Homebrew AM transistor radio
 

I'm attempting to build a small AM broadcast superhet transistor radio
from scratch and having a problem with the three IF stages oscillating
at the IF frequency (455KHz). The IF coils are salvaged from other
radios and the ground connections are to a common bus wire that
connects all the IF coil shields. It works reasonably well around 5
volts, but breaks into oscillation (at the IF frequency) if the supply
voltage is increased 1/2 volt and the signal drops off significantly if
the supply voltage is reduced 1/2 volt. So, it only works within a very
small supply voltage range of around 5 to 5.5 volts. The IF stages are
decoupled from the supply with a small resistor in series and bypass
cap to ground which helps, but doesn't solve the problem. I'm wondering
what can be done to stop oscillations and increase gain?

How did they manage to avoid the oscillation problems in the old tube
radios that were hand wired without any PC board?

-Bill

I don't know what you mean by "small resistor" in series with the V+ line.
Each amplifier stage should have a series resistor of 100 ohms and a bypass
capacitor of .01uF. Actually, you can bypass every other stage if you are
using common emitter or common source stages, but I would do it with every
amplifier stage.
As far as physical distance for each stage, you should allow 1/2 inch for
every 40dB of gain. You shouldn't have this type of problem. A ground plane
is always a good idea, but at the very least you should have quite a bit of
ground flood on the circuit board. Although it is possible if you have quite
a bit of copper foil tape around the board, you can build this circuit on
perfboard but you have to be very careful, as you have already discovered. I
don't know if they have perfboard that has copper on only one side, but this
is one solution.
You had a question about the old hand-wired tubed equipment. With a metal
chassis, you would keep the wiring down close to the chassis itself, which
would in effect provide that ground plane. A ground is very important to
provide a place for the return currents for all of the signals, whether
physical wires or copper traces are used.

Pete

wrote in message
ups.com...
I'm attempting to build a small AM broadcast superhet transistor radio
from scratch and having a problem with the three IF stages oscillating
at the IF frequency (455KHz). The IF coils are salvaged from other
radios and the ground connections are to a common bus wire that
connects all the IF coil shields. It works reasonably well around 5
volts, but breaks into oscillation (at the IF frequency) if the supply
voltage is increased 1/2 volt and the signal drops off significantly if
the supply voltage is reduced 1/2 volt. So, it only works within a very
small supply voltage range of around 5 to 5.5 volts. The IF stages are
decoupled from the supply with a small resistor in series and bypass
cap to ground which helps, but doesn't solve the problem. I'm wondering
what can be done to stop oscillations and increase gain?

How did they manage to avoid the oscillation problems in the old tube
radios that were hand wired without any PC board?

-Bill

I don't know what you mean by "small resistor" in series with the V+ line.
Each amplifier stage should have a series resistor of 100 ohms and a bypass
capacitor of .01uF.

I've tried several values, 220,330,510, ect. with 0.05uF bypass.
Doesn't make much difference.

As far as physical distance for each stage, you should allow 1/2 inch for
every 40dB of gain. You shouldn't have this type of problem. A ground plane
is always a good idea,

The IF cans are spaced 1 inch apart and I disconnected the 2nd IF stage
so I only have an oscillator, mixer and single IF amp stage. Same
problem, it always oscillates when the voltage is raised to obtain
reasonable gain.

Although it is possible if you have quite a bit of copper foil tape around the
board, you can build this circuit on perfboard but you have to be very careful,
as you have already discovered.

I took a look inside a AM/FM clock radio I have and checked the
ground connections. There are 3 RF transformers, oscillator, mixer and
single IF stage, and all of the metal shields of the transformers are
isolated. There are no physical connections between the shields of the
3 transformers.

Obviously, I am missing something. Grounding all the transformer
housings on a massive ground plane does not seem to be the answer.

-Bill

Bill,

Transistor amplifiers with inputs and outputs tuned to the same frequency
turn out to be oscillators, especially if the the tuned frequency is very
low compared to the transistor's Ft. If you're using a modern transistor
with an Ft of say, 800 Mhz, and you're tuning it to say, 500KHz, then you
could have available gain of over 60 dB in one stage. This is a sure-bet
oscillator waiting to happen.

