motorboating in cascaded AF amplifier chain, help!
 

Hi,

I've been trying to build my first superhet but I have been having problems.

What I have been doing is building seperate boards for each of the
stages and debugging them seperately. I worked on the audio stages
first.

My two AF boards a 1) a low level X100 amplifier, and 2) an LM386
based power amplifier connected to speaker by RC network. Power
supply is a 9V battery.

Each of these boards work fine seperately, I have tested them with an
audio signal source and a scope. I used ground plane construction with
each of these.

However, when I try to use capacitive or inductive coupling between
the stages, I get the classic symptoms of motorboating... 10 to 45 hz
out of the loudspeaker.

I have the ground planes of each respective board on a piece of aluminum
foil, and both boards are connected in parallel to the 9V battery.

Advice, anyone?

Thanks,

The Eternal Squire

Hi

On Sun, 24 Oct 2004 14:30:59 -0700, The Eternal Squire
wrote:

Hi,

I've been trying to build my first superhet but I have been having problems.

snip details


However, when I try to use capacitive or inductive coupling between
the stages, I get the classic symptoms of motorboating... 10 to 45 hz
out of the loudspeaker.

I have the ground planes of each respective board on a piece of aluminum
foil, and both boards are connected in parallel to the 9V battery.

Advice, anyone?

Thanks,

The Eternal Squire


Some wild guesses g

- not enough decoupling on the 9v rail of the audio power amp (a
couple of hundred uF's as close as possible to the chip)

- speaker leads running too close to the inputs to the x100 amp

- try running the x100 board at less than 9v (anything from some
resistance and a decent electrolytic in the 9v line up to some
'proper' voltage regulation)

- some kind of an earth loop ? - if the board are connected by coax
try lifting the braid at one end of the cable

Hope these help ....?

Adrian
Suffolk UK

- not enough decoupling on the 9v rail of the audio power amp (a
couple of hundred uF's as close as possible to the chip)

You mean short the power leads of the 386 with 220uf? Also, the
X100 amplifier is an AMD instrumentation amp, should I short those
with a cap as well?


- speaker leads running too close to the inputs to the x100 amp


The chain is in a single line, inputs on the left coming from
attenuated audio source, and about a foot to the right start the
leads to the speaker, thoseleads are only a few inches.

- try running the x100 board at less than 9v (anything from some
resistance and a decent electrolytic in the 9v line up to some
'proper' voltage regulation)


I had read an article today on solving motorboating in tube audio
chains, and the B+ in each of the chains have different voltages. The
final stage had +220 V, the stage just before that +200 V, the stage
before that had +180 V, and resistors were used as the voltage drops.
I do not understand how that solved motorboating.

I mean, what I would do is place an 8V regular between the battery
and my X100 stage, and run the final stage off the 9V battery itself.
But can you tell me why this should work?

- some kind of an earth loop ? - if the board are connected by coax
try lifting the braid at one end of the cable


You mean leave the coax floating at one end? Wouldn't that disable
common-mode signal rejection?

Hope these help ....?

Adrian
Suffolk UK

HI again

On Sun, 24 Oct 2004 15:25:07 -0700, The Eternal Squire
wrote:

- not enough decoupling on the 9v rail of the audio power amp (a
couple of hundred uF's as close as possible to the chip)

You mean short the power leads of the 386 with 220uf? Also, the
X100 amplifier is an AMD instrumentation amp, should I short those
with a cap as well?

Certainly you'd do well to get some capacitance across the 386.
The AMD amp is less critical - but you might like to put something a
bit smaller there as well. Can you put the circuits of the two modules
up on the web somewhere ?



- speaker leads running too close to the inputs to the x100 amp


The chain is in a single line, inputs on the left coming from
attenuated audio source, and about a foot to the right start the
leads to the speaker, thoseleads are only a few inches.

Sounds OK


- try running the x100 board at less than 9v (anything from some
resistance and a decent electrolytic in the 9v line up to some
'proper' voltage regulation)


I had read an article today on solving motorboating in tube audio
chains, and the B+ in each of the chains have different voltages. The
final stage had +220 V, the stage just before that +200 V, the stage
before that had +180 V, and resistors were used as the voltage drops.
I do not understand how that solved motorboating.

