mosfet RF amplifier with resistance

[jure]

Well, my question no longer applies since I found the feedback arrangement
is very old, 1963, and well-known, even though it is often misspelled

You can apply feedback to the base of the input stage, which lowers the
input impedance and requires a known and fixed source impedance, or you can
apply it to the emitter, which raises the input impedance and requires a
resistor in the emtter. Plus a lot of other details.

But I don't understand the point of your comment. It is obvious, but what
does it have to do with the feedback configuration?

Regards,

Mike Monett

Mike,

If the input common source (CS) transistor sees an almost 0 ohm load
( the input Z of the transimpedance pair , if the loop gain is high
enough ) the time constant of that node will be very minimized.
Also, because this first stage has no voltage gain, Cgd1 is not
multiplied by Miller feedback when reflected into the input.
The CS input device has been unilateralized ... and its output behaves
close to an ideal current source.

Then, if ( BIG IF ) the second and third transistors could be modeled
as a TIA, with gain Rm, the total gain at midband frequencies is
(very) approximately equal to gm1.Rm .

The problem has now been transfered to the proper design of the
combination of M2 and M3.
Yet another twist would be to have M2 and M3 as cascode pair, instead
of a cascade CS - CS.

Here is my second attempt at ASCII art of the AC circuit.

|-----------|
|----VWV---|-----|| M3 |
| RL2 | |-->--| ///
/// | |
| |
|----| |
|----|| M2 |
| |-->-| |
| | |
| /// |
| |
| |
|----|-----------VWV---------|
IN ------|| M1 Rm
|-->---|
|
///

Thanks , Jure Z.

 

[Mike Monett]

[...]

Mike,

If the input common source (CS) transistor sees an almost 0 ohm load
( the input Z of the transimpedance pair , if the loop gain is high
enough ) the time constant of that node will be very minimized.
Also, because this first stage has no voltage gain, Cgd1 is not
multiplied by Miller feedback when reflected into the input.
The CS input device has been unilateralized ... and its output behaves
close to an ideal current source.

Then, if ( BIG IF ) the second and third transistors could be modeled
as a TIA, with gain Rm, the total gain at midband frequencies is
(very) approximately equal to gm1.Rm .

The problem has now been transfered to the proper design of the
combination of M2 and M3.
Yet another twist would be to have M2 and M3 as cascode pair, instead
of a cascade CS - CS.

Here is my second attempt at ASCII art of the AC circuit.


Thanks , Jure Z.

Thanks, Jure. I see that.

The only transistor supplying voltage gain is M2. M1 and M3 basically kill
Miller Effect and provide a low impedance drive to the next stage.

I wonder if that is the most effective use of the gain capability of the
three devices. For example, my understanding is the Tek 7400-series scopes
made a dramatic change in the way gain was obtained. Instead of going for
voltage gain in each stage, the designer went for current gain instead.

This obviously killed Miller effect, and reduced power consumption since
the only place voltage swing was needed was the last stage driving the crt.
Correct me if I'm wrong - the schematics really didn't tell much about what
was going on inside the chips. But that seemed a very effective way of
getting the high gain-bandwidth needed with a flat response.

Another configuration is travelling-wave, where low-gain transistors are
placed along a transmission line and operate in parallel. There, each one
contributes to the output signal. Horribly wasteful of power, but it
apparently satisfies a need.

So how do you decide which is the most overall effective use of the gain
potential of each device? Assume a 50 ohm environment and a reasonably low
noise figure.

BTW, I like your ascii art. Other people use a lot of funny characters,
and I have to reprogram my editor so it doesn't think they are erroneous
punctuation marks and try to correct them

Regards,

Mike Monett

 

[jure]

[...]

Mike,
If the input common source (CS) transistor sees an almost 0 ohm load
( the input Z of the transimpedance pair , if the loop gain is high
enough ) the time constant of that node will be very minimized.
Also, because this first stage has no voltage gain, Cgd1 is not
multiplied by Miller feedback when reflected into the input.
The CS input device has been unilateralized ... and its output behaves
close to an ideal current source.

Then, if ( BIG IF ) the second and third transistors could be modeled
as a TIA, with gain Rm, the total gain at midband frequencies is
(very) approximately equal to gm1.Rm .

The problem has now been transfered to the proper design of the
combination of M2 and M3.
Yet another twist would be to have M2 and M3 as cascode pair, instead
of a cascade CS - CS.

Here is my second attempt at ASCII art of the AC circuit.

Thanks , Jure Z.

Thanks, Jure. I see that.

The only transistor supplying voltage gain is M2. M1 and M3 basically kill
Miller Effect and provide a low impedance drive to the next stage.

