NE602/612 vs 74HC4053?

Discussion in 'Homebrew and Kit Projects' started by KB1CKT, Sep 8, 2021.

ad: L-HROutlet
ad: l-rl
ad: L-MFJ
ad: L-Geochron
ad: Left-2
ad: Left-3
ad: MessiPaoloni-1
ad: abrind-2
ad: HRDLLC-2
  1. KB1CKT

    KB1CKT Ham Member QRZ Page

    Looking to get into homebrewing (bought the intro to RF design years ago, but life got in the way). I've built some kits over the years, and have done some ugly construction, so not unfamiliar with things here.

    The '602/612 chip is of course well established in the QRP community, no wonder as it has gain and a built-in oscillator, so it works great for low parts count and low power consumption. But it also is known for not handling large signals--fine for after a filter for conversion to baseband, but on the front end... it can suffer.

    I know there is the Dan Tayloe mixer, but I recently saw that KD1JV has a mixer using a 74HC4053 on the DCxxB series of transceivers. Might be in use elsewhere. Appears to run the RF differentially through it, kinda cool. My only concern with that is the high on resistance, but maybe that's not a big thing, a mixer is usually lossy. Might not be good beyond HF either, but it might be easier than a diode mixer and maybe it handles the big signals better than the 602/612?

    So the question is, where does this mixer fit in? Between 602 and a diode mixer? I think I have parts in the mail for making both '4053 and diode mixers, so I could always build both I suppose, but I'm wondering if others have settled this question.
  2. KB1CKT

    KB1CKT Ham Member QRZ Page

    It strikes me that the 74HC4053 might not have the intermodulation products that a diode mixer has: the FET on resistances shouldn't change (much?) but if they did, it'd be a linear thing. Unlike a diode where a large RF signal basically changes the diode voltage, which is exponential. BUT if this was true then we'd see more radios built this way. Beats me--but I might try my hand with '4053 first, see how it works. Only thing is, I don't have lab equipment to really test intermod performance, so maybe it's a flat wash.
  3. KV4PD

    KV4PD Ham Member QRZ Page

    No, it's always about empirical testing. Then you know what you've got.
  4. KB1CKT

    KB1CKT Ham Member QRZ Page

    True, not sure what I was thinking when I typed that. Still curious what others have determined about these two approaches.
  5. W1VT

    W1VT Ham Member QRZ Page

    Testing really important for IMD performance. It is quite possible to have all the "right parts" and screw things up with a bad design choice.

    Building two signal sources and a hybrid combiner isn't all that hard. These days you can look at the output on a computer soundcard based audio spectrum analyzer.
    Some homebrewers may be able to use previously built hardware as one or both of the signal sources. Merely dropping the voltage on the final amp may allow 100% duty cycle operation with minimal effort.

    Zak W1VT
  6. VK2CLF

    VK2CLF QRZ Member

    The NE602 (which is now really SA602/612 - original NE602 masks lost in a plant fire) are Gilbert cell mixers, which means that on the LO port they like switching levels, but on the other you have the logarithmic B-E junction. If you don't keep the signals small you get a lot of intermod. Regarding SA612, the data sheet was written to replace the NE6*2, but the actual performance is much better; the mixer itself will give acceptable performance at 900MHz, quite contrary to the data sheet limit. The LO won't easily run that high, of course, but you can inject an LO from elsewhere.

    The CMOS switch mixers stay linear at much higher levels, but suffer from charge injection from the gate switching transients - so you get a lot of harmonics of the LO on the output. Not much intermod because the through path is capable of much higher signal levels, and often consists of back-to-back devices whose resistance curves tend to cancel each other.
    US7IGN likes this.
  7. KL7AJ

    KL7AJ Ham Member QRZ Page

    I've been building Lock-In amplifiers using 4066 bilateral switches. The ~4053 is basically the same thing, but with a lot of extra coding you probably don't need. Of course, there's no gain in either of these....but for dynamic range, you probably don't want that anyway....use a little gain up front (and a LOT of gain after the fact, HI!)
  8. KB1CKT

    KB1CKT Ham Member QRZ Page

    Haven't used the 4066, have used the 405x (but at DC). Tempted to buy one of the DCxxB kits and use that for a basis, start with one and modify away, kinda hard to not buy some of the low cost kits rather than homebrewing everything (career is kinda in the way too).
  9. KX4OM

    KX4OM Ham Member QRZ Page

    I'm familiar with the 74HC4053 as a 1 to 2 multiplexer/switch used as pairs in some of KD1JV's transceiver designs, such as the 80/75 meter SSB simple transceiver designs. He uses them to switch the signal path for the IF, including the SSB filter between receive and transmit. I know that they have been employed as switch-mode mixers, but they are not as good (their on high resistance) as a CMOS CD4066 or the subsequent fast bus switches like the FST3253. I bought some of those after an online discussion with Wes, W7ZOI some years ago. The FST series were the first switch mode mixers, the development of which I read in SPRAT several years ago.

    Ted, KX4OM
    US7IGN likes this.
  10. KB1CKT

    KB1CKT Ham Member QRZ Page

    Is '4066 lower Rdson? I grabbed the first CD4066 datasheet, which isn't always the right thing to do; I suspect '4066 is easier to use in a 12V system. Without spending some time in the datasheets I'm not sure which switches faster and/or has lower crosstalk.

    CD4066 Datasheet. In a 5V system it's pretty high, but quickly drops with some decent rails.

    74HC405x datasheet. Not so bad; datasheet indicates for the non-HCT version that you could go as high as 10V on VCC, so it might get really low Rdson, but I suspect you will need more than 5V for a high logic level in order to make it switch if you do that (maybe run a voltage doubler? with an 8V rail? then a DDS/PLL at 3.3V might be enough signal for good switching, just a wild guess).

    edit; not doubler, just capacitively couple and maybe dc bias to half rail.
    Last edited: Sep 11, 2021
    US7IGN likes this.

Share This Page

ad: w5yi