I mean, what I would do is place an 8V regular between the battery
and my X100 stage, and run the final stage off the 9V battery itself.
But can you tell me why this should work?

I'm guessing (!) that your problems could be caused this way....

The power amp will pull a reasonably heavy current from the 9v battery
on audio peaks. This will result in the 9v 'dipping'. If the x100 amp
stage isn't decoupled itself then it may see the drop in the 9v rail
as a signal, which it will try to amplify. If it succeeds in doing
this then the power amp will pull a reasonably heavy current ... and
so it goes on. This could be the cuase of your oscillation..?
A big decoupling cap at the power amp will help to keep the 9v rail
constant, and mght cure the problem.


- some kind of an earth loop ? - if the board are connected by coax
try lifting the braid at one end of the cable


You mean leave the coax floating at one end? Wouldn't that disable
common-mode signal rejection?

All depends how the other wiring runs. Try it & see ?

Regards

Adrian
Suffolk UK

I'll certaily try your suggestions and let you know.

Side question: is there a way to design this problem out of
circuits?

For example, would it work if I started with 12V main supply
and placed a 9V regulator on each board? Or would I still get
the same current dipping problem?


Adrian Brentnall wrote:
HI again

On Sun, 24 Oct 2004 15:25:07 -0700, The Eternal Squire
wrote:


- not enough decoupling on the 9v rail of the audio power amp (a
couple of hundred uF's as close as possible to the chip)

You mean short the power leads of the 386 with 220uf? Also, the
X100 amplifier is an AMD instrumentation amp, should I short those
with a cap as well?


Certainly you'd do well to get some capacitance across the 386.
The AMD amp is less critical - but you might like to put something a
bit smaller there as well. Can you put the circuits of the two modules
up on the web somewhere ?


- speaker leads running too close to the inputs to the x100 amp


The chain is in a single line, inputs on the left coming from
attenuated audio source, and about a foot to the right start the
leads to the speaker, thoseleads are only a few inches.


Sounds OK


- try running the x100 board at less than 9v (anything from some
resistance and a decent electrolytic in the 9v line up to some
'proper' voltage regulation)


I had read an article today on solving motorboating in tube audio
chains, and the B+ in each of the chains have different voltages. The
final stage had +220 V, the stage just before that +200 V, the stage
before that had +180 V, and resistors were used as the voltage drops.
I do not understand how that solved motorboating.

I mean, what I would do is place an 8V regular between the battery
and my X100 stage, and run the final stage off the 9V battery itself.
But can you tell me why this should work?


I'm guessing (!) that your problems could be caused this way....

The power amp will pull a reasonably heavy current from the 9v battery
on audio peaks. This will result in the 9v 'dipping'. If the x100 amp
stage isn't decoupled itself then it may see the drop in the 9v rail
as a signal, which it will try to amplify. If it succeeds in doing
this then the power amp will pull a reasonably heavy current ... and
so it goes on. This could be the cuase of your oscillation..?
A big decoupling cap at the power amp will help to keep the 9v rail
constant, and mght cure the problem.


- some kind of an earth loop ? - if the board are connected by coax
try lifting the braid at one end of the cable


You mean leave the coax floating at one end? Wouldn't that disable
common-mode signal rejection?


All depends how the other wiring runs. Try it & see ?

Regards

Adrian
Suffolk UK

Side question: is there a way to design this problem out of
circuits?

E.S.-

I can't say for sure about your circuit, but motorboating in general happens
when peak current in the output stage develops an AC voltage across the power
supply impedance, that is amplified by an earlier stage.

Yes, a separate regulator should fix the problem, but seems like over-kill. If
you look at existing designs, they usually feed the output stage directly from
the power source, and feed each earlier stage via a series resistor with a
parallel capacitor across the previous stage.

Values of the resistors and capacitors depend on how much voltage drop the
earlier stage can tolerate, and the lowest frequency of operation (or
motorboating).

73, Fred, K4DII