Topics

uBitxr4 audio distortion fix

 

Got back to town and plugged in the uBitxr4 and found the audio distorted.

Poking around the finals I found that the two diodes D15 and D16 gave only a drop of 1.2V total
but the final transistors needed more like 1.5V. i.e. 0.75v each.

I simply cut the track between D5 and D16 and soldered a 100 ohms multiturn preset as shown
in picture below. (Preset leads used is center and one side)

Start with the two soldered connections of the preset set to ZERO ohms.

Power on and tune to a station at low volume and slowly increase the preset till the audio
becomes clear. In my board the RX power consumption increased by 25mA at the best sound.
Just check and see that the output transistors are not getting hot.

My board sounded great with 47 Ohms between D15 D16 and I run at 13.6V. If you run at the
same supply voltage then you can try a fixed resistor. For 12V supply you could try a higher value
like 56/62 or even 68 Ohms.

Cheers

--
Raj, vu2zap
Bengaluru, South India.

Emacs!

Ashhar Farhan
 

What if we use three diodes instead? 


On Fri, 29 Jun 2018, 13:35 Raj vu2zap, <rajendrakumargg@...> wrote:
Got back to town and plugged in the uBitxr4 and found the audio distorted.

Poking around the finals I found that the two diodes D15 and D16 gave only a drop of 1.2V total
but the final transistors needed more like 1.5V. i.e. 0.75v each.

I simply cut the track between D5 and D16 and soldered a 100 ohms multiturn preset as shown
in picture below. (Preset leads used is center and one side)

Start with the two soldered connections of the preset set to ZERO ohms.

Power on and tune to a station at low volume and slowly increase the preset till the audio
becomes clear. In my board the RX power consumption increased by 25mA at the best sound.
Just check and see that the output transistors are not getting hot.

My board sounded great with 47 Ohms between D15 D16 and I run at 13.6V. If you run at the
same supply voltage then you can try a fixed resistor. For 12V supply you could try a higher value
like 56/62 or even 68 Ohms.

Cheers

--
Raj, vu2zap
Bengaluru, South India.

 

Farhan,

It was too much! Q72 started to get hot but not Q73. Means they are not matched.
Iq went up almost 100mA. So I used a pot where 47 Ohms seems fine.

What may work that instead of these diodes use the same transistor BE junctions!
Then just two may work, need to experiment on this.

Raj


At 29-06-18, you wrote:
What if we use three diodes instead?Â

On Fri, 29 Jun 2018, 13:35 Raj vu2zap, < rajendrakumargg@...> wrote:
Got back to town and plugged in the uBitxr4 and found the audio distorted.

Poking around the finals I found that the two diodes D15 and D16 gave only a drop of 1.2V total
but the final transistors needed more like 1.5V. i.e. 0.75v each.

I simply cut the track between D5 and D16 and soldered a 100 ohms multiturn preset as shown
in picture below. (Preset leads used is center and one side)

Start with the two soldered connections of the preset set to ZERO ohms.

Power on and tune to a station at low volume and slowly increase the preset till the audio
becomes clear. In my board the RX power consumption increased by 25mA at the best sound.
Just check and see that the output transistors are not getting hot.

My board sounded great with 47 Ohms between D15 D16 and I run at 13.6V. If you run at the
same supply voltage then you can try a fixed resistor. For 12V supply you could try a higher value
like 56/62 or even 68 Ohms.

Cheers

--
Raj, vu2zap
Bengaluru, South India.



Content-Type: image/jpeg; name="121df75.jpg"
Content-Disposition: inline; filename="121df75.jpg"
Content-ID: <>
X-Attachment-Id:

Ashhar Farhan
 

we will probably have to increase the emitter resistors in that case too. The other option would be to use slightly higher bias voltage rectifiers like the 1N4001s.
- f

On Fri, Jun 29, 2018 at 4:17 PM, Raj vu2zap <rajendrakumargg@...> wrote:
Farhan,

It was too much! Q72 started to get hot but not Q73. Means they are not matched.
Iq went up almost 100mA. So I used a pot where 47 Ohms seems fine.

