High Reflected Power Puzzle

FM does it with frequency!
User avatar
jthorusen
Posts: 85
Joined: Sun Sep 02, 2018 4:35 am
Contact:

High Reflected Power Puzzle

Post by jthorusen »

Greetings to the Forum:

Here's a little puzzle for you. If you have any brilliant ideas, please let me know as the station has been off the air for more than 2 weeks. Fortunately, it is only a low power FM, but we are rapidly approaching the point where we will have to notify the FCC.

Equipment: MT300 300 watt capable transmitter, made in Italy and fronted by Richardson Electronics. BKG88 single bay circularly polarized antenna by NicomUSA on a roughly 20 foot pole affixed to the side of a two story building with the bottom of the pole at approximately the floor level of the second story. Feed line is 43 feet of Andrew FSJ4-50B super-flex hard line with type "N" connectors on each end. In-line with the antenna is a Bird 4522 dual slug panel watt meter (uses same slugs as Bird 43).

Initial problem: Transmitter failed and off the air. Main 48 volt switch mode power supply in over-current shutdown. SD2932 FET in the PA shorted. Ordered replacement SD2932 (took 3 days) and installed it. Transmitter made 175 watts into Bird dummy load with no difficulty. Attempted to place transmitter on air. 100 % reflected power observed; transmitter shutdown circuitry copes with lesser amounts of reflected power but logic folds up with such high level of reflection; not only does the transmitter not shut down automatically but the "Enable" switch on front panel becomes inoperative; one must make a grab for the AC mains switch to shut down the transmitter. This is presumed to be the cause of the initial FET failure.

Swept the antenna using a VNA and found the resonance point to be about 600 KHz low. However, the return loss at the carrier frequency was around 19 dB, which should have been adequate. Unfortunately, the return loss goes to 0 dB with more than a few watts applied. I conducted tests of the feed line for shorts with a megger; found none. DC continuity OK. Connected feed line to transmitter (through Bird in rack) and connected dummy load to the antenna end of the feedline through a Bird 43 with our one and only 250B slug installed outside on the ground. Transmitter perfectly happy; 200 watts delivered to the load according to the Bird 43. Connected the feed line directly to the dummy load and brought the 250B slug back inside and installed in panel watt meter. No reflected power; transmitter making a bit over 200 watts just fine. Flexed and shook line and connectors; no problems detected.

Attempted to tune antenna in parking lot; lots of traffic made consistent VNA readings problematic. Discovered that the feed strap was loose when the clamp bolt was fully tightened due to mis-positioning of the bolt holes at time of manufacture; could not be tightened securely. Cleaned joints and shimmed feed strap; no change... 100% reflected power although looked acceptable on VNA. Tried a short length of RG-8X cable as a feed in the parking lot with transmitter on tail gate of pickup truck. Same 100% reflected power... about 150 watts regardless of which direction the Bird slug arrow points.

Conclusion: Defective antenna. Ordered replacement from NicomUSA. It arrived a few days ago. Put it up and it exhibits exactly the same failure... 100% reflected power. Repeated tests of feed line with dummy load at far end; everything works well, transmitter makes normal power.

Observed transmitter output spectrum using a 40 dB loss directional coupler in a Bird 43 watt meter with 20 dB of additional pads into a Rigol DSA18TG spectrum analyzer. Sweeping from 0 to 1.5 GHz; no spurs observed; everything was about 60 dB below carrier with only the spectrum analyzer noise floor showing anywhere except at carrier frequency. This was done first on the bench when the FET was replaced using a dummy load. The test was repeated with the transmitter connected to the antenna and 100% reflected power on the theory that the PA might generate a spur with a slightly reactive load. Both spectra very clean.

The new antenna was supplied with a picture taken of their VNA showing 30 dB of return loss at the carrier frequency. We can't see better than 20 dB and we get two different values for the frequency of the best return loss depending on which VNA we are using. At no time can we reproduce the factory test results, although 19 to 20 dB of return loss should be acceptable. Unfortunately, that isn't what we see with power applied.

We are going to try a new feed line on Wednesday, even though the old one passed every test I can think of. Two antennas that can be "seen" by VNA's but not by the transmitter. No spurs or out of band signals that could account for the reflected power. I will try again with the spectrum analyzer set to 0 to 200 MHz on Wednesday to see if there might be a spur close to carrier that can't be seen with the 0 to 1.5 GHz sweep, but I really don't think there is.

Anyone have any ideas on this one? I'm flabbergasted.

Thanks,
James K. (Jim) Thorusen
KB6GM
Central Coast Electronics
www.centcoast.com
NW Oregon Consulting Bdcst Eng.

