¿ªÔÆÌåÓý

  1. BITX20
  2. Topics

Adding speech compression.

 

I have been reading stuff here about low power or audio for about 3 months now because my BITX40 was factory set to low power (drawing .024 mA at the final during wireup), output and also sounded like the audio should be louder. ?R136 was a mystery to everyone but me when I discovered it adjusts the PA current draw and power output. It seems to have been deleted or left out from the alignment steps at the bottom of the wire up page. Look at #5, it's blank. This is still tripping up noobs who just got one today, i'm sure.

There have been some people talking about "speech compressors". And I think it's time I try it. I never heard of it before. And some of what I have seen here including PDF's Seems to assume I already know how to pick it up and use it. I can make circuitry just fine. But it's how to make the two work together, mostly because speech compressors are totally new to me, that's holding me back from starting such a project.

Could some of you fine folks point me and probably some others of the gang, towards some basic info to bring me up to speed on using speech compression, and why I need it? And speech compressors with the BITX40 board? Some links would probably save you a bunch of writing time. But personal perspectives would be greatly appreciated too.?

Thanks in advance.
 

Correction on first line, that's 240mA max current draw on the PA in the beginning.
 

Hi John,


Adding speech compression can make a difference when trying to cut through noise or a heavy pileup, but tread lightly; it's easy to turn your good intentions into over-modulation or worse.? Here's a bit of introductory reading to get you started (be sure to check out the references):
 

Alright, I am up to speed now. It kinda seems like it isn't necessary and could make intelligibility worse and or annoying. I know what overmodulation sounds like, and you can't tell those people anything different. And make the final even hotter than it already is with 25V. ?Has anyone done this with the BITX40? What difference did it make? Would you recommend it?

Hello? (tap, tap, tap) Is this thing on?
 

Hi,

A _PROPERLY_ _ADJUSTED_ speech processor will probably make a world of
improvement in received signal quality, typically adding about +6dB of
audio punch---that's roughly equivalent to turning the 7 watts PEP all the
way up to 28 watts! This really, really helps on QRP rigs!

You are correct that this could make the IFR510 much, much hotter if
you're long-winded, perhaps needing a larger heatsink or other cooling
enhancement.

Here is an interesting audio clipper/compressor on ebay. I don't know how
well it works:

I again stress that PROPER processor level adjustment is the key here.
Setup right, they're a VERY useful addition for voice communications.

Oh, and NEVER EVER run a compressor/clipper with any digital mode!

Just my $0.02.

73, David KB4FXC



On Sun, 2 Apr 2017, John Smith via Groups.Io wrote:

Alright, I am up to speed now. It kinda seems like it isn't necessary
and could make intelligibility worse and or annoying. I know what
overmodulation sounds like, and you can't tell those people anything
different. And make the final even hotter than it already is with 25V.
??Has anyone done this with the BITX40? What difference did it make?
Would you recommend it?

Hello? (tap, tap, tap) Is this thing on?
 

When I think speech compression, I think of pop songs where they artificially make the quiet parts louder in an attempt to keep those with short attention spans from dozing off.
 

Yes, but those pop songs are usually coming across broadcast FM radio,
with maybe a 100KW ERP transmitter and very low noise floor....A lot
different than 7W PEP on 40M with a high and varying noise floor due to
QRM and QRN!




On Sun, 2 Apr 2017, Jerry Gaffke via Groups.Io wrote:

When I think speech compression, I think of pop songs where they
artificially make the quiet parts louder in an attempt to keep those with
short attention spans from dozing off.
 

The Analog Devices SSM2167 should make a great SSB speech compressor. It has a low level Microphone input (or just insert an attenuator network) and is completely adjustable for compression ratio. It is also readily available and inexpensive. It's single supply and runs on 3 to 5 volts. The big caveat is that it is in an MSOP ten pin package with 0.5 mm (~0.02 in) lead spacing. Similar devices that came in DIP packages were the SSM2165 (8 pin DIP) and SSM2166 (14 pin DIP) which are obsolete but still can be purchased on eBAY from merchants in China, but at high prices.

