Re: [funwithtransistors] QRP Power Amp Again


Jan 16, 2013

 


----------------------------

#1174 Jan 16, 2013

Hi all,



I'm trying to get a transistor RFPA to work and not having much

success. The usual route with these is to stick broadband ferrite

transformers on the inputs and outputs and subject the resulting

signal to heavy filtering. I'm trying to do this the old-fashioned way

as it was done in tube amplifiers. The problem is, the MOSFET with the

12v power supply wants to see a lower power than 50 ohms rather than a

higher one. Thus, I have to use a T matching circuit instead of the

familiar pi one. Basically, one L steps the impedance up above 50 ohms

and the second one takes it back down to 50. Since the two capacitors

would be in parallel, they can become one. Once I master this circuit,

I want to make a single-ended one that's just like it, using a 2E26

for about 25w. I'll need to fiddle with the output components so I can

match it to bands on 80 through 10. For now, this amplifier is only

going to operate on the 20m CW portion, which is 14.06 MHz.



I'm including all of the image files so I can illustrate what's going

on in my head and what I think will fix it. I want to do this the way

of the tube but I obviously require some assistance. Dave, WA4QAL, has

already helped me some but the people on QRZ have stopped talking to

me about it. Either there's something glaringly obvious that I'm

missing or I don't understand how MOSFETs operate after all. I don't

want to go the route of broadband transformers because I want to find

out the hard way to do this.



I appreciate anyone's suggestions and help.

Ed



----------------------------

#1181 Jan 17, 2013

--- On Wed, 1/16/13, J Ed jedwardsat1@...> wrote:

> From: J Ed jedwardsat1@...>

> Subject: [funwithtransistors] QRP Power Amp Again

> To: "funwithtransistors" funwithtransistors@yahoogroups.com>

> Date: Wednesday, January 16, 2013, 3:19 AM

>

> Hi all,

>

> I'm trying to get a transistor RFPA to work and not having much

> success. The usual route with these is to stick broadband ferrite

> transformers on the inputs and outputs and subject the resulting

> signal to heavy filtering.



And, there's a reason that that is usually done. ;-)

> I'm trying to do this the old-fashioned way as it was done in tube

> amplifiers.



The difference is that tube amplifiers usually have a fairly high

output impedance, while solid state amplifiers usually have a very low

output impedance.



Thus, while you can do things the old fashioned way, you need to

take special steps to ensure that the low output impedance doesn't

result in excessive resistive losses in the output circuit.

> The problem is, the MOSFET with the 12v power supply wants to see

> a lower power than 50 ohms rather than a higher one.



Understandable.

> Thus, I have to use a T matching circuit instead of the familiar

> pi one. Basically, one L steps the impedance up above 50 ohms

> and the second one takes it back down to 50.



Ok...

> Since the two capacitors would be in parallel, they can become one.



Correct. However, note that, due to the low output impedance, you'll

have some rather incredible currents flowing, so you'll have to

ensure that the capacitor (and inductor) being used are capable of

handling those rather incredible currents without incurring excessive

(I**2)*R losses. That usually argues for inductors with rather heavy

gauge wire (or even Silver plated Copper tubing), and variable

capacitors with geometries different from the traditional rotary

variable capacitor, which usually has excessive resistance in the

rotary joint (e.g., A "butterfly" variable capacitor would be better,

since the current doesn't have to transition across a movable joint.).

> Once I master this circuit, I want to make a single-ended one

> that's just like it, using a 2E26 for about 25w.



Which will require a completely different input, output, bias, and

power supply circuit topography (Umm, will any part of the circuit

remain the same? I don't think so!).

> I'll need to fiddle with the output components so I can

> match it to bands on 80 through 10.



Check the values you have specified in the circuit. I'm showing a

resonant peak at near 4 MHz, not near 14 MHz. In the simulation I

ran, the output at 14 MHz is WAY down from what it is near 4 MHz.

> For now, this amplifier is only going to operate on the 20m CW

> portion, which is 14.06 MHz.



Check your circuit values. I may have used the wrong value for some

of the parasitic elements, but I'm showing a resonance very far away

from 14 MHz.

> I'm including all of the image files so I can illustrate what's going

> on in my head and what I think will fix it.



You can build the original circuit in SPICE and run a frequency sweep

simulation on it to see what it does at various frequencies. I swept

my model from 1 MHz to 30 MHz (I think), and it was showing a resonance

near 4 MHz.

> I want to do this the way of the tube but I obviously require some

> assistance. Dave, WA4QAL, has already helped me some but the people

> on QRZ have stopped talking to me about it.



It's not commonly done this way, so there probably aren't very many

people who have had empirical experience with this circuit topology.

> Either there's something glaringly obvious that I'm missing or I

> don't understand how MOSFETs operate after all.



The circuit does seem to be somewhat sensitive to bias and drive levels.

