RCI-2970N2 CARRIER ON SSB ONLY INTO ANTENNA

I have a 2970N2 that works fine on both USB and LSB into a dummy load with no carrier...but puts out a carrier on both USB and LSB when hooked up to my shockwave ground plane OR my 5 element beam. The SWR on both antennas are below 1.2:1 and yes, modulation was turned down to zero.
I am experiencing approx. 50W of carrier. I have adjusted the emission VR7 pot so there is no carrier on either USB or LSB (into the dummy load) Does anyone have any idea what is causing this? All other radios work fine going into either antenna except this 2970N2. Thanks,
-75

Thank you for your reply, sir. I put a frequency counter on the unit and it reads exactly on frequency. I mean exact. The radio was re-aligned a few days ago. So the frequency counter is not showing me anything else that I should be concerned about. But when I hooked the 2970N2 to the spectrum analyzer, I noticed with the amp turned off ( I have it switched) it is fine. But when I turn the amp on, the IF frequency is out of range on the analyzer. So, I believe that you are correct about the radio PA section self oscillating on ssb. The previous owner had installed new IRF520N MOSFETS in the PA section because of a high bias issue which took the originals out (IRF520). The feedback boards are installed, all screws are tight on both the main board and PA board. I checked all small coax connections and they are secure. Cover for the PA amp is also secure. Bias for the driver and finals are set as per Ranger specs from Chip as well as the PA amp and I have the radio turned down on ssb to 150W PEP to drive my larger amplifier. Would the IRF520N replacement for the IRF520 have something to do about it? I believe the "N" is more robust with slightly higher gain...I'm not sure, but that is what the previous owner told me. Is there any other way, that you know of, to rectify this problem? Thanks in advance.
-75

The IRF530N has a much higher resistance (DS) when on and has a drain current of 17Amps. The IRF530 has less than 0.16 OHMS (DS) when on and a drain current of 14Amps. That is the first thing I noticed looking at IR's data sheets.

The IRF530 is a N channel as is the IRF530N but the IRF530 does not have the N suffix. Read my above post for definate differences between the two as per the data sheets.

Thanks for the reply, sp5it.
The thread link you posted is about adding caps and resistors to form a negative feedback loop. This amp has negative feedback boards with 0.01ufd caps and 2W 100 Ohm resistors already installed. Yet the problem still persists.

Thank you for your reply, Old Tech03. This unit does have the thicker ceramic insulators. If I am to replace the insulators with the thinner mica ones….perhaps I should install the IRF530's as well and remove the IRF530N MOSFETS. This will give me something new to try. I understand the mica insulators will sink the MOSFET for better heat transfer but did not think of the ceramic insulators causing the MOSFETS to self oscillate. Cheers! I will let you know how I make out

yes 520's...sorry.

I'm sorry that I forgot to update my findings. As it turned out, the surface mount 0.01uf caps on the feedback boards were the culprits. I purchased some 0.01uf 100v ceramic caps, removed the surface mount caps and soldered in these ceramic disc caps. (2 on ea. feedback board) Problem solved. Thank you for all the help.
-75