Joined: 31 Jan 2009 Posts: 114 Location: Ansbach, Germany
Posted: Sun May 17, 2009 7:49 Post subject: buffalo hp vs 1650
I've renamed the title since it appear's it's more of a buffalo hp vs the 1650 after doing more testing.
I had 4 devices, two buffalo hp 54g's flashed with the dd-wrt firmware built using the wifigator system. I then had two 1650's flashed with the Open-Mesh firmware. One each of the devices were placed on the second floor in the window of one house. The second set was placed 100 meters away in another building window. Between the two buildings is a huge brick house, tree, and other items so no line of site was possible. The buffalo routers had the stock antenna's, while the 1650's were using the internal directional antenna.
At the second location I plugged my laptop into the buffalo and did a wireless scan with dd-wrt, right off the bat I could see the main buffalo dd-wrt signal at around 70+ but not a hint of any signal from the 1650. The buffalo was able to connect and get online from the first one, very impressive. I'm about to start round two of testing where I'll place the omni antenna on the 1650 and see if that's the difference. Round three will be flashing my paid version of dd-wrt back onto the 1650 and see what I get. I have a feeling that dd-wrt is hands down going to beat out the current performance I'm getting with Open-Mesh.
The only reason I'm brining this up and giving input is in the hopes my request will be heard about a special build of the paid dd-wrt. I'm not the only one that would benefit, others that use dd-wrt for hotspots would also use this build. All I need is several scripts placed in the standard dd-wrt firmware to make everything possible. Otherwise it's way to painful and just not pratical to setup each and every device manual to comply with the needs of the system. Right now using the firmware mod kit with broadcom makes everything possible. I don't expect BrainSlayer to open up the firmware, just asking that he consider allowing a special build to suit a huge customer potential. I'll be back with round 2 and 3 findings, but I think it's going to be in dd-wrt favor.
Joined: 31 Jan 2009 Posts: 114 Location: Ansbach, Germany
Posted: Sun May 17, 2009 9:14 Post subject:
I went ahead and jumped forward a bit. I've now got dd-wrt back up and running on the 1650. I've made no changes in the firmware and attached the factory omni antenna. I've walked over to the house and stood on the porch with my laptop so far. I can do a scan and still see the buffalo hp signal with no problem, yet no signal what so ever from the 1650 running dd-wrt. I walked back towards the first house and got within line of sight on the devices. The buffalo hp is shoiwng almost full bars on vista while the 1650 is showing one bar. I then walked to the point where I could actually see the 1650 in the window, yet it still showed hardly any signal. What is going on here? I'm connecting another brand new 1650 and since I don't want to buy another dd-wrt I'll test open-mesh now with the omni. So far I'm unable to determine what is the cause of very poor signal. The only change I made to the buffalo was set the tx power to 100. I'm not sure what to set the 1650 at but I figured the factory dd-wrt settings would atleast outperform the buffalo even with 100 set. I'm off to read the transmit power section thread, see what I come up with.
edit: I'm not seeing any clue in the other post, other then maybe bumping up to 27 or so on dbm. I'm using the very latest firmware so it says I should only have to flip the switch on the outside case, which is set for external now. This is so far very disappointing signal. How is the buffalo HP outperforming this thing? Why do I have to pay for a key if the free home type Buffalo HP firmware can outperform the 1650? I have today to do further testing if anyone else has suggestions, but as of right now it's either the device or firmware that just plain stinks. I wish I had the factory firmware to truely do a comparision test, but looks like I cannot go back. I'll try the other 1650 in the meantime, rule out faulty hardware I guess.
do u have only 1 1650? i have a 2610 which is the same board just with other antennas and i receive indoor aps 200m away or outdoor units 2km away with nonlos.
maybe there are problems like some batch of 5610 have
BrainSlayer wrote:
in addition. the pre production samples from senao do show a 10 db less output power than the sales versions. so i hope you have the final hardware revision. but i believe you have not, since senao just released them some days ago
Joined: 06 Jun 2006 Posts: 7492 Location: Dresden, Germany
Posted: Sun May 17, 2009 16:03 Post subject:
first. the eoc1650 has not 27 dbm, but 23 dbm.
the eoc2610 has 27-28 dbm
beside this. which dd-wrt firmware version is installed on it? we resolved alot of antenna related issues in the last weeks _________________ "So you tried to use the computer and it started smoking? Sounds like a Mac to me.." - Louis Rossmann https://www.youtube.com/watch?v=eL_5YDRWqGE&t=60s
I'm sure your correct on the 23dbm stuff, I didn't change a thing and didn't have the device in front of me when I was typing. I just looked in the other post about signal strengh to see what was going on.
