Let's talk about PIM testing.....

[Marconi]

Let's start a discussion here about passive intermod. What is it? Why is it bad? How do we find it? More and more carriers are requiring PIM testing with the addition, replacement of antennas and with new site construction.

What is the most cost effective testing method?

[remote]

They say that PIM affects base station performance. How true is this? I agree that testing may ensure the quality of the installation, but what about the quality of those antennae, couplers, connectors, coax cables and jumpers. Are those internal to the manufacturers?

[leo]

What is it?

Am curious too. In my non-RF brain it sounds like something mechanical. Either a connection not tight (that lets the center pin move), a design that leaves part of the edge of the outer conductor unsupported, or some other mechanical item or surface that is being affected (causing movement, causing interference) by the RF signal.

Of course I've sniffed a lot of bitumastic paint in my field days so I'm probably way off.

[naazlee]

I was wondering what is the most effective way of PIM testing. Is it true that the PIM affects the base state performance?Why is passive intermod considered to be bad?

[farcanal]

Hi Folks,

I have been watching this post for a while and feel I can offer some info.

PIM is not a difficult test to understand. Finding the fault and predicting PIM is the hard part...

PIM testing has been around for 15-20 years, it is not new.

Filter mfg's have been PIM testing for 12+ years. I was originally from the filter world and first started PIM testing 15 years ago. We made our own test systems.
It is a normal and common test for the better filter and cable mfg's and has become a normal part of their production system.

All cell sites have PIM issues and have always done so. It was not an issue until we started pushing so much voice and data over spread spectrum modulation systems.
Spread spectrum is what we currently have and LTE and any standards going forward are likely to be the same. It is a very efficient system in terms of data speed and number of users but can become affected by PIM.

The principle of PIM is very simple - inject 2 high power signals into a device and if it is not good it will give a PIM signal which can be measured.
PIM is only bad when it is louder than the signal you want to recieve. You can have PIM and while ever it is weaker than the rx sensitivity it will not affect service.

PIM is broken down it to two parts
transmitters - 2x 20 watt signals are injected into an antenna system.
recieve - based to the Tx frequencies it is easy to calculate the PIM frequencies. In the USA we tend to look for the 3rd order on the low side as it is the one that interferes with our cell systems.
take 870MHz and 894Mhz these are 24MHz apart. if you subtract 24 from 870 this equals 846Mhz which is in the Rx band. If this condition exists the 3rd order PIM will interfere with the performance of the BTS if the PIM signal is powerful enough.

The engineering equation goes like this:
2xF1 +/- F2 = 3rd order
3x F1+/- 2F2 = 5th order and so on...

Consider that the Rx sensitivity of the BTS is generally around -105dBm, this means that any signal stronger than -105dBm is usable by the BTS. This is the level that the BTS will hand off or terminate the call/connection.

If PIM falls into the rx band, the BTS goes deaf and coverage area receeds. If your PIM level is -90dBm then the cell site needs to hand calls off early in terms of distance. This clogs the cells around the high PIM cell as they see more traffic than they should.
If it is a data connection, the speed will slow dramatically until the connection is lost.
Ultimately, the user experience is not good.

Within a cell site, PIM is not the real problem....
All components were designed with PIM performance in mind and may have actually passed a PIM test at some point. The problem is that all of these components are old and may not have been installed properly in the first place.
Continuous price pressure from outside the USA has forced prices and quality down.
In production, PIM is a difficult spec to meet. An antenna that needs to sell for $300 to be competitive with asian units is not going to be put together very well. When you throw shipping, hanging, temperature change, vibration, storms and underpaid tower people in the mix it is not suprising that these problems now exist.

I'm not trying to offend anyone who builds towers but if the customer does not pay well they usually get the minimum quality of work done.
In my experience with PIM testing on site, about half of the problems are due to faulty termination work and half is due to deterioration/poor quality components.
A good example is:

Verizon pays well
At&t's turf vendors do not - Who has the better network?

In the filter world, the company I worked for sold Lucent BTS filter for $1800 at the start of the contract and 6 months later we had to drop to $450 to be competitive with other suppliers. We had no choice but to get really cheap in order to compete. The antenna industry is worse..

