SOLVED - Signal generator idea ~ will it work?

[antenna]

I like working on CB radios but I never do the receiver alignments because I do not have a signal generator that can output that low of a modulated signal. My plan was to take an old 23ch CB, install a 1kHz oscillator, disable and bypass the final amp and route the buffer output to an adjustable attenuator and out to the coax jack.  I would then repurpose one of the knobs for modulation and another for attenuation.

I figured with the huge amount of attenuation I need, a pi pad with the one of the shunt resistors equal to the buffer impedance and the other 50ohm with the series resistor being a potentiometer of some high value.  Is this a bad idea? Should I be using some other means of attenuation? Varactor diodes maybe? I'm clueless, any suggestions would be appreciated.

[David Hess]

It takes more attenuation than a single stage will allow, and 100% shielded cables to keep the signal from leaking around the attenuators.

[radiolistener]

You can use some old signal generator which supports AM/FM modulation.

I figured with the huge amount of attenuation I need, a pi pad with the one of the shunt resistors equal to the buffer impedance and the other 50ohm with the series resistor being a potentiometer of some high value.  Is this a bad idea?

Yes, this is bad idea, because potentiometer has inductive impedance. And it will require several potentiometers and mechanic to rotate it together in order to keep 50 Ω impedance on the input and output of the attenuator.

You can use standard 60 dB attenuator for your needs. If you're needs more weak signal for testing sensitivity. You can add second 60 dB attenuator, so the total attenuation will be 120 dB. For precise power tune you can add more 1-2-3-6-10 dB attenuators for more precise tune. Since you exclude power amplifier, the output is low power. If it is less than 1 W you can just use standard 2 W attenuators.

Another way is to use variable attenuator from an old signal generator, but they are very expensive.

[antenna]

What if I terminated the radio in a dummy load with a capacitive sampler tee'd into the line? No coax, just solid connectors to the dummy load...

[radiolistener]

No, capacitive and inductive elements are not allowed if you're want to get flat frequency response. You're needs to use resistors.

There is also some note, if you're planning to bypass power amplifier, it needs to make sure that power amplifier use 50 Ω input. Because integrated power amplifier may use different impedance for transmission line between driver amplifier and power amplifier. In such case you're needs to take it into account for attenuator.

What is your goal for that generator? What power level you're want to get?

Also you can use powerful attenuator on the output of power amplifier in order to get the same result. It doesn't require any change in your radio, but powerful attenuator is expensive.

For example:
https://www.aliexpress.com/item/1005002293316978.html

You can add standard low power attenuators on the output of powerful attenuator to get precise required power value.
But if your power amplifier is too powerful this soulution is not ideal, because it will dissipate a lot of power into heat and powerful attenuator will be too expensive. The output power may vary due to high temperature change.

[antenna]

It does not have to have a flat response. I intend to use it on one frequency for receiver alignments on cb radios. The signal I need is 27.185MHz, 30% amplitude modulation at 1kHz, 50ohm output level variable between -110dBm and -130dBm. 

Edit: When I put that amount of attenuation into a calculator, the shunt resistors are so close to the impedance that one may as well assume they match.  If I assume the driver that fed the final was 100mW, that is 20dBm. I then need 130 to 150dB attenuation to get down to where the adjustments need to be made in the radio. Being the series resistor of the pi network will be so large, I assume I can make the other two match whatever circuit impedance is there (like 50ohm on one side and 600 on the other, or whatever it may be).  I am not looking for accuracy or figures, I simply want to be able to drive the receivers while decreasing the power until the signal cannot be detected.

[bob91343]

You are wasting time and energy trying to repurpose a junky radio as a signal generator.  Break down and get yourself a decent generator and be done with it.  Short cuts can turn into Pandora's boxes.

[radiolistener]

It does not have to have a flat response. I intend to use it on one frequency for receiver alignments on cb radios. The signal I need is 27.185MHz, 30% amplitude modulation at 1kHz, 50ohm output level variable between -110dBm and -130dBm.

This is a waste of time. Just use signal generator, for example Chinese FY6800 cost about 100 USD and give you much better precise and more function for other needs.

If you want more cheap solution you can build frequency source on cheap 3 USD si5351 module and some blue-pill or arduino MCU module for 2 USD. So the total cost will be about 5 USD

[antenna]

I decided to follow the advice here and start looking into better signal generators.

For the fun of it, just burning time this evening, I cut the power to the final, removed the collector pin, bridged the buffer to the output network without a single thought to matching them, and the output now provides a clean sine wave (not modulating yet) at about 20.3dBm. The signal looks nice, and so far nothing gets warm and the output is still giving me roughly 50ohm.  Then, using a coax switch that shorts the unused antenna port, I connected the radio to a 50ohm load. The shorted switch port (assuming that shorted contact is a pickup loop) was connected to a 30dB attenuator I had on hand, and now im down to -72.5dBm into the spectrum analyzer. I am almost there!!!! Yes, totally wasting my time, agreed.  It's been fun seeing how close I can get with whats available though.

[HB9EVI]

simple attenuators, specially for low power <=5W can be built by yourself; just don't exceed 30dB per attenuator, much more cannot be properly achieved with simple parts (crosstalk a.s.o. ) ; stack 3x 30dB in series and you're good to go

[David Hess]

I have done exactly what antenna is asking.  I use a 20 dB high power coaxial attenuator, and then the switched attenuator shown below.  I had to make up a set of doubled shielded coaxial patch cables and put the transmitter inside a metal enclosure to keep the signal from leaking around the attenuators.

[CaptDon]

How many CB's do you run into that need the receiver aligned?? Aligning a crystal plexer unit
that someone has diddled with is a pain in the butt. Most of the out of alignment chicken
bander units that I have come across were basket cases as some idiot tried to get more
power out of them. Isn't it amazing how when morons de-tune the second harmonic trap
they get another watt on the wattmeter. What a shame that extra watt is at 54mhz!!!
I got some radios that came to my bench for repair when someone "Tightened all those
loose things in the silver cans". You can align the receiver for a peak 'S' reading and
seemingly make it more sensitive than it was from the factory, but generally what you
have actually done is narrow the I.F. bandwidth by going for 'peak' and in some radios
that makes the receive and transmit audio sound like crap.
 

[Yansi]

Use just the proper low level signal RF generator.  It is not that expensive to get some basic one from ebay for the frequency range you are interested in.

[antenna]

Again, I intend to purchase a good one. I'll edit the title as solved. 

As for having fun with this, I decided to make a variable attenuator based on crosstalk using bean cans cut and soldered together. Sure, nowhere near linear, but at 27MHz, I get 50-70dB attenuation. Both sides are terminated in 50ohm and there is a window between the two that can be adjusted by rotating the cans thus controlling how much RF leaks into the adjacent side. I wonder if this would work for close range fox hunting....

[radiolistener]

I decided to make a variable attenuator based on crosstalk using bean cans cut and soldered together.

Such construction will suffers from a bad VSWR at 27 MHz.
I have old-school measurement dummy load in similar implementation. With usual 2W resistors. And it has about VSWR about 1.5 at 27 MHz. Another instance uses RF-grade 2W resistors С2-10 and it works much better with VSWR about 1.1

The main issue here is a resistor and it's long wires. In your case it also  use twisted wires, which will have even more inductance.

Such construction is usable for 3-7 MHz, but 27 MHz is too high.

[antenna]

SWR 1.19 at 50MHz. Looking in the other side was about the same.