.

I stand by my statement regarding amplifiers and distortion. Whether the gain is controlled by hardware AGC or by software control, the underlying amplifier has finite linearity and therefore produces unwanted components - this is distortion. It's not about the gain control method (unless the attenuators are badly designed), it's about the semiconductors which accomplish the gain.

From his description, the amplifier is hardly being inserted midline:

KR1 router -> 2' pigtail -> * amp amp -> 2' LMR400 -> lightning arrestor -> 30' LMR400 -> 9dbi gain omni 24" mast antenna 38' up on a rohn tower

it is located essentially at the router. The only thing the amplifier could do, were it ideal, is to remove the degradation to the C/N produced by the 2' pigtail. Also the 30' of LMR, if connectored properly, probably has under 2 dB of insertion loss at 1.9 GHz and in all likelihood even if the amplifier were placed at the antenna path variations would probably account for more degradation than that. For a downlink signal, unless the router receiver is not properly designed or the 2' pigtail has something very wrong, no amplifier of any kind is going to significantly help matters. It is possible the KR1 has insufficient gain, I've never measured one to know, but because low noise gain at 1.9 GHz is extremely cheap, I wouldn't expect the manufacturer to have short changed it here.
The presence of the additional receive gain, even it were essentially perfect, could cause distortion within the KR1. Rremember that the early stages of the receiver may have significant bandwidth so at issue is the total PCS power, not just the carrier or two from the 1xRTT and EV-DO systems.

Unless the KR1/2' pigtail is seriously sick, the receive preamp will only make things worse.

I think it would be interesting to find out what signal power is reported in the absence of the amplifier (and perhaps w/o the lightning arrestor). The -80-something reported is after the effects of the preamplification and probably doesn't tell us what C/N is. A -110 dBm signal, all but buried in the noise will look like -80 dBm after 30 dB of gain but it will still be buried in noise (also amplified by 30 dB) and any well designed receiver still won't be able to get high speed data from it.

Heck, try it , at worst you find out what signal the card sees and how it deals with that level against a background of -174 dBm/Hz noise.

n6gn

kaefer · EVDO User Joined: 05 Dec 2005 Posts: 65

n6gn -

Nice posts. I'm going to try removing the amp on my system with LMR400 cable and see what happens....

getut · EVDO Fledgling Joined: 16 Aug 2006 Posts: 11

Ok.. to clarify before I go spend more money...

My download speeds are not great, but many times faster than dialup at 45-50 kBps (400 kbps or so). My uploads are horrendous. In my first thread I mentioned that I am about 3.8 miles away from two different towers.

Here is a mostly typical dslreports test. The upload is a little FASTER than usual in this one.
dslreports.com speed test result on 2006-09-19 18:48:54 EST:
345 / 19
Your download speed : 345 kbps or 43.2 KB/sec.
Your upload speed : 19 kbps or 2.4 KB/sec.

If I run the S720 card in my laptop instead of in the KR1, I can force it to run in 1xRTT mode. If I run in 1xRTT mode, I can get 9 to 12 KB/s uploads, but only 15 KB/s or so on downloads. If I force it to EVDO mode, then I can get 45-50 KB/s downloads but only 1 to 2 KB/s uploads. That is all using the same setup I listed (but with one error corrected below in that I left the S720 out)

KR1 router OR laptop -> S720 card -> 2' pigtail -> * amp amp -> 2' LMR400 -> lightning arrestor -> 30' LMR400 -> 9dbi gain omni 24" mast antenna 38' up on a rohn tower.

I really don't see how I can get the amp any closer to the antenna. My NEMA enclosure is about 6' off the ground and I have bulkead lightning arrestors pentrating the enclosure. The 30' LMR attaches to the bulkhead N-female of the lightning arrestor.

In this photo the front LMR400 cable goes up to the 1900Mhz mast way up on the tower, the rear LMR400 goes to the flat panel that is clipped off at the top of the photo to shoot the signal via 802.11g to my house about 800' away. The flashing is riveted on to cover the cooling fan which is mounted to the top to suck hot air out of the box.

I really NEED to get this to work. I'm close, but still too slow on the upload.

MaximumSignal

n6gn · Posted: Tue Sep 19, 2006 5:44 pm Post subject:

If I run the S720 card in my laptop instead of in the KR1, I can force it to run in 1xRTT mode. If I run in 1xRTT mode, I can get 9 to 12 KB/s uploads, but only 15 KB/s or so on downloads. If I force it to EVDO mode, then I can get 45-50 KB/s downloads but only 1 to 2 KB/s uploads. That is all using the same setup I listed (but with one error corrected below in that I left the S720 out)

It sounds like your 1XRTT is working reasonably (presuming those are all "B's' and not 'b's")
The issue seems to be the EV-DO uplink. This is strange since 1XRTT and EV-DO uplinks should be almost identical. The difference is the carrier and probably the tower location. Perhaps the 1XRTT signal and site/path is a lot better for you than the EV-DO one. Around here only about one quarter or less of the 1XRTT sites also have EV-DO.

I still think you can learn a lot by measuring signals in each mode w/o any other goo on the line. If it turns out that the EV-DO signal is pretty weak, and this is why downlink works but uplink so marginal you'll then know it and probablyl want to go directly to a gain antenna. This is not a bad idea in any case since it will reduce self-interference as well.
Go for a reflector antenna (dish) rather than a yagi if you can. Your likely to be able to do better than 20 dB with a dish and probably not better than 15 dBd with a dipole, no matter what the manufacturer claims. Then you'll know which tower(s) have EV-DO and what to do to optimize the path. At 2 GHz, typical foliage runs on the order of .25 dB/foot insertion loss. A good sized tree can undo all the good your height and antenna gain would otherwise provide.

As I've said, I think there is little reason for an amplifier nearer the antenna, or probably an amplifier at all, given your cable. You'll get a lot more bang for your buck out of adding antenna gain.

I like your heatsink, from my experience, this is definitely a good idea.

Let us know what you find.

n6gn

getut · EVDO Fledgling Joined: 16 Aug 2006 Posts: 11

MaximumSignal · Posted: Tue Sep 19, 2006 6:22 pm Post subject:

n6gn · Posted: Tue Sep 19, 2006 6:57 pm Post subject: