Need mofset final and companion part info

[Handy Andy]

The LM317 is a LINEAR regulator, meaning that it takes a DC voltage that can vary a wide range - into a "strict" leveled and filtered (somewhat) steady voltage at a lower VOLTAGE and current...

A Linear Power supply using an LM317
- taking High Voltage down to a safer voltage,
using adjustable voltage-controller capable of some Amperage
​

Just remember the LM317 does not consume a lot of current - but it can generate a lot of heat - from the effort it has to do. It has to take a higher voltage and current - and turn it into a lower voltage - that causes work, some of it as power for the terminal leg, and the rest as dissipated heat from dropping power across itself to this new lower regulated voltage out, or work to be performed, because of this process, this limits it's own current thru the device.

A typical "strip" of MOSFETs' may need about 250mA of current at 3.8~2.7 volts - that's 1/4 of an AMP of current at that voltage - so you're more concerned about wattage ratings of your needed supply voltage divider. I SAID STRIP - like the LM317 can feed at least 10 devices needing 50mA to bias properly (That's' 10 MOSFET's X 50mA = 500mA - or 1/2 an amp of current)

Yes, a voltage divider...

Most and simplest voltage divider is two resistors of equal value - one has a lead to the POSITIVE rail, the other has a lead to the NEGATIVE rail (or Ground) - the two are tied together in the middle and if you measure with a VOM/DVM - the CENTER tied lead of the two resistors equals 1/2 the supply voltage.

The limitation of a divider is the ability to supply current at that specific voltage. The higher the ohmic values used, the less current is possible to be used thru the divider. Not very effective.

The LM317 performs similar-fashion by dropping power across itself - only it's variable, using the support circuitry and uses a method of "sense" - a type of feedback - to help it keep it's voltage steady, when the current demand increases. This is all done mostly inside the device - automatically - but you have to set up your side - the outside - to work with what's inside.​

DONT PANIC
It's just a circuit...
​

You can use whatever you want, I'm just showing a way to use the LM317 and protect you and your radio - especially when problems with wiring can occur - they increase in mobile installs where polarity reversals can occur - even remotely - obscurely, don't bet on it - kind of happenings... Murphy and his Laws are always willing to show up when you least expect it - in fact I believe that is his Murphy's Law #4 - Things can go wrong and if it can, will happen in the worse possible order...

So when you look for a supply to drive the MOSFET's better give those electron's a way to pass out of the circuit easily when things get a little too strong - like RF carrier or Audio drive - yelling into the Mike - type of recourse.

It's why I mentioned the Voltage divider - let's power swings occurring at one side or location near the divider - occur in a controlled fashion and yet your power feed stays steady and at the ready to keep things under control.

Remember they have a very High Impedance at the Gate - per Tallman (-THX BTW)

The following is a schematic of a typical Dual-MOSFET Final Galaxy class Radio...

The above is so I can show you why things are done like they are - they "divide" the voltage arriving to the BIAS section for these MOSFET devices. Your LM317 will be supplying to something like the above.

Not all are done like the above, but you have to be ready for even the unexpected...read on.

They also use Zener Diodes (D90 and D91 for Finals and D92 for Driver) as a means to keep BIAS from getting too high or getting out of control when the operator starts yelling into their Microphone and that Audio can and will get rectified, and the combined RF and Audio power result will BOOST the bias voltages up to a level that if not kept in check - ZAP and destroy the MOSFETs' . How? Look at the Maximum ratings - they can be destroyed, caused by excessive RF level drive that exceeded the MAXIMUM input levels the MOSFETs' were designed to take.

So take heed of precautions I'm trying to tell you - a simple divider can work, and LM317 supply will work well, and do the job - but remember to keep the divider principle in use for those MOSFET's - else they can be damaged by excessive power levels that even the LM317 can't handle.

 

[Tallman]

Tallman, any concerns with the gate voltage being present all the time?

I like to switch the bias voltage on for transmit only.

The bias is switched on and off via the keying circuits. Keeping on all the time is asking for disaster like /oscillations. That's also why I use pull down resistors.
The LM317 only needs 1.7 volts head room to do it;s job so running it off of the keyed circuit, but if you hook it to the place where the +8 volts is only 8volts when keyed won't work very well because there is a voltage there when not keyed right? I'm using the Hard Key circuit my Icom 746 uses to turn the bias on/off.