Ray
Posts: 75
Joined: Sat Mar 22, 2014 10:04 am

Re: High Reflected Power Puzzle

Post by Ray »

Problems like this can drive you crazy.

I'd check the integrity of the FSJ4 Type N connectors installation.

Pull on them and see if they come off.
If the cable is not trimmed carefully the connectors will not make good contact and will be intermittent.
I use the cable prep tool (the one you use with the drill). It is important to make sure the prep tool completely finished the cut before the connectors are installed.

Ray

Kelly
Posts: 548
Joined: Thu Aug 21, 2008 5:32 pm
Location: Washington D.C. Area

Re: High Reflected Power Puzzle

Post by Kelly »

We've all been bitten by this one: Check the center pins on both N-connectors. Depending on how they were installed and whether they were the solder-on type center pins or captive, the pin may have pulled back lightly into the connector. That could account for why sometimes it works and other times not.
Skype:kellyalford Twitter: @KellyAlford

TPT
Posts: 969
Joined: Mon Dec 03, 2007 3:18 pm
Location: St. Marys, WV

Re: High Reflected Power Puzzle

Post by TPT »

Had a similar problem in building a new LPFM. VS300, combination of flex line and regular 1/2". Everything looked good--but high VSWR.

Tower climber (also a ham) asked--what frequency is this station on? Me: 96.3 He: The antenna is marked 93.1.

If you could borrow another transmitter or even an exciter might be worth the exercise to see if you still have a VSWR problem with something different driving the antenna. Mr. Murphy still lives.

User avatar
jthorusen
Posts: 85
Joined: Sun Sep 02, 2018 4:35 am
Contact:

Re: High Reflected Power Puzzle

Post by jthorusen »

Thanks much to all who responded. I appreciate the time you spent to formulate your replies. In response, I will say that we hope to try a completely new feed line on Wednesday. This will be about 50 feet of LMR600. This should eliminate the connector problem if present. We are set to pull the trigger on new feedline and connectors if necessary.

I would love to have a second transmitter to try. And... we do, sort of. Someone donated a Ramsey PX1 transmitter. He was going to sell it to us, except that he decided to fire it up on his bench for a test before the sale and it died. So.... he didn't feel like bothering with a repair job and gave it to the station. It isn't much of a transmitter.... it was Ramsey's top of the line years back, but it is only capable of 30 watts max so it really doesn't qualify as a backup transmitter for the MT300.

Anyway, the Ramsey uses a switch mode supply to provide the 12 volts at 12 amps it needs to run and the supply output collapses under load. I got to looking at replacing the electrolytic caps in the thing and by the time I ordered those and paid for shipping, I was looking at something like twelve or thirteen dollars. There was a complete nearly identical supply brand new on eBay for $16 with free shipping, so I just ordered that. As soon as it gets here and I can get it installed (and assuming the Ramsey has no other problems), I will have a 30 watt signal to test with.

I also intend to look more closely at the spectrum of the MT300 when the reflected power is extreme. I have to be careful with this; I am surprised that I haven't blown up the new FET already. It occurred to me that if the PA is flying with a slightly reactive load (the antenna) then the antenna, being tuned to the carrier frequency, may be forming part of a tuned circuit causing self-oscillation. I intend to do a narrow bandwidth spectral analysis of the transmitter when in the failed mode and I also intend to pull off the exciter BNC and remove drive and see if the output persists. I'll know more Wednesday. By the way, the antenna is marked with the correct frequency and I have a scope camera photo of the antenna response at the factory showing the return loss dip at the correct frequency.

Thanks again, all. I'll keep you posted.

73,
James K. (Jim) Thorusen
KB6GM
Central Coast Electronics
www.centcoast.com
NW Oregon Consulting Bdcst Eng.

Lee_Wheeler
Posts: 125
Joined: Sun Jul 24, 2016 1:53 pm
Location: Kansas City

Re: High Reflected Power Puzzle

Post by Lee_Wheeler »

Are there grounding kits on the coax?

The outer conductor is so paper thin on coax these days that I have had trouble with the outer conductor actually coming apart at the point where a ground kit was installed due to the tower crew getting overly aggressive with the utility knife when they gut the jacket. In all cases it has shown up just after it gets cold and the line tries to contract in length. The load goes bad and you can't hardly see anything with a TDR.

Be sure to remind the tower guys, more than once, to take it easy when they install the ground kits.

...Lee

User avatar
jthorusen
Posts: 85
Joined: Sun Sep 02, 2018 4:35 am
Contact:

Re: High Reflected Power Puzzle

Post by jthorusen »

Thanks for the response, Lee! That would definitely be something to check. However, there is no tower involved; only a 20 or so foot pipe that hangs on the side of the building with two wall brackets. There are no grounding kits installed anywhere on the feedline. It is certainly possible that there is outer conductor damage; complete replacement of the feedline with a piece of different line with different connectors should rule out that cause; this should be done tomorrow, weather permitting.