You'll need to hit the MOSFET with a fairly large signal to drive it

from cutoff to full conduction, which is why most solid state amplifiers

have a broad-band step-up transformer, to produce a large voltage swing

at the gate of the MOSFET. A direct drive input circuit, with a low

power input signal, probably won't drive the gate hard enough to turn

the MOSFET from cutoff to full conduction. Maybe. Depending upon your

particular MOSFET part selection.

> I don't want to go the route of broadband transformers because I

> want to find out the hard way to do this.



You'll find that, when doing things the hard way, that they're hard

to do.

> I appreciate anyone's suggestions and help.



Run a SPICE simulation, ideally using LTSpice. I'll even toss in

the sloppily constructed model I built yesterday.



www.linear.com/designtools/software/

> Ed



Dave







----------------------------

#1183 Jan 17, 2013

Ed. I meant to answer this when you first posted it but it slipped my so

called mind. All you nee to do to make a pi net to step up impedance is to

design as if it were a step down and then turn it around. The

plate/collector/drain impedance is 50 ohms and the load is the drain

impedance of the FET.



Regards.



Max. K 4 O DS.



Email: max@...



Transistor site www.funwithtransistors.net

Vacuum tube site: www.funwithtubes.net

Woodworking site

www.angelfire.com/electronic/funwithtubes/Woodworking/wwindex.html

Music site: www.maxsmusicplace.com



To subscribe to the fun with transistors group send an email to.

funwithtransistors-subscribe@yahoogroups.com



To subscribe to the fun with tubes group send an email to,

funwithtubes-subscribe@yahoogroups.com



To subscribe to the fun with wood group send a blank email to

funwithwood-subscribe@yahoogroups.com

----- Original Message -----

From: "Dave" wa4qal@...>

To: funwithtransistors@yahoogroups.com>

Sent: Thursday, January 17, 2013 12:46 PM

Subject: Re: [funwithtransistors] QRP Power Amp Again





> --- On Wed, 1/16/13, J Ed jedwardsat1@...> wrote:

>

>> From: J Ed jedwardsat1@...>

>> Subject: [funwithtransistors] QRP Power Amp Again

>> To: "funwithtransistors" funwithtransistors@yahoogroups.com>

>> Date: Wednesday, January 16, 2013, 3:19 AM

>>

>> Hi all,

>>

>> I'm trying to get a transistor RFPA to work and not having much

>> success. The usual route with these is to stick broadband ferrite

>> transformers on the inputs and outputs and subject the resulting

>> signal to heavy filtering.

>

> And, there's a reason that that is usually done. ;-)

>

>> I'm trying to do this the old-fashioned way as it was done in tube

>> amplifiers.

>

> The difference is that tube amplifiers usually have a fairly high

> output impedance, while solid state amplifiers usually have a very low

> output impedance.

>

> Thus, while you can do things the old fashioned way, you need to

> take special steps to ensure that the low output impedance doesn't

> result in excessive resistive losses in the output circuit.

>

>> The problem is, the MOSFET with the 12v power supply wants to see

>> a lower power than 50 ohms rather than a higher one.

>

> Understandable.

>

>> Thus, I have to use a T matching circuit instead of the familiar

>> pi one. Basically, one L steps the impedance up above 50 ohms

>> and the second one takes it back down to 50.

>

> Ok...

>

>> Since the two capacitors would be in parallel, they can become one.

>

> Correct. However, note that, due to the low output impedance, you'll

> have some rather incredible currents flowing, so you'll have to

> ensure that the capacitor (and inductor) being used are capable of

> handling those rather incredible currents without incurring excessive

> (I**2)*R losses. That usually argues for inductors with rather heavy

> gauge wire (or even Silver plated Copper tubing), and variable

> capacitors with geometries different from the traditional rotary

> variable capacitor, which usually has excessive resistance in the

> rotary joint (e.g., A "butterfly" variable capacitor would be better,

> since the current doesn't have to transition across a movable joint.).

>

>> Once I master this circuit, I want to make a single-ended one

>> that's just like it, using a 2E26 for about 25w.

>

> Which will require a completely different input, output, bias, and

> power supply circuit topography (Umm, will any part of the circuit

> remain the same? I don't think so!).

>

>> I'll need to fiddle with the output components so I can

>> match it to bands on 80 through 10.

>

> Check the values you have specified in the circuit. I'm showing a

> resonant peak at near 4 MHz, not near 14 MHz. In the simulation I

> ran, the output at 14 MHz is WAY down from what it is near 4 MHz.

>

>> For now, this amplifier is only going to operate on the 20m CW

>> portion, which is 14.06 MHz.

>

> Check your circuit values. I may have used the wrong value for some

> of the parasitic elements, but I'm showing a resonance very far away

> from 14 MHz.

>

>> I'm including all of the image files so I can illustrate what's going

>> on in my head and what I think will fix it.

>

> You can build the original circuit in SPICE and run a frequency sweep

> simulation on it to see what it does at various frequencies. I swept

> my model from 1 MHz to 30 MHz (I think), and it was showing a resonance

> near 4 MHz.

