Topics

uBitx V5 Audio improvement

Philip
 

Here is a way to improvement the audio on the uBitx V5, connect a 10 micro F cap to the LM386 + to pin1 - to pin 8. solder to C75.
This will give a bit more gain, and bass.

Philip G7JUR

Jerry Gaffke
 

Adding Philip's 10uF cap across the LM386 pins 1 and 8 on the uBitx v5 looks like a very good idea.
This can either replace the stock 1uF cap at C75, or be placed in parallel with it.
This may also be of interest to Bitx40 owners.
I have not actually tried this and compared the audio, just a quick check of the math.

The LM386 datasheet says to expect a gain of 26dB without anything between pins 1 and 8,
and that adding a 10uF cap increases the gain to around 46dB.  
    http://www.ti.com/lit/ds/symlink/lm386.pdf

Here's a more in depth look at fiddling with that LM386 feedback network:
  http://www.arduinotehniq.com/old/lm386/Unleash386_LA3ZA.pdf
Using the first equation there, we compute the LM386 gain after finding the impedance of our cap
at the frequency of interest in parallel with the 1.35k internal resistor that lies between pins 1 and 8. 

The uBitx crystal filter passes audio frequencies between roughly 300hz and 2300hz,
The v5 uBitx has a 1uF cap across pins 1 and 8 of the LM386,
but that is not large enough to give consistent gain across the audio spectrum of interest.

With the stock 1uF cap, I figure the LM386 voltage gain at 2300 hz to be 3.36 times greater than the gain at 300hz.
Using a 10uF cap instead, the LM386 voltage gain at 2300 hz is around 1.07 times greater than the gain at 300hz.

Here's python code for the calculations, uses the complex arithmetic facility that is native to python.
Let me know if you spot any errors here.

import math as m
def av(c, f):
  z15 = 15000.0 + 0j
  z18 = 1350.0 + 0j
  zc = 0 - 1j/(2*m.pi*f*c)
  z18 = z18*zc/(z18+zc)
  return(2*z15/(150+z18))
print "gain with  1uF at 2300Hz:", abs(av( 1e-6,2300))
print "gain with  1uF at  300Hz:", abs(av( 1e-6, 300))
print "gain with 10uF at 2300Hz:", abs(av(10e-6,2300))
print "gain with 10uF at  300Hz:", abs(av(10e-6, 300))
print " 1uF ratio", abs(av( 1e-6, 2300))/abs(av( 1e-6, 300))
print "10uF ratio", abs(av( 10e-6,2300))/abs(av( 10e-6,300))
 
Here's an old discussion from two years ago on how the LM386 works,
we were mostly concerned then with various schemes for muting the audio to avoid Bitx40 PTT pops.
    https://groups.io/g/BITX20/topic/5727291

Jerry, KE7ER



On Mon, Apr 8, 2019 at 02:41 PM, Philip wrote:
Here is a way to improvement the audio on the uBitx V5, connect a 10 micro F cap to the LM386 + to pin1 - to pin 8. solder to C75.
This will give a bit more gain, and bass.

Philip G7JUR

Ashhar Farhan
 

LM386 distortion really increases with increased gain. gain is not sensitivity. a better idea might be to bypass pin 7 to ground with the 10 uf.
either way, keep experimenting. I tried an external TDA2030 audio amp connected directly from the volume control. It can blow your ears if you are using earbuds, but it was really cool for speaker listening.
- f 

On Tue, Apr 9, 2019 at 5:14 AM Jerry Gaffke via Groups.Io <jgaffke=yahoo.com@groups.io> wrote:
Adding Philip's 10uF cap across the LM386 pins 1 and 8 on the uBitx v5 looks like a very good idea.
This can either replace the stock 1uF cap at C75, or be placed in parallel with it.
This may also be of interest to Bitx40 owners.
I have not actually tried this and compared the audio, just a quick check of the math.

The LM386 datasheet says to expect a gain of 26dB without anything between pins 1 and 8,
and that adding a 10uF cap increases the gain to around 46dB.  
    http://www.ti.com/lit/ds/symlink/lm386.pdf

Here's a more in depth look at fiddling with that LM386 feedback network:
  http://www.arduinotehniq.com/old/lm386/Unleash386_LA3ZA.pdf
Using the first equation there, we compute the LM386 gain after finding the impedance of our cap
at the frequency of interest in parallel with the 1.35k internal resistor that lies between pins 1 and 8. 

The uBitx crystal filter passes audio frequencies between roughly 300hz and 2300hz,
The v5 uBitx has a 1uF cap across pins 1 and 8 of the LM386,
but that is not large enough to give consistent gain across the audio spectrum of interest.

With the stock 1uF cap, I figure the LM386 voltage gain at 2300 hz to be 3.36 times greater than the gain at 300hz.
Using a 10uF cap instead, the LM386 voltage gain at 2300 hz is around 1.07 times greater than the gain at 300hz.