The trouble is not intermittent, however. It is 100% failed every time we try to put power into the antenna. The original feedline passes 200 watts to the far end as long as it is connected to a dummy load. We have access to the entire length of the feedline and have pulled and pushed, flexed and wiggled it pretty much end to end and there are no hiccups as long it is terminated in a purely resistive load, i.e. the Bird dummy load. It is only when we go to the antenna that we have problems.

I am leaning more and more toward the idea that the PA is oscillating on its own, but it only does it with the antenna present. It is perfectly calm when resistively terminated. In order to set and test trip points in the VSWR protection circuitry, I built a gadget which is nothing more than a variable capacitor connected across a coax connector. This is connected into the line to the dummy load with a "T" fitting. I calibrated the thing for return loss using my network analyzer. When placed in the line between the load and the transmitter, it introduces -j reactance into the load connection. I can control the magnitude of that reactance with a knob on the outside of the box. For initial connection, it has no effect on the transmitter even though it introduces some reactance even with the variable cap fully open. As I increase the capacity, the change in reactance causes the return loss of the "VSWR Inserter" and dummy load combination to decrease (increasing VSWR). I can therefore insert anywhere from a fairly small decrease in return loss to a large decrease in return loss. As I do this, the transmitter obediently first folds back its power output (measured with a Bird 43) and then shuts it down when the VSWR gets too great (return loss too small).

This test and setup was performed on the bench after the first 100% reflected power episode. With test results like this, I felt confident that the transmitter should behave if the antenna was correct. This impression was so strong on my part that I recommended purchase of a new antenna. Imagine my chagrin when the new antenna behaved exactly the same way as the old one.

So, weather permitting, I hope to have more data after we are through tomorrow.

Thanks again,
James K. (Jim) Thorusen
KB6GM
Central Coast Electronics
www.centcoast.com
NW Oregon Consulting Bdcst Eng.

DaveSt
Posts: 56
Joined: Mon Jul 23, 2018 2:21 pm
Location: UK

Re: High Reflected Power Puzzle

Post by DaveSt »

The problem with the design of antenna you have is that they are extremely narrow band. That can be quite challenging for any transmitter to drive. It increases the likelihood of instability in the transmitter PA. That may also have been made worse after the change in transistor. The instability will show up as spurs close to the operating frequency, maybe 0.5-2MHz from the carrier. Your capacitor box may help with this. Another trick is to add a few feet of cable at the transmitter output. Try a few lengths.

With the antennas, you have to make sure they are installed exactly as the manufacturer recommends. Make sure the pole they are on is the correct diameter and extends above and below the antennas sufficiently. Also make sure no other metalwork is near the antennas. And make sure the vertical spacing is correct. Any thing that is not exactly right will make the tuning change. When this is correct, you can expect better than 30dB return loss. If the antennas are detuned to the extent you have 20dB at the carrier frequency, remember you need to allow for the modulation, so look at the return loss 100kHz away too. It will be much worse.

From what you have said about the transmitter, it sounds like you should be budgeting for a replacement soon.

Ray
Posts: 75
Joined: Sat Mar 22, 2014 10:04 am

Re: High Reflected Power Puzzle

Post by Ray »

I was thinking you had the BKG77 slant vee dipole design which is much more broadband (entire band). I use it as a standby antenna.
I also have a BKG88 in storage, I only want to use it as a last resort because of the narrow bandwidth. If you get ice/snow on the antenna it will detune. There are some off-shore aluminum slant vee antennas for about $400 that work well also. There was a pair of Nicom BKG77s that sold last month on ebay for about $500. I've found most 300 watt transmitters do not like VSWR > 1.5:1 and start sending alarms and folding back power.

I'm wondering if anyone has modified the Nautel VS300 setup for a higher return power threshold. When the return power exceeds approx 10 watts, the Nautel VS300 starts sending alarms with the default setup. There is an advanced menu (I've never used) where you can increase the alarm level threshold. It seems like the 300 watt Mosfet should be able to handle more than 10 watts return power.

Ray

ncradioeng
Posts: 77
Joined: Wed Nov 18, 2015 3:58 pm

Re: High Reflected Power Puzzle

Post by ncradioeng »

How far is the transmitter from the antenna? If there is poor shielding in the TX, you might have an RF feedback situation causing some issue.

