Topics

Automatic RF Attenuation Mod for Bitx40

jpkellyburbank@...
 

I've come up with a mod to automatically RF attenuate strong signals.  It limits the audio at the volume knob to about half a volt peak to peak for stations that are 40db over S9.  For stations S9 and under, there is no RF attenuation.  Now I can listen to the nets without needing to adjust the volume control.  And strong stations that used to be distorted are now clean.

I  enjoy using my Bitx40 a lot more now.

Jim/ND6P

Jim/ND6P

Norberto Modanesi
 

Hi Jim

I was looking for such AGC: Thanks for sharing.-

 

73 – LU5DNM

 

De: BITX20@groups.io [mailto:BITX20@groups.io] En nombre de jpkellyburbank@...
Enviado el: domingo, 27 de agosto de 2017 01:16 p.m.
Para: BITX20@groups.io
Asunto: [BITX20] Automatic RF Attenuation Mod for Bitx40

 

I've come up with a mod to automatically RF attenuate strong signals.  It limits the audio at the volume knob to about half a volt peak to peak for stations that are 40db over S9.  For stations S9 and under, there is no RF attenuation.  Now I can listen to the nets without needing to adjust the volume control.  And strong stations that used to be distorted are now clean.

I  enjoy using my Bitx40 a lot more now.

Jim/ND6P

Jim/ND6P

Jerry Gaffke
 

That looks pretty good to me, may try it.
The Bitx40 probably has more RF gain than is required (if you have a decent antenna hooked up to it).
So a simple attenuator like this is a good choice, rather than trying to incorporate a variable gain amp.

Since we are attenuating incoming RF, as the agc kicks in for a strong signal the receiver should quiet down
nicely, leaving very clear audio.  

You have three diodes labeled BA4790, but I can't find such a part number with a web search.
Is that a special PIN diode, or will a regular silicon diode such as a 1n914 suffice?

I doubt you need the new relay, just insert the RX to RX part of this circuit between K2-14 (input from antenna)
and K1-12 (output to the Bitx40 receiver) of the existing Bitx40 schematic.  The stock Bitx40 just has a trace
between K2-14 and K1-12, so cut that trace and then insert this circuit.

The chokes must be large (100uH) to avoid conducting the audio frequency signal we are sensing,
but I would be a bit worried that capacitive coupling between windings might bleed off a fair bit of the incomming RF.
Perhaps add a 10uH choke in series with each 100uH choke if that proves to be an issue?
(It's an issue if adding this agc circuit to the radio makes it signiicantly less sensitive for weak signals.)

Is this a circuit that you cooked up on your own?
If you borrowed the circuit from somewhere else, where did you find it?

Jerry, KE7ER



On Sun, Aug 27, 2017 at 09:15 am, <jpkellyburbank@...> wrote:
I've come up with a mod to automatically RF attenuate strong signals.  It limits the audio at the volume knob to about half a volt peak to peak for stations that are 40db over S9.  For stations S9 and under, there is no RF attenuation.  Now I can listen to the nets without needing to adjust the volume control.  And strong stations that used to be distorted are now clean.

Jerry Gaffke
 

Scratch that, the chokes are there only to block the incoming RF signal.


On Mon, Aug 28, 2017 at 01:54 pm, Jerry Gaffke wrote:
The chokes must be large (100uH) to avoid conducting the audio frequency signal we are sensing,
but I would be a bit worried that capacitive coupling between windings might bleed off a fair bit of the incomming RF.

jpkellyburbank@...
 

Hi Jerry,

The diodes are BA479G and are available from Mouser.  They are PIN diodes, designed for RF attenuation.  The voltage controlled RF attenuator in the diagram is variable, attenuating between zero and about 30 db based on the incoming audio level.  The attenuation does not start until the audio is sufficiently strong (about S9), so weak signals are not affected.  I can hear a 1 uv signal after installing the mod.  The 1N914 diode will not substitute.

I used the relay because I did not want to modify the circuit board; however, insertion of the attenuator between K2-14 (input from antenna) 
and K1-12 (output to the Bitx40 receiver) should work nicely.

The 100 uhy chokes are used to block RF, not audio.  22 uhy chokes would probably have worked but I went with 100 uhy because I had them and they provide a higher RF impedance.  I tested the 100 uhy chokes with an antenna analyzer and they worked fine at 7 Mhz, not develping self resonance until 136 Mhz.

