Date   

Re: 2sc1815 interesting stable bjt vfo schematic

Ashhar Farhan <farhan@...>
 

over years, i have visited kazuhiro's site. his work output continues to amaze me each time. he experiments fearlessly, then documents it all too. quite a volume of work. i find his designs interesting but they take minimalism to an extent where reproducability and performance become difficult to sustain.

my favourite from his site is the no-power AM transmitter which is literally powered by your speech. extreme qrp!

...

the general idea of using transistors with feedback is that the amount of gain extracted is much less than what the transistor is capable of. In such cases as BITX20, there are just two really important parameters to be considered: the Ft should be above 150 Mhz, the noise figure should be less than 6dB at low currents. Using a 1GHz transister will not make a difference as the gain of the stage is pegged at 16dB regardless of the actual Ft or the beta of the transister. It might fall below 16dB if the Ft is any lower, but wont exceed 16dB in anycase.

The inherent stability (and hence the reproducability) of the BITX is due to the fact that transistor types dont matter much, and the negative feedback assures that the individual stages dont self-oscillate.

- farhan

On Fri, 26 Aug 2005, Max wrote:

JF1OZL in most of the constructions even on 50mhz is using this type
of trasistor. As with the increasing of colector current ft getting
higher. Also comon base ft is several times bigger than comon emiter
circuit in the contrarry with comon colector design.

More information about 2sc1815 and implementation of generall purpose
trasistors.

http://www.intio.or.jp/jf10zl/basic.htm


--- In BITX20@..., "ajparent1" <kb1gmx@a...> wrote:
The 2SC1815 is a good device but it's Ft is only 80mhz. For
HF VFOs that is adaquate but for the BITx if it may be a bit
weak. Other devices of the era are the 2sc784, 2sc930 and 2sc945
all good and easily found in salvaged equipment.

As to transistor capacitance in VFOs, it's is easy to bury the
device in configurations that overwhelm the devices capacitance.
By doing so you make the VFO less dependent on the device used and
it can aid stability. The series tuned Colpits VFO used in
orginal BIT-x is such an oscillator. The 500pf across Base to
Emitter is far larger tha the transistors internal capacitance.
The second 500pf capacitor from emitter to ground (collector is
at RF ground due to the .1uf from Collector to ground) swamps any
capacitance from Collector to Emitter.

One trick that I've used for VFOs that helps with stability is
a form of compensation. For most BJT VFOs there is a voltage
where the initial drift is neutral or nearly so. Often this is
easily found by using a LM317 variable voltage regulator and
starting at the design voltage and noting startup drift. Then by
adjustment of the voltage to the VCO it's possible to reduce or
even eliminate that drift. What it achieves is to find a point
where the transistors internal heating due to feedback and junction
heating is minimal or at least stable.

NOTE 1: This is not temperature compensation. There will still
be drift from warming or cooling of other components. The
inductor, capacitors and if used Varactor(Varicap) diodes are
still contributors to thermally drift. The Varactor diode is a
significant source of thermal drift in most VFOs.

NOTE 2: from testing I've found three terminal regulators
provide a more stable voltage than zener diodes due to temerature.


Allison
KB1GMX

--- In BITX20@..., "Max" <m_orwell@y...> wrote:
2sc1815 is toshiba pct process based transistor. so this mean
lower
capacitance and noise figure. this japanese design remind on
comon
base vfo used in the 70's in swan trx. schematics are in file
section.

73 de 4N1ZM






Yahoo! Groups Links






Re: 2sc1815 interesting stable bjt vfo schematic

ajparent1 <kb1gmx@...>
 

Hello,

Yes 'OZL is a fantastic and very active constructor. Indeed I've
been aware of his projects for many years now. The site for JF1OZL
is one of about 40 I have bookmarked as they represent both active
homebrewed (autoconstructed) and good application of theory.

www.qrp.pops.net is another worth looking along with www.k8iqy.com.

As to the 2SC1815, I do have a good quantity of them and it is a
very good and useful device. It only one of many I chose from
for various projects. I have an unusually good selection of new
devices as I used to provide service in commercial space for a
variety of electronic gear. I also have databooks for most major
makers of transistors as well as a detailed on line data sheet
archive. For BITX I don't think it is as good a choice for the
feedback amplifiers. For tuned amplifiers it is very good.

Most BJT devices can be used in common (grounded) base circuits
as the Alpha cutoff is much higher. I used to take advantage of
this ways back when transistors capable of HF or low VHF were very
hard to get.

Allison
KB1GMX

--- In BITX20@..., "Max" <m_orwell@y...> wrote:
JF1OZL in most of the constructions even on 50mhz is using this type
of trasistor. As with the increasing of colector current ft getting
higher. Also comon base ft is several times bigger than comon emiter
circuit in the contrarry with comon colector design.