You can lower the gain by 'loading' the input and output tuned circuits with
parallel resistors (try 1K for starters and work up from there if it stops
the oscillation. Just remember that this technique, in addition to lowering
the gain, will also load the tuned circuit and degrade selectivity.

You can also try a small resistor (less than a couple hundred ohms) in
series with the collector lead of the transistor, or in series with the base
lead. This technique will lower the stage gain and actually increase the
impedance seen by the tuned circuits, improving selectivity.

Keep in mind that raising the selectivity too much can cause tuned circuit
losses, just as parallel loading it will lower selectivity and increase
losses.

Joe
W3JDR


wrote in message
oups.com...
I don't know what you mean by "small resistor" in series with the V+
line.
Each amplifier stage should have a series resistor of 100 ohms and a
bypass
capacitor of .01uF.

I've tried several values, 220,330,510, ect. with 0.05uF bypass.
Doesn't make much difference.

As far as physical distance for each stage, you should allow 1/2 inch for
every 40dB of gain. You shouldn't have this type of problem. A ground
plane
is always a good idea,

The IF cans are spaced 1 inch apart and I disconnected the 2nd IF stage
so I only have an oscillator, mixer and single IF amp stage. Same
problem, it always oscillates when the voltage is raised to obtain
reasonable gain.

Although it is possible if you have quite a bit of copper foil tape
around the
board, you can build this circuit on perfboard but you have to be very
careful,
as you have already discovered.

I took a look inside a AM/FM clock radio I have and checked the
ground connections. There are 3 RF transformers, oscillator, mixer and
single IF stage, and all of the metal shields of the transformers are
isolated. There are no physical connections between the shields of the
3 transformers.

Obviously, I am missing something. Grounding all the transformer
housings on a massive ground plane does not seem to be the answer.

-Bill

On 17 Aug 2005 19:13:03 -0700, wrote:

I don't know what you mean by "small resistor" in series with the V+ line.
Each amplifier stage should have a series resistor of 100 ohms and a bypass
capacitor of .01uF.

I've tried several values, 220,330,510, ect. with 0.05uF bypass.
Doesn't make much difference.

As far as physical distance for each stage, you should allow 1/2 inch for
every 40dB of gain. You shouldn't have this type of problem. A ground plane
is always a good idea,

The IF cans are spaced 1 inch apart and I disconnected the 2nd IF stage
so I only have an oscillator, mixer and single IF amp stage. Same
problem, it always oscillates when the voltage is raised to obtain
reasonable gain.

Although it is possible if you have quite a bit of copper foil tape around the
board, you can build this circuit on perfboard but you have to be very careful,
as you have already discovered.

I took a look inside a AM/FM clock radio I have and checked the
ground connections. There are 3 RF transformers, oscillator, mixer and
single IF stage, and all of the metal shields of the transformers are
isolated. There are no physical connections between the shields of the
3 transformers.

Obviously, I am missing something. Grounding all the transformer
housings on a massive ground plane does not seem to be the answer.

-Bill

I will assume good wiring, lead dress and stage bypassing have been
attended to as even small amounts of feedback regardless the cause
will insitgate oscillation.

Transistors have internal capacitance from collector to base and in
high impedence RF amplifiers that serves to couple input to output
and make a fine oscillator. The gain attainable before oscillation
was usually listed as MUG (max usable gain) and is usually specified
with a test circuit. Often that circuit was neutralized (more later)
to allow best gain without oscillation. For a .455khz IF a stage gain
of more than around 35DB from one transistor is begging for
instability.

There are several solutions for that. One is reduce the gain, not
always a good solution but sometimes required. Another is to
neutalize the amplifier. Neutralization is to feed back some of the
signal out of phase back to the base lead to balance out the signal
comming back via the collector base capacitance. Like anything
enough is good too much is bad.

As it turns out most ransistor radios of the 50s up till the do it all
on a chip era had to cope with unintentional feedback. Most applied
both of the mentioned solutions, reduce the per stage gains and apply
neutralization.

