…and possibly a circuit schematic recommendation.

A disclaimer here. I am not an EE major, though I have dabbled in electrical and electronic compositions. By no means am I an expert, nor am I ready to advance to microcontrollers due to the expense of the investment in time and decent equipment cost. Also I am a Mac user, and most microcontroller H/W I've seen advertised usually interfaces with PC's either as an option or as the main programming interface. I'll leave the microcontroller battle for another day.

For now, I wish to assemble a function generator, possibly from transistors and associated RLC components. I have seen Wein bridge circuit schematics, and understand their appeal for sine wave accuracy and good stability, so I am leaning towards that type. However, I don't know which designs are best or elegant in their simplicity. More importantly, none of the schematics I have seen provide the maximum frequency output. Of course, I'm aware that RLC component values affect this, but also the type of transistors or IC's (or tubes, even) limit the the top frequency, or at least the quality of the waveform. Furthermore, I am clueless as to the secret of said transistor/IC/ tube correlation to the top frequency. As stated in the subject line, The max I would like to achieve is 2Mhz, though it doesn't matter if it goes higher. I would like to be able to vary the output frequency with a rotary knob or two or three or four, etc., and switch(es) depending on if I need to stage the frequency bands, starting as low as I can. One hertz, maybe? Output voltage is not critical, about 1 and 1/2 to 3 volts is adequate. Amperage can be trivial. Although sine wave generation is important, I would also like to generate square waves. This could possibly be some components the sine wave feeds into. Triangle and sawtooth would be icing on the cake, but not required at this point.


Any takers?

Thanks in advance,

Mark

 

Well, the Max038 was the way to do that - http://www.maxim-ic.com/quick_view2.cfm/qv_pk/1257.

But, like the audio frequency ICL8038, they're getting hard to find.

 

Yeah, I noticed the Max038 disappearing trick. Not only are they hard to find, but pricey too. This is one of the reasons I'm looking into a more old fashioned approach.

Thanks anyway.

Mark

 

This kit will go from 1Hz to 1MHz:
http://www.electronickits.com/kit/complete/surface/ck1301.htm
From page:
http://www.electronickits.com/kit/complete/complete.htm

Not bad for $30.
Kitsrus used to have a MAX038-based signal gen as DIY101, but I heard that the plant that made that IC was destroyed years ago in a typhoon. The kit didn't come close to utilizing the MAX038's potential. You might luck out and find one at an electronics swap meet or the like; perhaps some new old stock in a mom & pop store.

 

Say, that looks interesting. I wonder if the kit contains a Max038 chip. Hopefully, they still have some left. Well, at any rate, good call. Thanks Sarge.

Mark

 

No, I can assure you that 1Hz-1MHz kit doesn't have a MAX038 in it, or they would be selling it for a LOT more than they are.

Just the IC's alone have been fetching more than the price of the kit in online auctions.

If you're looking for something in the range of 100kHz-450MHz, Elenco has units (not kits) they're selling for under $300. Not fancy, but in the world of test equipment, it's very inexpensive for a brand-new unit.

 

Just my luck, they are out of them anyway. Sniff.

Mark

 

I just ordered the FG-500 kit from electronickits. Even though it sez 100 KHz, I looked at the specs and lo and behold, THEY say “Frequency: 1Hz - 1MHz in 6 decade steps with variable ranges.” I downloaded the pdf for it, and it uses the XR-2206 chip. So I got a data sheet for that, and by golly, the frequency range is from .01 to more than 1 MHz. A little more pricey, but it looks like it will fit the bill. I know very little about the chips available, and I certainly didn't know about the XR-2206. Just dumb luck that I looked at the misleadingly named kit.

Mark

 

The XR2206 might still be available as an NTE892 for an horrific price.
ICL8038/NTE864 may still be available. NTE stuff is ALWAYS far more expensive than the original item. I think the ICL8038 was developed as a response to the MAX038, as if it could be a comparable item. It wasn't even close! The MAX038 is more than an order of magnitude faster than the ICL8038.

You can pick up ten original LM3914's (10-step analog level detector) online for less than ONE NTE1507 at retail, which is their equivalent replacement.

But at least they're available.

 

I believe the ICL8038 was in production before Maxim was in existence. I think the price took an increase from NTE just last year. I had been getting them for around $8.00, which was less than the ICL8038CCJD ceramic packages. Exar chips have been hard to find for a very long time.

 

Gee, I just discovered Jameco has the XR2206 for $3.59 each.
https://www.jameco.com/webapp/wcs/s...storeId=10001&catalogId=10001&productId=34972

But it's running out of steam real fast if you get anywhere near 1MHz.

