ATV WHY THE M57762 BRICK BLOWS UP
HOW TO MAKE IT LINEAR AND ONLY NEED 90m/W OF DRIVE
The commonly used 18 Watt PA for 1255MHz seems to have an expensive short life and put me off the air for a few years until I found a local source of supply.... You must be aware of the problems of Matching and Frequency roll off.... Without proper equipment it is very difficult to find out what is going on so I wish to share my experiences and my own conclusions as to why out of the blue they blow up.....
I obtained last year a G1MFG 23c/m TX that gives out about 100m/W of RF.... Then I wrote my own software to go up and down the band and to auto-save on power down so it would come back on at the frequency it was last set to.... However I replaced my blown up PA.... That had been overdriven in RF... So I fed in the 100m/W from the G1MFG exciter and low and behold power was coming out..... I then did a plot of the Output power with reference to Frequency and was very interested to find it peaked up at 1300MHz and gave "FULL POWER"... Ivor G1IXF informed me that's what they originally was designed for not ATV.... On coming down in Frequency it slowly drops off then about1270MHZ starts to drop off very sharply indeed there was a big difference between 1255 and 1249MHz... Alas here is the problem... I suspect if the input matching is wrong then it will not be 50 OHMS input impedance at the lower frequencies hence changing the length of co-ax between the Exciter and the PA will introduce bad effects..... History has show that it is when ATV guys change things around it blows or try tuning up the Exciter Wrongly.... The real problem is the PA brick itself... Tests I carried out showed that its Every stage in the Brick that is a problem.... Just place a plastic object on any of the active lines brings the power up with a low frequency fed in.... I will try the PTFE sheet next to see if that will bring each stage up enough..... It did work on the 10GHz interdigital filter on the LNB ... I did try trimmers but what I have are not good enough for this frequency and power.... I will effect a cure so watch this space.... Another thing I noticed was the big heat sink at 1300MHz was cool with full output coming out and drawing about 4.5 AMPS........ I would be interested in your finding as well !.. I do not own a Network Analyzer YET ....... G3RFL
AND THERE IS A SOLUTION 16/8/2001
I did promise to look into getting the brick down in frequency and it requires just three trimmers adding to the capacitors that feed the bases in-between each stage and one on the output cap about 3pf trimmers.....
I first fed in a small signal +9dbm and plotted the results the thing had two peaks one at 1320MHz and another at 1296MHz but after 1270MHz dropped off very fast indeed...
I found that if I moved the input lead had a big effect on the output so I recited the input at the opposite side to the output making the input lead shorter and away from the output but alas still poor 1249MHz gain about 1watt coming out and 10Watts at 1296MHz....
I fitted the first trimmer across the caps on the first stage then that went up to 2.25Watts the next trimmer took it up to 4Watts and lastly the output cap took it up to 9Watts this was with a 16.8Volt PSU.....(was the only only supply I had at the time not recommended its near the blow up point.)
The Signal source was a level controlled 50ohms Signal Generator that I checked for being a flat response..
One trouble I had was a SHORT CIRCUIT trimmer and yet after 12AMPS going through the PA stage it did survive....HI.....I switched off very fast indeed....the short was a solder splash it had been removed from some old equipment....
NOTE:-40m/W in at 1249MHz will drive it up to 10Watts output after the mods and it was drawing just under 2 Amps....
I tested 90 m/W is all that is now required to fully drive said brick....22Watts output......3 amps DC.....
It now has a flat response (better 1db) from 1249 to 1270MHz and falls off any higher unlike before......
Another side effect is it runs "COOLER".
SO IT CAN BE DONE !...........G3RFL