As to how to apply neutalization... Varies depending on circuit.
Assuming your using common IF cans and generic transistor circuit
it's usually easy. The output IF can has three terminals on the
tuned side, one is meant for DC power, another is collector and the
remaining is the out of phase feedback to the base via a capacitor
in the range of 1 to 10pf. Which pin is which, depends of the can and
who made it. Another way if the IF can is not tapped is to feed
back some fo the secondary winding to the base. The tricky part here
is if the secondary is backward it will make oscillation worse and you
will have to reverse the seconday IF transformer leads. What you are
shooting for is a stable amplifier.

One of the side effects is that if the applied neutralization is too
small the stage will have some regeneration but not oscillate
and that is also unstable though a very high gain state to be in.
Strive for stability. The stage should not oscillate at any
reasonable voltage or tuning adjustment.

All of the proceeding is predicated on the transistor stage
oscillating at the IF frequency. It is possible due to various
factors that the transistor is oscillation at VHF and a number of
other frequencies as well. This is trickier, as you need to
suppress that and then deal with the IF oscillation if still present.
VHF oscillation can be killed with a 100ohm resistor in series with
the collector lead. The resistor and the collector lead should be
short. There is no measurable gain lost at IF frequencies to that
resistor as it's size is very small compared to the load.


Having built more than few radios with high gain IF stages I've had
to deal with this problem more than once. Hope it helps.

Allison
KB1GMX

A ground plane should make the system more stable. The radios that I have
been designing over the years have gains in the I.F. systems that range from
90 to 120dB, depending on the topology. You are definitely getting signal
feedback through your power supply rails. 100 to 120 ohms is about as high
as you want to go as far as decoupling resistor values. .05uF is too high of
a value for a decoupling cap..............01uF is what you should use for
local decoupling.
You should also have at least one bulk decoupling cap that has a value in
the range of 10 to 47uF. The larger decoupling cap cancels out the
inductance of your power supply wiring while the smaller local decoupling
caps cancel out the inductance of the bulk cap.

Pete

wrote in message
oups.com...
I don't know what you mean by "small resistor" in series with the V+
line.
Each amplifier stage should have a series resistor of 100 ohms and a
bypass
capacitor of .01uF.

I've tried several values, 220,330,510, ect. with 0.05uF bypass.
Doesn't make much difference.

As far as physical distance for each stage, you should allow 1/2 inch for
every 40dB of gain. You shouldn't have this type of problem. A ground
plane
is always a good idea,

The IF cans are spaced 1 inch apart and I disconnected the 2nd IF stage
so I only have an oscillator, mixer and single IF amp stage. Same
problem, it always oscillates when the voltage is raised to obtain
reasonable gain.

Although it is possible if you have quite a bit of copper foil tape
around the
board, you can build this circuit on perfboard but you have to be very
careful,
as you have already discovered.

I took a look inside a AM/FM clock radio I have and checked the
ground connections. There are 3 RF transformers, oscillator, mixer and
single IF stage, and all of the metal shields of the transformers are
isolated. There are no physical connections between the shields of the
3 transformers.

Obviously, I am missing something. Grounding all the transformer
housings on a massive ground plane does not seem to be the answer.

-Bill

Well, I don't have an exact dB gain figure, but I hooked up the mixer
stage by itself and I get about 300 to 400 mV p-p with a generator
connected to the antenna coil using 3 turns of wire around the coil and
a 50 ohm resistor in series. The generator is set to 400mV p-p and I
get about the same amount out which is fairly stable from 9 volt supply
to 6 volts.

The first stage seems to be working right, no oscillations,
and stable output as the supply changes, and reasonable gain.

The transistor is a 2N4123 with a Ft of 250 MHz and
hfe of 50-200. It is biased using a 1k emitter resistor
at 1 volt, or about 1 mA.

I assume this stage works ok and will investigate
number 2.

-Bill

Correction to other post. The output is 3 volts p-p.

I sucessfully hooked up two stages with no oscillation problems. I
added a 560 ohm load resistor across the output winding of the first
stage. The LO oscillator is a separate circuit (JFET) and there are no
physical connections from the the oscillator to the mixer stage. It
seems to mix and produce the IF from just the radiation from the LO
capacitor to the station tuning capacitor.