Larry Gruber
K7LJG
 

Hi
A few notes on speech processing.
Speech processing increased the average value when compared to the peak value of the speech waveform.
Usually, it also limits the peak value.
The best method to adjust a speech processor is to use an osciloscope where you can also see the adjustments being made.
However, the old method of adjusting a speech processor if no oscilloscope was available was to first monitor the Tx or Xcvr ALC and adjust the Tx or Xcvr mic gain until you get an indication on the meter or according to instructions.
Then plug in the speech processor and adjust its gain until it reaches the same value without touching the TX or Xcvr mic gain.
If you use too much gain in the processor it will simply increase distortion and increase the background noise.
However, done properly, it will increase the average power.
As you said, this means more dissipation in the power amplifier and the power supply must also be able to withstand the increased demand.
Increased cooling will be required.
With modern Tx or Xcvr if it has some form of ALC and if it is monitored, the same adjustment can be made.
If not, put a meter in line with the supply and adjust the Tx or Xcvr audio gain as instructed and then the speech processor is adjusted until the meter approaches the peak reading that you get without the processor.
Remember that your meter will not be showing the true value due to mechanical inertia and the true reading will be approximately double the meter reading depending upon your voice characteristics.
Furher details can be found in radio handbooks.
Regards
Lawrence


On Mon, Apr 3, 2017 at 2:56 AM, John Smith via Groups.Io <johnlinux77@...> wrote:

Alright, I am up to speed now. It kinda seems like it isn't necessary and could make intelligibility worse and or annoying. I know what overmodulation sounds like, and you can't tell those people anything different. And make the final even hotter than it already is with 25V.? Has anyone done this with the BITX40? What difference did it make? Would you recommend it?

Hello? (tap, tap, tap) Is this thing on?
 

I am still wanting to find out if anyone has done something like this with the BITX40. I am still not sure if this pig is worth that much lipstick. Did microphone gain help folks with low audio input. Is seems people have plans, but don't do or talk about them for some reason. I haven't even connected the audio jacks because there is nothing to do with it at this point. Digital modes are what most people are wanting. Someone should think about a BITX for digital modes only. ?I'd buy one, if it's cheap. But yea, I have read a lot about this as a solution for folks who think it doesn't have enough grunt. Me too. But I have been happy with getting more power out as is. But with an average output around 10-12W while 25W steady from a tone generator is possible. I would really like to increase my average output a bit. Within reason.?

I look forward to reading your results as I begin sourcing parts to help decide on a circuit to start with. Someone emailed this one to me. It seems easy to make, but I don't know what type of speech processor it is. I'll probably start here.

http://www.rason.org/Projects/speech/speech.htm
 

the easy way is to just speak close to the mic and allow the mic amp to do all the audio clipping. audio clipping introduces artifacts inside the audio passband (harmonics of a clipped 500 Hz are at 1000Hz, 1500 Hz, 2500Hz, etc.)
the better way is to clip the RF. That way, the harmonics of 12 Mhz are at 24, 36 etc that are filtered away. So, the easy way to do that is to put two back to back germanium diodes at the crystal filter towards the BFO side.?
- f

On Mon, Apr 3, 2017 at 11:53 PM, John Smith via Groups.Io <johnlinux77@...> wrote:

I am still wanting to find out if anyone has done something like this with the BITX40. I am still not sure if this pig is worth that much lipstick. Did microphone gain help folks with low audio input. Is seems people have plans, but don't do or talk about them for some reason. I haven't even connected the audio jacks because there is nothing to do with it at this point. Digital modes are what most people are wanting. Someone should think about a BITX for digital modes only.? I'd buy one, if it's cheap. But yea, I have read a lot about this as a solution for folks who think it doesn't have enough grunt. Me too. But I have been happy with getting more power out as is. But with an average output around 10-12W while 25W steady from a tone generator is possible. I would really like to increase my average output a bit. Within reason.?

I look forward to reading your results as I begin sourcing parts to help decide on a circuit to start with. Someone emailed this one to me. It seems easy to make, but I don't know what type of speech processor it is. I'll probably start here.
Jack Purdum
 

Dennis and I are working on an AGC module built around the AD8367 and speech compressor using the SM2167. The goal is to have a daughter PCB similar to that for the Forty-9er (see: www.farrukhzia.com/k2zia) that handles those functions plus the TFT display, the AD9850 VFO, all using a Mega 2560 Pro Mini which replaces the Raduino board. The board will be released at the FDIM conference that takes place just before the Dayton Hamfest. We're not sure of the board's final size yet, but we plan to sell the bare board for under $10. Dennis is checking to see what it would cost to have the SMT parts set as part of fabrication. Too soon to tell about that.