I wonder if that is the most effective use of the gain capability of the
three devices. For example, my understanding is the Tek 7400-series scopes
made a dramatic change in the way gain was obtained. Instead of going for
voltage gain in each stage, the designer went for current gain instead.

This obviously killed Miller effect, and reduced power consumption since
the only place voltage swing was needed was the last stage driving the crt.
Correct me if I'm wrong - the schematics really didn't tell much about what
was going on inside the chips. But that seemed a very effective way of
getting the high gain-bandwidth needed with a flat response.

Another configuration is travelling-wave, where low-gain transistors are
placed along a transmission line and operate in parallel. There, each one
contributes to the output signal. Horribly wasteful of power, but it
apparently satisfies a need.

So how do you decide which is the most overall effective use of the gain
potential of each device? Assume a 50 ohm environment and a reasonably low
noise figure.

BTW, I like your ascii art. Other people use a lot of funny characters,
and I have to reprogram my editor so it doesn't think they are erroneous
punctuation marks and try to correct them

Regards,

Mike Monett

Mike,
I think that you may have read the book by Jim Williams:
Analog Circuit Design: Art, Science and Personalities . True ?

There is a chapter by John Addis: 14. Good
Engineering and Fast Vertical Amplifiers.

the T-Coil, the ft doubler, the cascomp and other amplifier
topologies are referenced, very interesting !

Thanks , Jure Z.

 

[Mike Monett]

Mike, I think that you may have read the book by Jim Williams:

Analog Circuit Design: Art, Science and Personalities. True ?

There is a chapter by John Addis: 14. Good Engineering and Fast
Vertical Amplifiers.

the T-Coil, the ft doubler, the cascomp and other amplifier
topologies are referenced, very interesting !

Thanks, Jure Z.

Hi Jure,

I haven't read it yet, but I thank you for the recommendation. It
looks very interesting. It is not very expensive compared to most:

http://www.amazon.com/Science-Analog-Circuit-Design-
Engineers/dp/0750670622

Regards,

Mike Monett

 

[jure]

[...]

Mike,
If the input common source (CS) transistor sees an almost 0 ohm load
( the input Z of the transimpedance pair , if the loop gain is high
enough ) the time constant of that node will be very minimized.
Also, because this first stage has no voltage gain, Cgd1 is not
multiplied by Miller feedback when reflected into the input.
The CS input device has been unilateralized ... and its output behaves
close to an ideal current source.

Then, if ( BIG IF ) the second and third transistors could be modeled
as a TIA, with gain Rm, the total gain at midband frequencies is
(very) approximately equal to gm1.Rm .

The problem has now been transfered to the proper design of the
combination of M2 and M3.
Yet another twist would be to have M2 and M3 as cascode pair, instead
of a cascade CS - CS.

Here is my second attempt at ASCII art of the AC circuit.

Thanks , Jure Z.

Thanks, Jure. I see that.

The only transistor supplying voltage gain is M2. M1 and M3 basically kill
Miller Effect and provide a low impedance drive to the next stage.

I wonder if that is the most effective use of the gain capability of the
three devices. For example, my understanding is the Tek 7400-series scopes
made a dramatic change in the way gain was obtained. Instead of going for
voltage gain in each stage, the designer went for current gain instead.

This obviously killed Miller effect, and reduced power consumption since
the only place voltage swing was needed was the last stage driving the crt.
Correct me if I'm wrong - the schematics really didn't tell much about what
was going on inside the chips. But that seemed a very effective way of
getting the high gain-bandwidth needed with a flat response.

Another configuration is travelling-wave, where low-gain transistors are
placed along a transmission line and operate in parallel. There, each one
contributes to the output signal. Horribly wasteful of power, but it
apparently satisfies a need.

So how do you decide which is the most overall effective use of the gain
potential of each device? Assume a 50 ohm environment and a reasonably low
noise figure.

BTW, I like your ascii art. Other people use a lot of funny characters,
and I have to reprogram my editor so it doesn't think they are erroneous
punctuation marks and try to correct them

Regards,

Mike Monett

Mike,

it seems that you may have read the book by Jim
Williams : Analog Circuit Design: Art, Science and Personalities.

there is a chapter by John Addis : 14 Good Engineering
and Fast Vertical Amplifiers.

There are references to the topics : T-Coil , distributed
amplifier, Ft doubler,Battjes, cascomp, Quinn, Hallen, Traa,
"Traashcomp".

these names may also be used as keywords in freepatents
online ...

Thanks, Jure Z.

