Completely removing (filtering) a 4th order harmonic (458MHz) from FM TX

[radiofish]

Need to get rid of a 4th order harmonic, already down 80 dB below carrier as required, but interferes with co-located receiver on tower. Any suggestions on filters to do this? Low pass harmonic filters?

Already have a harmonic filter on the transmitter output. How about putting a filter after the exciter, before the transmitter IPA/PA?

 

[Tom Wells]

If a filter has already been placed in the TX output (-80 db), the next place to filter is the line to the receiver.
At 458 Mhz, filters become small and reasonable for receivers, as opposed to transmitters. You don't say what the receiver or application is. It may be possible to simply to loss out the problem with a padder. Another -20db might not hurt the desired signal yet remove your
problem with the 4th harmonic.

A filter between the exciter and PA will be useless if the products are all from the transmitter.
If they are from the transmitter, a good look at the schematic will show if sufficient RF-bypassing is provided.
The BIG electrolytics in the filter outputs are good for low freq AC/RF, but as they age, do less and less to bypass stray RF and harmonics.
It may be necessary to add/augment RF bypassing with .01 mfds liberally in the DC distribution ckts.

 

[fugazi]

Use a simple but very effective Stub filter.
Can be made with a T connection and a piece of 7/8" cellflex coax.

Is the transmitter in the same chamber as the transmitter?

regards,

Evert

 

[rfry]

Need to get rid of a 4th order harmonic, already down 80 dB below carrier as required, but interferes with co-located receiver on tower.

First off, if your 4th harmonic lands on 458 MHz, your FM tx would have to be on 114.5 MHz, which it probably isn't. So assuming that was a typo...

This could be the result of front-end overload of the UHF rx by your fundamental, creating the 4th harmonic in the receiver itself. In that case a shorted stub on the rx input, cut for an electrical 1/4-wave at 1/2 your FM frequency might work.

Good cabinet shielding and r-f grounding of the UHF receiver and its antenna system, and of your own tx can be important here also.

Other possibilities are that harmonics are being generated in the UHF antenna itself, and/or the structural components on the tower (by contact rectification). Those can be tough to find and fix.
//

 

[rfry]

In that case a shorted stub on the rx input, cut for an electrical 1/4-wave at 1/2 your FM frequency might work.

Sorry, my error, make that an open-circuited stub. The length needs be trimmed for maximum reduction of the FM tx signal, at which time it will also reject its odd harmonics -- while not much affecting the 4th harmonic of the FM that the rx is tuned to.

Another approach might be an appropriate notch or bandpass filter for the UHF rx, such as those supplied by Microwave Filter Company and others.
//

 

[fugazi]

A bandpass would be quite expensive, depending on the used RF power.

A single (open ) stub filter will add an additional 30 dB attenuation. This should be quite adequate.
This is a quarter lambda stub which is exactly a 1/4 lambda ( electrical length ).

Formula L=(Frequency (mhz) / ( 4 x 300 )) x shortening factor of the coax.
frequency is frequency of harmonic to be filtered.

The filter looks like this:

50Ecoax 50ETsplit 50Ecoax 50E = 50 ohm
TRANSMITTER ---------------T-------------Antenna
|
|
|
| 50 ohms 1/4 lambda stub with open end

The best way is to adjust the stub filter on a network analyzer or an spectrum analyzer with an tracking generator.
A signal source with an RF (low) power meter will also do.

If there is no harmonic relation between the signals, then the counter party has a problem. It's called intermodulation and can be solved by installing a bandpass filter before the receiver of the counter party. But that is not your problem.



Regards,


Evert

 

[rfry]

A bandpass would be quite expensive, depending on the used RF power.

You might want to double-check the issues here. My suggestion applied to a stub at the input of the UHF receiver, where the power level on the desired frequency will be very low.

If the UHF receiver is generating the interference due to overload by the fundamental of the FM transmitter, and/or the interference is caused by contact rectification in nearby mechanical structures, and/or by ineffective cabinet shielding/grounding -- then inserting additional harmonic filtering at the output of the transmitter will not be useful.
//

 

[radiofish]

Shively Labs has proposed fabricating a stub filter that will be inserted after the existing harmonic filter and directional coupler on the transmitter and be tuned to provide an additional 30 dB of attenuation at 458.5 MHz. We may try a microwave company low pass FM harmonic filter between the exciter and the transmitter as well.

Modulation is present at the harmonic frequency (around 458.5 MHz) when a spectrum analyzer is connected to the monitoring outputs after the directional coupler and harmonic filter.

The two-way antennas with inputs around 458MHz may need to moved farther from the FM antenna if allowed.

Thanks for your suggestions.

 

[rfry]

Modulation is present at the harmonic frequency (around 458.5 MHz) when a spectrum analyzer is connected to the monitoring outputs after the directional coupler and harmonic filter.

Before spending money on additional filtering for your tx, you might want to substitute a dummy load for your antenna, and see if your ~458 MHz signal is still there. That ~458 MHz signal could be an rf-intermod product generated in the output stage of your tx from mixing with an external signal picked up by your FM antenna. Clearly it is not a harmonic of your FM carrier, unless you are transmitting at ~114.5 MHz.

If the ~458 MHz problem is the result of an external signal getting into your tx, probably it will be less expensive to reduce/remove that interfering source than to filter the intermod product(s) it produces in your tx. It might even be free, if the interfering signal shouldn't be there -- because that correction could be the responsibility of someone else.