Here's python code for the calculations, uses the complex arithmetic facility that is native to python.
Let me know if you spot any errors here.

import math as m
def av(c, f):
  z15 = 15000.0 + 0j
  z18 = 1350.0 + 0j
  zc = 0 - 1j/(2*m.pi*f*c)
  z18 = z18*zc/(z18+zc)
  return(2*z15/(150+z18))
print "gain with  1uF at 2300Hz:", abs(av( 1e-6,2300))
print "gain with  1uF at  300Hz:", abs(av( 1e-6, 300))
print "gain with 10uF at 2300Hz:", abs(av(10e-6,2300))
print "gain with 10uF at  300Hz:", abs(av(10e-6, 300))
print " 1uF ratio", abs(av( 1e-6, 2300))/abs(av( 1e-6, 300))
print "10uF ratio", abs(av( 10e-6,2300))/abs(av( 10e-6,300))
 
Here's an old discussion from two years ago on how the LM386 works,
we were mostly concerned then with various schemes for muting the audio to avoid Bitx40 PTT pops.
    https://groups.io/g/BITX20/topic/5727291

Jerry, KE7ER



On Mon, Apr 8, 2019 at 02:41 PM, Philip wrote:
Here is a way to improvement the audio on the uBitx V5, connect a 10 micro F cap to the LM386 + to pin1 - to pin 8. solder to C75.
This will give a bit more gain, and bass.

Philip G7JUR

Adrian Chadd
 

Hi!

Would you consider adding some pads / 0 ohm links for some experimenting with AGC and other things?

I'm totally cool on cutting up the board and such to do dirty tricks but having some 0 ohm links in strategic places would make tapping in series for transmit and receive really useful. Eg, if there was a 0 ohm link in the RX path (with a ground plane pad nearby) someone could add a SMD socket or just plain coax there for receive filtering.

I only ask because I'm about to cut the IF path up for AGC and the RX input path up for broadcast filtering and a hacky preselector :-P



-adrian


On Mon, 8 Apr 2019 at 20:08, Ashhar Farhan <farhanbox@...> wrote:
LM386 distortion really increases with increased gain. gain is not sensitivity. a better idea might be to bypass pin 7 to ground with the 10 uf.
either way, keep experimenting. I tried an external TDA2030 audio amp connected directly from the volume control. It can blow your ears if you are using earbuds, but it was really cool for speaker listening.
- f 

On Tue, Apr 9, 2019 at 5:14 AM Jerry Gaffke via Groups.Io <jgaffke=yahoo.com@groups.io> wrote:
Adding Philip's 10uF cap across the LM386 pins 1 and 8 on the uBitx v5 looks like a very good idea.
This can either replace the stock 1uF cap at C75, or be placed in parallel with it.
This may also be of interest to Bitx40 owners.
I have not actually tried this and compared the audio, just a quick check of the math.

The LM386 datasheet says to expect a gain of 26dB without anything between pins 1 and 8,
and that adding a 10uF cap increases the gain to around 46dB.  
    http://www.ti.com/lit/ds/symlink/lm386.pdf

Here's a more in depth look at fiddling with that LM386 feedback network:
  http://www.arduinotehniq.com/old/lm386/Unleash386_LA3ZA.pdf
Using the first equation there, we compute the LM386 gain after finding the impedance of our cap
at the frequency of interest in parallel with the 1.35k internal resistor that lies between pins 1 and 8. 

The uBitx crystal filter passes audio frequencies between roughly 300hz and 2300hz,
The v5 uBitx has a 1uF cap across pins 1 and 8 of the LM386,
but that is not large enough to give consistent gain across the audio spectrum of interest.

With the stock 1uF cap, I figure the LM386 voltage gain at 2300 hz to be 3.36 times greater than the gain at 300hz.
Using a 10uF cap instead, the LM386 voltage gain at 2300 hz is around 1.07 times greater than the gain at 300hz.

Here's python code for the calculations, uses the complex arithmetic facility that is native to python.
Let me know if you spot any errors here.

import math as m
def av(c, f):
  z15 = 15000.0 + 0j
  z18 = 1350.0 + 0j
  zc = 0 - 1j/(2*m.pi*f*c)
  z18 = z18*zc/(z18+zc)
  return(2*z15/(150+z18))
print "gain with  1uF at 2300Hz:", abs(av( 1e-6,2300))
print "gain with  1uF at  300Hz:", abs(av( 1e-6, 300))
print "gain with 10uF at 2300Hz:", abs(av(10e-6,2300))
print "gain with 10uF at  300Hz:", abs(av(10e-6, 300))
print " 1uF ratio", abs(av( 1e-6, 2300))/abs(av( 1e-6, 300))
print "10uF ratio", abs(av( 10e-6,2300))/abs(av( 10e-6,300))
 
Here's an old discussion from two years ago on how the LM386 works,
we were mostly concerned then with various schemes for muting the audio to avoid Bitx40 PTT pops.
    https://groups.io/g/BITX20/topic/5727291

Jerry, KE7ER



On Mon, Apr 8, 2019 at 02:41 PM, Philip wrote:
Here is a way to improvement the audio on the uBitx V5, connect a 10 micro F cap to the LM386 + to pin1 - to pin 8. solder to C75.
This will give a bit more gain, and bass.