What may work that instead of these diodes use the same transistor BE junctions!
Then just two may work, need to experiment on this.

Raj

At 29-06-18, you wrote:
What if we use three diodes instead?Â

On Fri, 29 Jun 2018, 13:35 Raj vu2zap, < rajendrakumargg@...> wrote:
Got back to town and plugged in the uBitxr4 and found the audio distorted.

Poking around the finals I found that the two diodes D15 and D16 gave only a drop of 1.2V total
but the final transistors needed more like 1.5V. i.e. 0.75v each.

I simply cut the track between D5 and D16 and soldered a 100 ohms multiturn preset as shown
in picture below. (Preset leads used is center and one side)

Start with the two soldered connections of the preset set to ZERO ohms.

Power on and tune to a station at low volume and slowly increase the preset till the audio
becomes clear. In my board the RX power consumption increased by 25mA at the best sound.
Just check and see that the output transistors are not getting hot.

My board sounded great with 47 Ohms between D15 D16 and I run at 13.6V. If you run at the
same supply voltage then you can try a fixed resistor. For 12V supply you could try a higher value
like 56/62 or even 68 Ohms.

Cheers

--
Raj, vu2zap
Bengaluru, South India.

Emacs!

Content-Type: image/jpeg; name="121df75.jpg"
Content-Disposition: inline; filename="121df75.jpg"
Content-ID: <>
X-Attachment-Id:


Gordon Gibby <ggibby@...>
 

Don’t forget those emitter resistors, that Add additional voltage drop.

Someone suggested using a bit of feedback, I didn’t look at the circuits, but wouldn’t that be one solution?

On Jun 29, 2018, at 06:47, Raj vu2zap <rajendrakumargg@...> wrote:

Farhan,

It was too much! Q72 started to get hot but not Q73. Means they are not matched.
Iq went up almost 100mA. So I used a pot where 47 Ohms seems fine.

What may work that instead of these diodes use the same transistor BE junctions!
Then just two may work, need to experiment on this.

Raj

At 29-06-18, you wrote:
What if we use three diodes instead?Â

On Fri, 29 Jun 2018, 13:35 Raj vu2zap, < rajendrakumargg@...> wrote:
Got back to town and plugged in the uBitxr4 and found the audio distorted.

Poking around the finals I found that the two diodes D15 and D16 gave only a drop of 1.2V total
but the final transistors needed more like 1.5V. i.e. 0.75v each.

I simply cut the track between D5 and D16 and soldered a 100 ohms multiturn preset as shown
in picture below. (Preset leads used is center and one side)

Start with the two soldered connections of the preset set to ZERO ohms.

Power on and tune to a station at low volume and slowly increase the preset till the audio
becomes clear. In my board the RX power consumption increased by 25mA at the best sound.
Just check and see that the output transistors are not getting hot.

My board sounded great with 47 Ohms between D15 D16 and I run at 13.6V. If you run at the
same supply voltage then you can try a fixed resistor. For 12V supply you could try a higher value
like 56/62 or even 68 Ohms.

Cheers

--
Raj, vu2zap
Bengaluru, South India.



Content-Type: image/jpeg; name="121df75.jpg"
Content-Disposition: inline; filename="121df75.jpg"
Content-ID: <>
X-Attachment-Id:

Tim Gorman
 

Three diodes won't fix the problem that the bias on the base of Q73 is
highly dependent on Q71 collector voltage.

If you ever redesign this circuit I would use capacitor coupling from
the collector of Q71 to the bases of Q72 and Q73. The base bias
of Q73 would then be set with a resistor to ground. Then everything is
symmetric and distortion should be minimized.

tim ab0wr



On Fri, 29 Jun 2018 15:45:51 +0530
"Ashhar Farhan" <farhanbox@...> wrote:

What if we use three diodes instead?

On Fri, 29 Jun 2018, 13:35 Raj vu2zap, <@Raj>
wrote:

Got back to town and plugged in the uBitxr4 and found the audio
distorted.

Poking around the finals I found that the two diodes D15 and D16
gave only a drop of 1.2V total
but the final transistors needed more like 1.5V. i.e. 0.75v each.