Steve Brown
Posts: 15
Joined: Tue Feb 21, 2017 4:23 pm

Re: High Reflected Power Puzzle

Post by Steve Brown »

I don't know anything about the transmitter you're using, and maybe this is a long shot, but several years ago I ran into a similar problem with an offshore exciter/PA amp combination. Into a dummy load at the end of the line, no problem, but terminated into the antenna the VSWR rose, not to a terrible level, BUT the output of the amp was VERY spurious. After a return to the European factory and a list of changes they claimed to have made to the amp, the problem persisted. Finally, with no help from the factory or the equipment manuals, I figured out that I was controlling the power output from the exciter, not the amp. I had the exciter running about 5 watts and had the PA amp output set to deliver 500 watts. After I wised up and ran the exciter up to about 20 watts and then adjusted the PA amp for 500 watts all was good.

Steve B

User avatar
jthorusen
Posts: 85
Joined: Sun Sep 02, 2018 4:35 am
Contact:

Re: High Reflected Power Puzzle

Post by jthorusen »

Thanks for all the replies. I have found the problem. Now, I have to figure out what to do about it.

The problem is indeed a spur.... but one considerably removed from the carrier frequency. The carrier freq is 98.7; the spur (extreme high power) is at 90 MHz almost exactly.

I couldn't see it before because I couldn't tell the difference between 90 and 98.7 when looking from zero to 1500 MHz. When the PA takes off, there is a little output at 98.7, but most of it is at 90 and there are no other emissions above the noise floor so everything looked OK. It still works beautifully into a dummy load.

DaveSt seems to have hit the nail on the head. Unfortunately, the antenna does not work well because it is too low and in too close proximity to the building and other structures. The return loss is only about 19 dB and that is not quite at carrier. We are scraping by the RF exposure limits by the skin of our teeth. The antenna really needs to be 12 meters higher than it is.... but funding constraints make that difficult.

I am going to try a band-aid approach.... insert a 1/4 wave stub filter for 90 MHz and see if that quiets it down. I suspect it may just move the spur somewhere else, but I will give it a try. We have already tried a number of different lengths and types of feed line without success.

I will try to get the station to consider doing a fund-raiser to replace the antenna with a BKG77 and get it up higher. Unfortunately, that's not an option for now as we will have to file a construction permit to change the antenna type or to raise it. The station has been streaming only for over two weeks now and I don't know what the FCC ramifications would be if we tell them we need a construction permit to get back on the air.

I wanted to post a couple of spectrum analyzer screen captures, but apparently this site does not allow them to be embedded in the text; I don't want to go to the bother of uploading them to some web site just to get them on here. Anyway, for DaveSt, the null in return loss is very broad; not sharp at all as in the factory test data. 250 KHz either side of the null, the return loss has only degraded to 17 dB or so.

Running on 4 hours sleep so I am off to bed. Thanks again for all the helpful input. It is greatly appreciated.

73,
James K. (Jim) Thorusen
KB6GM
Central Coast Electronics
www.centcoast.com
NW Oregon Consulting Bdcst Eng.

ChuckG
Posts: 1068
Joined: Tue Mar 11, 2008 11:12 pm
Location: Moo
Contact:

Re: High Reflected Power Puzzle

Post by ChuckG »

If that's the transmitter I am thinking, it would have two light blue modules in the exciter? This was about 15 years ago.... I wonder how far gone the PSU and caps are in that box if it's that old?
<><><><><><><><><>
Chuck Gennaro
Central Wisconsin
www.northwoodstech.com

DaveSt
Posts: 56
Joined: Mon Jul 23, 2018 2:21 pm
Location: UK

Re: High Reflected Power Puzzle

Post by DaveSt »

It looks like the PA is oscillating by itself, which is not unusual in some PA designs. If the drive to the PA section is removed, most likely it will carry on oscillating. When the MOSFET was changed, if the bias current was not set correctly then that can make the PA more likely to take off like this. It is also worth changing any electrolytic capacitors in the DC supply on the PA PCB. Use the best low ESR type you can get.

I was thinking you had a pair of antennas, but I re-read your original post and see it is just one. There will not be much you can do with the single antenna other than getting it clear of the surrounding objects. Even a few feet would help. You could try tuning it where it is. If you are using a PC based VNA, you can get a VNC connection to the PC from a smart phone. I have done that before with this type of antenna and it worked well.

fjstuden
Posts: 6
Joined: Thu May 29, 2014 5:18 pm

Re: High Reflected Power Puzzle

Post by fjstuden »

A 1/4 wave piece of coax line in series with feeder to move the impedance seen out of band might stop it until you can find the real solution. Or, a 3 dB inline attenuator, either a real one or made from lots of coax. Half power is better than nothing and a temp authorization for lower power is easier than moving antennas.

Post Reply