More information about 2sc1815 and implementation of generall purpose
trasistors.

http://www.intio.or.jp/jf10zl/basic.htm


--- In BITX20@..., "ajparent1" <kb1gmx@a...> wrote:
The 2SC1815 is a good device but it's Ft is only 80mhz. For
HF VFOs that is adaquate but for the BITx if it may be a bit
weak. Other devices of the era are the 2sc784, 2sc930 and 2sc945
all good and easily found in salvaged equipment.

As to transistor capacitance in VFOs, it's is easy to bury the
device in configurations that overwhelm the devices capacitance.
By doing so you make the VFO less dependent on the device used and
it can aid stability. The series tuned Colpits VFO used in
orginal BIT-x is such an oscillator. The 500pf across Base to
Emitter is far larger tha the transistors internal capacitance.
The second 500pf capacitor from emitter to ground (collector is
at RF ground due to the .1uf from Collector to ground) swamps any
capacitance from Collector to Emitter.

One trick that I've used for VFOs that helps with stability is
a form of compensation. For most BJT VFOs there is a voltage
where the initial drift is neutral or nearly so. Often this is
easily found by using a LM317 variable voltage regulator and
starting at the design voltage and noting startup drift. Then by
adjustment of the voltage to the VCO it's possible to reduce or
even eliminate that drift. What it achieves is to find a point
where the transistors internal heating due to feedback and junction
heating is minimal or at least stable.

NOTE 1: This is not temperature compensation. There will still
be drift from warming or cooling of other components. The
inductor, capacitors and if used Varactor(Varicap) diodes are
still contributors to thermally drift. The Varactor diode is a
significant source of thermal drift in most VFOs.

NOTE 2: from testing I've found three terminal regulators
provide a more stable voltage than zener diodes due to temerature.


Allison
KB1GMX

--- In BITX20@..., "Max" <m_orwell@y...> wrote:
2sc1815 is toshiba pct process based transistor. so this mean
lower
capacitance and noise figure. this japanese design remind on
comon
base vfo used in the 70's in swan trx. schematics are in file
section.

73 de 4N1ZM


Re: 2sc1815 interesting stable bjt vfo schematic

Max
 

JF1OZL in most of the constructions even on 50mhz is using this type
of trasistor. As with the increasing of colector current ft getting
higher. Also comon base ft is several times bigger than comon emiter
circuit in the contrarry with comon colector design.

More information about 2sc1815 and implementation of generall purpose
trasistors.

http://www.intio.or.jp/jf10zl/basic.htm


--- In BITX20@..., "ajparent1" <kb1gmx@a...> wrote:
The 2SC1815 is a good device but it's Ft is only 80mhz. For
HF VFOs that is adaquate but for the BITx if it may be a bit
weak. Other devices of the era are the 2sc784, 2sc930 and 2sc945
all good and easily found in salvaged equipment.

As to transistor capacitance in VFOs, it's is easy to bury the
device in configurations that overwhelm the devices capacitance.
By doing so you make the VFO less dependent on the device used and
it can aid stability. The series tuned Colpits VFO used in
orginal BIT-x is such an oscillator. The 500pf across Base to
Emitter is far larger tha the transistors internal capacitance.
The second 500pf capacitor from emitter to ground (collector is
at RF ground due to the .1uf from Collector to ground) swamps any
capacitance from Collector to Emitter.

One trick that I've used for VFOs that helps with stability is
a form of compensation. For most BJT VFOs there is a voltage
where the initial drift is neutral or nearly so. Often this is
easily found by using a LM317 variable voltage regulator and
starting at the design voltage and noting startup drift. Then by
adjustment of the voltage to the VCO it's possible to reduce or
even eliminate that drift. What it achieves is to find a point
where the transistors internal heating due to feedback and junction
heating is minimal or at least stable.

NOTE 1: This is not temperature compensation. There will still
be drift from warming or cooling of other components. The
inductor, capacitors and if used Varactor(Varicap) diodes are
still contributors to thermally drift. The Varactor diode is a
significant source of thermal drift in most VFOs.

NOTE 2: from testing I've found three terminal regulators
provide a more stable voltage than zener diodes due to temerature.


Allison
KB1GMX

--- In BITX20@..., "Max" <m_orwell@y...> wrote:
2sc1815 is toshiba pct process based transistor. so this mean
lower
capacitance and noise figure. this japanese design remind on
comon
base vfo used in the 70's in swan trx. schematics are in file
section.

73 de 4N1ZM


Re: 2sc1815 interesting stable bjt vfo schematic

Arv Evans K7HKL <arvevans@...>
 

Hi

The 2SC1815 is the 2N2222 of transistors in Japan. It is even more plentiful and readily available there than the ubiquitous 2N2222 is in other countries. While they are usually of quite high quality, I suspect the main reason that Sunamura JF1OZL uses so many of them is that they are plentiful and relatively inexpensive.