I got the RF attenuator design from https://www.eeweb.com/blog/circuit_projects/rf-signal-attenuator-up-to-30db-for-1mhz-500mhz-input and simplified it a bit.  The rest of the design is my own, developed after doing a lot of testing with the Bitx40 to see how much RF it could handle before distorting at the detector and over-driving the audio.

Jim/ND6P

Jerry Gaffke
 

Very nice.  Good complete explanation. Thanks!

I've never quite figured out the difference between PIN and regular diodes.
Apparently a PIN goes into some sort of free-wheeling mode where it continues to conduct
for a short time once it becomes reverse biased.
So a straight silicon diode would tend to clip the sine waves coming in,
whereas a PIN diode takes its time when going from conducting to non-conducting.

Jerry


On Mon, Aug 28, 2017 at 02:49 pm, <jpkellyburbank@...> wrote:
The diodes are BA479G and are available from Mouser.

Henning Weddig
 

according to the data sheet the BA479G is siutable for frequencies above 10 MHz.

There are other diodes which will work from 1 MHz on.

Please have a look on the theoretical function of a PIN diode, the main difference is the I layer in between the p and n layer.

At least, tom put a cap directly from the output of an opamp to ground is a bad idea as it may and probably will intorduce another pole in the frequency domain and therefore give the possibility of oscialltions.

Henning DK5LV


Am 29.08.2017 um 01:10 schrieb Jerry Gaffke via Groups.Io:

Very nice.  Good complete explanation. Thanks!

I've never quite figured out the difference between PIN and regular diodes.
Apparently a PIN goes into some sort of free-wheeling mode where it continues to conduct
for a short time once it becomes reverse biased.
So a straight silicon diode would tend to clip the sine waves coming in,
whereas a PIN diode takes its time when going from conducting to non-conducting.

Jerry

On Mon, Aug 28, 2017 at 02:49 pm, <jpkellyburbank@...> wrote:
The diodes are BA479G and are available from Mouser.

Jerry Gaffke
 

Looking at the schematic here:
    https://www.eeweb.com/blog/circuit_projects/rf-signal-attenuator-up-to-30db-for-1mhz-500mhz-input

we can probably do without  C4 L2 D1 C2 C6 C7
and just drive L1 from about 4 volts

Those parts seem to be there to rectify incoming RF and use that rectified signal to further attenuate the signal,
providing some local AGC action.  
Given the RC time constant of (10+10uF)*1k = 0.02 seconds, our audio derived AGC control voltage can be about as fast.
(Though Jim did away with most of that capacitance, his circuit should react much faster to changes in the RF level.)

Alternately, keep that original scheme, but add some amplification of the RF before it goes into D1.
So have RF derived control for our AGC, and forget about the audio derived control.

The notes state it was designed to deal with up to 100mW of incoming RF, or 2.2 volts rms into 50 ohms.
That's way bigger than we should have coming into this receiver, as that would fry Q13.
I'd just use Raj's back-to-back diode trick to clip incoming RF if it gets near to that level:  https://groups.io/g/BITX20/message/19105
Perhaps include back-to-back diodes somewhere in the audio amp also, to limit max volume into the headphones.
 
Just some random thoughts, still half baked.

Jerry


On Tue, Aug 29, 2017 at 03:41 am, Henning Weddig wrote:

according to the data sheet the BA479G is siutable for frequencies above 10 MHz.

There are other diodes which will work from 1 MHz on.

Please have a look on the theoretical function of a PIN diode, the main difference is the I layer in between the p and n layer.

At least, tom put a cap directly from the output of an opamp to ground is a bad idea as it may and probably will intorduce another pole in the frequency domain and therefore give the possibility of oscialltions.

 

jpkellyburbank@...
 

Hello Henning,

My first choice for the Pin diode was the BA389, but I could only find it at www.rf-microwave.com where they have a 30 Euro minimum order.  I also tried the MA27B from RF parts and it worked as well as the BA479G, so I assumed the 10 Mhz minimum in the specification is not a hard limit.  I also have some MI105 pin diodes coming from Asia that I will be trying, but so far the pin diode choice does not seem to be critical.

The op amp stage that has the capacitor on the output is functioning as a voltage follower.  I added the cap as a precaution for any RF problems, but I suppose it is not actually necessary. Anyway, I have built and tested and am not noticing any oscillations.

Jim/ND6P

Henning Weddig
 

Jim,

for the PIN diode to work properly as a variable resistor, and not rectify the RF the so called carrier lifetime n(minority carriers ??)  is an important factor.

