You can now help support WorldwideDX when you shop on Amazon at no additional cost to you! Simply follow this Shop on Amazon link first and a portion of any purchase is sent to WorldwideDX to help with site costs.
The Feb 2025 Radioddity Giveaway Results are In! Click Here to see who won!

Reply to thread

Message

Another thing to consider about bias is that many of the test circuits given in transistor data sheets are simplified to the point of allowing the transistor to just produce good results under a specific test. Bias was considered just long enough to show the transistor having good gain. Not that it will hold the transistor in class AB operation over a wide range of drive or temperature fluctuations.

If the impedance of the bias supply is too high (excessive series resistance), the RF drive can now modulate the DC bias level. High quality designs will hold this impedance below 100 ohms and usually require a buffer transistor between the bias control and the gate input to do this.

The diode and two resistors Palomor is using as a "companion" part on the gates of their transistors is interesting too. It appears to intentionally maximize the effect of modulating the DC bias by rectifying RF and applying it as additional bias. The gain on these parts is so low at 30 MHz. that I suspect this setup is done to improve swing or PEP output. You'll also notice they add a 1000 ohm resistor to the back of the input transformer feeding the fixed DC bias in and supplying the RF ground to the center tap.

The "sponginess" of that 1000 ohm resistor on the input transformer is about the only thing keeping these switch mode transistors alive in "not so linear" use. They have hidden the guts of their companion part in an effort to make you think it's required. True, their FET will burn up if you leave it out but that's only because it forms a voltage divider leg to ground, with that 1000 ohm resistor in series. Now if you leave it out, the anticipated voltage drop across the 1000 ohm resistor is no longer present and you apply too much bias voltage!

The added RF rectified bias in driver stages can sometimes kill final stages by applying too much peak drive. RM Italy gets around this problem by leaving the RF bias diode out of their companion circuit on driver stages. They still keep all of the extra resistors intact in an effort to protect the gates with bias applied to them. Something you do not have to do with linear rated RF MOSFET's.

<blockquote data-quote="Shockwave" data-source="post: 699735" data-attributes="member: 10882">Another thing to consider about bias is that many of the test circuits given in transistor data sheets are simplified to the point of allowing the transistor to just produce good results under a specific test. Bias was considered just long enough to show the transistor having good gain. Not that it will hold the transistor in class AB operation over a wide range of drive or temperature fluctuations.If the impedance of the bias supply is too high (excessive series resistance), the RF drive can now modulate the DC bias level. High quality designs will hold this impedance below 100 ohms and usually require a buffer transistor between the bias control and the gate input to do this.The diode and two resistors Palomor is using as a &quot;companion&quot; part on the gates of their transistors is interesting too. It appears to intentionally maximize the effect of modulating the DC bias by rectifying RF and applying it as additional bias. The gain on these parts is so low at 30 MHz. that I suspect this setup is done to improve swing or PEP output. You&#039;ll also notice they add a 1000 ohm resistor to the back of the input transformer feeding the fixed DC bias in and supplying the RF ground to the center tap.The &quot;sponginess&quot; of that 1000 ohm resistor on the input transformer is about the only thing keeping these switch mode transistors alive in &quot;not so linear&quot; use. They have hidden the guts of their companion part in an effort to make you think it&#039;s required. True, their FET will burn up if you leave it out but that&#039;s only because it forms a voltage divider leg to ground, with that 1000 ohm resistor in series. Now if you leave it out, the anticipated voltage drop across the 1000 ohm resistor is no longer present and you apply too much bias voltage!The added RF rectified bias in driver stages can sometimes kill final stages by applying too much peak drive. RM Italy gets around this problem by leaving the RF bias diode out of their companion circuit on driver stages. They still keep all of the extra resistors intact in an effort to protect the gates with bias applied to them. Something you do not have to do with linear rated RF MOSFET&#039;s.</blockquote>

[QUOTE="Shockwave, post: 699735, member: 10882"] Another thing to consider about bias is that many of the test circuits given in transistor data sheets are simplified to the point of allowing the transistor to just produce good results under a specific test. Bias was considered just long enough to show the transistor having good gain. Not that it will hold the transistor in class AB operation over a wide range of drive or temperature fluctuations. If the impedance of the bias supply is too high (excessive series resistance), the RF drive can now modulate the DC bias level. High quality designs will hold this impedance below 100 ohms and usually require a buffer transistor between the bias control and the gate input to do this. The diode and two resistors Palomor is using as a "companion" part on the gates of their transistors is interesting too. It appears to intentionally maximize the effect of modulating the DC bias by rectifying RF and applying it as additional bias. The gain on these parts is so low at 30 MHz. that I suspect this setup is done to improve swing or PEP output. You'll also notice they add a 1000 ohm resistor to the back of the input transformer feeding the fixed DC bias in and supplying the RF ground to the center tap. The "sponginess" of that 1000 ohm resistor on the input transformer is about the only thing keeping these switch mode transistors alive in "not so linear" use. They have hidden the guts of their companion part in an effort to make you think it's required. True, their FET will burn up if you leave it out but that's only because it forms a voltage divider leg to ground, with that 1000 ohm resistor in series. Now if you leave it out, the anticipated voltage drop across the 1000 ohm resistor is no longer present and you apply too much bias voltage! The added RF rectified bias in driver stages can sometimes kill final stages by applying too much peak drive. RM Italy gets around this problem by leaving the RF bias diode out of their companion circuit on driver stages. They still keep all of the extra resistors intact in an effort to protect the gates with bias applied to them. Something you do not have to do with linear rated RF MOSFET's. [/QUOTE]

Top Bottom