Arv
_._

On Thu, 25 Aug 2005 19:06:45 -0600, Max <m_orwell@...> wrote:

2SC1815
--
Life is too short to be serious. 8-)


Re: 2sc1815 interesting stable bjt vfo schematic

ajparent1 <kb1gmx@...>
 

The 2SC1815 is a good device but it's Ft is only 80mhz. For
HF VFOs that is adaquate but for the BITx if it may be a bit
weak. Other devices of the era are the 2sc784, 2sc930 and 2sc945
all good and easily found in salvaged equipment.

As to transistor capacitance in VFOs, it's is easy to bury the
device in configurations that overwhelm the devices capacitance.
By doing so you make the VFO less dependent on the device used and
it can aid stability. The series tuned Colpits VFO used in
orginal BIT-x is such an oscillator. The 500pf across Base to
Emitter is far larger tha the transistors internal capacitance.
The second 500pf capacitor from emitter to ground (collector is
at RF ground due to the .1uf from Collector to ground) swamps any
capacitance from Collector to Emitter.

One trick that I've used for VFOs that helps with stability is
a form of compensation. For most BJT VFOs there is a voltage
where the initial drift is neutral or nearly so. Often this is
easily found by using a LM317 variable voltage regulator and
starting at the design voltage and noting startup drift. Then by
adjustment of the voltage to the VCO it's possible to reduce or
even eliminate that drift. What it achieves is to find a point
where the transistors internal heating due to feedback and junction
heating is minimal or at least stable.

NOTE 1: This is not temperature compensation. There will still
be drift from warming or cooling of other components. The
inductor, capacitors and if used Varactor(Varicap) diodes are
still contributors to thermally drift. The Varactor diode is a
significant source of thermal drift in most VFOs.

NOTE 2: from testing I've found three terminal regulators
provide a more stable voltage than zener diodes due to temerature.


Allison
KB1GMX

--- In BITX20@..., "Max" <m_orwell@y...> wrote:
2sc1815 is toshiba pct process based transistor. so this mean lower
capacitance and noise figure. this japanese design remind on comon
base vfo used in the 70's in swan trx. schematics are in file section.

73 de 4N1ZM


Re: 6M power chain working

ajparent1 <kb1gmx@...>
 

--- In BITX20@..., Rahul Srivastava <vu3wjm@y...> wrote:
Hi!

A DG MOSFET would certainly be an improvement. I had been thinking
on only BJT lines.

A good BJT can be excellent but most BJT designs are aimed at
high IP amplifiers. I've built one radio that used them MRF5xx
series in a norton push pull config with IP in the +35dbm range.
Nice but 100mA for the RF amp is high power drain for a portable.

At 6m a cascode connected pair of MPF102 jfets will work very well
with a good noise figure. I'm testig that to see if I'll use it.
The appeal is common jfets, easy to find, easy substitution.

Rather than purchase them from mkt I normaly salvage them from old
defuct TV tuners. The SMD BF989 or the 2SK series works well in most
cases.

My favortie source. Also old VCRs are a good source for balun cores
and PIN diodes.

Allison
KB1GMX


2sc1815 interesting stable bjt vfo schematic

Max
 

2sc1815 is toshiba pct process based transistor. so this mean lower
capacitance and noise figure. this japanese design remind on comon
base vfo used in the 70's in swan trx. schematics are in file section.

73 de 4N1ZM


Re: 6M power chain working

Rahul Srivastava
 

Hi!
 
A DG MOSFET would certainly be an improvement. I had been thinking on only BJT lines.
 
Rather than purchase them from mkt I normaly salvage them from old defuct TV tuners. The SMD BF989 or the  2SK series works well in most cases.
 
73
 
Rahul VU3WJM

at a Norton amp as they are quieter for the same BJT
though they tend to run lower gain.  I can get a U310 common gate
fet to a below 1db noise figure and 12db gain with less effort.
However I'd need two stages to overcome the DBM and image filter
losses. Based on earlier work with DBMs at 6m if I want a good NF
and MDS a dual gate MOSFET with around 20db gain is successful. 
The idea is enough RF gain to overcome the image filter and DBM
losses (typically -10 to -12DB) with a little extra. The only
problem is that is a more difficult amplifier due to high gain.
The positive is the IP for a dual gate mosfet is very good for the
power required.

>


To help you stay safe and secure online, we've developed the all new Yahoo! Security Centre.


Re: question re performance of bidirectional amplifier (corrected)

ajparent1 <kb1gmx@...>
 

HI,

Catch up how? I have two 20m SSB home brew radios one being BITx
and the BITx for 6M makes the fifth homebrew for that band alone.
I can add a few tube design I've done as well for HF. I enjoy
building and trying new ideas.

Generally I find Ashar's design a very good one and remarkably
reproducable. It represents a well executed minimalist design
that doesn't require exotic components or as some of us call
"unobtainium", those rare parts that only few can find. When
built with good quality parts and a few minor mods the performance
at 20m is excellent. That and it assembled in a few nights as
ugly form with minimal startup problems is impressive.

