MRA

[poynt99]

In response to DS's request for input on the MRA, I'm submitting a bit of circuit design I did a while ago.

The document contains the designs for two low power (relatively clean) sine-wave oscillators intended to drive the MRA, but the circuits can be used anywhere a sine oscillator is required. Also included is a circuit for a low power square/pulse generator and a conceptual method and arrangement for COP>1 testing.

Those skilled in the art can certainly take the designs and with some minor modifications, make the frequency range wider and/or the amplitude range larger. These circuits are cheap to build and about as simple as it gets.

.99

[darkspeed]

Thanks .99

[MileHigh]

There is a cool chip, a voltage-controlled oscillator that also outputs a very decent sine wave:

http://pdf1.alldatasheet.com/datasheet-pdf/view/67443/INTERSIL/ICL8038.html

>>>  Sorry, I fought to drill down to the real link, but just checked and it doesn't work.  I don't know why these datasheet web sites make you click like crazy hoping to get the actual .pdf but that's life.  Click and read and after a few attempts you will find the datasheet.  Sheesh.

I know it as the "XR8038."  It's an old chip from the 70s.

There is another chip, but the name escapes me now.  The LM-something.  I may come to me later.

A good search is "VCO" or "VCO chip" or "voltage-controlled oscillator."

MileHigh

P.S.:  A voltage-controlled oscillator chip is a really cool thing.  You just connect up a pot set up to sweep through a certain voltage range and you adjust the frequency like that.  That means a lower-frequency oscillator voltage output could be connected to the main VCO input to modulate the frequency of the main VCO output and you can go crazy if you want to.

[poynt99]

Hi MH.

The 8038 and 2206 (probably the other one you were thinking about) never were very good oscillators imho. The distortion could be trimmed to a point, but they consume a relatively large amount of current. They are ok for a poor man's function generator, but I avoided them years ago after noting the poor sine quality.

A large part of the design I did went toward low power consumption so that COP>1 testing was more readily available. Low distortion is also a bonus with the Wein bridge topology.

.99

[BEP]

Use the 8038ACP and go to 200kHz  

THD <1% and very low power consumption with a clean build.

The chip is hard to beat, for simplicity and flexibility. I have three of them in one project  

[darkspeed]

I have some MAX038 chips. how about those?

[BEP]

I would say that beats the 8038, just a bit 

[darkspeed]

The evaluation board cam out when you know who was doing you know what back in 94  :

http://www.selectronic.fr/includes_selectronic/pdf/Maxim/MAX038CCP.pdf

[BEP]

I wonder if 'you know who' held off from the MAX038 for the same reasons I did? My max theoretical injection frequency is 175kHz.

[poynt99]

Use the 8038ACP and go to 200kHz  

THD <1% and very low power consumption with a clean build.

The chip is hard to beat, for simplicity and flexibility. I have three of them in one project  

It's a difference of opinion I guess.

Simple, Cheap Oscillator P_D ~ 10mW
xx8038 P_D from ~ 240mW (min) to ~ 400mW (max)

There's no comparison in P_D, and the Wein Bridge design distortion (properly built) will always out-do the all-in-one chips by a mile. 

DS, the MAX038 was a great chip, and you're lucky to have a few. They go for big bucks now days. I seriously contemplated re-engineering a discrete design version of the MAX038 some time ago, but then when I saw the potential DDS chips had, I quickly came to my senses

If you're looking for a truly low power, clean, cheap, and simple sine wave oscillator, it's pretty hard to beat the two in that document, imho.

.99

[darkspeed]

Good point..

They will stay in their box and one day go to some collector in mint condition... 

[BEP]

.99,

I'll agree with you on power usage and clean output.

The truth is.... I rarely use anything more complex than a Fet. It is rare that I need it 

I've been known to whip-up an oscillator with a reverse connected 2n2222 and a couple of passive components. My favorite was the NE2 relaxation oscillator (Ring NRO).

Three NE2's, a few caps and resistors and weird operation. Place them in a circle. The rotation direction is random.

One direction and remove power. The blasted things flash in reverse with no power connected(one time around). The other direction, they stop when the power is removed.
Never figured that one...

[MileHigh]

Well thanks everybody, Poynt especially for the information.  You are right, the 2206!

We used to use them as signal generators for doing LASER light shows.  I am shocked because it was 30 years ago.  I feel so out of it sometimes!  Like I assume that the "IC Master" book doesn't mean too much to some of you out there.  I have to assume they stopped printing it years ago.  They are phasing out the White Pages also, I just heard that on the radio yesterday.

Poynt's designs are more interesting and like he said they are better.  You actually build it yourself instead of having a chip do it for you.

It's interesting because it is a real challenge to produce a variable-frequency sine wave generator.  It's really really hard to do and a pure sine wave is extremely important for doing a serious analysis of how a circuit responds.  You want the purest possible sine wave because then it looks like a single infinitely thin spike on your spectrum analyzer (or two spikes if you count the negative frequency).  That way you can plot the response of the circuit vs. frequency.  Once you know that then you in theory can know exactly know how the circuit will respond to any arbitrary excitation waveform, including square waves.  And on the contrary, square waves are about the worst excitation signal you can feed into a circuit to see how it responds to changing frequency.

MileHigh

[BEP]

 You actually build it yourself instead of having a chip do it for you.

It's interesting because it is a real challenge to produce a variable-frequency sine wave generator.  It's really really hard to do and a pure sine wave is extremely important for doing a serious analysis of how a circuit responds.  You want the purest possible sine wave because then it looks like a single infinitely thin spike on your spectrum analyzer (or two spikes if you count the negative frequency).  That way you can plot the response of the circuit vs. frequency.  Once you know that then you in theory can know exactly know how the circuit will respond to any arbitrary excitation waveform, including square waves.  And on the contrary, square waves are about the worst excitation signal you can feed into a circuit to see how it responds to changing frequency.

MH,

Wow! Now I'm impressed. You really are with it 

[Paul-R]

It's interesting because it is a real challenge to produce a variable-frequency sine wave generator.

Is there an argument for thinking in terms of an analogue solution rather than a digital one?
Paul-R

[poynt99]

If one has the capability and the tools, imo, no. Digital would be better for many applications.

There are pros and cons to both analog and digital, so it depends on the application.

.99