Jack, W8TEE


From: John Smith via Groups.Io <johnlinux77@...>
To: [email protected]
Sent: Monday, April 3, 2017 2:23 PM
Subject: Re: [BITX20] Adding speech compression.

I am still wanting to find out if anyone has done something like this with the BITX40. I am still not sure if this pig is worth that much lipstick. Did microphone gain help folks with low audio input. Is seems people have plans, but don't do or talk about them for some reason. I haven't even connected the audio jacks because there is nothing to do with it at this point. Digital modes are what most people are wanting. Someone should think about a BITX for digital modes only. ?I'd buy one, if it's cheap. But yea, I have read a lot about this as a solution for folks who think it doesn't have enough grunt. Me too. But I have been happy with getting more power out as is. But with an average output around 10-12W while 25W steady from a tone generator is possible. I would really like to increase my average output a bit. Within reason.?
I look forward to reading your results as I begin sourcing parts to help decide on a circuit to start with. Someone emailed this one to me. It seems easy to make, but I don't know what type of speech processor it is. I'll probably start here.
http://www.rason.org/Projects/speech/speech.htm
 

Jack, Sounds like a great project! Looking forward to the availability of the board! Just need a couple more components to come in from China and I'll be getting started on the antenna analyzer!

Roy
WA0YMH

On Apr 3, 2017 2:27 PM, "Jack Purdum via Groups.Io" <econjack=[email protected]> wrote:
Dennis and I are working on an AGC module built around the AD8367 and speech compressor using the SM2167. The goal is to have a daughter PCB similar to that for the Forty-9er (see: ) that handles those functions plus the TFT display, the AD9850 VFO, all using a Mega 2560 Pro Mini which replaces the Raduino board. The board will be released at the FDIM conference that takes place just before the Dayton Hamfest. We're not sure of the board's final size yet, but we plan to sell the bare board for under $10. Dennis is checking to see what it would cost to have the SMT parts set as part of fabrication. Too soon to tell about that.

Jack, W8TEE


From: John Smith via Groups.Io <johnlinux77=yahoo.com@groups.io>
To: [email protected]
Sent: Monday, April 3, 2017 2:23 PM
Subject: Re: [BITX20] Adding speech compression.

I am still wanting to find out if anyone has done something like this with the BITX40. I am still not sure if this pig is worth that much lipstick. Did microphone gain help folks with low audio input. Is seems people have plans, but don't do or talk about them for some reason. I haven't even connected the audio jacks because there is nothing to do with it at this point. Digital modes are what most people are wanting. Someone should think about a BITX for digital modes only.? I'd buy one, if it's cheap. But yea, I have read a lot about this as a solution for folks who think it doesn't have enough grunt. Me too. But I have been happy with getting more power out as is. But with an average output around 10-12W while 25W steady from a tone generator is possible. I would really like to increase my average output a bit. Within reason.?
I look forward to reading your results as I begin sourcing parts to help decide on a circuit to start with. Someone emailed this one to me. It seems easy to make, but I don't know what type of speech processor it is. I'll probably start here.
Jack Purdum
 

The BITX40 project has really been enjoyable. Can't wait for the microBITX.

Jack, W8TEE


From: Roy Appleton <twelveoclockhigh@...>
To: [email protected]
Sent: Monday, April 3, 2017 3:31 PM
Subject: Re: [BITX20] Adding speech compression.

Jack, Sounds like a great project! Looking forward to the availability of the board! Just need a couple more components to come in from China and I'll be getting started on the antenna analyzer!