It's that reason I built the first BIX as 20M just to see it work.
The second for 6M is an experiment and also because I want a small
transceiver of modest power for portable ops. Besides building for
VHF is interesting for it's own sake.

Then I get into the wish list mods..


Allison
KB1GMX

--- In BITX20@..., "eternalesquire" <eternalsquire@c...>
wrote:
Thanks, Ashar!

I think that Allison might need to catch up to you, too.

The Eternal Squire


Re: 6M power chain working

ajparent1 <kb1gmx@...>
 

Hello,

--- In BITX20@..., Rahul Srivastava <vu3wjm@y...> wrote:
Hi!

As far as BITX design is concerned it is well done. The IF gain
issue arises only due to varying characteristics of same devices
available worldwide.

Even within the USA there are poor spec parts problems and
availability issues.

Where noise figure is concerned under ideal components there is a
limitation to what we can get in a negative feedback config. BITX
nature and paralleling of the amps further effects the feedback and
consequently affects the noise figure on the whole.

Very true, for the IF stages the is not a big problem unless poor
devices are used. The Q1/Q13 stage has more than adaquate noise
figure for 20m or lower HF bands. For 15/17/12/10M bands I'd
want a better device than the 2n3904/2n2222 in the recieve path
a 2sc2570 or 2n5179 would do well there. At VHF we want better
as the band noise is much lower.

Another device that needs some attention is mixer driver Q7. At
above 25mhz the 2n3904 is getting near it's limits though a
2n4124 worked well at 41mhz.

Where BJT preamps are concerned a grounded base cofig provides
better performance though at a lesser gain. A Norton noiseless preamp
provides very low NF probably the best in BJT. Even the Icom IC7800
uses it.

Grounded base BJT is stable but low gain and generally requires
a propper input match for good noise performance. I've looked
at a Norton amp as they are quieter for the same BJT
though they tend to run lower gain. I can get a U310 common gate
fet to a below 1db noise figure and 12db gain with less effort.
However I'd need two stages to overcome the DBM and image filter
losses. Based on earlier work with DBMs at 6m if I want a good NF
and MDS a dual gate MOSFET with around 20db gain is successful.
The idea is enough RF gain to overcome the image filter and DBM
losses (typically -10 to -12DB) with a little extra. The only
problem is that is a more difficult amplifier due to high gain.
The positive is the IP for a dual gate mosfet is very good for the
power required.

Maybe we ought to look at the unique switching arrangement used in
RU design Desna already under files for improved NF.

I am considering splitting the path at the mixer with a optimized RX
path to mirror the TX. It's a trivial thing to use a diode switch
at that point. Building a second filter for the RX is trivial.

Other considerations is balancing power consumption as this is for
portable ops and making fit the intended package. That and keeping
it simple to reproduce at least in part. The PLL I use may be a
parts procurement problem for some.


Allison


Re: 6M power chain working

Rahul Srivastava
 

Hi!
 
As far as BITX design is concerned it is well done. The IF gain issue arises only due to varying characteristics of same devices available worldwide.
 
Where noise figure is concerned under ideal components there is a limitation to what we can get in a negative feedback config. BITX nature and paralleling of the amps further effects the feedback and consequently affects the noise figure on the whole.
 
Where BJT preamps are concerned a grounded base cofig provides better performance though at a lesser gain. A Norton noiseless preamp provides very low NF probably the best in BJT. Even the Icom IC7800 uses it.
 
Maybe we ought to look at the unique  switching arrangement used in RU design Desna  already under files for improved NF.
 
Just my suggestions.
 
73
 
Rahul VU3WJM
 
PS: BTW , BITX and ARRL PP mosfet amp. is becoming big hit here in N India.
 

ajparent1 wrote:
This morning I tested the 6m power chain.  Results are very good
with more power than expected.  When integrated on the bench
a two tone SSB signal nets 4.5W.  I made the first local contact
after connecting to the antenna. Reports were good as expected as
I first tested with a local reciever before putting it on the air.
A check on the spectrum analyser says harmonics and spurs at least
-50dbc. That will require a bit of minor work to the output low pass
filter get it in the range I consider clean.  Also I think a bit of
ALC for TX would help. The ciruit for that would be rectify the RF and
use that voltage to throttle the gain of the microphone amplifier.
This will help prevent overdrive.