>

>> I want to do this the way of the tube but I obviously require some

>> assistance. Dave, WA4QAL, has already helped me some but the people

>> on QRZ have stopped talking to me about it.

>

> It's not commonly done this way, so there probably aren't very many

> people who have had empirical experience with this circuit topology.

>

>> Either there's something glaringly obvious that I'm missing or I

>> don't understand how MOSFETs operate after all.

>

> The circuit does seem to be somewhat sensitive to bias and drive levels.

> You'll need to hit the MOSFET with a fairly large signal to drive it

> from cutoff to full conduction, which is why most solid state amplifiers

> have a broad-band step-up transformer, to produce a large voltage swing

> at the gate of the MOSFET. A direct drive input circuit, with a low

> power input signal, probably won't drive the gate hard enough to turn

> the MOSFET from cutoff to full conduction. Maybe. Depending upon your

> particular MOSFET part selection.

>

>> I don't want to go the route of broadband transformers because I

>> want to find out the hard way to do this.

>

> You'll find that, when doing things the hard way, that they're hard

> to do.

>

>> I appreciate anyone's suggestions and help.

>

> Run a SPICE simulation, ideally using LTSpice. I'll even toss in

> the sloppily constructed model I built yesterday.

>

> www.linear.com/designtools/software/

>

>> Ed

>

> Dave

>

>

---------------

>

> Always keep electrons and holes flowing in opposite directions.Yahoo!

> Groups Links

>

>

>

>







----------------------------

#1184 Jan 17, 2013

The pi network would allow me to factor in the output capacitance of

the fet which is something like 150pf. I couldn't find a drain

impedance for the fet but i chose 6 ohms so i could get 12w with a 12v

supply. Is it really that simple?

On 1/17/13, Max Robinson max@...> wrote:

> Ed. I meant to answer this when you first posted it but it slipped my so

> called mind. All you nee to do to make a pi net to step up impedance is to

>

> design as if it were a step down and then turn it around. The

> plate/collector/drain impedance is 50 ohms and the load is the drain

> impedance of the FET.

>

> Regards.

>

> Max. K 4 O DS.

>

> Email: max@...

>

> Transistor site www.funwithtransistors.net

> Vacuum tube site: www.funwithtubes.net

> Woodworking site

> www.angelfire.com/electronic/funwithtubes/Woodworking/wwindex.html

> Music site: www.maxsmusicplace.com

>

> To subscribe to the fun with transistors group send an email to.

> funwithtransistors-subscribe@yahoogroups.com

>

> To subscribe to the fun with tubes group send an email to,

> funwithtubes-subscribe@yahoogroups.com

>

> To subscribe to the fun with wood group send a blank email to

> funwithwood-subscribe@yahoogroups.com

>

> ----- Original Message -----

> From: "Dave" wa4qal@...>

> To: funwithtransistors@yahoogroups.com>

> Sent: Thursday, January 17, 2013 12:46 PM

> Subject: Re: [funwithtransistors] QRP Power Amp Again

>

>

>> --- On Wed, 1/16/13, J Ed jedwardsat1@...> wrote:

>>

>>> From: J Ed jedwardsat1@...>

>>> Subject: [funwithtransistors] QRP Power Amp Again

>>> To: "funwithtransistors" funwithtransistors@yahoogroups.com>

>>> Date: Wednesday, January 16, 2013, 3:19 AM

>>>

>>> Hi all,

>>>

>>> I'm trying to get a transistor RFPA to work and not having much

>>> success. The usual route with these is to stick broadband ferrite

>>> transformers on the inputs and outputs and subject the resulting

>>> signal to heavy filtering.

>>

>> And, there's a reason that that is usually done. ;-)

>>

>>> I'm trying to do this the old-fashioned way as it was done in tube

>>> amplifiers.

>>

>> The difference is that tube amplifiers usually have a fairly high

>> output impedance, while solid state amplifiers usually have a very low

>> output impedance.

>>

>> Thus, while you can do things the old fashioned way, you need to

>> take special steps to ensure that the low output impedance doesn't

>> result in excessive resistive losses in the output circuit.

>>

>>> The problem is, the MOSFET with the 12v power supply wants to see

>>> a lower power than 50 ohms rather than a higher one.

>>

>> Understandable.

>>

>>> Thus, I have to use a T matching circuit instead of the familiar

>>> pi one. Basically, one L steps the impedance up above 50 ohms

>>> and the second one takes it back down to 50.

>>

>> Ok...

>>

>>> Since the two capacitors would be in parallel, they can become one.

>>

>> Correct. However, note that, due to the low output impedance, you'll

>> have some rather incredible currents flowing, so you'll have to

>> ensure that the capacitor (and inductor) being used are capable of

>> handling those rather incredible currents without incurring excessive

>> (I**2)*R losses. That usually argues for inductors with rather heavy

>> gauge wire (or even Silver plated Copper tubing), and variable

>> capacitors with geometries different from the traditional rotary

>> variable capacitor, which usually has excessive resistance in the

>> rotary joint (e.g., A "butterfly" variable capacitor would be better,

>> since the current doesn't have to transition across a movable joint.).