Roy
WA0YMH

On Apr 3, 2017 2:27 PM, "Jack Purdum via Groups.Io" <econjack=[email protected]> wrote:
Dennis and I are working on an AGC module built around the AD8367 and speech compressor using the SM2167. The goal is to have a daughter PCB similar to that for the Forty-9er (see: ) that handles those functions plus the TFT display, the AD9850 VFO, all using a Mega 2560 Pro Mini which replaces the Raduino board. The board will be released at the FDIM conference that takes place just before the Dayton Hamfest. We're not sure of the board's final size yet, but we plan to sell the bare board for under $10. Dennis is checking to see what it would cost to have the SMT parts set as part of fabrication. Too soon to tell about that.

Jack, W8TEE


From: John Smith via Groups.Io <johnlinux77=yahoo.com@groups. io>
To: [email protected]
Sent: Monday, April 3, 2017 2:23 PM
Subject: Re: [BITX20] Adding speech compression.

I am still wanting to find out if anyone has done something like this with the BITX40. I am still not sure if this pig is worth that much lipstick. Did microphone gain help folks with low audio input. Is seems people have plans, but don't do or talk about them for some reason. I haven't even connected the audio jacks because there is nothing to do with it at this point. Digital modes are what most people are wanting. Someone should think about a BITX for digital modes only.? I'd buy one, if it's cheap. But yea, I have read a lot about this as a solution for folks who think it doesn't have enough grunt. Me too. But I have been happy with getting more power out as is. But with an average output around 10-12W while 25W steady from a tone generator is possible. I would really like to increase my average output a bit. Within reason.?
I look forward to reading your results as I begin sourcing parts to help decide on a circuit to start with. Someone emailed this one to me. It seems easy to make, but I don't know what type of speech processor it is. I'll probably start here.
 

Right Farhan
Your idea is good and cheap to try.
RF clippers are the best.
Simple AF processing is second best as the artifacts are inside the audio passband..
If I remember correctly, there was an RF clipper in the old ARRL SSB Handbook that clipped the DSB before the filter to use the filter for filtering the sideband and resulting artifacts.
I will try to find it and post the article to get the idea but it used a valve not a transistor.


On Mon, Apr 3, 2017 at 9:14 PM, Ashhar Farhan <farhanbox@...> wrote:
the easy way is to just speak close to the mic and allow the mic amp to do all the audio clipping. audio clipping introduces artifacts inside the audio passband (harmonics of a clipped 500 Hz are at 1000Hz, 1500 Hz, 2500Hz, etc.)
the better way is to clip the RF. That way, the harmonics of 12 Mhz are at 24, 36 etc that are filtered away. So, the easy way to do that is to put two back to back germanium diodes at the crystal filter towards the BFO side.?
- f

On Mon, Apr 3, 2017 at 11:53 PM, John Smith via Groups.Io <johnlinux77=yahoo.com@groups.io> wrote:

I am still wanting to find out if anyone has done something like this with the BITX40. I am still not sure if this pig is worth that much lipstick. Did microphone gain help folks with low audio input. Is seems people have plans, but don't do or talk about them for some reason. I haven't even connected the audio jacks because there is nothing to do with it at this point. Digital modes are what most people are wanting. Someone should think about a BITX for digital modes only.? I'd buy one, if it's cheap. But yea, I have read a lot about this as a solution for folks who think it doesn't have enough grunt. Me too. But I have been happy with getting more power out as is. But with an average output around 10-12W while 25W steady from a tone generator is possible. I would really like to increase my average output a bit. Within reason.?

I look forward to reading your results as I begin sourcing parts to help decide on a circuit to start with. Someone emailed this one to me. It seems easy to make, but I don't know what type of speech processor it is. I'll probably start here.
 

@Ashhar,

So, if I understand well, the two germanium diodes should be placed here?


73, Allard PE1NWL
 

Very elegant if 2 diodes work well enough. ?Might want variable gain at the mike amp so you can set the compression level. ?And then adjust R136 to determine the output power.

Jerry, KE7ER


On Mon, Apr 3, 2017 at 01:08 pm, Allard PE1NWL wrote:

So, if I understand well, the two germanium diodes should be placed here?

?
 

¿ªÔÆÌåÓý

Jack

Tracking with anticipation of your unveiling of this effort.?