Reciever proved to be a bit weak during testing. The problem was
not an IF gain issue.  The 6m band requires a lower noise front
end than HF.  The changes were using a 2n5179 for Q1 initially.
I also tried a MAR-6 MIMIC (20dbgain 3db noise figure), it did
not improve matters and verified the bandpass filter loss
was significant.  Later I added a U310 JFET preamp before the
bandpass filter.  The loss through the filter (at 6m) is great
enough that the noise figure of the reciever is compromized without
12db of gain before it.  Measured MDS was -130dbm with the preamp
without it it was a very weak -118Dbm. For 6M operation _my_ goal is
-137Dbm or better.  Either way this will recieve some work as there
is to much wide band gain before the crystal filter.  This is
important on 6M because of the sometimes very strong signals that
appear during band openings.  I also have commercial television
broacast less than 10 miles away on US CH-2 (56mhz) so intermodulation
performance is important. A  better layout would be preselector
filter, RF amp, image filter, then mixer.  The preselector filter
would be low loss and the image reject filter would be the higher
order bandpass.  The revised topology would permit applying AGC
to the RF amp if a dual gate MOSFET such as 2sk122 or BFR988 is
used. I am also considering cascode MPF102 JFET.

Experiments with the mixer indicate the two transformer doubly
balanced works better at 6m and substituting a commercial DBM
(Minicircuits TUF-1) gave similar performance. Transformers
were wound using two hole ferrite balun cores using 6 turns of
#32 wire x3 (trifilar). Diodes used were 1n4148 type matched
for same forward voltage at 1mA.

This makes the second operational BITx. I need to fine tune the
design for best performance on 6M and package it so it's portable.
I tend to be fussy about reciever performance and transmitted signal
quality but so far this version is behaving like the 20m version
allowing for the differences needed for 6M operation.


Allison
KB1GMX



Send instant messages to your online friends http://uk.messenger.yahoo.com


6M power chain working

ajparent1 <kb1gmx@...>
 

This morning I tested the 6m power chain. Results are very good
with more power than expected. When integrated on the bench
a two tone SSB signal nets 4.5W. I made the first local contact
after connecting to the antenna. Reports were good as expected as
I first tested with a local reciever before putting it on the air.
A check on the spectrum analyser says harmonics and spurs at least
-50dbc. That will require a bit of minor work to the output low pass
filter get it in the range I consider clean. Also I think a bit of
ALC for TX would help. The ciruit for that would be rectify the RF and
use that voltage to throttle the gain of the microphone amplifier.
This will help prevent overdrive.

Reciever proved to be a bit weak during testing. The problem was
not an IF gain issue. The 6m band requires a lower noise front
end than HF. The changes were using a 2n5179 for Q1 initially.
I also tried a MAR-6 MIMIC (20dbgain 3db noise figure), it did
not improve matters and verified the bandpass filter loss
was significant. Later I added a U310 JFET preamp before the
bandpass filter. The loss through the filter (at 6m) is great
enough that the noise figure of the reciever is compromized without
12db of gain before it. Measured MDS was -130dbm with the preamp
without it it was a very weak -118Dbm. For 6M operation _my_ goal is
-137Dbm or better. Either way this will recieve some work as there
is to much wide band gain before the crystal filter. This is
important on 6M because of the sometimes very strong signals that
appear during band openings. I also have commercial television
broacast less than 10 miles away on US CH-2 (56mhz) so intermodulation
performance is important. A better layout would be preselector
filter, RF amp, image filter, then mixer. The preselector filter
would be low loss and the image reject filter would be the higher
order bandpass. The revised topology would permit applying AGC
to the RF amp if a dual gate MOSFET such as 2sk122 or BFR988 is
used. I am also considering cascode MPF102 JFET.

Experiments with the mixer indicate the two transformer doubly
balanced works better at 6m and substituting a commercial DBM
(Minicircuits TUF-1) gave similar performance. Transformers
were wound using two hole ferrite balun cores using 6 turns of
#32 wire x3 (trifilar). Diodes used were 1n4148 type matched
for same forward voltage at 1mA.

This makes the second operational BITx. I need to fine tune the
design for best performance on 6M and package it so it's portable.
I tend to be fussy about reciever performance and transmitted signal
quality but so far this version is behaving like the 20m version
allowing for the differences needed for 6M operation.


Allison
KB1GMX


Re: question re performance of bidirectional amplifier (corrected)

eternalesquire <eternalsquire@...>
 

Thanks, Ashar!

I think that Allison might need to catch up to you, too.

The Eternal Squire


Re: Bidirectional amplifier, DC measurements.

ajparent1 <kb1gmx@...>
 

--- In BITX20@..., "eternalesquire" <eternalsquire@c...>
wrote:
Allison,

Here are my AC and DC measurements for the Q3 amplifier circuit. I
had used a scope to measure rather than my voltmeter, so measurements
are eyeballed.
Not useful.

With R set to 12V from an external power supply.
Anode of diode is 9 V
Cathode of diode/collector of Q3 = 8 V
Base = 3 V
Emitter = 3 V (probably a tad less due to BE voltage drop)
If you'd used a accurate voltmeter or DVM you would see more
like 2.7V emitter and 3.4V on the base and that small difference
can be telling.