>>

>>> Once I master this circuit, I want to make a single-ended one

>>> that's just like it, using a 2E26 for about 25w.

>>

>> Which will require a completely different input, output, bias, and

>> power supply circuit topography (Umm, will any part of the circuit

>> remain the same? I don't think so!).

>>

>>> I'll need to fiddle with the output components so I can

>>> match it to bands on 80 through 10.

>>

>> Check the values you have specified in the circuit. I'm showing a

>> resonant peak at near 4 MHz, not near 14 MHz. In the simulation I

>> ran, the output at 14 MHz is WAY down from what it is near 4 MHz.

>>

>>> For now, this amplifier is only going to operate on the 20m CW

>>> portion, which is 14.06 MHz.

>>

>> Check your circuit values. I may have used the wrong value for some

>> of the parasitic elements, but I'm showing a resonance very far away

>> from 14 MHz.

>>

>>> I'm including all of the image files so I can illustrate what's going

>>> on in my head and what I think will fix it.

>>

>> You can build the original circuit in SPICE and run a frequency sweep

>> simulation on it to see what it does at various frequencies. I swept

>> my model from 1 MHz to 30 MHz (I think), and it was showing a resonance

>> near 4 MHz.

>>

>>> I want to do this the way of the tube but I obviously require some

>>> assistance. Dave, WA4QAL, has already helped me some but the people

>>> on QRZ have stopped talking to me about it.

>>

>> It's not commonly done this way, so there probably aren't very many

>> people who have had empirical experience with this circuit topology.

>>

>>> Either there's something glaringly obvious that I'm missing or I

>>> don't understand how MOSFETs operate after all.

>>

>> The circuit does seem to be somewhat sensitive to bias and drive levels.

>> You'll need to hit the MOSFET with a fairly large signal to drive it

>> from cutoff to full conduction, which is why most solid state amplifiers

>> have a broad-band step-up transformer, to produce a large voltage swing

>> at the gate of the MOSFET. A direct drive input circuit, with a low

>> power input signal, probably won't drive the gate hard enough to turn

>> the MOSFET from cutoff to full conduction. Maybe. Depending upon your

>> particular MOSFET part selection.

>>

>>> I don't want to go the route of broadband transformers because I

>>> want to find out the hard way to do this.

>>

>> You'll find that, when doing things the hard way, that they're hard

>> to do.

>>

>>> I appreciate anyone's suggestions and help.

>>

>> Run a SPICE simulation, ideally using LTSpice. I'll even toss in

>> the sloppily constructed model I built yesterday.

>>

>> www.linear.com/designtools/software/

>>

>>> Ed

>>

>> Dave

>>

>>

---------------

>>

>> Always keep electrons and holes flowing in opposite directions.Yahoo!

>> Groups Links

>>

>>

>>

>>

>

>







----------------------------

#1185 Jan 18, 2013

And, there's a reason that that is usually done. ;-)



I get that. Still, I'm soon going to be doing this with 400-ish volts

and a 2E26, so I want to approximate the design with something a

little smaller and safer first.

>The difference is that tube amplifiers usually have a fairly high

>output impedance, while solid state amplifiers usually have a >very low output impedance.



I get that. That's because tubes tend to push voltage instead of

current and transistors do the opposite. I am intending to make a

couple of transistor power amplifiers first and then start on my tube

project.

>Thus, while you can do things the old fashioned way, you need >to take special steps to ensure that the low output impedance >doesn't result in excessive resistive losses in the output circuit.



Yep. Resistive losses can occur in a toroid core just as they can in a

pi-L network. I have been winding my coils with 20 AWG with no core

material in order to keep losses low. I am looking at silver-mica

capacitors online for the capacitors, because someone (I think it was

you) said that ceramic caps can be lossy and drift at higher RF

powers.

>Correct. However, note that, due to the low output impedance, >you'll have some rather incredible currents flowing, so you'll have >to ensure that the capacitor (and inductor) being used are >capable of handling those rather incredible currents without >incurring excessive (I**2)*R losses. That usually argues for >inductors with rather heavy gauge wire (or even Silver plated >Copper tubing), and variable capacitors with geometries different >from the traditional rotary variable capacitor, which usually has >excessive resistance in the rotary joint (e.g., A "butterfly" >variable capacitor would be better, since the current doesn't >have to transition across a movable joint.).



This is all true. I'm using rather heavy wire for the current levels

I'm employing. There won't be a demand for more than a few watts. The

purpose of this amplifier is to increase the power of my small QRPP

transceiver to rise up above the noise floor. If I ever sell or trade

that transceiver then I'll also sell the power amp with it.