 

[jure]

Hi Jure,

I haven't read it yet, but I thank you for the recommendation. It
looks very interesting. It is not very expensive compared to most:

http://www.amazon.com/Science-Analog-Circuit-Design-
Engineers/dp/0750670622

Regards,

Mike Monett

Here is a patent list ( my compilation) of the Ft doubler , cascomp
amplifier and derivatives
( use freepatents online)

US Patent No. Inventor Descr
3633120 Battjes
3868580 Battjes bootrastrapped amp .pdf
4146844 Quinn Feed forward amp.pdf
4236119 Battjes monolithic amplifier.pdf
4267516 Traa CE Ft doubler .pdf
4267516 Traa Ft doubler.pdf
4322688 Schlotzhauer Cascode Feed Forward.pdf
4484147 Metz shunt fb amp.pdf
4491803 Metz current limiting .pdf
4528515 Gross HF diff amp.pdf
4528515 Winthorp HF Diff amp.pdf
4528517 Schlotzhauer.pdf
4692712 Quinn Diff Z neutralization .pdf
4720685 Garuts FET amp.pdf
4730124 Battjes high gm PNP .pdf
4779057 Woo cascode w improved transient resp.pdf
4835488 Garuts Wideband Lin Emitter FB amp.pdf
4862102 LaVoie Minimizing Self heating Diff amp.pdf
4890067 Lamb CB Ft multiplier.pdf
5307024 Metz linearized level shifting amp.pdf
5331289 Battjes Translinear ft multiplier.pdf
5399988 Knierim Ft doubler.pdf
5424510 Gusinov Thermal Compensation.pdf
5495201 Alini Gm stage .pdf
5506536 Yeung Diff amp w-exp gain.pdf
5587689 Bowers voltage controlled amp .pdf
5677646 Entrikin Diff Pair amp.pdf
6320467 Thor Hallen Ft multiplier.pdf

Thnks , Jure Z.

 

[Mike Monett]

Here is a patent list (my compilation) of the Ft doubler, cascomp
amplifier and derivatives (use freepatents online)

US Patent No. Inventor Descr

3633120 : Battjes
3868580 : Battjes bootrastrapped amp
4146844 : Quinn Feed forward amp
4236119 : Battjes monolithic amplifier
4267516 : Traa CE Ft doubler
4267516 : Traa Ft doubler
4322688 : Schlotzhauer Cascode Feed Forward
4484147 : Metz shunt fb amp
4491803 : Metz current limiting
4528515 : Gross HF diff amp
4528515 : Winthorp HF Diff amp
4528517 : Schlotzhauer
4692712 : Quinn Diff Z neutralization
4720685 : Garuts FET amp
4730124 : Battjes high gm PNP
4779057 : Woo cascode w improved transient resp
4835488 : Garuts Wideband Lin Emitter FB amp
4862102 : LaVoie Minimizing Self heating Diff amp
4890067 : Lamb CB Ft multiplier
5307024 : Metz linearized level shifting amp
5331289 : Battjes Translinear ft multiplier
5399988 : Knierim Ft doubler
5424510 : Gusinov Thermal Compensation
5495201 : Alini Gm stage
5506536 : Yeung Diff amp w-exp gain
5587689 : Bowers voltage controlled amp
5677646 : Entrikin Diff Pair amp
6320467 : Thor Hallen Ft multiplier

Thnks, Jure Z.

Hi Jure,

Thanks very much for the above list. I'll bet that took a lot of
work to compile. I will download these files and study them.

Also, thank you very much for the ones you sent by email. That was
very nice of you, but I'm on dialup and they took a long time to
download. Also, I noticed afterwards my email client missed decoding
about half of them, so I was planning on downloading them anyway.

These are very interesting patents. They show many ways to bypass
the conventional limits on bandwidth. They also show there were a
lot of very smart people back in the old days

Regards,

Mike Monett

 

[jure]

3633120 : Battjes
3868580 : Battjes bootrastrapped amp
4146844 : Quinn Feed forward amp
4236119 : Battjes monolithic amplifier
4267516 : Traa CE Ft doubler
4267516 : Traa Ft doubler
4322688 : Schlotzhauer Cascode Feed Forward
4484147 : Metz shunt fb amp
4491803 : Metz current limiting
4528515 : Gross HF diff amp
4528515 : Winthorp HF Diff amp
4528517 : Schlotzhauer
4692712 : Quinn Diff Z neutralization
4720685 : Garuts FET amp
4730124 : Battjes high gm PNP
4779057 : Woo cascode w improved transient resp
4835488 : Garuts Wideband Lin Emitter FB amp
4862102 : LaVoie Minimizing Self heating Diff amp
4890067 : Lamb CB Ft multiplier
5307024 : Metz linearized level shifting amp
5331289 : Battjes Translinear ft multiplier
5399988 : Knierim Ft doubler
5424510 : Gusinov Thermal Compensation
5495201 : Alini Gm stage
5506536 : Yeung Diff amp w-exp gain
5587689 : Bowers voltage controlled amp
5677646 : Entrikin Diff Pair amp
6320467 : Thor Hallen Ft multiplier


Hi Jure,

Thanks very much for the above list. I'll bet that took a lot of
work to compile. I will download these files and study them.