 

[Tallman]

A typical "strip" of MOSFETs' may need about 250mA of current at 3.8~2.7 volts - that's 1/4 of an AMP of current at that voltage - so you're more concerned about wattage ratings of your needed supply voltage divider. I SAID STRIP - like the LM317 can feed at least 10 devices needing 50mA to bias properly (That's' 10 MOSFET's X 50mA = 500mA - or 1/2 an amp of current)

There should be ZERO DC current flow into the gates of the FET's. Only the DC charges the gate capacitance. Fets are Voltage controlled not current controlled. That's why there no gain number involved with fets. The gain figures quoted for Bipolar transistors is a ratio of base current versus collector current, since there no Gate current there will not be a gain figure.
The in rush current of charging the gates is small in comparision since you are only X amount of Pico farads.

 

[LeapFrog]

.... but if you hook it to the place where the +8 volts is only 8volts when keyed won't work very well because there is a voltage there when not keyed right? I'm using the Hard Key circuit my Icom 746 uses to turn the bias on/off.

The voltage goes to about 0 during receive at the point where I grab it.
8 Volts during TX, this spot would be a good place to connect hard key circuit relay, and feed a jack on the rear of the radio, for those candle boxes i'm fond of.
(thanks for the idea Tallman!)

 

[Handy Andy]

Wow, you forget one important point...

Where will that charge drain off - and will you await the Gate or does the MOSFET have an on off switch in it?

You forget, they have still the need to drain their charge....

Hence the 50mA draw in a typical biasing scheme to apply enough voltage and to sustain the current needed to maintain voltage as RF pull in and drains off charge.

There should be ZERO DC current flow into the gates of the FET's. Only the DC charges the gate capacitance. Fets are Voltage controlled not current controlled. That's why there no gain number involved with fets. The gain figures quoted for Bipolar transistors is a ratio of base current versus collector current, since there no Gate current there will not be a gain figure.
The in rush current of charging the gates is small in comparision since you are only X amount of Pico farads.

In a TX situation the TX line itself has enough to draw down the power in the line by simple loading

But to forget the Upgrade method keeps the original biasing - we use it to our advantage and in this instance it provides the stability like as you had suggested to someone else...

The bias is switched on and off via the keying circuits. Keeping on all the time is asking for disaster like /oscillations. That's also why I use pull down resistors.

So I don't know why you posted the response when the very next one posted a contradiction - as in the TX line to toggle the gate by itself versus leaving a charge on that bias line all the time.

So that divider circuit is also for that very purpose - to remove or stabilize the gate to a known ground reference when the part is supposed to be turned off.

There is another effort the divider provides is a means to take off RF noises and other intangible effects of noise from Source and Drain and their inherited effects on the gates ability to remain in the off state - so putting the Gate to a divider is safer practice than to leave the gate to fend for itself at the TX lines whims.

 

[Tallman]

Wow, you forget one important point...

Where will that charge drain off - and will you await the Gate or does the MOSFET have an on off switch in it?

You forget, they have still the need to drain their charge....

Hence the 50mA draw in a typical biasing scheme to apply enough voltage and to sustain the current needed to maintain voltage as RF pull in and drains off charge.



In a TX situation the TX line itself has enough to draw down the power in the line by simple loading

But to forget the Upgrade method keeps the original biasing - we use it to our advantage and in this instance it provides the stability like as you had suggested to someone else...



So I don't know why you posted the response when the very next one posted a contradiction - as in the TX line to toggle the gate by itself versus leaving a charge on that bias line all the time.

So that divider circuit is also for that very purpose - to remove or stabilize the gate to a known ground reference when the part is supposed to be turned off.

There is another effort the divider provides is a means to take off RF noises and other intangible effects of noise from Source and Drain and their inherited effects on the gates ability to remain in the off state - so putting the Gate to a divider is safer practice than to leave the gate to fend for itself at the TX lines whims.

No I did not forget,you just did not read all of my other post relating to this whole process.
Read post number 14 where I address the pull down resistor to halt the latch up problem.
There should not be any gate current in the off state unless you are using defective FET's.
The gate bias is off when un-keyed.

 

[Handy Andy]

IF you're referring to this...

The input impedance to the FET is extremely high(There should be Zero current continuous, only gate charging) Current flow would be dictated by how many stages, and the resistor setup around the gate. I always use pull down resistors to assure the transistors will not "Latch up." 220 K ohms or higher down to 10 K ohms depending on circuit use.