From BFO through an attenuator, I am applying 10.7 Mhz signal of
5mV pp seen by scope injected at junction of input capacitors

25 mV pp same frequency seen at junction of output capacitors.
25/5=5, way low. for 5MV in I'd expect 100mV or more.

I get no amplification whatever when power is removed or reversed,
which is what we want.
What happens if you reverse the signal (TX direction) same?


Except for the AC gain, this appears to be a properly DC biased small
signal transistor amplifier.

On the other sets of amplifier circuits, I see 50 mV pp at junction of
output capacitors.

Could it be that I am using the wrong techniques to view the signal
with my scope, rather than there being a bug in the circuit? I'm
halfway tempted to try putting this circuit in LTSpice as a sanity
check.

I do question scope calibration from range to range and at that frequency.

Me too. I'm building my second and it's mostly a straight up deal.
Sure I have made mods and used ferrite loaded rather than washers
for the transformers but those are not going to affect the amps.
I may add that I've built a large number of these style (RC feedback)
amps and they work or are broken. the latter is often due to bad part
or the occasional use of 100ohm where I meant to pull 10ohm from the
parts bin.

Just for reference I haywaired that stage on the desk to try it
standalone. Used 2n3904s as they were handy, MSP2222 and 2N2222
have roughly the same DC beta so the biasing will be similar.
With Q3/Q11 biasing (470ohm emitter reistor) at 10mhz I got a
voltage gain of just under 7.2 ( IE: 5mV in and 36 out). Voltages
from ground were emitter 2.4V, Base 3.1V, Collector 10.0V and other
side of diode 10.7V. I was using nominal 13.8V supply for this. When
I changed the biasing to Q1/Q13 (220 ohm) I got a bit more gain 8.4
as measured. Gain fall off at lower current is not unusual and the
overall design of BITx is to not use a lot of RF gain as it favors
stability and resistance to overload. The later is important as
theres no AGC in the base design so it must handle big signals
without distortion. So all in all those MPS2222s are usable but
maybe not the best choice.


I used a 50 ohm loaded scope probe and precision 50ohm step
attenuator for the tests. So the gains are likely lower than
in circuit as a result of the 50ohm loading. By calculation
and eyeball the numner I got numbers look right for that
loading at 10mhz.


Allison
Kb1GMX


Re: question re performance of bidirectional amplifier (corrected)

ajparent1 <kb1gmx@...>
 

Hi,

I've found the MSP2222, PN2222, 2n2222A to be interchangeable
assuming good parts. However, depending on the source and how
they got in to the product stream they can be out of spec devices.
A good 2N2222a has an Ft of not less than 350mhz and good gain
at 5ma Ic. The 2n3904 however is a little quieter and slightly
better gain even though the Ft is the same. A 2n5179
or other VHF device will not achieve much more gain as the
amplifiers are of RC feedback type and component values set
the gain not the transistor used unless the transistor is
poor quality or simply not reasonable for the frequency
in use.

I have some bad commercial experience with parts that just were
not up to spec. I had this once back around '78 with MPS1764s,
bought a bag of them from Motorola and tested them with 100%
yeild to spec. Purchasing bought another 1000 from a cheaper
source and the failure rate was 25% to spec. Showed Motorola
and they wanted to know where they came from. Seems they were
purchased as known offspec for a non critical app the excess
were being resold as full spec. I've been similary burned with
2n3866s and 2n5109s (low Ft and beta), 2n5179s (noisy and low Ft)
to name a few.

So when talking parts like transitors simple DC tests do not tell
all. Measured in an RF test circuit it's easy to spot.


Allison
KB1GMX

--- In BITX20@..., Arv Evans <arvevans@e...> wrote:
Hi

On my first BITX (a BITX20) I used PN2222s and also found the gain
to be quite
low. After substituting a handfull of 2N3904s my gain became much
better.

Another early BITX builder found that his VFO was unstable, until he
replaced
a 2N2222 with a 2N3904.

There have been recent suggestions that VHF transistors be used in
the IF & RF
sections of the BITX. There seems to be some merit in that idea.

It seems that some providers of 2N2222 & PN2222 maybe assume that any
transistor that is NPN and not something else, is a 2N2222. My
2N2222s that
were purchased from DigiKey were all good in the BITX but those that
were
purchased from another source (100 each 2N2222s for $1.50 USD) show
very poor
HF gain, but they are just fine for audio applications.

Arv K7HKL
_._


On Sunday 21 August 2005 21:28, eternalesquire wrote:
1) Checked. These are 10 ohm resistors. I've made 2 stages so far
and had double checked all values of resistors going in.

2) I had been using 0.1 microfarad (value code 104) for all of the
capacitors of these units.

3) On a hunch that this was a bad type of transistor, I substituted a
2N2222A transistor. Gain was acutally worse, X6 rather than X10,
so I
put it back.

I strongly suspect that whatever mistake I am making, I am making it
consistently.

