I just installed a 20 dB RF AGC modification in the BITX40. I have not installed it in the uBITX yet but intend to soon. I have posted the project on It's a simple circuit and replaces the S meter circuitry, too. 73, Don

Don,

Where in the ubitx do you plan to tap into the RF part of the receiver?

There doesn't appear to be a good point to introduce an attenuator
without upsetting the termination insensitive amplifiers which are
critical to the double-balanced mixers terminations.

tim ab0wr

On Mon, 26 Mar 2018 06:52:48 -0700
"Don, ND6T via Groups.Io" <nd6t_6@...> wrote:

I just installed a 20 dB RF AGC modification in the BITX40. I have
not installed it in the uBITX yet but intend to soon. I have posted
the project on nd6t.com. ( ) It's a
simple circuit and replaces the S meter circuitry, too. 73, Don

Tim,
I opened the receive trace between K1 and K3 for the RF gain control. This board is mounted on, and connected to, that pot. That way it has no problems with the mixers and amplifiers. I've been using that mod on BITX units for more than a year now without issues so I feel safe in doing it. 73, Don

Don,

Good idea!

I expect the mosfet will withstand any reasonable RF signal. Let the
group know how it works out!

tim ab0wr

On Mon, 26 Mar 2018 07:48:47 -0700
"Don, ND6T via Groups.Io" <nd6t_6@...> wrote:

Tim,
I opened the receive trace between K1 and K3
here?( ) for the RF gain control.
This board is mounted on, and connected to, that pot. That way it has
no problems with the mixers and amplifiers. I've been using that mod
on BITX units for more than a year now without issues so I feel safe
in doing it. 73, Don

Oh, it works. Been using it for more than a week. I would like more dynamic AGC range (20 dB isn't quite enough for me) but it is the best I have tried so far. -Don

?

Don,

What circuit are you using to obtain the agc voltage?

Even a simple 741 op amp would give more dynamic range than this (I
assume the use of a noisy op amp isn't a problem in an agc loop like
this).

Is the problem the dynamic range of the input signal?

tim ab0wr

On Mon, 26 Mar 2018 11:45:52 -0700
"Don, ND6T via Groups.Io" <nd6t_6@...> wrote:

Oh, it works. Been using it for more than a week. I would like more
dynamic AGC range (20 dB isn't quite enough for me) but it is the
best I have tried so far. -Don

As you can see, I am using a single BJT to amplify the sample from the hot side of the volume control. The "problem" isn't the detection, rather the range of a single control element (the MOSFET in this instance). 20 dB is pretty good but I shall always crave more. Probably another MOSFET up the receive chain is my next experiment. First, though, I want to tame the uBITX by installing the board there (providing an S meter there, too) and keep moving onward. No, I don't believe that AGC action is necessary below S8 or S9. My background noise level is usually S5 to S7. Instead, I am adverse to those loud excursions described in the article. I tried more amplification (quite simple to do) and it does not help things. I usually run the RF gain a bit low. All of the BITX series have very good gain and there is no reason to go flat out unless one enjoys listening to hiss.
More gain is easily accomplished by adding another BJT in cascade (another 2 cents). A wider control voltage range is available by adding 2 diodes and 2 capacitors for a voltage doubler (10 cents). Neither increases the dynamic attenuation range. Without an RF pre-amp stage we need to either improve the control element or add more of them up the receive chain. VVAs at this frequency either have excessive insertion loss or cost just too darn much. A single 2N7002 (2 cents) as a shunt shows no insertion loss at idle. Simple is good.
I'm just saying that there is always room for improvement.

A BAP64Q pin diode attenuator gives 60dB of dynamic range:??/g/BITX20/message/32066
Attenuating back in RF?gets around the limited dynamic range that Henning points out in the first post of that thread.
Note that the control voltage is inverted with respect to the 2n7002 FET, higher voltages give less attenuation.
Could get a slightly lower noise figure for the receiver if the attenuator is inserted at a later stage of the RF chain.

