EMRFD Message Archive 9716

Message Date From Subject
9716 2014-02-11 16:32:41 Ephemeral Investigating a Norton feedback amplifier for battery operation
I have been investigating a Norton noiseless feedback amplifier. The published results seem
to all describe high power very high intercept amplifiers using transistors like a 2N5109 with
a heat sink.  I am interested in battery operation so my power budget is about 5 mA at 5 V. I have
been two-tone testing at 27 MHz. One source was crystal oscillator as per the book and the
other was my signal generator; they were combined with a 6 dB hybrid.
 
The transformer ratio was 1:19:5 yielding a nominal gain of 14 dB. I was a bit concerned that at
27 MHz the winding capacitance would be too high and performance would drop, but actually
performance seems to hold up to 40 MHz or so before dropping off slightly. I tried both a
BN43-2402 and a BN61-2402 binocular core and did not detect any difference in performance.
(It might be that a toroid would be better, to reduce capacitance. Alternatively a Trask design
would give high gain without such a high collector impedance).
 
Given that the current gain of our favourite transistors (e.g. 2N3904) is dropping off at 27 MHz
I preferred transistors with higher Ft. I tried 3 different transistor types, all of them in SOT-23
packages:
 
  MMBTH10 (similar to MPSH10)
 
I was a bit disappointed with this. The gain was about 1 dB lower than the others, which suggests
to me that the negative feedback was not working too well. The distortion was nothing special.
 
  DSC2G02 (high linearity and low noise, they say)
 
The gain was OK but the distortion was marginally worse than the MMBTH10. I suspect it is
designed for low voltages and low collector currents so it might give relatively better results
for a really low power preamp.
 
  MMBT5179 (similar to 2N5179)
 
The gain was OK. The distortion was nearly 20 dB less than the MMBTH10. So this was the clear
winner of the 3. The Ft is the highest but no stability issues were observed. Upping the voltage to
8 V (with a corresponding rise in current) reduced the distortion by about another 10 dB; further
increasing the voltage caused it to deteriorate again and get noticeably hot.
 
If anyone has any suggestions for other transistors to try, I would be interested to hear you. I would
also like to test at 5 MHz, but when I tried this using my function generator as the second source
the results were dire. So I need to build another crystal source. At 5 MHz other transistors could be
tried: MMBT3904 and MMBT4401. I also intend to test the noise figure.
 
 
Adrian
 
 
 
9717 2014-02-11 18:39:20 Russell Shaw Re: Investigating a Norton feedback amplifier for battery operation
9718 2014-02-11 19:40:08 Harold Smith Re: Investigating a Norton feedback amplifier for battery operation
I don't know if the MPS-H17 is still made, or if there is/was a version in a surface mount package, but we used to use it in TV video IF amps.  It falls between the H10 and the 5179 as far as Ft, and had better distortion performance than either, at least the way we used them.  Sorry I have no more information, but I hope this might help.

de KE6TI, Harold


9719 2014-02-11 23:58:13 kerrypwr Re: Investigating a Norton feedback amplifier for battery operation

I built a two-stage Norton amplifier with MRF581A devices; they are available on ebay which is where I bought my "stash".


They are in a "helicopter" package which is convenient for "ugly" construction;


http://i62.tinypic.com/2n19j4x.jpg


I used rather more current than you propose with these results;


http://i58.tinypic.com/dcbfcp.jpg


I think that there are resonances going-on in the transformers at very high frequencies.


Kerry VK2TIL.



9720 2014-02-12 02:13:26 Roelof Bakker Re: Investigating a Norton feedback amplifier for battery operation
Hello Adrian,

I have used a BFQ19 to replace the 2N5109.
It might be worthwhile to try it.

At the moment I am testing a BFG591 in a Norton amplifier.
My aim is to get a high IP3, which is not your goal.
Just like the BFQ19 this is a transistor with an Ft above 5 GHz.
It provides better strong signal handling performance than the 2N5109.
The IP3-in is approaching +40 dBm and my test setup can't handle this
properly.