Right now I'm making some capacitors so I can build a QRO tuner. I'm

not sure whether or not I want to buy that MFJ tuner. It would give me

a good point of reference.

>Which will require a completely different input, output, bias, and

>power supply circuit topography (Umm, will any part of the circuit

>remain the same? I don't think so!).



Differences: the grid will require a negative bias voltage instead of

a positive one. The input may be routed to the cathode of the tube

instead of the grid, depending on which one works better. The screen

grid will require its own regulated supply. The pi network will need

to match a higher impedance instead of a low.



Similarities: there will still be a plate choke (to cover down to

160m, I'd need one with an inductance of 4,000 uH). There will still

be a plate blocking choke and a pi network. The input network may be

similar as it would have to match a similar impedance.

>Check the values you have specified in the circuit. I'm showing a

>resonant peak at near 4 MHz, not near 14 MHz. In the >simulation I ran, the output at 14 MHz is WAY down from what >it is near 4 MHz.



Hm. I had been considering pulling the T network and trying again with

a different type. Maybe I will have more luck with the pi network.

>Check your circuit values. I may have used the wrong value for >some of the parasitic elements, but I'm showing a resonance >very far away from 14 MHz.



That would make sense. When I prodded the output coils the output

level moved slightly. The center capacitor I used wasn't the correct

value...I was supposed to have a 350 pF and the one I found was

something like 380. Maybe it was the other way around.

>You can build the original circuit in SPICE and run a frequency >sweep simulation on it to see what it does at various >frequencies. I swept my model from 1 MHz to 30 MHz (I think), >and it was showing a resonance near 4 MHz.



I will have to do that. In TINA-TI, I ran the circuit as built and got

2.2 watts out. Maybe I didn't enter in some parasitic capacitances

that exist in the circuit.

>It's not commonly done this way, so there probably aren't very many

>people who have had empirical experience with this circuit topology.



Which is why there needs to be people like me. I'm intending on

retaining the knowledge to pass on to future electronics people.

>The circuit does seem to be somewhat sensitive to bias and drive levels.

>You'll need to hit the MOSFET with a fairly large signal to drive it

>from cutoff to full conduction, which is why most solid state amplifiers

>have a broad-band step-up transformer, to produce a large voltage swing

>at the gate of the MOSFET. A direct drive input circuit, with a low

>power input signal, probably won't drive the gate hard enough to turn

>the MOSFET from cutoff to full conduction. Maybe. Depending upon your

>particular MOSFET part selection.



That's what I was thinking at first...I put in a bias circuit so I can

put the MOSFET gate right under the point of conduction. The Rockmite

puts out about 5 volts peak so that should be enough to kick the FET

into heavy current flow.

>You'll find that, when doing things the hard way, that they're hard to do.



Yes. When I make another amplifier with some ferrite cores I ordered,

I'll appreciate how easy it is to get raw power with a pair of

2N2222As in push-pull into a FT37-61 core. Raw power is a relative

term, but that approach will be much less frustrating than doing it

with a tuned output network. I think that's the point, though, so I

can learn how to make a tuned output network. The next few times

should be easier.

>Run a SPICE simulation, ideally using LTSpice. I'll even toss in

>the sloppily constructed model I built yesterday.

>www.linear.com/designtools/software/



I'll hold on to that. I think, one day, I'll run some experiments on

those ferrite transformer cores that can be scrapped from power

supplies. They're pretty large so it stands to reason I could make a

100w output coupler using one.



Ed







----------------------------

#1186 Jan 18, 2013

Source impedance is Ed/Id. It's really that simple.



Regards.



Max. K 4 O DS.



Email: max@...



Transistor site www.funwithtransistors.net

Vacuum tube site: www.funwithtubes.net

Woodworking site

www.angelfire.com/electronic/funwithtubes/Woodworking/wwindex.html

Music site: www.maxsmusicplace.com



To subscribe to the fun with transistors group send an email to.

funwithtransistors-subscribe@yahoogroups.com



To subscribe to the fun with tubes group send an email to,

funwithtubes-subscribe@yahoogroups.com



To subscribe to the fun with wood group send a blank email to

funwithwood-subscribe@yahoogroups.com

----- Original Message -----

From: "J Ed" jedwardsat1@...>

To: funwithtransistors@yahoogroups.com>

Sent: Thursday, January 17, 2013 3:55 PM

Subject: Re: [funwithtransistors] QRP Power Amp Again





> The pi network would allow me to factor in the output capacitance of

> the fet which is something like 150pf. I couldn't find a drain

> impedance for the fet but i chose 6 ohms so i could get 12w with a 12v

> supply. Is it really that simple?

>

> On 1/17/13, Max Robinson max@...> wrote:

>> Ed. I meant to answer this when you first posted it but it slipped my so

>> called mind. All you nee to do to make a pi net to step up impedance is

>> to

>>

>> design as if it were a step down and then turn it around. The

>> plate/collector/drain impedance is 50 ohms and the load is the drain

>> impedance of the FET.