Philip G7JUR

Jerry Gaffke
 

My math shows about the same gain at 2300 hz for 1uF (voltage gain of 178) as for 10uF (200) caps.
But at 300 hz, the 1uF cap has a voltage gain that's down by over a factor of 3
whereas the 10uF cap has flat gain from 300 to 2300 hz.
I haven't tried it with real parts, and I could have slipped up on the math.
But does seem that the 10uF cap would give a much higher fidelity response.

There might be reasons to not want audio gain that's flat over frequency.
Maybe lows tend to distort more than highs.
Perhaps some voices are more intelligible without all that low end response.
Better lows would mean more 50/60hz supply hum getting into those high gain audio amps.

Jerry, KE7ER


On Mon, Apr 8, 2019 at 08:07 PM, Ashhar Farhan wrote:
LM386 distortion really increases with increased gain. gain is not sensitivity. a better idea might be to bypass pin 7 to ground with the 10 uf.
either way, keep experimenting. I tried an external TDA2030 audio amp connected directly from the volume control. It can blow your ears if you are using earbuds, but it was really cool for speaker listening.
- f 

 

Farhan,

For ear buds you should put a 100 ohms or so in series with the speaker output line. That will prevent blowing ears and ear buds!

Raj

At 09/04/2019, you wrote:
LM386 distortion really increases with increased gain. gain is not sensitivity. a better idea might be to bypass pin 7 to ground with the 10 uf.
either way, keep experimenting. I tried an external TDA2030 audio amp connected directly from the volume control. It can blow your ears if you are using earbuds, but it was really cool for speaker listening.

Jerry Gaffke
 

For LM386 gain that is flat across 300 to 2300 Hz 
and is more than 26dB (no cap) but less than 46dB (10uF cap),
this is done by adding a resistor in series with that 10uF cap.

Ashhar Farhan
 

Exact!


On Tue 9 Apr, 2019, 12:09 PM Jerry Gaffke via Groups.Io, <jgaffke=yahoo.com@groups.io> wrote:
For LM386 gain that is flat across 300 to 2300 Hz 
and is more than 26dB (no cap) but less than 46dB (10uF cap),
this is done by adding a resistor in series with that 10uF cap.

Jim Tibbits
 

Attached couple of mods I've been using ...One variable audio filter and one mod Re 386 ...
The 386 shows what I measured on the ic with a 10 uF cap pin 1-8 ...flat from near dc to daylight ..What I heard was LOTS of wideband noise ..Cut it with the installed L/C filter ... then installed the op-amp variable filter ..found I did not need the 386 bandpass mod with the variable filter ...both mounted on same board and plugged into U1 socket on V3 board ...dono what would be needed to adapt to ver 4/5

Jim
LM386 Bandpass Filter.jpeg


On Tue, Apr 9, 2019 at 9:26 PM Ashhar Farhan <farhanbox@...> wrote:
Exact!

On Tue 9 Apr, 2019, 12:09 PM Jerry Gaffke via Groups.Io, <jgaffke=yahoo.com@groups.io> wrote:
For LM386 gain that is flat across 300 to 2300 Hz 
and is more than 26dB (no cap) but less than 46dB (10uF cap),
this is done by adding a resistor in series with that 10uF cap.

Philip
 

I have found that a 470 ohm in series with the 10 uF, seems to work well.
Philip G7JUR

John Owens
 

Philip,
Where does the 470/10uf go across for audio improvement on the uBitx.
Thanks John Owens K6TVY

On 14 Apr 2019, at 2:32 PM, Philip <philip.g7jur@...> wrote:

I have found that a 470 ohm in series with the 10 uF, seems to work well.
Philip G7JUR

Steve MW0KST
 

John, as far as I understand, it's between pins 1 and 8 of the LM386 chip. You'd need to lift C75 and put these in series between the C75 pads.

Ashhar Farhan
 

Yes, with diodes across the ep jack. This is to be done off-board.


On Tue 9 Apr, 2019, 12:04 PM Raj vu2zap, <rajendrakumargg@...> wrote:
Farhan,

For ear buds you should put a 100 ohms or so in series with the speaker output line. That will prevent blowing ears and ear buds!

Raj

At 09/04/2019, you wrote:
>LM386 distortion really increases with increased gain. gain is not sensitivity. a better idea might be to bypass pin 7 to ground with the 10 uf.
>either way, keep experimenting. I tried an external TDA2030 audio amp connected directly from the volume control. It can blow your ears if you are using earbuds, but it was really cool for speaker listening.