EMRFD Message Archive 12091

Message Date From Subject
12091 2015-12-28 04:59:29 peter_dl8ov High Level RF Mixing
All of the work I can find on mixers assumes that I am working on a receiver and that the signal level on at least one of the ports is at a very low level. So what about transmitters? Assume that I have two signals at (say) 10.7 MHz and 3.6 MHz, what is the best technology to use when combining them to create a 7.1 MHz signal?

Peter DL8OV

12093 2015-12-28 11:49:04 iq_rx Re: High Level RF Mixing
Mixers have been used in transmitter signal paths for nearly a century.  There are many examples in EMRFD.  Generally the same rules apply--a good mixer for transmitters is governed by the same rules as for receivers.  Which is best depends on specifics of your application.  A good start for your reading might be VHF transverters.

These days the regulations for spurious outputs are tight enough that most of us who design transmitters using mixers in the signal path routinely use a spectrum analyzer on the bench.  I built one years ago using guidance from Wes and Terry White (the W7ZOI Spectrum Analyzer).

I learned a lot about shielding, high level mixing, signal path integrity etc. from that spectrum analyzer project.  After that, designing and building clean transmitters was easy.

Best Regards,

Rick KK7B
12096 2015-12-28 19:16:51 Ashhar Farhan Re: High Level RF Mixing
This is a tricky question. Let's pick up a diode mixer. the standard issue  ADE-1. You will want high level output so you can drive the linear chain well enough  and you want to keep the IMD down.
At +15dbm IIP3, if we shoot for -50dbc of IMD, the input has to be kept at -10dbm. The output is going to be around -17dbm. That is way down low. From there to the 1 watt driver level, you will need 47dbm of gain. It IMD will rapidly build up in each stage.
Most transmitter portions regularly achieve -30dbc of in-band distortion, but the extra 10db at -40dbc makes a sparkling difference to the quality of your voice on the air. This needs a careful control of number of stages, gain and losses. If you have too much loss in the following bandpass filter or attenuation (that is needed to keep the chain stable) you  will end up adding more distortion after the mixer.

Probably what does make sense is to use a higher level mixer. 

Two 2N7000s in a KISS configuration are my current favourites. At +30bm IIP3, this mixer can output at least 0dbm without significant IMD. This is followed by a +30dbm post mix amplifier and then the band pass filter. You will need fairly good shielding around the band pass filter. For me, only air coils work at these levels. With the output of over 10 dbm from this band pass filter, you can afford to drown the signal down by a 6db pad before feeding it to the linear chain.

The classic design for a low distortion transmitter is the KK7B's T2. I havent heard anything else like that. It isnt' just the IIP3. There are by-products from the microphone onwards that need dealing with.

- f

12097 2015-12-29 01:55:09 peter_dl8ov Re: High Level RF Mixing
Thank you for the replies so far but everyone appears to be missing the original point of the question. I wish to mix high level signals, maybe 3V peak to peak from two oscillators, can this be done or do I need to drop the levels down to 0dB and then amplify again?

Peter DL8OV
12098 2015-12-29 03:35:03 kb1gmx Re: High Level RF Mixing
Oddly my favorite is the classic CA3028 dif-pair makes a good high level single balanced mixer
and can be done with discrete devices too.  

Oddly the 602/612 produced a good high level output and fairly clean.  Just watch the 
input levels. its hard to hate that 15db of gain.

Diode mixer for TX is not so bad.  Try using one with 1n914/4148s as they need more 
drive but produce more output.  They work well with about 12-15dbm drive.  

The problem with Kiss with 2n7000 is second harmonic and third harmonics in the output.
I tried it with a 9008 monolythic pair of low RDS-on mosfets not much better.  The D-G 
capacitance is a problem even on that part (very low).

Switching mixers using bus switches do well but are upper frequency limited.

At VHF and up diode rings are hard to beat.

12099 2015-12-29 05:18:20 neomag_magneo Re: High Level RF Mixing

According to my knowledge (and experience), good linearity with diode mixers can be obtained only if the other signal (e.g. ssb from the balanced modulator) is considerably smaller (20 dB or less) than the high level signal (e.g. VFO). It means that very often one needs to attenuate the lower level signal first, then mix it with high level signal, do the bandpass filtering and start amplification from the level which may be 10 dB smaller than the original. In this respect the ssb transmitter chain is no different from the rx if chain. 

Heikki (OH2LZI)

12100 2015-12-29 05:22:03 Russell Shaw Re: High Level RF Mixing
12101 2015-12-29 05:30:33 Russell Shaw Re: High Level RF Mixing
12102 2015-12-29 05:47:16 Russell Shaw Re: High Level RF Mixing
12103 2015-12-29 05:52:46 kb1gmx Re: High Level RF Mixing
For diode mixers that's true.   The bounds for that are covered in EMRFD.

You can push a little but I've found if the SSB signal is more (larger than) 
than 10db below the LO distortion grows.  Generally that means making 
sure there is solid drive to the mixer (level 7 with  7dbm applied or more).

With active mixers the proportions may be very different.   So some testing 
is required if there dis limited data available.

12104 2015-12-29 05:55:28 Chris Trask Re: High Level RF Mixing
> The problem with Kiss with 2n7000 is second harmonic and third harmonics
> in the output. I tried it with a 9008 monolythic pair of low RDS-on mosfets
> not much better. The D-G capacitance is a problem even on that part (very low).

When I first designed the KISS mixer I used discrete MOSFETs, but they did not turn on quickly enough so as to produce good IMD performance. I set that idea aside and used a few of the dual bus switches as others had done and found the FSA5157 to give the best performance.


,----------------------. High Performance Mixers and
/ What's all this \ Amplifiers for RF Communications
/ extinct stuff, anyhow? /
\ _______,--------------' Chris Trask / N7ZWY
12109 2015-12-30 06:01:08 i7swx Re: High Level RF Mixing
Hi Peter,

There is an interesting mixer that could be used, a digital mixer, using a D flip-flop. This was published on "Ham Radio" magazine beginning of 70s.
Having 3 Vpp signal, the solution could be simple (bias net), the highest frequency should be sent to the clock and the lowest to the data, the output will be square and the difference of the two frequencies.

There is one problem, data Fm (3.6MHz) should be < than clock FM (10.7MHz), and IF out (Q) should be less than half clock FM, so the required IF of 7.1MHz is not possible. 74AC74, at CMOS output. I tried this digital mixer beginning of 90s using 74S74, it was working but I had an IF respecting the limit vs clock.

73 and a Happy New Year 2016 to all


12112 2015-12-30 08:57:32 Ashhar Farhan Re: High Level RF Mixing
One could even use an XOR gate. all these produce fairly high degree of spurs.

- f