>>

>> Regards.

>>

>> Max. K 4 O DS.

>>

>> Email: max@...

>>

>> Transistor site www.funwithtransistors.net

>> Vacuum tube site: www.funwithtubes.net

>> Woodworking site

>> www.angelfire.com/electronic/funwithtubes/Woodworking/wwindex.html

>> Music site: www.maxsmusicplace.com

>>

>> To subscribe to the fun with transistors group send an email to.

>> funwithtransistors-subscribe@yahoogroups.com

>>

>> To subscribe to the fun with tubes group send an email to,

>> funwithtubes-subscribe@yahoogroups.com

>>

>> To subscribe to the fun with wood group send a blank email to

>> funwithwood-subscribe@yahoogroups.com

>>

>> ----- Original Message -----

>> From: "Dave" wa4qal@...>

>> To: funwithtransistors@yahoogroups.com>

>> Sent: Thursday, January 17, 2013 12:46 PM

>> Subject: Re: [funwithtransistors] QRP Power Amp Again

>>

>>

>>> --- On Wed, 1/16/13, J Ed jedwardsat1@...> wrote:

>>>

>>>> From: J Ed jedwardsat1@...>

>>>> Subject: [funwithtransistors] QRP Power Amp Again

>>>> To: "funwithtransistors" funwithtransistors@yahoogroups.com>

>>>> Date: Wednesday, January 16, 2013, 3:19 AM

>>>>

>>>> Hi all,

>>>>

>>>> I'm trying to get a transistor RFPA to work and not having much

>>>> success. The usual route with these is to stick broadband ferrite

>>>> transformers on the inputs and outputs and subject the resulting

>>>> signal to heavy filtering.

>>>

>>> And, there's a reason that that is usually done. ;-)

>>>

>>>> I'm trying to do this the old-fashioned way as it was done in tube

>>>> amplifiers.

>>>

>>> The difference is that tube amplifiers usually have a fairly high

>>> output impedance, while solid state amplifiers usually have a very low

>>> output impedance.

>>>

>>> Thus, while you can do things the old fashioned way, you need to

>>> take special steps to ensure that the low output impedance doesn't

>>> result in excessive resistive losses in the output circuit.

>>>

>>>> The problem is, the MOSFET with the 12v power supply wants to see

>>>> a lower power than 50 ohms rather than a higher one.

>>>

>>> Understandable.

>>>

>>>> Thus, I have to use a T matching circuit instead of the familiar

>>>> pi one. Basically, one L steps the impedance up above 50 ohms

>>>> and the second one takes it back down to 50.

>>>

>>> Ok...

>>>

>>>> Since the two capacitors would be in parallel, they can become one.

>>>

>>> Correct. However, note that, due to the low output impedance, you'll

>>> have some rather incredible currents flowing, so you'll have to

>>> ensure that the capacitor (and inductor) being used are capable of

>>> handling those rather incredible currents without incurring excessive

>>> (I**2)*R losses. That usually argues for inductors with rather heavy

>>> gauge wire (or even Silver plated Copper tubing), and variable

>>> capacitors with geometries different from the traditional rotary

>>> variable capacitor, which usually has excessive resistance in the

>>> rotary joint (e.g., A "butterfly" variable capacitor would be better,

>>> since the current doesn't have to transition across a movable joint.).

>>>

>>>> Once I master this circuit, I want to make a single-ended one

>>>> that's just like it, using a 2E26 for about 25w.

>>>

>>> Which will require a completely different input, output, bias, and

>>> power supply circuit topography (Umm, will any part of the circuit

>>> remain the same? I don't think so!).

>>>

>>>> I'll need to fiddle with the output components so I can

>>>> match it to bands on 80 through 10.

>>>

>>> Check the values you have specified in the circuit. I'm showing a

>>> resonant peak at near 4 MHz, not near 14 MHz. In the simulation I

>>> ran, the output at 14 MHz is WAY down from what it is near 4 MHz.

>>>

>>>> For now, this amplifier is only going to operate on the 20m CW

>>>> portion, which is 14.06 MHz.

>>>

>>> Check your circuit values. I may have used the wrong value for some

>>> of the parasitic elements, but I'm showing a resonance very far away

>>> from 14 MHz.

>>>

>>>> I'm including all of the image files so I can illustrate what's going

>>>> on in my head and what I think will fix it.

>>>

>>> You can build the original circuit in SPICE and run a frequency sweep

>>> simulation on it to see what it does at various frequencies. I swept

>>> my model from 1 MHz to 30 MHz (I think), and it was showing a resonance

>>> near 4 MHz.

>>>

>>>> I want to do this the way of the tube but I obviously require some

>>>> assistance. Dave, WA4QAL, has already helped me some but the people

>>>> on QRZ have stopped talking to me about it.

>>>

>>> It's not commonly done this way, so there probably aren't very many

>>> people who have had empirical experience with this circuit topology.