The IF modulated output is about 2 volts p-p on a strong station, 600
mV on a average station, and only 50mV on weak stations.

It's almost usable, but not quite. I need to raise the weak stations
another 12 dB or so. Also, the LO oscillator is leaking into the output
at the high end of the band. I get 100mV of noise from the LO toward
the high end.

Is it a good idea to try and raise the gain of the 2 existing stages,
or try to add a third stage to raise the gain?

I've noticed some transistor radios use only one IF amplifier stage
while others use 2.

What are the advantages of the third stage if enough gain can be
obtained from just 2?

-Bill

Bill,

Good to hear that you tamed the oscillation, but the 560 ohm load resistor
seems a bit severe...have you tried something in the few kohms range?

50 mv out of the mixer on a 'weak' station sounds pretty good. What is the
combined gain of the two IF stages? Are they possibly overloading ?

Also, how are you coupling the antenna into the mixer?

Joe
W3JDR

wrote in message
oups.com...
I sucessfully hooked up two stages with no oscillation problems. I
added a 560 ohm load resistor across the output winding of the first
stage. The LO oscillator is a separate circuit (JFET) and there are no
physical connections from the the oscillator to the mixer stage. It
seems to mix and produce the IF from just the radiation from the LO
capacitor to the station tuning capacitor.

The IF modulated output is about 2 volts p-p on a strong station, 600
mV on a average station, and only 50mV on weak stations.

It's almost usable, but not quite. I need to raise the weak stations
another 12 dB or so. Also, the LO oscillator is leaking into the output
at the high end of the band. I get 100mV of noise from the LO toward
the high end.

Is it a good idea to try and raise the gain of the 2 existing stages,
or try to add a third stage to raise the gain?

I've noticed some transistor radios use only one IF amplifier stage
while others use 2.

What are the advantages of the third stage if enough gain can be
obtained from just 2?

-Bill

On 19 Aug 2005 22:49:15 -0700, wrote:

I sucessfully hooked up two stages with no oscillation problems. I
added a 560 ohm load resistor across the output winding of the first
stage.

You could load the primary instead, values in the range of 10k to 80k
would be tried.

The LO oscillator is a separate circuit (JFET) and there are no
physical connections from the the oscillator to the mixer stage. It
seems to mix and produce the IF from just the radiation from the LO
capacitor to the station tuning capacitor.

Not uncommon, one circuit copuples to another due to adjacent
components. However, the Mixer may not be seeing the optimum
LO level for best gain or overload characteristics.

The IF modulated output is about 2 volts p-p on a strong station, 600
mV on a average station, and only 50mV on weak stations.

It's almost usable, but not quite. I need to raise the weak stations
another 12 dB or so. Also, the LO oscillator is leaking into the output
at the high end of the band. I get 100mV of noise from the LO toward
the high end.

Any number of factors here. More gain means more stanility problems.

Also, the LO oscillator is leaking into the output
at the high end of the band. I get 100mV of noise from the LO toward
the high end.

The LO could be noisy due to design error. The front end (loopstick
in AM BCB radios) may be tuned better at that end of the band and you
hearing (seeing) more noise. Or you are seeing better coupling of the
LO (increses with frequency) and better mixer gain. The MW range
(300kc to 3mhz) has a very high backgound noise both teresterial and
man made.
..
Is it a good idea to try and raise the gain of the 2 existing stages,
or try to add a third stage to raise the gain?

NO, they are only conditonally stable as is. That load on the one
transformer is both robbing gain and also giving a little stability.

I've noticed some transistor radios use only one IF amplifier stage
while others use 2.

Depends on perfomance. Those with one are pretty tepid in the
sensitivity department.

What are the advantages of the third stage if enough gain can be
obtained from just 2?

Many AM radios apply AGC to one or more stages to "level"
the strong and weak stations for even audio output and to prevent
later states from overloading. More stages allow for more applied AGC
and overall better perfomance.


Allison KB1GMX