Now, the original schematic specifies 220 Ohm emitter bias resistors
in parallel with the emitter degeneration leg, for 2 of these units.
A 440 Ohm resistor is specified for the unit after the crystal
filter.

Were the original values wrong?

Should I make DC voltage readings and show them to you?

Thanks,

The Eternal Squire

--- In BITX20@..., "ajparent1" <kb1gmx@a...> wrote:
> Edited. inline comments.
>
> --- In BITX20@..., "eternalesquire"
<eternalsquire@c...>
>
> wrote:
> > I wired up a second stage and checked it against the
schematic as I
> > was doing it, every trace is accounted for.
>
> Another thought..
>
> One possible is that the 10 ohm emitter resistor (in series with
> capacitor) is not 10 ohms! It is sometimes easy to get 100ohm
> and 10 ohm by error. Rough calculation suggests that might be the

error.

> Another source of error is if any of the capacitors are not
> the correct value (too small).
>
> I'd also check the transistors used. Is the gain going TX and
RX the
> same? It could be possible one transistor is shorted collector to
> base internally.
>
> Just thinking out loud.
>
> Allison
> KB1GMX

YAHOO! GROUPS LINKS


Visit your group "BITX20" on the web.

To unsubscribe from this group, send an email to:
BITX20-unsubscribe@...

Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service.


Re: question re performance of bidirectional amplifier (corrected)

ajparent1 <kb1gmx@...>
 

--- In BITX20@..., "eternalesquire" <eternalsquire@c...>
wrote:
1) Checked. These are 10 ohm resistors. I've made 2 stages so far
and had double checked all values of resistors going in.
Ok, there must be something...

2) I had been using 0.1 microfarad (value code 104) for all of the
capacitors of these units.
Ok,

3) On a hunch that this was a bad type of transistor, I substituted
a 2N2222A transistor. Gain was acutally worse, X6 rather than X10,
so I put it back.
Should have either made no difference or better. Worse is totally
unespected!

I strongly suspect that whatever mistake I am making, I am making it
consistently.

Now, the original schematic specifies 220 Ohm emitter bias resistors
in parallel with the emitter degeneration leg, for 2 of these
units. A 440 Ohm resistor is specified for the unit after the
crystal filter.

When you say 440 did you really mean 470? Thats what was on the
original schematics.

Were the original values wrong?
No they are well chosen.

Should I make DC voltage readings and show them to you?
Ok, just the preselector stage. Use the same component call outs
(IE: Q1 and Q13) as I will use the publish drawings.

Allison
KB1GMX


Re: question re performance of bidirectional amplifier (corrected)

Arv Evans <arvevans@...>
 

Hi

On my first BITX (a BITX20) I used PN2222s and also found the gain to be quite
low. After substituting a handfull of 2N3904s my gain became much better.

Another early BITX builder found that his VFO was unstable, until he replaced
a 2N2222 with a 2N3904.

There have been recent suggestions that VHF transistors be used in the IF & RF
sections of the BITX. There seems to be some merit in that idea.

It seems that some providers of 2N2222 & PN2222 maybe assume that any
transistor that is NPN and not something else, is a 2N2222. My 2N2222s that
were purchased from DigiKey were all good in the BITX but those that were
purchased from another source (100 each 2N2222s for $1.50 USD) show very poor
HF gain, but they are just fine for audio applications.

Arv K7HKL
_._

On Sunday 21 August 2005 21:28, eternalesquire wrote:
1) Checked.  These are 10 ohm resistors.   I've made 2 stages so far
and had double checked all values of resistors going in. 

2) I had been using 0.1 microfarad (value code 104) for all of the
capacitors of these units. 

3) On a hunch that this was a bad type of transistor, I substituted a
2N2222A transistor.  Gain was acutally worse, X6 rather than X10, so I
put it back.

I strongly suspect that whatever mistake I am making, I am making it
consistently.  

Now, the original schematic specifies 220 Ohm emitter bias resistors
in  parallel with the emitter degeneration leg, for 2 of these units.
A 440 Ohm resistor is specified for the unit after the crystal filter.

Were the original values wrong?

Should I make DC voltage readings and show them to you?

Thanks,

The Eternal Squire

--- In BITX20@..., "ajparent1" <kb1gmx@a...> wrote:
> Edited. inline comments.
>
> --- In BITX20@..., "eternalesquire" <eternalsquire@c...>
>
> wrote:
> > I wired up a second stage and checked it against the schematic as I
> > was doing it, every trace is accounted for.
>
> Another thought..
>
> One possible is that the 10 ohm emitter resistor (in series with
> capacitor) is not 10 ohms!  It is sometimes easy to get 100ohm
> and 10 ohm by error.  Rough calculation suggests that might be the

error.

> Another source of error is if any of the capacitors are not
> the correct value (too small).
>
> I'd also check the transistors used.  Is the gain going TX and RX the
> same?  It could be possible one transistor is shorted collector to
> base internally.
>
> Just thinking out loud.
>
> Allison
> KB1GMX

YAHOO! GROUPS LINKS


 Visit your group "BITX20" on the web.
 
