[Amps] 160m mosfet linear amplifier problem with ferrite cores

[sasas asasas]

hi. i am not use insulators between mosfet and heatsink in order to have the minimum thermal resistance. I am using only silicon thermal conductive paste to decrease the  thermal resistance. So, the heatsink is connected with the drain of each mosfet and the heatsink has the drain voltage.
I am using 12 Siliconix's IRFP360 mosfets, but i have order and soon i will replace them with International Rectifier IRFP360 mosfets. Propably this is the reason, that the amplifier don't give many watts at outupt.
The other reason, is that the material of ferrite cores (43 material) is not the proper for 1-2mhz frequences. What do you think??

--- On Sun, 4/18/10, David Cutter <d.cutter at ntlworld.com> wrote:

From: David Cutter <d.cutter at ntlworld.com>
Subject: Re: [Amps] 160m mosfet linear amplifier problem with ferrite cores
To: "sasas asasas" <tzitzikas_ee at yahoo.com>, amps at contesting.com
Date: Sunday, April 18, 2010, 3:50 AM



 
I see you are using hex socket screws.  It is very 
easy to over-tighten screws to fix plastic devices to a heatsink, I've seen them 
crack after a little time in the field due to expansion and contraction against 
an over-tight fixing.  The crack can be invisible to the naked eye 
initially but causes failure of the device.  Mounting torque is 1.13Nm 
which is quite low.  
 
The mounting tab is the drain and you could look at the 
heatsink isolation. Is there an isolation insulator under each device 
or is the whole heatsink isolated?   
 
I could be drifting off the subject, but fixing small 
errors can often chase away the big ones.
 
David
G3UNA
 
  

  ----- Original Message ----- 
  From: 
  sasas 
  asasas 
  To: amps at contesting.com ; David 
  Cutter 
  Sent: Sunday, April 18, 2010 10:41 
  AM
  Subject: Re: [Amps] 160m mosfet linear 
  amplifier problem with ferrite cores
  

  
    
    
      Some additional photos from my linear:

http://tzitzikas.webs.com/fet20.JPG
http://tzitzikas.webs.com/linear20.JPG
http://tzitzikas.webs.com/linear21.JPG
http://tzitzikas.webs.com/linear22.JPG
http://tzitzikas.webs.com/fet21.JPG
http://tzitzikas.webs.com/pompos20.JPG


--- 
        On Sun, 4/18/10, David Cutter <d.cutter at ntlworld.com> 
        wrote:

        
From: 
          David Cutter <d.cutter at ntlworld.com>
Subject: 
          Re: [Amps] 160m mosfet linear amplifier problem with ferrite 
          cores
To: "sasas asasas" <tzitzikas_ee at yahoo.com>, 
          amps at contesting.com
Date: Sunday, April 18, 2010, 2:35 AM


          
          

          I've looked again at the jpg and I think you 
          have used trifilar windings, so, that's my first idea more or 
          less gone.  Imbalance can be due to poor coupling between the 
          windings themselves or imbalance in the driving currents.  With 
          high voltage supplies, balance error should be small (compared to say 
          a 12V amplifier where a small difference in voltages can show a large 
          imbalance %).  If you could measure the current in each half of 
          the primary that would tell you the whole story: you will need a HF 
          current probe, preferably 2 probes and dual trace scope which will 
          also tell you about cross-conduction.  
           
          Another thought: you have 6 transistors in 
          parallel in each half and emitter (sorry - 
          source) degeneration.  Did you try matching the FETs for 
          gain or phase delay?  If not, it's conceivable that one side is 
          conducting harder or for longer than the other side and causing an 
          imbalance.  I don't know if the source resistor is the optimum 
          value for mis-matched FETs, but a higher value would increase 
          negative feedback and help with this; you have oodles of gain to 
          sacrifice.  Along the same lines, the input drive to each FET may 
          need optimising for the same reason.  Check all your 
          resistor values, one might be a dud.
           
          This is all armchair speculation, I don't have 
          enough experience for better detail.  Someone with modelling 
          experience could probably do this quite 
          quickly.  
           
          Alex has probably got the answer regarding core 
          material.  
           
          David
          G3UNA
          
            ----- 
            Original 

            
              
              
                And how can i repair this problem???

                  



                    You might consider imbalance in each 
                    side of the primary which will cause a net dc to pass, 
                    saturation and over-heating.  This may be rubbish but 
                    it could get others thinking along a different 
                    path.

David
G3UNA


> HI. i have 
                    constructed this linear amplifier 500w rms (2kw pep) (50ohm) 
                    with 12 mosfet irfp360, http://tzitzikas.webs.com/linear500w.jpg 
                    for 160m band.
> When i gave 3watts of driving r.f 
                    power, it gave to output only 190w at 106VDC (6A current). 
                    Two radio amateurs who have construct this linear claim tha 
                    it gives 500w
> r.f power at 110vdc.
> But when 
                    i tried to give 4watts of driving r.f power the ferrite 
                    Cores (43 material)
> of transformer T3 broken! Which 
                    do you think is the problem??
>