Quadrature Demodulation at 90 KHz

naumank · Feb 19, 2016

Hello all

I have a linear frequency modulated (LFM) signal having 90 KHz center frequency and 20 KHz bandwidth around it. My problem is to demodulate it at base band to get I and Q both (phase information required). I have searched a lot at internet but mostly found quadrature demodulator ICs have starting input frequency range in MHz. AD8333 was what i found having lowest input frequency but even it does not work for 90 KHz (as i discussed with analog devices technical staff).

Can any one kindly guide me how to solve my problem?.

Thanks in advance

Best Regards
Nauman

Tony Stewart · Feb 20, 2016

https://discriminator.nl/ic/

You will need a to make a custom 90kHz IF filter.

Mikebits · Feb 20, 2016

Maybe the guys at Analog Devices were being lazy the day you called because I found one of their devices that meets your needs AD8339 (It is quad mixer also, but it is DC - 50MHz and you can leave unused mixers simply terminated just to keep them happy). https://www.analog.com/media/en/technical-documentation/evaluation-documentation/AD8339.pdf

naumank · Feb 20, 2016

Thanks for your concern. Actually problem was the minimum frequency of sine wave at 4LOP pin which is 2MHz in case of AD8333. In AD8339, minimum frequency range for sine wave at 4LOP is not define or i could not found it in its data sheet.

In case of using square wave at 4LOP pin, will not the harmonics/noise be significant in I/Q output?

Tony Stewart · Feb 21, 2016

https://jaunty-electronics.com/blog/2012/09/how-it-works-quadrature-detector/

Lots of choices for IC's
https://discriminator.nl/ic/

But with a a wide ratio of deviation to carrier ratio (1/3 vs <1%) you must have good LO rejection and signal noise filter without group delay distortion on band edges.

If this is a challenge for you then you must have high SNR to begin with.
e.g. IF's in AM radio's use 455kHz for 7 to 15 kHz BW or 11 MHz for FM etc.

Make sense?

Mikebits · Feb 21, 2016

naumank said:
Thanks for your concern. Actually problem was the minimum frequency of sine wave at 4LOP pin which is 2MHz in case of AD8333. In AD8339, minimum frequency range for sine wave at 4LOP is not define or i could not found it in its data sheet.

In case of using square wave at 4LOP pin, will not the harmonics/noise be significant in I/Q output?

From page 3 of the data sheet:
Local Oscillator (LO) Frequency Range: Min - .01MHz Max- 200MHz

For LO leakage a filter is normally used. A active filter would be easy to implement.
The mixer (analog multiplier) is a Gilbert Cell topology. You could build your own using matched transistors for the differential pairs. I think there are transistor array IC's.

naumank · Feb 22, 2016

Thanks. Is not this LO frequency range for square wave or it is also applicable for sine wave also?

Mikebits · Feb 22, 2016

naumank said:
Thanks. Is not this LO frequency range for square wave or it is also applicable for sine wave also?

This directly from the data sheet:
LO Inputs. No internal bias; optimally biased by an LVDS driver. For best performance, these inputs should be driven differentially. If driven by a single-ended sine wave at 4LOP or 4LON, the signal level should be >0 dBm (50 Ω) with external bias resistors.

Naumank, I don't mind helping you, but please read the data sheet first before posting your questions, as most of the answers to your questions can be found there. If you can't be bothered to read the data sheet then why should I. If your understanding of English is a problem, then I completely understand that and in that case will continue to assist you.

You seem to be overly concerned with using a square wave for a LO. In the case of a Gilbert Cell (most quad mixers), the LO acts as a switch control so a square wave works best for this application. Besides, even if you input a sine wave, it will become a square wave after the internal LO divider. I suggest this again, you should research the Gilbert Cell Mixer.
Have a look at the following link to get an idea of how these mixers work. The video has a good explanation.

naumank · Feb 23, 2016

Mikebits said:
This directly from the data sheet:
LO Inputs. No internal bias; optimally biased by an LVDS driver. For best performance, these inputs should be driven differentially. If driven by a single-ended sine wave at 4LOP or 4LON, the signal level should be >0 dBm (50 Ω) with external bias resistors.

Besides, even if you input a sine wave, it will become a square wave after the internal LO divider. I suggest this again, you should research the Gilbert Cell Mixer.

Thanks for pointing towards right part of data sheet. I actually read this part but as i could not found any mentioning of sine wave frequency range explicitly (as given in AD8333 data sheet), i was confused.

As for as your second point is concerned, it means that if we apply square wave at 4LOP pin or sine wave at this pin, the I/Q outputs will be same. Am i right?

Mikebits · Feb 23, 2016

I almost missed your question as it was hidden in the quote (not a good idea to do that). Now I am going to answer your question with a question.

1. As for as your second point is concerned, it means that if we apply square wave at 4LOP pin or sine wave at this pin, the I/Q outputs will be same. Am i right?

Looking at the block diagram, what do you think?
Also be advised, the LO must be driven differentially as per the data sheet.

LO INPUT
The LO input is a high speed, fully differential analog input that responds to differences in the input levels (and not logic levels). The LO inputs can be driven with a low common-mode voltage amplifier, such as the National Semiconductor® DS90C401 LVDS driver. The graph in Figure 22 shows the range of common-mode voltages. Logic families such as TTL or CMOS are unsuitable for direct coupling to the LO input.

See the eval board schematic.

naumank · Feb 24, 2016

Thanks for all your help. I think i need to study its data sheet more in-depth.

Mikebits · Feb 24, 2016

I can't emphasis this enough, try and learn about the Gilbert cell mixer, in fact try and build one, you will learn a great deal by doing so.

Mikebits · Feb 29, 2016

I guess I need to work on my teaching style, I ran the guy off.

naumank · May 9, 2016

Mikebits said:
I guess I need to work on my teaching style, I ran the guy off.

Hello

Your teaching style is perfect, actually i was unfortunately busy in other matters. I have few more queries for you can kindly provide some answers.

1. How can i simulate these ICs?. I have Orcad 9.2 but it has no support for this IC. Is there any free simulation software for this type of ICs?

2. This question may be stupid but i am wondering that this demodulation can be done using principle of "under sampling" i.e. chose the sampling frequency (lower than the nyquist criteria) such that alias of desired frequency (e.g. 420 K Hz) comes in low frequency region?

Thanks

Dick Cappels · May 17, 2016

As an alternative, you might want to try Dan Tayloe's quadrature detector

https://www.google.co.th/url?sa=t&r...3a.pdf&usg=AFQjCNFyRZEd_YqA63XpziWRH5n6iWbI6g.

Quadrature Demodulation at 90 KHz

naumank

Member

Tony Stewart

Well-Known Member

Mikebits

Well-Known Member

naumank

Member

Tony Stewart

Well-Known Member

Mikebits

Well-Known Member

naumank

Member

Mikebits

Well-Known Member

naumank

Member

Mikebits

Well-Known Member

naumank

Member

Mikebits

Well-Known Member

Mikebits

Well-Known Member

naumank

Member

Dick Cappels

Active Member

Similar threads

Quadrature Demodulation at 90 KHz

Member

Well-Known Member

Well-Known Member

Member

Well-Known Member

Well-Known Member

Member

Well-Known Member

Member

Well-Known Member

Member

Well-Known Member

Well-Known Member

Member

Active Member

Similar threads

Privacy & Transparency

Privacy & Transparency