Both transistors are cheap and widely available.

73,
Roelof Bakker, pa0rdt
9721 2014-02-12 02:51:44 sm5glc Re: Investigating a Norton feedback amplifier for battery operation

Roelof,

David Norton did write an nice application note while working for Anzac, describing how the feedback works and what trade-off to be made. Cannot find it right now though, somwhere in the pile of paper :)

Still there  is an article written by David N in Microwave Journal may 1976 which may be of interest, covering pretty much the same as in the application note. This articel is available if you need it :)


Ulrich Rohde did show a two transistor version with BFT66 and BFR34 which did cover up to 500 MHz. Design was made by some other german (DK1NV???) and I think (not sure) this was presented either in UKW berichte or VHF com late 70 or early 80. Rohde did write abut it in his book "!Digital PLL Frequency Synthesizers",


/Lasse SM5GLC

9722 2014-02-12 03:05:39 Weddig, Henning-C... Re: Investigating a Norton feedback amplifier for battery operation
I remember that Michael Martin DJ7VY first wrote about the Norton amp
in the UKW Berichte in the late 70´s
first an article about the amp itself, then these amps were used iun a 2
m frontend

ULR has "refinded" the design, e.g. base emitter feedback see e.g. his
book "communication receivers DSP Software Radios and design McGraw Hill
page 288 ff

73
Henning


Am 12.02.2014 11:51, schrieb lasse.moell@webdog.se:
> Roelof,
>
> David Norton did write an nice application note while working for Anzac,
> describing how the feedback works and what trade-off to be made. Cannot
> find it right now though, somwhere in the pile of paper :)
>
> Still there is an article written by David N in Microwave Journal may
> 1976 which may be of interest, covering pretty much the same as in the
> application note. This articel is available if you need it :)
>
>
> Ulrich Rohde did show a two transistor version with BFT66 and BFR34
> which did cover up to 500 MHz. Design was made by some other german
> (DK1NV???) and I think (not sure) this was presented either in UKW
> berichte or VHF com late 70 or early 80. Rohde did write abut it in his
> book "!Digital PLL Frequency Synthesizers",
>
>
> /Lasse SM5GLC
>
>
9724 2014-02-12 09:44:58 Ephemeral Re: Investigating a Norton feedback amplifier for battery operation
Thanks for all the interesting suggestions. I am only really interested in SM devices
so I am unlikely to order in older devices for experiments. However I do have some
BFY90 transistors in stock, which might be worth testing some day.
 
I have been measuring the noise figure of my Amplifier. I tried 5 MHz and 27 MHz with
very little difference between the two frequencies. Power supply was 5V and current
about 5mA. I tried with various transistors:
 
  MMBT3904. NF=9.9 dB. I find this result surprisingly bad. It is possible of course
that my transistor was a particularly bad specimen. I don't know whether the figure
varies much between production batches of between manufactures.
 
  DSC2G02 NF=4.9 dB.
 
  MMBTH10 NF=4.9 dB.
 
  MMBT5179 NF=2.9 dB.
 
  MMBT4401 NF=2.6 dB.
 
I tested the MMBT4401 for distortion at 27MHz and got results  similar but a little
worse than the MMBT5179. I should like to test it at 5 MHz some day in case it has
potential, but for this project the MMBT5179 seems to be giving me the results I desire.
 
Adrian
 
 
 
9725 2014-02-12 10:24:17 chutton12000 Re: Investigating a Norton feedback amplifier for battery operation

Adrian:


Have you investigated the low power option of the Z10042A at http://cliftonlaboratories.com/z10042a_norton_amplifier.htm ?


It draws only 25 mA for a design with good IP3 - check out the excellent documentation.

Jack has been down the route of investigating different devices for Norton amps.

Also, Dallas Lankford has done some work, although I don't have a specific reference or source.