>>>

>>>> Either there's something glaringly obvious that I'm missing or I

>>>> don't understand how MOSFETs operate after all.

>>>

>>> The circuit does seem to be somewhat sensitive to bias and drive levels.

>>> You'll need to hit the MOSFET with a fairly large signal to drive it

>>> from cutoff to full conduction, which is why most solid state amplifiers

>>> have a broad-band step-up transformer, to produce a large voltage swing

>>> at the gate of the MOSFET. A direct drive input circuit, with a low

>>> power input signal, probably won't drive the gate hard enough to turn

>>> the MOSFET from cutoff to full conduction. Maybe. Depending upon your

>>> particular MOSFET part selection.

>>>

>>>> I don't want to go the route of broadband transformers because I

>>>> want to find out the hard way to do this.

>>>

>>> You'll find that, when doing things the hard way, that they're hard

>>> to do.

>>>

>>>> I appreciate anyone's suggestions and help.

>>>

>>> Run a SPICE simulation, ideally using LTSpice. I'll even toss in

>>> the sloppily constructed model I built yesterday.

>>>

>>> www.linear.com/designtools/software/

>>>

>>>> Ed

>>>

>>> Dave

>>>

>>>

---------------

>>>

>>> Always keep electrons and holes flowing in opposite directions.Yahoo!

>>> Groups Links

>>>

>>>

>>>

>>>

>>

>>

>

>

---------------

>

> Always keep electrons and holes flowing in opposite directions.Yahoo!

> Groups Links

>

>

>

>







----------------------------

#1187 Jan 18, 2013

Cool. When I gravitate back home I will retry the device. If it works out well I think I'm going to work on a DC vibrator supply so I can run some tubes. With any luck, I might be able to go HF mobile by hooking the supply up to a vehicle.

Ed

On Thu, Jan 17, 2013 at 8:50 PM, Max Robinson max@...> wrote:

.Source impedance is Ed/Id. It's really that simple.



Regards.



Max. K 4 O DS.



Email: max@...



Transistor site www.funwithtransistors.net

Vacuum tube site: www.funwithtubes.net

Woodworking site

www.angelfire.com/electronic/funwithtubes/Woodworking/wwindex.html

Music site: www.maxsmusicplace.com



To subscribe to the fun with transistors group send an email to.

funwithtransistors-subscribe@yahoogroups.com



To subscribe to the fun with tubes group send an email to,

funwithtubes-subscribe@yahoogroups.com



To subscribe to the fun with wood group send a blank email to

funwithwood-subscribe@yahoogroups.com



----- Original Message -----

From: "J Ed" jedwardsat1@...>

To: funwithtransistors@yahoogroups.com>

Sent: Thursday, January 17, 2013 3:55 PM

Subject: Re: [funwithtransistors] QRP Power Amp Again



> The pi network would allow me to factor in the output capacitance of

> the fet which is something like 150pf. I couldn't find a drain

> impedance for the fet but i chose 6 ohms so i could get 12w with a 12v

> supply. Is it really that simple?

>

> On 1/17/13, Max Robinson max@...> wrote:

>> Ed. I meant to answer this when you first posted it but it slipped my so

>> called mind. All you nee to do to make a pi net to step up impedance is

>> to

>>

>> design as if it were a step down and then turn it around. The

>> plate/collector/drain impedance is 50 ohms and the load is the drain

>> impedance of the FET.

>>

>> Regards.

>>

>> Max. K 4 O DS.

>>

>> Email: max@...

>>

>> Transistor site www.funwithtransistors.net

>> Vacuum tube site: www.funwithtubes.net

>> Woodworking site

>> www.angelfire.com/electronic/funwithtubes/Woodworking/wwindex.html

>> Music site: www.maxsmusicplace.com

>>

>> To subscribe to the fun with transistors group send an email to.

>> funwithtransistors-subscribe@yahoogroups.com

>>

>> To subscribe to the fun with tubes group send an email to,

>> funwithtubes-subscribe@yahoogroups.com

>>

>> To subscribe to the fun with wood group send a blank email to

>> funwithwood-subscribe@yahoogroups.com

>>

>> ----- Original Message -----

>> From: "Dave" wa4qal@...>

>> To: funwithtransistors@yahoogroups.com>

>> Sent: Thursday, January 17, 2013 12:46 PM

>> Subject: Re: [funwithtransistors] QRP Power Amp Again

>>

>>

>>> --- On Wed, 1/16/13, J Ed jedwardsat1@...> wrote:

>>>

>>>> From: J Ed jedwardsat1@...>

>>>> Subject: [funwithtransistors] QRP Power Amp Again

>>>> To: "funwithtransistors" funwithtransistors@yahoogroups.com>

>>>> Date: Wednesday, January 16, 2013, 3:19 AM

>>>>

>>>> Hi all,

>>>>

>>>> I'm trying to get a transistor RFPA to work and not having much

>>>> success. The usual route with these is to stick broadband ferrite

>>>> transformers on the inputs and outputs and subject the resulting

>>>> signal to heavy filtering.