I simply cut the track between D5 and D16 and soldered a 100 ohms
multiturn preset as shown
in picture below. (Preset leads used is center and one side)

Start with the two soldered connections of the preset set to ZERO
ohms.

Power on and tune to a station at low volume and slowly increase the
preset till the audio
becomes clear. In my board the RX power consumption increased by
25mA at the best sound.
Just check and see that the output transistors are not getting hot.

My board sounded great with 47 Ohms between D15 D16 and I run at
13.6V. If you run at the
same supply voltage then you can try a fixed resistor. For 12V
supply you could try a higher value
like 56/62 or even 68 Ohms.

Cheers

--
Raj, vu2zap
Bengaluru, South India.

[image: Emacs!]


ajparent1/KB1GMX <kb1gmx@...>
 

Tim no.

The problem is if the first transistor Q71 is not biased to the half power supply
point the output are both unbalanced and the bias is wrong.  IF Q71 collector
is not at about 6V (for 12V supply) full stop.  Adjust R72 until it is.  Then adjust
the bias if needed to the complementary pair.  

The most common variation is to derive the bias from the common point of both
emitters as then it tries to self center at 6V or half the supply.  The cost of of doing
that is lower gain but higher fidelity.

See the design from this site (attached) and explanation.
http://www.deeptronic.com/electronic-circuit-design/small-audio-power-amplifier-using-3-transistors/

Do not let their published 3V power scare you it works well at 12V too.
The key is how Q1 is biased and the resulting bias to the output string (3904 and 3906).

Allison

It should be possible to change the V4 circuit to that.

ajparent1/KB1GMX <kb1gmx@...>
 

Another that more closely resembles the V4 circuit with improvements.

The two NPN can be 2n3904 and the pnp is 2n3906.

I've used this flavor from 5V though 20V.  At 12V it will put 250mW in to a 8ohm speaker.

Allison

Jack, W8TEE
 

Thanks, Allison...looks interesting.

Jack, W8TEE

On Friday, June 29, 2018, 10:09:15 PM EDT, ajparent1/KB1GMX <kb1gmx@...> wrote:


Another that more closely resembles the V4 circuit with improvements.

The two NPN can be 2n3904 and the pnp is 2n3906.

I've used this flavor from 5V though 20V.  At 12V it will put 250mW in to a 8ohm speaker.

Allison

 

I tried a mod like this and the gain dropped so I split the feedback resistor by 2
and bypassed to ground with a 10uF at the center. Sounds great now.

At 30-06-18, you wrote:
Another that more closely resembles the V4 circuit with improvements.

The two NPN can be 2n3904 and the pnp is 2n3906.

I've used this flavor from 5V though 20V. At 12V it will put 250mW in to a 8ohm speaker.

Allison

Tim Gorman
 

Allison,

This circuit does the same thing I mention, just in a different manner.
This circuit uses feedback to drive the collector of TR1 to the proper
voltage.

My method eliminates the need for the feedback.

Both will work to enhance the fidelity from the amplifier. I have used
my method in several audio amps and had no problem.

tim ab0wr

On Fri, 29 Jun 2018 19:02:34 -0700
"ajparent1/KB1GMX" <kb1gmx@...> wrote:

Tim no.

The problem is if the first transistor Q71 is not biased to the half
power supply point the output are both unbalanced and the bias is
wrong.  IF Q71 collector is not at about 6V (for 12V supply) full
stop.  Adjust R72 until it is.  Then adjust the bias if needed to the
complementary pair.  

The most common variation is to derive the bias from the common point
of both emitters as then it tries to self center at 6V or half the
supply.  The cost of of doing that is lower gain but higher fidelity.

See the design from this site (attached) and explanation.
http://www.deeptronic.com/electronic-circuit-design/small-audio-power-amplifier-using-3-transistors/

Do not let their published 3V power scare you it works well at 12V
too. The key is how Q1 is biased and the resulting bias to the output
string (3904 and 3906).

Allison

It should be possible to change the V4 circuit to that.


Tim Gorman
 

You can also fix this by adding a bias resistor from the base of the
PNP transistor to ground and feeding the audio into the midpoint of the
diodes with an isolating capacitor.