Art N2AJO

Sent from my iPhone

On Apr 3, 2017, at 3:26 PM, Jack Purdum via Groups.Io <econjack@...> wrote:

Dennis and I are working on an AGC module built around the AD8367 and speech compressor using the SM2167. The goal is to have a daughter PCB similar to that for the Forty-9er (see: ) that handles those functions plus the TFT display, the AD9850 VFO, all using a Mega 2560 Pro Mini which replaces the Raduino board. The board will be released at the FDIM conference that takes place just before the Dayton Hamfest. We're not sure of the board's final size yet, but we plan to sell the bare board for under $10. Dennis is checking to see what it would cost to have the SMT parts set as part of fabrication. Too soon to tell about that.

Jack, W8TEE


From: John Smith via Groups.Io <johnlinux77@...>
To: [email protected]
Sent: Monday, April 3, 2017 2:23 PM
Subject: Re: [BITX20] Adding speech compression.

I am still wanting to find out if anyone has done something like this with the BITX40. I am still not sure if this pig is worth that much lipstick. Did microphone gain help folks with low audio input. Is seems people have plans, but don't do or talk about them for some reason. I haven't even connected the audio jacks because there is nothing to do with it at this point. Digital modes are what most people are wanting. Someone should think about a BITX for digital modes only. ?I'd buy one, if it's cheap. But yea, I have read a lot about this as a solution for folks who think it doesn't have enough grunt. Me too. But I have been happy with getting more power out as is. But with an average output around 10-12W while 25W steady from a tone generator is possible. I would really like to increase my average output a bit. Within reason.?
I look forward to reading your results as I begin sourcing parts to help decide on a circuit to start with. Someone emailed this one to me. It seems easy to make, but I don't know what type of speech processor it is. I'll probably start here.
 

I would like to read a bit more from Farhan about clipping RF and what the diodes will do. And does Allard have them in the right place? Will it give me a little higher average power output? Getting up to speed is a relative term. Especially in this group. But at this speed I think I have failed to appreciate the simple design of a QRP SSB only transceiver. I expected too much based on big hacking dreams, and I caused my own frustration.

At this point I half wonder if I should accept it as it is and shut up. Don't tell anyone I said that, they won't believe you.

I couldn't help but notice no one said they did this hack.
Jack Purdum
 

We'll keep you posted!

Jack, W8TEE


From: Art Olson via Groups.Io <olsonaj6927@...>
To: [email protected]
Sent: Monday, April 3, 2017 5:12 PM
Subject: Re: [BITX20] Adding speech compression.

Jack

Tracking with anticipation of your unveiling of this effort.?

Art N2AJO

Sent from my iPhone

On Apr 3, 2017, at 3:26 PM, Jack Purdum via Groups.Io <econjack@...> wrote:

Dennis and I are working on an AGC module built around the AD8367 and speech compressor using the SM2167. The goal is to have a daughter PCB similar to that for the Forty-9er (see: ) that handles those functions plus the TFT display, the AD9850 VFO, all using a Mega 2560 Pro Mini which replaces the Raduino board. The board will be released at the FDIM conference that takes place just before the Dayton Hamfest. We're not sure of the board's final size yet, but we plan to sell the bare board for under $10. Dennis is checking to see what it would cost to have the SMT parts set as part of fabrication. Too soon to tell about that.

Jack, W8TEE


From: John Smith via Groups.Io <johnlinux77@...>
To: [email protected]
Sent: Monday, April 3, 2017 2:23 PM
Subject: Re: [BITX20] Adding speech compression.

I am still wanting to find out if anyone has done something like this with the BITX40. I am still not sure if this pig is worth that much lipstick. Did microphone gain help folks with low audio input. Is seems people have plans, but don't do or talk about them for some reason. I haven't even connected the audio jacks because there is nothing to do with it at this point. Digital modes are what most people are wanting. Someone should think about a BITX for digital modes only. ?I'd buy one, if it's cheap. But yea, I have read a lot about this as a solution for folks who think it doesn't have enough grunt. Me too. But I have been happy with getting more power out as is. But with an average output around 10-12W while 25W steady from a tone generator is possible. I would really like to increase my average output a bit. Within reason.?
I look forward to reading your results as I begin sourcing parts to help decide on a circuit to start with. Someone emailed this one to me. It seems easy to make, but I don't know what type of speech processor it is. I'll probably start here.

1 - 20 of 27