Another thought about this is that your FM antenna is no doubt a very poor match at ~458 MHz, and much of such energy would not be radiated, but reflected back to your tx in many round trips, until it was dissipated in system losses. That's a good reason to look carefully outside of your own r-f system for the real source of the problem.

//

 

[mr_elbows]

Two thoughts come to mind here

1) It can NOT be a '4th harmonic' unless you are on 114.5 MHz

2) Is the 458 MHz frequency the input to a repeater ? If so, you need to find the cause (most likely intermodulation) as opposed to trying to 'filter it': 'filtering' at the RX input will knock down the desired frequency

Any other transmitters at the site ? If so what are frequencies ?

 

[Stone Johnson]

This is a very timley post for me. I am currently working with a situation where a UHF (429, 459, and 468) and 900 (896-900) trunk system has been on the tower since the mid 90's and a new high power FM antenna bay (26Kw input to antenna, EIRP 76Kw) has been installed just above the RX antenna tip (12'). Consultants feel the tower top preamps and filtering will be ok, but an increse in noise floor will likly be an issue.

We have run some testing to see what "breaks" the RX system, and just a 10db increase in the noise floor renders the 900 basicly useless. The UHF seems to a little more forgiving , but 15db beats it up pretty bad.

I've been staying neutral between the two sides, but I can see where any increase in the noise floor would be detremental to the overall performance of the RX systems. I am in strange waters here, and look to solve this with both sides happy. Any "usefull" input would be welcome.

 

[fugazi]

When it is intermodulating(higher order), fundamentals WILL mix together with other signals into the uhf band, no matter if they are harmonically related.
The only way to see this is to monitor the uhf receiver's mixer stage with an spectrum analyzer. Just putting filters in front of the receiver or at the end of the transmitter will only cost you a lot of money without the ensurance of the problem being solved.
Noise increments in the receiver is in most cases due to intermodulation especially when the received frequency is hundreds of mhz's away.
BUT, intermodulation can be caused in the transmitter as well, so, again the only way to "see" what the problem is, is to measure the receivers mixer stage.


regards,


evert

 

[radiofish]

Still fighting harmonics from FM TX

To continue this post, the grounding has been much improved, a stub filter tuned to the 5th harmonic frequency (458.5 MHz) has been added after the transmitter's external harmonic output filte (SWR). The harmonic emission at the fifth harmonic, and possibly the 2nd or 3rd harmonic are unacceptable. Based on spectrum analyzer measurments it appears that the problem's source is isolated to the transmitter (prior to the -50 dB tuned stub filter - i.e., the harmonic level is 20 to 30 dB higher before the stub filter, "upstream" toward the transmitter). It appears that the unwanted signal components may be somehow correlated to amount of drive of the transmitter, relatively worse at lower levels. The only other odd observation is that the grid current seems to run higher than normal at full TPO (used to be around 1AMP, now around 1.5AMP grid current).

The transmitter is a grounded grid, 30kW, consisting of 100 watt exciter, 2kW solid state IPA, and 3CX15000A7 PA.

Any suggestions on how to trouble shoot this?

 

[Tom Wells]

The increased grid current is an important clue. As input power "feeds right through" a grounded grid stage, higher grid currents
suggest that the drive to the PA also contains the unwanted harmonics. If not, they are produced in the PA tubes.
If the grounded grid is truly straight to RF common, suspect the filament circuit. If there are any components (particularly capacitors)
anywhere in the grid circuit, they are suspect, as any capacitance decreases here will likely cause exactly this problem.
Any RF bypassing in the filament circuit is also suspect.
Any failing of any semiconductors in the 2kw driver will also cause harmonics. Is the drive signal clean?

 

[radiofish]

Haven't been able to check the directional coupler monitoring port at the output of the IPA just before it goes into the PA.

The capacitors that match the IPA and the PA (tuning/loading) are in fact VERY suspect. This is the only real modification that has been done to the transmitter. The transmitter, about 6 months of operation, began showing lower output, low grid current, and high VSWR indicating on the IPA's directional coupler meter (driver stage output). There are 3 door knob caps, two in parallel to ground, one in series (RF path to PA), that were suggested to be swapped (10, 15, 25 PF, etc.) to obtain a lower VSWR (by trial and error) which would presumably be a better impedance match to the PA indicated by higher grid current, low VSWR on the driver stage, and overall higher TPO. This did in fact help, though over time, the grid current would fall, the VSWR would rise, and the TPO drop again. No amount of tuning the variable capacitor on the PA input tuning could yield a minimal VSWR. Changing the tube to a new tube, on two different times, seems to bring things to "normal" which is high TPO, low VSWR on the driver output meter, and higher grid current.

Here is one more tidbit of a clue: a low-pass harmonic filter was purchased from Microwave Filter and put on the output of the 100 watt exciter. Before the filter was inserted, the reflected power on the exciter (reflecting back from the IPA/driver) was about 4.5-5.6 watts at 100 watts output. Inserting the filter made the exciter's reflect power drop noticeably in half. I would assume since this is a harmonic filter, it was cutting out harmonic frequency signal that was coming back toward the exciter. Does this suggest bad things are happening in the driver, driver/PA stage?

This is a frustrating, complicated, and expensive situation. Thanks for your help.