Also, thank you very much for the ones you sent by email. That was
very nice of you, but I'm on dialup and they took a long time to
download. Also, I noticed afterwards my email client missed decoding
about half of them, so I was planning on downloading them anyway.

These are very interesting patents. They show many ways to bypass
the conventional limits on bandwidth. They also show there were a
lot of very smart people back in the old days

Regards,

Mike Monett

Mike,
Sorry for having sent the files without asking first.
I thought that I was the last one on dialup when I changed to DSL 16
months ago... (grin)

Thanks, Jure Z.

 

[Mike Monett]

Mike, Sorry for having sent the files without asking first.

I thought that I was the last one on dialup when I changed to DSL
16 months ago. (grin)

Thanks, Jure Z.

Hi Jure,

Thanks - I really appreciate your help and all the info you have
posted. I sure am glad you are interested in this topic and have
collected all these patents. They will keep me busy for a while.

There's not much choice for highspeed services here in Midland, so
the prices are high. Sympatico has switched over to Microsoft
software, so you never know when the system will go down or for how
long. If you call tech support to ask about any problem, they say
it's your computer and you are the only one having trouble. Later,
you find on the forums that everyone was having the same problem,
and they lied about where the problem was.

Typical Microsoft crap. They couldn't tell the truth if it was
printed on their eyeballs. I will never use any service that depends
on Microsoft. I try to stay with services that run on Linux. Much
more reliable, any problems are fixed in minutes instead of weeks,
and they tell you the truth.

Rogers cable has service here, but the price is too high. It is
bundeled with television, which I never watch. I don't even have a
TV set. So it is way too expensive for just the high speed.

There is another service I just heard about yesterday. It's
$29.95/month for 5M/800K for residential, which may be ok.

http://www.teksavvy.com/

I'll let my friends use it for a while and see how they like it,
then look at their contract and maybe consider switching over. In
the meantime, dialup is very reliable, so any new service I try will
have to meet very high standards.

Regards,

Mike Monett

 

[Tony Williams]

I'll let my friends use it for a while and see how they
like it, then look at their contract and maybe consider
switching over. In the meantime, dialup is very reliable, so
any new service I try will have to meet very high standards.

ADSL is well worth it Mike, it transforms the
browser performance and things like that DoS
attack on SED last night...... 5776 articles,
50MB to download! On ADSL that was just a
timewaster rather than a Denial of Service.

IMO reliability with ADSL starts with your
internal cabling and the quality of the line
to the DSLAM in the exchange. This is helped
by having a good Router, one where you can
see the line diagnostics for yourself. In the
UK, Netgear seems to have a good reputation for
being able to handle marginal lines.

Don't be tempted by the highest speeds. A rock
solid 2Mb/s is far better than a flakey 5Mb/s.

 

[Mike Monett]

ADSL is well worth it Mike, it transforms the browser performance
and things like that DoS attack on SED last night. 5776 articles,
50MB to download! On ADSL that was just a timewaster rather than a
Denial of Service.

Hi Tony, thanks for posting this info.

Unlike Firefox, Opera is very fast on dialup. I'm sure I will
eventually go to ADSL, as soon as I can find a reputable vendor (not
Symatico) that doesn't gouge too much and can keep the system up for
long periods.

The DOS attack on SED was no problem for XNews. Newsposts are stored
on two separate servers - the headers are on one, and the body text
on another. Xnews downloads the headers which contain the subject.
This is very fast. It was a simple matter to quickly scan the topics
and see they were all crap, then tell XNews to mark them all as
read. So I never had to download the 50MB of actual text.

IMO reliability with ADSL starts with your internal cabling and
the quality of the line to the DSLAM in the exchange. This is
helped by having a good Router, one where you can see the line
diagnostics for yourself. In the UK, Netgear seems to have a good
reputation for being able to handle marginal lines.

Thanks - that is good info to keep in mind. I didn't know that and
will look into it when I switch over.

Don't be tempted by the highest speeds. A rock solid 2Mb/s is far
better than a flakey 5Mb/s.

A flakey anything is always worse than nothing

Regards,

Mike Monett