Jerry Gaffke proposed another circuitry of the same topology but a different diode, i.e. the BA595 from Infineon. From the data sheet the carrier lifetime is specified 1600 ns 1/1600 ns= 625 kHz, so this diode is really suited for attenuation i.e. acting as a resistor for frequencies above 1 MHz.

Opamp: if You put a resistor (about 100 ohms or so) in series with the output of the opamp and then the cap tor ground the possibility of an oscillating opamp  will be gone.

I am not an expert in control theory, but to put the actuator in front of  a time delay element (in our case xtal filter) can render the control loop into instabilities i.e. can oscillate.

From my simulations the BITX40 has one big drawback: the high gain AF preamp (LTSPICE XVII tells me 46 dB gain) , so every AF AGC after this amp will limit the dynamic range, as the AF preamp  is the "weak" point.
Plöease have a look on my level diagramme in the DK5LV folder.
73
Henning DK5LV



Am 29.08.2017 um 20:10 schrieb jpkellyburbank@...:

Hello Henning,

My first choice for the Pin diode was the BA389, but I could only find it at www.rf-microwave.com where they have a 30 Euro minimum order.  I also tried the MA27B from RF parts and it worked as well as the BA479G, so I assumed the 10 Mhz minimum in the specification is not a hard limit.  I also have some MI105 pin diodes coming from Asia that I will be trying, but so far the pin diode choice does not seem to be critical.

The op amp stage that has the capacitor on the output is functioning as a voltage follower.  I added the cap as a precaution for any RF problems, but I suppose it is not actually necessary. Anyway, I have built and tested and am not noticing any oscillations.

Jim/ND6P

Jerry Gaffke
 

Datasheet of the BA479G says the "minority carrier lifetime" is typically 4 uS.
That's the average time it takes for the hole or electron to get swallowed up.
You would have to take it down to 1/(2*4uS) = 125 khz before you had 4 uS between
peak forward current and peak reverse current, at which point I assume
the reverse current would diminish by about half.  

Perhaps AM broadcast signals in the 550-1700 mhz range would be distorted some,
creating harmonics that add to the noise at HF.
But if the Bitx40 receiver sounds as well on weak signals with this attenuator inline
as it does without the attenuator, then I'd think the BA479G is plenty good enough.

I could be wrong about any of the above, PIN diodes are not  something I claim to know much about.


I had previously said:

"we can probably do without  C4 L2 D1 C2 C6 C7
and just drive L1 from about 4 volts
Those parts seem to be there to rectify incoming RF and use that rectified signal to further attenuate the signal,
providing some local AGC action."

If that is true, then it is a feature we don't want here.
We don't want an out of band signal such as a strong local AM broadcaster to throttle back our 40m signals. 

Jerry


On Tue, Aug 29, 2017 at 11:10 am, <jpkellyburbank@...> wrote:
My first choice for the Pin diode was the BA389, but I could only find it at www.rf-microwave.com where they have a 30 Euro minimum order.  I also tried the MA27B from RF parts and it worked as well as the BA479G, so I assumed the 10 Mhz minimum in the specification is not a hard limit.  I also have some MI105 pin diodes coming from Asia that I will be trying, but so far the pin diode choice does not seem to be critical.

Jerry Gaffke
 

Jim first referenced that other circuitry in this post:    https://groups.io/g/BITX20/message/31875

But yes, I agree, I'd expect these parts to work ok at least down to 3.5mhz, 
and I don't have much interest in going any lower.


On Tue, Aug 29, 2017 at 12:04 pm, Henning Weddig wrote:

Jerry Gaffke proposed another circuitry of the same topology but a different diode, i.e. the BA595 from Infineon. From the data sheet the carrier lifetime is specified 1600 ns 1/1600 ns= 625 kHz, so this diode is really suited for attenuation i.e. acting as a resistor for frequencies above 1 MHz.

 

jpkellyburbank@...
 

I've modified my Version 1, changing the coupling between the op amps from AC to DC.  This is the best performer now but my Version 2 works quite well also.  If you need a drawing that you can see more clearly, email me and I will send it to you.

-Jim/ND6P

VU3WMJ@gmail.com Virtualview
 

Pls send me ur attenuatr diagram for me regards vu3wmj 73s...


On Sep 5, 2017 8:57 PM, <jpkellyburbank@...> wrote:
I've modified my Version 1, changing the coupling between the op amps from AC to DC.  This is the best performer now but my Version 2 works quite well also.  If you need a drawing that you can see more clearly, email me and I will send it to you.

-Jim/ND6P