Re: Mixer question

by dare4444 Thu Mar 25, 2021 10:31 pm

Ladder filter is gone. The phase splitter has two bifilar windings. In spice, even when there was no coupling, it still outputted 0 and 90 degree RF. Can be driven by a sine wave source. The two mosfets are also fed 0 and 90 phase shift audio from TL072 opamps connected to laptop. The 1.13uh value will give the 90 degree required shift on 40m. It can be squeezed or stretched a little for good SSB shape and the bifilar winding can be wound on 1cm ballpen plastic core. The 90 degree signals are combined.. one sideband is cancelled out and we have ssb. What do you think?

by dare4444 Sun Mar 21, 2021 9:27 pm

Paper design. Soldering iron cold haha it's strange. This should output 100mW

by dare4444 Tue Mar 16, 2021 7:28 pm

Tested today. I used 680ohm resistors from 7404 to the diode switch. It's picking up CW transmission. The 22pf trimmer nulled out LO at RF port.

5/1360ohm = 3.7ma shared by 3 diodes each side when on.

The TX for testing is a simple 5V 7.2MHz single transistor oscillator powered by Arduino nano and coded with CW word - Te s t i n g and kept in my kitchen with 1 foot antenna.

by dare4444 Tue Mar 16, 2021 4:55 pm

Hi Ivan,
I guess it's 1.4V peak for 2 diodes conducting,
1.4 * 1.4 / 200ohm load x 2
= 5mW / 0.7v per diode
= 7ma per diode = 14ma for 2 diodes conducting out of 4. It's coming down to 7ma of peak RF current for full turn ON.

4 diodes conducting out of 8 = current of 28ma

Am I right?

3 x CD4001 NOR gates on each side with 180ohm series resistor would drive the 4 diodes out of 8. Each gate can supply 10ma current max at 12V.

The transistor totem pole driver could be removed then. Two CD4001 would be required.

The CD series are dirt cheap and widely available.

by **Ivan** Tue Mar 16, 2021 3:57 pm

Hi,
there is no reason to thank to me. We help each other here.

I have never dealt with ring mixers to such details, I am sorry. Generally, signal levels in the devices I developed were very low, so RF currents were never an issue. Maybe someone else in this forum knows more.

VBR from Ivan

by dare4444 Mon Mar 15, 2021 3:08 pm

"Now it is a ring modulator with the LO transformer replaced by a H bridge. Are you sure the current buffers are necessary?"

Thank you for taking the time to reply and your valuable guidance. I added the buffers in order to sink 10V / 180ohm = 55ma / 4 diodes = approx 13ma per diode to switch it on hard.

What's is the RF peak current per diode in a four diode DBM?

by dare4444 Mon Mar 15, 2021 3:04 pm

The article does give port isolation figures. Half H mode looks interesting for my HF transceiver where the IF is 24MHz. Any LO leakage at the IF port should be knocked out by the ladder filter.

by **Ivan** Mon Mar 15, 2021 12:19 pm

Hi Dare,

"Check the above article"
It is quite much brain food! The article is interesting, but it uses one criterion - the upper limit of the dynamic range (IP3 and 1 dB compression points). It does not take other parameters into account, such as noise, port isolation etc.

Much time has passed since I played with mixers. My favourite type was a Gilbert cell on integrated bipolar transistors.

This is a rough schematic of HF bands SSB transceiver. I need to design the VFO/BFO and RF amp section. Check the product detector, it's a diode ring driven by a 74HC74. The front end is a double balanced switching mixer with Mosfets. The IF amp has 10dB gain at 24MHz.
Why not?

I've a new IF amp with feedback so different transistors can be used. Ic is 22ma and gain is 15dB. Too much gain? Can just one IF amplifier be used after the ladder filter?
It is hard to judge this way. As long as the IF amplifier is stable, linear and keeps good noise factor, the gain can be increased.

As per your suggestion I cut the number of diodes to four. Then I added extra diodes in series in an attempt to increase IP3. It's now powdered by 12V with current buffers and driven by CD4001 NOR gates. It's max trigger frequency is around 15MHz (propagation delay = 65nS). 90ohm resistor on each side limits the current. CD4001 can deliver up to 8ma at 12V regulated. Does this look good? Your thoughts?

Now it is a ring modulator with the LO transformer replaced by a H bridge. Are you sure the current buffers are necessary?

VBR from Ivan

by dare4444 Sun Mar 14, 2021 9:11 pm

Hi Ivan,

As per your suggestion I cut the number of diodes to four. Then I added extra diodes in series in an attempt to increase IP3. It's now powdered by 12V with current buffers and driven by CD4001 NOR gates. It's max trigger frequency is around 15MHz (propagation delay = 65nS). 90ohm resistor on each side limits the current. CD4001 can deliver up to 8ma at 12V regulated. Does this look good? Your thoughts?

by dare4444 Sat Mar 13, 2021 7:26 pm

I see, Ivan. Mixers are fascinating.

https://www.hindawi.com/journals/aee/2016/8039679/

Check the above article, using a pair of 4 diodes in rectifier type mixer is better?

This is a rough schematic of HF bands SSB transceiver. I need to design the VFO/BFO and RF amp section. Check the product detector, it's a diode ring driven by a 74HC74. The front end is a double balanced switching mixer with Mosfets. The IF amp has 10dB gain at 24MHz.

I've a new IF amp with feedback so different transistors can be used. Ic is 22ma and gain is 15dB. Too much gain? Can just one IF amplifier be used after the ladder filter?

In the new IF amp, max input is 20mV peak with 180mV peak output, second harmonic is -40dBc. At 0.3uV input, output is 1.8uV. Voltage gain remains the same at 6.

by **Ivan** Sat Mar 13, 2021 9:20 am

Hi Dare,
you use two ring mixers in your design. IMHO it is a real overkill. One diode ring would be sufficient in its typical circuit. It seems to be possible to omit the diodes at all, too. I would use open drain invertors in that case to avoid injecting Vcc to the transformer. It is just a raw idea, not proven to work.

VBR from Ivan OK1SIP

by dare4444 Sat Mar 13, 2021 12:16 am

Are the diodes really needed? We can just connect the output of 74hc04 to the transformer winding.

by dare4444 Fri Mar 12, 2021 11:16 pm

It's a switching (mode) mixer. Is this configuration going to work? I've seen online versions where each half of the transformer is grounded by an electronic switch like 74HC4066. I've put some diodes driven by a 74HC04.

by Sponsored content