 To unsubscribe from this group, send an email to:
 BITX20-unsubscribe@...
 
 Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service.


Re: question re performance of bidirectional amplifier (corrected)

Ashhar Farhan <farhan@...>
 

I dont see the problem at all.
If you inject a 3mv signal, and you get a 20mv signal, the circuit is working just fine.

here is how, you need to check the power gain, not the voltage gain.

Now, a voltage gain of 6.6 translates to 6.6 x 6.6 = 44.4 power gain. (assuming that the input and output impedances are the same). That is smack down 16db as promised.

- farhan

On Mon, 22 Aug 2005, eternalesquire wrote:

1) Checked. These are 10 ohm resistors. I've made 2 stages so far
and had double checked all values of resistors going in.

2) I had been using 0.1 microfarad (value code 104) for all of the
capacitors of these units.

3) On a hunch that this was a bad type of transistor, I substituted a
2N2222A transistor. Gain was acutally worse, X6 rather than X10, so I
put it back.

I strongly suspect that whatever mistake I am making, I am making it
consistently.

Now, the original schematic specifies 220 Ohm emitter bias resistors
in parallel with the emitter degeneration leg, for 2 of these units.
A 440 Ohm resistor is specified for the unit after the crystal filter.

Were the original values wrong?

Should I make DC voltage readings and show them to you?

Thanks,

The Eternal Squire



--- In BITX20@..., "ajparent1" <kb1gmx@a...> wrote:
Edited. inline comments.

--- In BITX20@..., "eternalesquire" <eternalsquire@c...>
wrote:
I wired up a second stage and checked it against the schematic as I
was doing it, every trace is accounted for.
Another thought..

One possible is that the 10 ohm emitter resistor (in series with
capacitor) is not 10 ohms! It is sometimes easy to get 100ohm
and 10 ohm by error. Rough calculation suggests that might be the
error.

Another source of error is if any of the capacitors are not
the correct value (too small).

I'd also check the transistors used. Is the gain going TX and RX the
same? It could be possible one transistor is shorted collector to
base internally.

Just thinking out loud.

Allison
KB1GMX





Yahoo! Groups Links






Re: Bidirectional amplifier, DC measurements.

Rahul Srivastava
 

Hi! 
 
Just keep in mind 2N2222 metal and PN2222 plastic have different lead configration.
 
Its C B E  L to R for metal where as it is E B C     L to R for plastic when viewed from front. Leads going down.
 
73
 
Rahul VU3WJM
 


eternalesquire wrote:
Allison,

Here are my AC and DC measurements for the Q3 amplifier circuit.  I
had used a scope to measure rather than my voltmeter, so measurements
are eyeballed.

With R set to 12V from an external power supply.
Anode of diode is 9 V
Cathode of diode/collector of Q3 = 8 V
Base = 3 V
Emitter = 3 V (probably a tad less due to BE voltage drop)

From BFO through an attenuator, I am applying 10.7 Mhz signal of
5mV pp seen by scope injected at junction of input capacitors

25 mV pp same frequency seen at junction of output capacitors.

I get no amplification whatever when power is removed or reversed,
which is what we want.

Except for the AC gain, this appears to be a properly DC biased small
signal transistor amplifier.

On the other sets of amplifier circuits, I see 50 mV pp at junction of
output capacitors.

Could it be that I am using the wrong techniques to view the signal
with my scope, rather than there being a bug in the circuit?  I'm
halfway tempted to try putting this circuit in LTSpice as a sanity check.

Scratching my head in confusion,

The Eternal Squire




Yahoo! Messenger NEW - crystal clear PC to PC calling worldwide with voicemail


Bidirectional amplifier, DC measurements.

eternalesquire <eternalsquire@...>
 

Allison,

Here are my AC and DC measurements for the Q3 amplifier circuit. I
had used a scope to measure rather than my voltmeter, so measurements
are eyeballed.

With R set to 12V from an external power supply.
Anode of diode is 9 V
Cathode of diode/collector of Q3 = 8 V
Base = 3 V
Emitter = 3 V (probably a tad less due to BE voltage drop)

From BFO through an attenuator, I am applying 10.7 Mhz signal of
5mV pp seen by scope injected at junction of input capacitors

25 mV pp same frequency seen at junction of output capacitors.

I get no amplification whatever when power is removed or reversed,
which is what we want.

Except for the AC gain, this appears to be a properly DC biased small
signal transistor amplifier.

On the other sets of amplifier circuits, I see 50 mV pp at junction of
output capacitors.

Could it be that I am using the wrong techniques to view the signal
with my scope, rather than there being a bug in the circuit? I'm
halfway tempted to try putting this circuit in LTSpice as a sanity check.

Scratching my head in confusion,

The Eternal Squire