If a pigtail is loosing 6-7db it gets thrown in the trash. It doesn't have to be lost in an RF standpoint it could be lost through heat also.
Kurt Fankhauser WAVELINC P.O. Box 126 Bucyrus, OH 44820 419-562-6405 www.wavelinc.com -----Original Message----- From: wireless-boun...@wispa.org [mailto:wireless-boun...@wispa.org] On Behalf Of Tom DeReggi Sent: Thursday, April 30, 2009 2:53 PM To: WISPA General List Subject: Re: [WISPA] RB333/433 eliminating self-interference test Greg, Excellent comments and questions... And My answer to you is..... "all of the above". Scott, And I agree, each card brand and model will have its own properties. I just used one high power brand in a CPE for the last time (which I will not mention for professional courtesy) that as a CPE can hear a -50 signal awesome at 54mb modulation, but the AP receiver can't make out the CPE's signal and CPE stay associated unless the CPE transmits at 12mbps modulation or lower. Clearly distortion from the transmit amp, considering all the CPEs with CM9s can successfully transmit at 54mbps modulation. If we extended this converstaion to full relevence, we'd extend it to ask the questions for each and every manufacturer's cards. What I'm looking for is establishing the best choices to optimize success. Its not a black and white world here. I have systems in the field that have two 5.x cards in them and operate fine on 10Mhz channels with only one channel seperation in between. But I had a XR9 and DCMA82 (5.x) card in a 2 port AP System, where I had to reduce the 5.X card's power down to 10-12 db, in order for the 900Mhzcard to associate with its client. I can give examples of where details may help us.... 2 antenna ports cards are more available. But I may want to buy single antenna port cards, if it helps reduce noise from other cards in teh system, IF antenna port is a place of noise injections. But I may chose a low power card instead, if Amps is a place of injections greater than that of a second unused antenna port. Sure two cards can co-locate, but why not install in the method taht will minimize self interference, but still meet the minimum need of teh deployment? Today... for example.... I build a Dual Pol MIkrotik PTP w/433, and decided to put a second mpci card in the unit, but it only was going to use single pol, since the case only had space for one external pigtail. What card would be best to isntall, not to interfere with the first primary more important PTP link? I chose to make a isolation plate between the cards. I took a peice of cardboard wrapped it with tin foil, and put it inside a 3"x3" static bag. I then stuck it inbetween the two stacked Mpci slots. Because MT has 3 stacked slots, and I used the top and bottom one, there was plenty of room to insert the shielding without restricting airflow to cool the cards. It seemed to help. (although didn't record exact before after results). lastly, amped and non amp'd cards are not equal in design. For example, when the amplication is done in a single device there is no connection between two devices. With a second add-on amp embedded on the card, there is a second path entering into that amp, where noise can be induced to the amp. AMPs are also designed to work at a specific power level to acheive the best noise reduction, when it does not operate at that level, there also becomes a situation where the amp is underloaded or overloaded, causing more distortion. When there are two amps working togeather, there are now two points and more vaiables to configuring cards to be operating at the least amount of distortion. For example, an Amp may work best if its input is 13db, but the first amp may not output 13db constantly. Actually, its one of the reasons here were pre-amps in hgih end hi-fi gear, to make sure the signal all amps where working at their optimal powers and optimal signals injected into them. So at the end of the day, I guess all