Then even telling that person to say it is ok to use 20K as pull down seems a little on the high side.

The 1230 can generate enough voltage to spike and ruin even a IRF520 or 13N10 due to the oblique use of the same input capacitor of say 270pF in a typical Cobra 25 or even 220pF of a 29 when they don't use a shunt cap - just leave the resistor of 12 ~ 22 ohm and that is what is supposed to be removed and the EN1230 put in its' place. So there is nothing else to shunt, remove, or otherwise dissipate the power to the gate - except for what is present at the EN1230.

This kid is trying to make his own EN parts...

And he wants to use an LM317...

He will need a divider to help stabilize the LM317 as well as the Gate(s) it's powering...

 

[Tallman]

IF you're referring to this...


Then even telling that person to say it is ok to use 20K as pull down seems a little on the high side.

The 1230 can generate enough voltage to spike and ruin even a IRF520 or 13N10 due to the oblique use of the same input capacitor of say 270pF in a typical Cobra 25 or even 220pF of a 29 when they don't use a shunt cap - just leave the resistor of 12 ~ 22 ohm and that is what is supposed to be removed and the EN1230 put in its' place. So there is nothing else to shunt, remove, or otherwise dissipate the power to the gate - except for what is present at the EN1230.

This kid is trying to make his own EN parts...

And he wants to use an LM317...

He will need a divider to help stabilize the LM317 as well as the Gate(s) it's powering...

He has changed his design and intent.
Andy I have always considered you to be a brilliant engineer,(Although a bit verbose) but for some reason you aren't grasping the fundamentals of the FET capacitance.
380 PF will not store much of a charge and the input impedance is probably over 100 Meg ohms.
A resistance of 100 k ohms to the ground will knock that charge down to nothing, or at least to the point conduction will cease.
I designed and assisted in the design of switch mode power supplies that were operating at 20 MHz back in 1985, and guess how we solved the fet latch up problem.
A simple 100 Kohm from gate to source or ground if you prefer.

 

[Handy Andy]

It's ok, I'd rather see the effort start small then to go "too big" and he still has yet to find out how static build up can even destroy all the work even before he even powered the device up.

I've had radios come across my desk wanting to have the final replaced only to find that they did a mod that is more like what is being shown on youtube so they try and copy it not knowing/understanding what they are doing.

Then they scream "foul" when it doesn't work right - and those; like you and I know, very budding radio enthusiasts seek us out and we have to rescue the radio from the hapless operator and get it to work properly so they can pay the $$$ and learn their lessons that way...

I'll cut the losses right now, you can take the lead and responsibility and accountability of this thread...

 

[Tallman]

I have had some major problems with esd damage from the people I bought parts from in the past.
If you check my albums look at the traces from my component tracer.
Six MS-1051 transistors two were DOA fresh from Italy. These were replacements for transistors that blew up on the first key up test. Testing for the auto band switching circuit in this HL300.
I'm sorry we crossed swords but sometimes the old fart who has been in the trenches still has functional knowledge. Sorry if I offended.

 

[LeapFrog]

Good morning y'all!

To reiterate, if i use a regulator, I should have a (large value) discharge resistor to ground, like 20K or higher?

 

[NZ8N]

I think you will be fine with or without your LM317 circuitry, the resistors used just in the bias section alone provide more than enough resistance to drop the +8v TX voltage to zero all by themselves, not counting the rest of the TX circuitry.

 

[LeapFrog]

I think you will be fine with or without your LM317 circuitry, the resistors used just in the bias section alone provide more than enough resistance to drop the +8v TX voltage to zero all by themselves, not counting the rest of the TX circuitry.View attachment 27891

Fair enough, why add more parts than necessary!
I've been specifically playing with AM only radios that came with BJT final transistors, and adding the MosFet and bias circuitry, and I agree I see no reason to over complicate things.

The adjustability of the regulator is very "smooth" IMHO and very stable, I seem to remember that if things warmed up the voltage would change a tiny bit when simply using a cheap trimmer potentiometer and voltage divider.
(probably due to the poor quality trimmer pot I was using)

 

[Tallman]

If you notice with the simple bias diagram the resistors that discharge the DC gate voltage is 100 k ohms and with the other resistor is 101 K ohms. The advantage of the LM317 is the rock solid voltage and adjustability.