NE602 NE612 double balanced mixer - amplitude modulation (AM) generation

[SouthPark]

Hi all! Anybody here used these Gilbert cell double-balanced mixers before for generating AM signals?

I have managed to generate AM signals with the NE612 very nicely. However, I don't know if my set-up is the 'correct' approach (even though it is working very nicely).

The situation is - the data sheets seem to indicate that the internal differential amplifier circuits are internally biased. And, usually a double balanced mixer takes in AC voltage inputs. And the data sheet doesn't seem to indicate how we can add some DC voltage level to the low frequency AC input signal (eg. audio signal).

For generating AM, we usually need to add a certain amount of DC to the message signal before mixing this combo with the local oscillator (carrier) signal.

Anybody know if the NE602 NE612 actually generates AM by default, without needing us to actually do anything extra (ie. no need for us to add our own DC to the message)?

Details about the NE602 and NE612 mention ability to generate AM, but I was assuming these ICs are meant to produce double sideband (DSB-SC, no carrier) by default. And I assumed that it would be necessary to somehow follow some procedure to offset the message with a DC voltage in order to then generate AM. So far, I haven't found procedures for setting a DC offset on the message.

So I thought I'd try my luck on EEVblog forums to see if anybody has generated AM with the NE602/612, and share some of the nice advice about generating AM with this IC - namely, where do we apply our own DC offset to the AC audio message signal? Obviously if the inputs have capacitors on the inputs, we can't add DC before the capacitors. But we also can't add DC after the AC coupling capacitor - because the data sheets say that external bias should not be applied.

Thanks in advance!

[w2aew]

Even though the datasheet says not to apply any DC bias to the RF inputs, you can safely apply a small DC offset to re-insert the carrier and get AM instead of DSB-SC.  Just keep the signal levels small, etc.

You may have already seen my video on the Gilbert Cell and DSB-SC / AM signal generation:

[Zero999]

What's the difference between the NE602 and NE612? I've looked at the data sheets but they seem the same to me.

[w2aew]

They are basically the same - now...

The longer story is:  Signetics initially designed the NE602 for the 45MHz FM wireless telephone market.  Some time later, the design was refreshed to allow the internal oscillator to operate up to 200MHz and the RF path to work to 500MHz.  This part was designated the NE612.  The '612 was intended to replace the '602. But, customer demand for the '602 forced Signetics to label some devices as 602 and some as 612, even though they were now the same part.  This is why the current datasheets list the same specs for both.

[SouthPark]

Thanks very much Alan! This particular situation had me stumped. I was thinking of deliberately applying a tad of offset at the immediate base input (for offsetting the message), but didn't go ahead with it - due to the data sheet. But I'll definitely try that out now that you kindly indicated it can be done. I'll just add a tad, following your recommendation.

Btw..... it's a big world, but yet a small one! I'm the same one that very recently asked you about the long tailed pair - regarding the Q2 turning off for relatively small rise of voltage at the base of Q1.  I didn't know you were on this forum too. Very nice to catch you here too! And thanks for helping me.

For the NE612, I was only able to get the AM signal after following some very old forum discussion (which I wasn't involved in) about putting a potentiometer (eg 500K pot) between pin 1 and 2, with the wiper part of the pot connected to ground. This was mainly to fix up the asymmetry between first half and second half cycles of the AM signal (like due to uneven gains on each half). The pot certainly helped, but I think the number of turns on my pot wasn't high enough (ie. I used less than 500K pot).... so it was very sensitive around the spot where I needed to turn it. After that, I noticed some small AC signal getting through to the output when I had the message signal connected (for the case with no LO signal applied). So I used that same potentiometer to null it out as much as possible. And then, after that, I applied LO and message signals (with no external DC offset applied), and out came nice clean symmetric AM signal. The screenshot I loaded was for the case where I reduced the modulation depth, as it was at around 100 percent when I first noticed that I was getting AM at the output. I'm thinking that changed internal bias due to the external pot has allowed the AM to come through. But I was really preferring to add my own DC offset. Thanks again for helping and showing what to do Alan. I certainly will be looking at the Gilbert cell video! I hadn't come to that one yet, but I noticed it, and it was going to be the next part of watching, after understanding what you taught about the long-tailed pair. Thanks Alan!