that really matters is.... How much distortion the specific card solution transmits or hears. Sure its possible that the Hi-power cards could be designed to be more resilient to distortion. Maybe that is a reason why they have much higher receive sensitivities? But then again, that was not the case when XR2s were compared to 200mw Prism cards. ManyWISPs reported better results from the PRism, regardless of what the spec sheets said. The pigtails also could be acting as antennas. I wonder how much loss they have. Some people reported pigtails having as much as 6-7 db loss if they were made poorly, and that energy loss all goes to somewhere, probably RF interference. For example, I wonder if crossing the pigtails or running them parallel can effect how much self interference betwee n the cards exist? Tom DeReggi RapidDSL & Wireless, Inc IntAirNet- Fixed Wireless Broadband ----- Original Message ----- From: "." <os10ru...@gmail.com> To: "WISPA General List" <wireless@wispa.org> Sent: Wednesday, April 29, 2009 10:38 PM Subject: Re: [WISPA] RB333/433 eliminating self-interference test > I'm curious to know if you're taking about transmit amps or receive > amps. When you say noise are you meaning transmitted noise (meaning > spectral impurity such as distortion, or do you mean unintentional > radiation of the desired transmitted signal?) or do you mean receiver > noise such as a higher noise floor, or signals considered to be noise > which are being picked up by the higher sensitivity receiver? I'm > assuming you mean transmitted noise of some kind as a result of the > transmit amp but I just want to clarify. Thanks! > > What's the point of these router boards that have multiple radio card > slots if you can't have the radio cards that close together? > > Greg > > On Apr 29, 2009, at 9:20 PM, Tom DeReggi wrote: > >> There is nothing wrong with lowering the power on them. >> I personally love SR5s. >> >> The facts are though that cards with add-on amps embedded have the >> potential >> to be noisier than one that does not. >> How much noisier, I can't say. That was part of tthe goal, to >> determine if >> XR5s are as clean as CM9s, and if there is a distinguishable >> difference or >> not. >> >> Tom DeReggi >> RapidDSL & Wireless, Inc >> IntAirNet- Fixed Wireless Broadband >> >> >> ----- Original Message ----- >> From: "Kurt Fankhauser" <k...@wavelinc.com> >> To: <wireless@wispa.org> >> Sent: Wednesday, April 29, 2009 6:30 PM >> Subject: Re: [WISPA] RB333/433 eliminating self-interference test >> >> >>> >>> Whats wrong with using XR5's and lowering the TX power on them? >>> They are >>> more rugged and have better RX sensitivity than many other cards. >>> -- >>> Kurt Fankhauser >>> WAVELINC >>> P.O. Box 126 >>> Bucyrus, OH 44820 >>> 419-562-6405 >>> www.wavelinc.com >>> >>> >>> --------- Original Message -------- >>> From: WISPA General List <wireless@wispa.org> >>> To: WISPA General List <wireless@wispa.org> >>> Subject: Re: [WISPA] RB333/433 eliminating self-interference test >>> Date: 04/29/09 16:31 >>> >>>> >>>> The first question is "why are the 4 mpci cards in teh RB600 >>>> seeing >>> each >>>> other so loudly"? >>>> There lies the problem needing fixed, because of course we want to >>>> use >>>> one >>> >>>> RB433, instead of 3 RB433s, to accommodate 3 mpci cards. >>>> (even if different channels and freqs). >>>> >>>> First question to you... "am I assuming correct that you >>>> still kept >>> the >>>> dummy loads on each of the mPCI cards, when testing all in teh same >>> RB600"? >>>> >>>> What is a bot disturbing is that you said you used a XR5. That >>>> means the >>>> card had a single antenna connector and a MMCX style, which is >>>> supposed >>>> to >>> >>>> give better isolation. >>>> >>>> (note: some have advocated that Ufl is as good as mmcx, regarding >>>> rssi >>> loss, >>>> stating that the UFl connector itself has less loss than the gain >>>> MMCX >>> adds >>>> by enabling thicker pigtail cable. I always still prefer MMCX >>>> because it >>> is >>>> more rugged abd less likely to break pigtails connectors in things >>>> like >>>> Rootenas that are not easy to access with short pigtails. But >>>> surely I >>>> thought mmcx would also add better shielding/isolation from outside >>>> sources.) >>>> >>>> So using XR5s, it would infer that the cards saw each other because >>>> either >>> >>>> loss from pigtail cable, loss from mmcx connector, or simply the >>>> cards >>>> electronics. >>>> The next relevent info might be to determine if it is the amp >>>> circuitry >>>> driving this interference. Just like a pair of PC speakers can >>>> sometimes >>>> pickup music radio. >>>> >>>> For years Lonnie (StarOS) gave teswtimonials for lower power CM9s >>> performing >>>> better than Amplified cards (SR5) for short range applications, >>>> because >>> they >>>> were quieter. >>>> >>>> So there is about 8db difference between a SR5 and a CM9. I wonder >>>> if you >>>> repeated your tests, but used CM9's instead (no ext embedded amps), >>> whether >>>> you'd just hear the other adjacenet radios at 8db lower, >>>> proportional to >>> the >>>> spec of the radios, or if you hear the otehr radio much much less, >>>> because >>> >>>> it doesn;t have the amp to pcikup the interference? >>>> >>>> >>>> Tom DeReggi >>>> RapidDSL & Wireless, Inc >>>> IntAirNet- Fixed Wireless Broadband >>>> >>>> >>>> ----- Original Message ----- >>>> From: "Kurt Fankhauser" <k...@wavelinc.com> >>>> To: "'WISPA General List'" <wireless@wispa.org> >>>> Sent: Wednesday, April 29, 2009 10:33 AM >>>> Subject: [WISPA] RB333/433 eliminating self-interference test >>>> >>>> >>>> > About a week ago there was some discussion about 5ghz radio's >>>> being >>>> > installed in the same board and causing self-interference on >>>> adjacent >>>> > channels and possible even on the entire band thus decreasing >>> throughput >>>> > on >>>> > backhauls. Because even if you were operating on frequency's >>>> 5745 >>>> and >>> 5825 >>>> > the two radio's would have side lobe harmonics that if >>>> installed in >>> the >>>> > same >>>> > enclosure they would still "hear" each other at >>>> that short >>> of separation. >>>> > I >>>> > decided to combat this problem and find a solution and share my >>> experience >>>> > with the list. >>>> > >>>> > >>>> > >>>> > I installed a single XR5 card into 3 different RB433's with >>>> indoor >>>> > enclosures. I also installed foil tape which I obtained from >>>> the >>> local >>>> > True >>>> > Value store for $2.49 on all the vent holes and unused bulkhead >>> connector >>>> > holes. This was done in order to prevent RF side lobe leaks >>>> from the >>> three >>>> > radio's that would escape from the indoor enclosures >>>> themselves. >>> Having >>>> > only >>>> > 1 card inside each enclosures I should not have a heat >>>> problem as >>>> the >>>> > outdoor box will not be in direct sunlight. >>>> > >>>> > >>>> > >>>> > I then stacked all 3 enclosures on top of each other with dummy >>>> loads >>> on >>>> > each of the N-bulkhead connectors and did some testing. This >>>> is what >>> I >>>> > found: >>>> > >>>> > >>>> > >>>> > I set the bottom board as AP and the middle board as Client on >>> frequency >>>> > 5825. Even with this close of separation the two XR5's could >>>> only >>>> see >>> each >>>> > other at -83 on the same channel. With the top board >>>> connecting to >>> the >>>> > bottom board they could only see each other at -90. Keep in >>>> mind >>>> this >>> is >>>> > on >>>> > the same frequency so adjacent channels should be much less >>>> than >>>> that >>>> > possibly even in the -100 ranges. Wish I had a spectrum >>>> analyzer. >>> With two >>>> > boards separating the AP and Client there was no link at all. >>>> The >>>> two >>> >>>> > boards >>>> > could not even see each other in an AP scan. >>>> > >>>> > >>>> > >>>> > Just