Chuck


9727 2014-02-12 13:00:11 Alberto I2PHD Re: Investigating a Norton feedback amplifier for battery operation
9728 2014-02-13 10:21:10 Ephemeral Re: Investigating a Norton feedback amplifier for battery operation
I have been checking my noise results. I discovered I had extra attenuation in my analyser so
the noise after the Norton amplifier was higher than I thought so I was getting numbers too
high. I also realised that the inductance of the RFC in the emitter bias was too to work properly
at 5MHz. So I have revised the numbers:
 
Norton amplifier 1:19:5 on BN43-2402, Vce = 4.8 V, Icc = 4.5 mA. 27MHz (except MMBT4401)
 
9.8 dB MMBT3904 (!?)
4.5 dB DSC2G02
4.5 dB MMBTH10
2.2 dB MMBT5179
1.8 dB MMBT4401 (5 MHz)
 
When I increased the supply voltage to 8V the noise level for the MMB4401 rose by about 0.4 dB.
I don't know what happens with the other transistors, but you should understand that other
transistors may perform better at different power levels. YMMV. Some day perhaps I will check
out the BFQ19 and NE46134.
 
It is possible that the noise at 27 MHz might drop by a fraction of a dB when the core is changed
to type 61 ferrite. I not going to bother to test it at this time because I don't feel a great need to
know.
 
The results obtained are now in line with other published results and meet my design goals. So
that concludes this series of my experiments. Now it is time to get the rest of the receiver working.
 
Thanks again for all the replies and sorry for the inaccurate numbers.
 
Adrian
 
 
 
9731 2014-02-13 13:26:14 kb1gmx Re: Investigating a Norton feedback amplifier for battery operation
Two things increasing the supply voltage will increase the operating current, for some devices the nose will go up and other down.  

Part of the reason is the specific devices characteristics over a range of currents.
The other is in some case the device will exhibit a noticeable different set of 
impedance at the terminals as the current changes.  This changes the match or 
mismatch and gain.

That MMBT3904 number seems high as I've used that in resistive feedback amps
and I usually get around 5-6db noise figure but at 10ma.

My experience with norton amps is they give the characteristic device noise plus a little
as the emitter bias resistor is still in the input noise path.   The otehr is they tend like 
the resistive feedback amp suffer the output load affects input impedance. 

Allison
9737 2014-02-14 00:48:36 Weddig, Henning-C... Re: Investigating a Norton feedback amplifier for battery operation
could it be that the MMBT3904 was oscilalting at some high (UHF)
freqeuncies?

I myself observed such spurs when using the Norton in a 144 MHz preamp
with a _BRF91 oscillalting at about 800 MHz. Remedy: a ferrite bead on
the collector-- but a "wrong" ferrite material deterioted the IMD
performance...

73
Henning Weddig

Am 13.02.2014 19:21, schrieb Ephemeral:
> I have been checking my noise results. I discovered I had extra
> attenuation in my analyser so
> the noise after the Norton amplifier was higher than I thought so I was
> getting numbers too
> high. I also realised that the inductance of the RFC in the emitter bias
> was too to work properly
> at 5MHz. So I have revised the numbers:
> Norton amplifier 1:19:5 on BN43-2402, Vce = 4.8 V, Icc = 4.5 mA. 27MHz
> (except MMBT4401)
> 9.8 dB MMBT3904 (!?)
> 4.5 dBDSC2G02
> 4.5 dB MMBTH10
> 2.2 dB MMBT5179
> 1.8 dB MMBT4401 (5 MHz)
> When I increased the supply voltage to 8V the noise level for the
> MMB4401 rose by about 0.4 dB.
> I don't know what happens with the other transistors, but you should
> understand that other
> transistors may perform better at different power levels. YMMV. Some day
> perhaps I will check
> out the BFQ19 and NE46134.
> It is possible that the noise at 27 MHz might drop by a fraction of
> a dB when the core is changed
> to type 61 ferrite. I not going to bother to test it at this time
> because I don't feel a great need to
> know.
> The results obtained are now in line with other published results
> and meet my design goals. So
> that concludes this series of my experiments. Now it is time to get the
> rest of the receiver working.
> Thanks again for all the replies and sorry for the inaccurate numbers.
> Adrian
>
>