 

The 1 MHz is a compromise, but I got the FG500 kit anyway, because I was afraid there might not be anything faster at a reasonable price. Maybe there is a chip similar to the XR2206 that goes to 2 MHz, but like I said I know so little about the available types.

Sarge! Thats a great find, price-wise. Looks like I should have waited.

Mark

 

Well, you're getting a complete kit! Having all the parts, a pre-made PCB and assembly instructions will save you lots of time and tinkering, not to mention shipping costs from different vendors you might've had to rack up doing it piecemeal.

Kits are a great way to go when you're first getting started; you have a better-than-average chance of getting it working right away. Sure, the function generator IC is a basic building block, but you need a number of other parts connected to get anything useful accomplished with it.

It's a little different when you've been accumulating parts for decades - but even then, they're always coming out with new stuff.

Once you get into microcontrollers, you'll find that you can make a really spiffy arbitrary waveform generator by just burning the waveforms into tables, and use that to drive a DAC at up to whatever speed the MCU is capable of.

 

Of course you are right on all points, Sarge. I don't feel so bad about it. But it is also good to know I can get replacement chips at a low price should I blow the one in the kit.

Mark

 

You will find that with all these function generator chips, the output becomes increasingly rounded as you approach the upper frequency limits. So the square,triangle and sine waves look increasingly similar on a scope.

You will also find that if you calibrate your instrument in terms of period, not frequency you can get several accurate and stable decades of steps using current programming of the XR2206 or 8038 generators.

 

Hi,


The ICL8038 (made by Intersil) is from way back in the seventies, just about any electronics mag and kit producer did a project with that chip but most of them made a bad job of it - all those function generator chips, (including the XR2206, XR2207 etc.) need a well designed output buffer and a good power supply.

Exar then made a clone of ICL8038 and called it XR8038.
Maxim made the MAX038 as a substantially beefed up version of the same chip later (around the early nineties IIRC)
What is left of Both Intersil and Exar are under the same holding these days.

The reason they haven't designed a new one is not so hard to see, when everybody with a need is using Direct Digital Synthesis (DDS) - only the amateurs have any use for dinosaur chips, since it's way cheaper and better to use DDS - Some DDS chips of present days can be used for the same purpose as the old analog ICL/XR/MAX-types (and then some), if a µcontroller app. is not justified/wanted/whatever, but I fear they will dissapear soon(ish), since µcontrollers are cheaper than DDS-chips.

 

Hi,

You will also find that if you calibrate your instrument in terms of period, not frequency you can get several accurate and stable decades of steps using current programming of the XR2206 or 8038 generators.

I fail to see what f contra 1/f calibration has to do with stability??
That is just a matter of preference of user interface.
Stability in the range is determined by a precise and stable reference for the FM input, not how you read (or smoke ) your pot.

 

Well, as luck would have it, I stumbled onto yet some MORE information. Apparently the TLC555 IC will go up to 2 MHz. Man, finding information is sometimes like pulling hens' teeth.

Mark

 

Tell us more about you goals; you may get more useful suggestions than just discussing any proposed chips.

What functions are important?
What purity of waveform is required?
What frequency range is required?
What degradation can be tolerated at the range extremes?

 

I do appreciate your contributions, studiot. Not to offend, but as I mentioned in my first post, sine and square wave forms are important, referenced the Wein bridge circuit due to it's accuracy and stability (so as much purity as I can achieve reasonably), and gave the frequency range of 1 Hz to 2 MHz.

But true, I did not mention the amount of degradation I can tolerate in my first post. I honestly don't know if I am able to quantify this, other than to say as little degradation as is reasonably possible.

I've read that Hewlitt-Wein bridge circuits were often used as the sine wave signal generator in much early tube radio equipment, so I suspect a Wein bridge circuit could meet both the accuracy and degradation criteria through the frequency range *IF* the transistors or tubes or other IC's and passive components chosen have specs that, when used in concert, would allow such an achievement. I just don't have that component knowledge. And I don't have a good knowledge of or savvy with the plethora of IC's available. Clearly, during the discourse of this thread, there was (is?) at least one chip that looked sweet and simple to use without actually building a Wein bridge, but now, it is an endangered species and costly if found. Truly, I don't mind building a Wein bridge since it really doesn't look that complicated, but the thread hasn't unfolded in that direction, and that's O.K. too. Hey, I'll take whatever info I can get. Of course, there is at least one disadvantage of the Wein bridge, which is having to add circuitry for the square wave option. But again, if it is simple enough and accurate enough, I don't mind.

Mark