as a comparison with the same radio's installed all on a >>>> RB600 >>> not on >>>> > top of each other but in the adjacent mpci slots the radio's >>>> were >>>> all >>> >>>> > seeing >>>> > each other at -30's. So I gained roughly -55 db of separation >>>> by >>> doing it >>>> > this way. So all I would have to do now is make sure that the >>> antennas on >>>> > the tower have at least 10 foot of vertical separation and the >>>> > self-interference problem should be gone and I should be >>>> enjoying >>> much >>>> > more >>>> > throughput! >>>> > >>>> > >>>> > >>>> > Thoughts anyone? >>>> > >>>> > >>>> > >>>> > >>>> > >>>> > Kurt Fankhauser >>>> > WAVELINC >>>> > P.O. Box 126 >>>> > Bucyrus, OH 44820 >>>> > 419-562-6405 >>>> > www.wavelinc.com >>>> > >>>> > >>>> > >>>> > >>>> > >>>> > >>>> > >>>> > >>>> >>>> >>>> >>> ---------------------------------------------------------------------------- ---- >>>> >>>> >>>> > >>>> > >>>> > >>> ---------------------------------------------------------------------------- ---- >>>> > WISPA Wants You! Join today! >>>> > http://signup.wispa.org/ >>>> > >>> ---------------------------------------------------------------------------- ---- >>>> > >>>> > WISPA Wireless List: wireless@wispa.org >>>> > >>>> > Subscribe/Unsubscribe: >>>> > http://lists.wispa.org/mailman/listinfo/wireless >>>> > >>>> > Archives: http://lists.wispa.org/pipermail/wireless/ >>>> >>>> >>>> >>>> >>> ---------------------------------------------------------------------------- ---- >>>> WISPA Wants You! Join today! >>>> http://signup.wispa.org/ >>>> >>> ---------------------------------------------------------------------------- ---- >>>> >>>> WISPA Wireless List: wireless@wispa.org >>>> >>>> Subscribe/Unsubscribe: >>>> http://lists.wispa.org/mailman/listinfo/wireless >>>> >>>> Archives: http://lists.wispa.org/pipermail/wireless/ >>>> >>>> >>>> >>>> >>> >>> >>> >>> ---------------------------------------------------------------------------- ---- >>> WISPA Wants You! Join today! >>> http://signup.wispa.org/ >>> ---------------------------------------------------------------------------- ---- >>> >>> WISPA Wireless List: wireless@wispa.org >>> >>> Subscribe/Unsubscribe: >>> http://lists.wispa.org/mailman/listinfo/wireless >>> >>> Archives: http://lists.wispa.org/pipermail/wireless/ >> >> >> >> ---------------------------------------------------------------------------- ---- >> WISPA Wants You! Join today! >> http://signup.wispa.org/ >> ---------------------------------------------------------------------------- ---- >> >> WISPA Wireless List: wireless@wispa.org >> >> Subscribe/Unsubscribe: >> http://lists.wispa.org/mailman/listinfo/wireless >> >> Archives: http://lists.wispa.org/pipermail/wireless/ > > > > ---------------------------------------------------------------------------- ---- > WISPA Wants You! Join today! > http://signup.wispa.org/ > ---------------------------------------------------------------------------- ---- > > WISPA Wireless List: wireless@wispa.org > > Subscribe/Unsubscribe: > http://lists.wispa.org/mailman/listinfo/wireless > > Archives: http://lists.wispa.org/pipermail/wireless/ ---------------------------------------------------------------------------- ---- WISPA Wants You! Join today! http://signup.wispa.org/ ---------------------------------------------------------------------------- ---- WISPA Wireless List: wireless@wispa.org Subscribe/Unsubscribe: http://lists.wispa.org/mailman/listinfo/wireless Archives: http://lists.wispa.org/pipermail/wireless/ -------------------------------------------------------------------------------- WISPA Wants You! Join today! http://signup.wispa.org/ -------------------------------------------------------------------------------- WISPA Wireless List: wireless@wispa.org Subscribe/Unsubscribe: http://lists.wispa.org/mailman/listinfo/wireless Archives: http://lists.wispa.org/pipermail/wireless/