The BAP64Q is relatively expensive at $0.50 single piece,
the frugal among us will note it's down at $0.20 if you buy a few thousand.?
Mouser and Digikey both stock it, Mouser points you to the wrong BAP64* datasheet.
There are other similar small signal pin diode attenuators out there from other manufacturers.

Jerry, KE7ER

Thanks Jerry! I'll check it out. Shouldn't be any problem going the other way with control voltages. Guess I had better hold off and try that one. Great catch! Or was that a "save"? -Don

开云体育

Here is the application circuit - Ron W7HD

开云体育

Here it is with component values.

Ron W7HD

Thank you. I pulled up the data sheet and it looks like I need to order some to try. I should have asked; Has any one used this as an AGC control yet? How well did it work. I kinda hate losing nearly half the incoming signal but it might be worth it. Still, no sense re-inventing the wheel. -Don

I think the glory is all yours, most of us here just talk.

Don,

I forgot you gave the circuit in the first mesage. Sorry.

The mosfet should be able to be driven from zero attenuation, Ids=0,
to almost a direct short or infinite attenuation, Vds=0, correct?

Is the problem the range of Vgs needed to go from Ids=0 to Vds=0? That
this only provides a 20db dynamic range?

Is it not possible for the mosfet to reach Vds=0?

tim ab0wr





On Mon, 26 Mar 2018 13:18:58 -0700
"Don, ND6T via Groups.Io" <nd6t_6@...> wrote:

As you can see, I am using a single BJT to amplify the sample from
the hot side of the volume control. The "problem" isn't the
detection, rather the range of a single control element (the MOSFET
in this instance). 20 dB is pretty good but I shall always crave
more. Probably another MOSFET up the receive chain is my next
experiment. First, though, I want to tame the uBITX by installing the
board there (providing an S meter there, too) and keep moving onward.
No, I don't believe that AGC action is necessary below S8 or S9. My
background noise level is usually S5 to S7. Instead, I am adverse to
those loud excursions described in the article. I tried more
amplification (quite simple to do) and it does not help things. I
usually run the RF gain a bit low. All of the BITX series have very
good gain and there is no reason to go flat out unless one enjoys
listening to hiss. More gain is easily accomplished by adding another
BJT in cascade (another 2 cents). A wider control voltage range is
available by adding 2 diodes and 2 capacitors for a voltage doubler
(10 cents). Neither increases the dynamic attenuation range. Without
an RF pre-amp stage we need to either improve the control element or
add more of them up the receive chain. VVAs at this frequency either
have excessive insertion loss or cost just too darn much. A single
2N7002 (2 cents) as a shunt shows no insertion loss at idle. Simple
is good. I'm just saying that there is always room for improvement.

The 2n7002 has an Rds(on) of 7.5 ohms.
This is in a 50 ohm environment, so reducing the voltage by roughly a factor of 10 seems about right, giving -20dB.
Perhaps other FET's have a lower on resistance, but either capacitance or price would go up.

PIN diodes have an inherent low frequency limit determined by the charge carrier lifetime.
This will make a pin diode attenuator at the receiver front end not so good a choice
if interested in anything below 80m.?

A better position for the attenuator might be in either the 45mhz or 12mhz IF.
This would also prevent the attenuator from degrading the receiver noise figure, which should be
pretty well established by that first amp at 45mhz.

If you don't like losing 2dB, could get the audio pre-amp to run a little hotter.? With AGC now in front
of that pre-amp, the pre-amp's dynamic range is no longer an issue.

Jerry


Assuming we can afford the loss of a couple dB,?

The pi network PIN diode attenuators try to maintain 50 ohms,
so they might work well with the BiDi amps of the uBitx.??
resistance of 7.2 ohms

With no amplifier in front of it that 2 dB comes at the expense of the overall noise figure.taking it from 13 up to 15 dB or so. That won't affect the lower frequencies but 10 meters would be certainly affected.