The shoddy contractors out there have learnt how to pass a line sweep and get paid. PIM is an excellent test of construction quality and finds loose connectors easily.
A line sweep is a measure of impedance over frequency and is still as important now as it ever was. PIM is not frequency dependant and good PIM at 800MHz will indicate good PIM at 1900MHz, in most cases when frequency dependant devices do not exist, such as an TTLNA.

There are a few companies making PIM test gear and most are small, do not really offer a test equipment background and the after sales service is poor to say the least.
Rumors are abound of new companies and technologies entering the market.
Current testing is done by a process of elimination - bypass stuff until it PIMs good and you have found your problem.

If you are not PIM testing I would suggest becoming familiar with the technology as it is definately here to stay. Most operators expect/ask for PIM test results and in the case of Verizon, you will not work for them unless you can PIM test. AT&T is about to go the same way. Cricket has been PIM testing for 3 years or more in some markets.

I hope this helps and I am happy to add more if anyone wants to hear it.

thanks,

Farcanal

[Marconi]

Wow, great input farcanal. Thanks for posting.

Last year at CTIA there was maybe one or two PIM test equipment vendors showing equipment. This year there was 10+. You are right PIM testing is here to stay. But if the carriers are going to require it they better be ready to pay for it.

Thanks again for your input.

[SuperTech]

Yea, great information. Thanks for taking the time to post.

[farcanal]

You are all very welcome. I have watched what has been going on here for a while and would like to see this site become popular. LTE and PIM seem to get some hits!

Marconi, you are correct, PIM is a bid deal and it needs to be paid for somehow. There is very little consistancy around the country and between carriers on price.
Some regions buy the equipment, loan it to the contractor and still pay $4k per site.

Others insist the contractor owns the equipment and expects it to be amortized into the job, and this is a shame. There needs to be a balance.

Turf vendores are charging carriers for PIM but seem reluctant to pay the contractors anything meaningfull for the actual service.

The cost of PIM can be put into real figures. I have personally performed experiments where the baseline PIM value for a particular 1900MHz EVDO line was -125dBm/167dBc.
A preliminary drive test was conducted by the owner/operator of the site and download speeds were mapped and recorded.
We then installed adaptors into the line and bought the baseline up to -107dBm/150dBc.
These figures are not considered failures at all, most cellular operators spec -97dBm/140dBc as the pass mark.
A drive test was conducted with the new reduced PIM performance in place, EVDO back on line with some percentage of users connected.
The results showed a 20% drop in download speed, and this is a remarkable effect for such a minor change in PIM level.

The same effect is true for 850MHz users. Sites that are tested and repaired for PIM behave more like the cell system design software said they would....

I once had a friend tell me about a site, 1900MHz in Michigan where the operator had spent over $120K trying to get it to perform correctly. They had basically replaced everything outside of the BTS, which was next on the list.
The guys I know went in, perfomed PIM tests, replaced a couple of antennas and connectors and left.
The perfomance stayed the same until I told them to force a calibration.
After the site recalibrated, it has never had a problem since. It only took a few hours to do the actual work.
It is most important to remember to force the BTS to recalibrate after you perform PIM testing. This is the mistake most folks make. You think you have fixed the problem only to find the stats are the same the next day.

The BTS monitors the Rx noise floor day and night and averages this figure. Most BTS manufacturures do this a little differently but at the end of the day, this noise floor figure is assumed. If the BTS thinks it is -105dBm and PIM interference is causing it to be -90dBm, the BTS will adjust power levels to keep itself on air and this causes LNA's to lower or raise gain. It also tells you handset to wind the Tx power up.
Have you ever had a day where you phone just seemed to burn its battery up in no time? I certainly have and it is because it is transmitting "louder" to get past the noise.

Sorry to ramble on again, I just find this stuff interesting.

Thanks

[Marconi]

More great info farcanal, thanks for posting.

So let's take a look at whats out there for test equipment. A quick Google search came up with the following three manufacturers. Please add to the list if I have missed any brands. There has to be more.

Summitek

Boonton

Rosenberger

[SuperTech]

I looked at the Summitek equipment at CTIA earlier this year. Nice looking package but you have to buy one for each frequency range I believe. Also saw the Rosenberger.

I am sure that at CTIA 2011 the number of PIM testing equipment manufacturers will double.

Hey farcanal....do you work for one of the equipment companies or some engineering company?