Using Q71 as a current source for Q72 and Q73 will require either
negative feedback or an adjustable resistor in the Q71 base bias or
even both, otherwise differences in the transistors used can cause
problems. You may even need to add a balance resistor between the bases
of Q72 and Q73 to eliminate cross-over distortion.

With my solution these adjustments aren't needed as much.

tim ab0wr

On Fri, 29 Jun 2018 19:02:34 -0700
"ajparent1/KB1GMX" <kb1gmx@...> wrote:

Tim no.

The problem is if the first transistor Q71 is not biased to the half
power supply point the output are both unbalanced and the bias is
wrong.  IF Q71 collector is not at about 6V (for 12V supply) full
stop.  Adjust R72 until it is.  Then adjust the bias if needed to the
complementary pair.  

The most common variation is to derive the bias from the common point
of both emitters as then it tries to self center at 6V or half the
supply.  The cost of of doing that is lower gain but higher fidelity.

See the design from this site (attached) and explanation.
http://www.deeptronic.com/electronic-circuit-design/small-audio-power-amplifier-using-3-transistors/

Do not let their published 3V power scare you it works well at 12V
too. The key is how Q1 is biased and the resulting bias to the output
string (3904 and 3906).

Allison

It should be possible to change the V4 circuit to that.


ajparent1/KB1GMX <kb1gmx@...>
 

Yes it does but insufficient and is very sensitive to the DC beta(HFE) of the particular device.

The two posted do not exhibit this.  I can list a small high of gear that use
similar rather than an IC without issues. 

Allison

Glenn
 

This amp works very well, with very low distortion. http://darrenyates.com.au/electronics/archives/31

I have built this amp to verify, it but note the voltage range up to 12v. I found the final pair can get very hot over 12v.  Possibly some re-biasing of the first stage might help that. Or use a BD139/140 pair in the output stage.  I got around 600mW out into 10R resistive load with 240mVrms input which is the max my audio gen will give.

glenn

lkorahul
 

Diz has one on his site for historical reference. Maybe for topics like these.





73

Rahul VU3WJM

lkorahul
 

 

Tried another uBitxr4 board and had the same problem.

I recommend this mod for everyone. Solder the pot and ZERO it and at low volume with the
band background noise increase the preset and you will notice the volume increasing. Leave
it a little more than the point you get max volume of noise. I am sure that will work for the
best.

Raj


At 29-06-18, you wrote:
Got back to town and plugged in the uBitxr4 and found the audio distorted.

Poking around the finals I found that the two diodes D15 and D16 gave only a drop of 1.2V total
but the final transistors needed more like 1.5V. i.e. 0.75v each.

I simply cut the track between D5 and D16 and soldered a 100 ohms multiturn preset as shown
in picture below. (Preset leads used is center and one side)

Start with the two soldered connections of the preset set to ZERO ohms.

Power on and tune to a station at low volume and slowly increase the preset till the audio
becomes clear. In my board the RX power consumption increased by 25mA at the best sound.
Just check and see that the output transistors are not getting hot.

My board sounded great with 47 Ohms between D15 D16 and I run at 13.6V. If you run at the
same supply voltage then you can try a fixed resistor. For 12V supply you could try a higher value
like 56/62 or even 68 Ohms.

Cheers

--
Raj, vu2zap
Bengaluru, South India.

Emacs!  

Andy V. Borisenko
 

all the same it's a problem of v4 ...
I had to shoot a video and post it on YouTube to indicate the problem.
Raj, please draw on the diagram your mods.

Sudanthiram N
 

Dear Raj,
I  added one 100 ohms preset (not trimmer) in between the diode d15 and D16 (after cut the track) the audio is ok and volume is increased.

Good result.
vu2nsk 

 

Sundar,

It should be best sound around half. Just check if final audio tr's are too hot!

Raj

At 03-11-18, you wrote:
Dear Raj,
I added one 100 ohms preset (not trimmer) in between the diode d15 and D16 (after cut the track) the audio is ok and volume is increased.

Good result.
vu2nsk