The performance of the 2N7002 was determined by test results by me. I actually got up to a wee bit more than 25 dB attenuation on some individual devices below 15 MHz but I like to keep my estimates on the conservative side. I can open my receive path by just unplugging the RF gain from the board. When I do, signals drop more than 90 dB. Quite promising. All I really need is 50 dB of dynamic range.

Currently a 50 microvolt signal (S9) produces a 1.5 volt control, giving a 1 dB drop. 20 dB more input raises my control to 3.7 volts and my resulting audio stays the same. Above that, the maximum attenuation has been reached so the audio level resumes tracking the input. The RF gain control still works independently so that easily handles my real needs. But I am never satisfied. Give me paucity and I strive for abundance. Give me abundance and I wish for simplicity. Hey, everybody needs a hobby. -Don

Maybe two 2n7002's in parallel?

2x 2n7002 is only another 6db, but definitely simple.

Don,

Hmmm... Looking at the 2n7002 datasheet it appears that Rds doesn't
change much for a change in Vgs. A change in Rds from 6ohm to about
2.4ohm for a Vgs from 2.5v to 10v. After 10v the Rds doesn't change
much.

Is this what is limiting the dynamic range?

Based on this admittedly cursory look I'm surprised you can get a 20db
range on attenuation!

I think you are correct, it might be needed to add more agc action
somewhere down the chain.

Take a look at this article. It might trigger a thought process that
leads somewhere.

tim ab0wr


On Mon, 26 Mar 2018 13:18:58 -0700
"Don, ND6T via Groups.Io" <nd6t_6@...> wrote:

As you can see, I am using a single BJT to amplify the sample from
the hot side of the volume control. The "problem" isn't the
detection, rather the range of a single control element (the MOSFET
in this instance). 20 dB is pretty good but I shall always crave
more. Probably another MOSFET up the receive chain is my next
experiment. First, though, I want to tame the uBITX by installing the
board there (providing an S meter there, too) and keep moving onward.
No, I don't believe that AGC action is necessary below S8 or S9. My
background noise level is usually S5 to S7. Instead, I am adverse to
those loud excursions described in the article. I tried more
amplification (quite simple to do) and it does not help things. I
usually run the RF gain a bit low. All of the BITX series have very
good gain and there is no reason to go flat out unless one enjoys
listening to hiss. More gain is easily accomplished by adding another
BJT in cascade (another 2 cents). A wider control voltage range is
available by adding 2 diodes and 2 capacitors for a voltage doubler
(10 cents). Neither increases the dynamic attenuation range. Without
an RF pre-amp stage we need to either improve the control element or
add more of them up the receive chain. VVAs at this frequency either
have excessive insertion loss or cost just too darn much. A single
2N7002 (2 cents) as a shunt shows no insertion loss at idle. Simple
is good. I'm just saying that there is always room for improvement.

Jerry,

Looking at the datasheet the active part of the attenuator runs from a
control voltage of .3v to 2v with a loss of 5db to 60db. Is that 5db
baseline loss acceptable after the 45Mhz filter?

tim ab0wr

On Mon, 26 Mar 2018 13:56:01 -0700
"Jerry Gaffke via Groups.Io" <jgaffke@...> wrote:

A BAP64Q pin diode attenuator gives 60dB of dynamic
range:??/g/BITX20/message/32066 Attenuating back in
RF?gets around the limited dynamic range that Henning points out in
the first post of that thread. Note that the control voltage is
inverted with respect to the 2n7002 FET, higher voltages give less
attenuation. Could get a slightly lower noise figure for the receiver
if the attenuator is inserted at a later stage of the RF chain.

The BAP64Q is relatively expensive at $0.50 single piece,
the frugal among us will note it's down at $0.20 if you buy a few
thousand. Mouser and Digikey both stock it, Mouser points you to the
wrong BAP64* datasheet. There are other similar small signal pin
diode attenuators out there from other manufacturers.

Jerry, KE7ER

On Mon, Mar 26, 2018 at 01:18 pm, Don, ND6T wrote:

The "problem" isn't the detection, rather the range of a single
control element (the MOSFET in this instance). 20 dB is pretty good
but I shall always crave more.