>>>

>>> And, there's a reason that that is usually done. ;-)

>>>

>>>> I'm trying to do this the old-fashioned way as it was done in tube

>>>> amplifiers.

>>>

>>> The difference is that tube amplifiers usually have a fairly high

>>> output impedance, while solid state amplifiers usually have a very low

>>> output impedance.

>>>

>>> Thus, while you can do things the old fashioned way, you need to

>>> take special steps to ensure that the low output impedance doesn't

>>> result in excessive resistive losses in the output circuit.

>>>

>>>> The problem is, the MOSFET with the 12v power supply wants to see

>>>> a lower power than 50 ohms rather than a higher one.

>>>

>>> Understandable.

>>>

>>>> Thus, I have to use a T matching circuit instead of the familiar

>>>> pi one. Basically, one L steps the impedance up above 50 ohms

>>>> and the second one takes it back down to 50.

>>>

>>> Ok...

>>>

>>>> Since the two capacitors would be in parallel, they can become one.

>>>

>>> Correct. However, note that, due to the low output impedance, you'll

>>> have some rather incredible currents flowing, so you'll have to

>>> ensure that the capacitor (and inductor) being used are capable of

>>> handling those rather incredible currents without incurring excessive

>>> (I**2)*R losses. That usually argues for inductors with rather heavy

>>> gauge wire (or even Silver plated Copper tubing), and variable

>>> capacitors with geometries different from the traditional rotary

>>> variable capacitor, which usually has excessive resistance in the

>>> rotary joint (e.g., A "butterfly" variable capacitor would be better,

>>> since the current doesn't have to transition across a movable joint.).

>>>

>>>> Once I master this circuit, I want to make a single-ended one

>>>> that's just like it, using a 2E26 for about 25w.

>>>

>>> Which will require a completely different input, output, bias, and

>>> power supply circuit topography (Umm, will any part of the circuit

>>> remain the same? I don't think so!).

>>>

>>>> I'll need to fiddle with the output components so I can

>>>> match it to bands on 80 through 10.

>>>

>>> Check the values you have specified in the circuit. I'm showing a

>>> resonant peak at near 4 MHz, not near 14 MHz. In the simulation I

>>> ran, the output at 14 MHz is WAY down from what it is near 4 MHz.

>>>

>>>> For now, this amplifier is only going to operate on the 20m CW

>>>> portion, which is 14.06 MHz.

>>>

>>> Check your circuit values. I may have used the wrong value for some

>>> of the parasitic elements, but I'm showing a resonance very far away

>>> from 14 MHz.

>>>

>>>> I'm including all of the image files so I can illustrate what's going

>>>> on in my head and what I think will fix it.

>>>

>>> You can build the original circuit in SPICE and run a frequency sweep

>>> simulation on it to see what it does at various frequencies. I swept

>>> my model from 1 MHz to 30 MHz (I think), and it was showing a resonance

>>> near 4 MHz.

>>>

>>>> I want to do this the way of the tube but I obviously require some

>>>> assistance. Dave, WA4QAL, has already helped me some but the people

>>>> on QRZ have stopped talking to me about it.

>>>

>>> It's not commonly done this way, so there probably aren't very many

>>> people who have had empirical experience with this circuit topology.

>>>

>>>> Either there's something glaringly obvious that I'm missing or I

>>>> don't understand how MOSFETs operate after all.

>>>

>>> The circuit does seem to be somewhat sensitive to bias and drive levels.

>>> You'll need to hit the MOSFET with a fairly large signal to drive it

>>> from cutoff to full conduction, which is why most solid state amplifiers

>>> have a broad-band step-up transformer, to produce a large voltage swing

>>> at the gate of the MOSFET. A direct drive input circuit, with a low

>>> power input signal, probably won't drive the gate hard enough to turn

>>> the MOSFET from cutoff to full conduction. Maybe. Depending upon your

>>> particular MOSFET part selection.

>>>

>>>> I don't want to go the route of broadband transformers because I

>>>> want to find out the hard way to do this.

>>>

>>> You'll find that, when doing things the hard way, that they're hard

>>> to do.

>>>

>>>> I appreciate anyone's suggestions and help.

>>>

>>> Run a SPICE simulation, ideally using LTSpice. I'll even toss in

>>> the sloppily constructed model I built yesterday.

>>>

>>> www.linear.com/designtools/software/

>>>

>>>> Ed

>>>

>>> Dave

>>>

>>>

---------------

>>>

>>> Always keep electrons and holes flowing in opposite directions.Yahoo!

>>> Groups Links

>>>

>>>

>>>

>>>

>>

>>

>

>

---------------

>

> Always keep electrons and holes flowing in opposite directions.Yahoo!

> Groups Links

>

>

>

>







This Site's Privacy Policy
Google's privacy policies

S
e
n
i
o
r
T
u
b
e
.
o
r
g