I spent hours doing tests the best way I could. I have two 1650's and the other is still running Open-Mesh but performs just as bad. I was going to flash it with dd-wrt but ran out of time. I figure either both devices are faulty(hard to beleive), something is up with both dd-wrt/open-mesh, or the buffalo 54g hp is just a bad ass piece of hardware making a connection that surprised me. Either way, the distance on the 1650 was extremly poor to the point my German Netgear router would probably outperform it. Maybe I will have to get a 2610 and try it out.
The open-mesh site has people finding poor performance out of the 1650, but there is one guy saying he's getting good results with dd-wrt and not open-mesh.
note: Wish I could speak german so I could present you what I wanted to input into a special build for the 1650, 2610, pico, and other devices that have atheros and work with the FON autoap, but it looks like it's a lost cause
Joined: 06 Jun 2006 Posts: 7492 Location: Dresden, Germany
Posted: Mon May 18, 2009 1:12 Post subject:
for sure you can speak german. we are all germans here in this thread. what you have to consider is the german power restrictions of 20 dbm for sure. but you can select a different country to override this. (new zealand for example)
unlike the ns2 or eoc2610, the eoc1650 has no poweroffset and will be limited to 20 dbm if you dont play with the regulatory limits. the superchannel feature would disable this limit too, but i dont think that its usefull to spend money into some dbm. so i can just advise you to change the country to new zealand for testing _________________ "So you tried to use the computer and it started smoking? Sounds like a Mac to me.." - Louis Rossmann https://www.youtube.com/watch?v=eL_5YDRWqGE&t=60s
Joined: 31 Jan 2009 Posts: 114 Location: Ansbach, Germany
Posted: Mon May 18, 2009 5:57 Post subject:
I will try that and for sure I will have to follow the 20dbm law. I'm curious now how much the buffalo hp is outputting to make the distance I describe. I'm still thinking maybe this equipment is faulty but I'll give it another chance. I'm located in Germany(up by the Rhine today) but can really only speak and understand english, that's what I meant before about wanting to communicate with you guys better.
It's highly unlikely but would you consider allowing a custom version of your paid level firmware to be built? During your building process I need some scripts hard coded in and maybe you could do it in such a way they would only run after activated? This would elimante any worry it would effect your activation program but still allow everything I need to run after buying your key? Sorry for asking so many times, i'm not sure if i'm asking something wrong or just not making sense. Just know a lot of people would buy a custom version of your firmware for these types of devices. I can send you a copy of each script, about 5-6 total. I need to adjust them for atheros but I think you could see what it's doing.
Joined: 31 Jan 2009 Posts: 114 Location: Ansbach, Germany
Posted: Mon May 18, 2009 17:16 Post subject:
I've sent you a link, please use the firmware modkit to break it down and you will see each and every current script that's input into the broadcom version we use. To run this like you mention would be impossible with out doing each and every build. I'm thining a modified version of the scripts for atheros and do away with the server ip and have that be the only thing that's entered into chillispot and required for the rest to kick into motion.
Joined: 31 Jan 2009 Posts: 114 Location: Ansbach, Germany
Posted: Mon May 18, 2009 17:23 Post subject:
More positive news for dd-wrt, this is another person showing good results showing dd-wrt is more powerful then open-mesh.
Another little update.
Quote:
I have two EOC-1650 units. One flashed with DD-WRT with EXT omni happening. The other EOC-1650 is also using an omni with open-mesh. I've alternated between the two units with the software to determine if there was any hardware issue. But I am coming out with the same results... as listed below.
DD-WRT EOC unit is set for Channel 2 while the other is on 5.
To give a little edge, I place the open-mesh flashed unit in the window facing the AP (EOC-2610) which is about 200 feet away with minor foliage about. The other unit is about 25 feet away opposite the room increasing the distance a minor amount. Handicapped... if you will.
EOC 2610 running a wireless survey is reporting the following:
DD-WRT flashed unit: RSSI -71 Noise: -95
Mesh flashed unit: RSSI -86 Noise: -95
Joined: 31 Jan 2009 Posts: 114 Location: Ansbach, Germany
Posted: Wed May 20, 2009 19:48 Post subject:
I'm not for sure so I will do further testing. If it was only one device doing it then maybe, but both are very underpowered atleast in the testing I've done. I'll mess around with them further since others are saying they got them to perform with dd-wrt. I'm now trying to grasp olsr on dd-wrt and mimick the VAP process with chilli and still have them all meshing.
