Mini DIY signal generator

[PRR]

Many links have rotted. Here is the project on Merlin's site:
http://www.valvewizard.co.uk/siggen.html

[PRR]

> What *does* determine the maximum ƒ?

When the open-loop gain is not >> than required closed-loop gain. Because of the resistance ratios, you may need a really fast opamp to get out of the audio band.

However Merlin has far more hands-on experience with this plan and may have further insight.

[tubegeek]

Thanks, PRR. I think it's a TL072 per the schematic, maybe try a 4558? Certainly swapping a chip is within my skill set. I'll look at the op amp comparison shootout page in The Art for clues....

[tubegeek]

... seems like the TL072 is not the worst choice based on the features? "High" slew rate of 16, not sure what I'd find that'd better. I hope Merlin has a thought....

[merlinb]

I didn't measure level BTW, just sent it into my laptop so I could count peaks per second and determine ƒ's. I was using a 96K sampling rate in my audio interface so I should have seen something above 15K if it was there, I guess.

Hmm I'm not convinced a standard laptop soundcard is going to have a bandwidth beyond 20kHz?

[tubegeek]

It's an external (USB) interface, only a cheap Behringer 2-in model but I'd expect it to go higher than 15K with a 96K sample rate. It's a reasonable question though - I'll hook the mini sig gen up to a scope (or run a different sig gen into the same interface) to see if I can narrow that question down. And using the scope I can get a reasonably accurate voltage measurement too.

[PRR]

'4558 is *slow*. Errors (lack of excess gain) will mount above 1kHz. (It will "pass audio" much higher, but that's an easier job.)

I've seen $13 USB sound "cards" pass 19KHz fine, but I agree it is a point to check. A classic method is to jack line-out to line-in and run tones. I have used RMAA but it may be getting old. PassMark SoundCheck is reputable but I'm not sure what tests it does.

[tubegeek]

A classic method is to jack line-out to line-in and run tones.

Great suggestion. Tried that with the computer playing tones out through the interface in one program, jacked into the interface and recording in another program. 50K was no problem, that's where I stopped.

With a fresh battery, I was able to get the Mini Sig Gen as high as 23K and record it but after that the signal again dropped precipitously. Does that tell us anything?

[Kevin Mitchell]

Here's a layout of the bare bone version.
Link to PDF with transfer layout and bottom view



-KM

[charbot]

Hey, old thread but Im just posting this for posterity... incase anyone stumbles across this looking for info:   I just built one of these, working great, but I found one small error on the DIY PCB layout:   the frequency pot is wired backwards.  The response is "normal "(clockwise=up/higher) and all the range is crammed in the last portion of the dial.    It is "correct" in the schematic and the perfboard diagram, but on the PCB, pins 2 & 3 are used to set the freq, but it should be 1 & 2, to get the intended behavior (pot to work "backwards" and use the log pot as a reverse log to spread out the freq in a more linear manner.)   The fix is to simply cut the connecting trace between pins 1 & 2, then wire a jumper between pins 2 & 3. Now using a log 10k pot gives the "reverse" response (clockwise= down/lower) and the freq are spread evenly.   Alternately, you could use a 10K C rev log pot, (but the knob behavior will still be "normal",  reversed from what is on the panel graphics)


I hope that this helps future builders of this great little project. 

[R.G.]

Given the stuff Electric Druid has done with multi-waveform LFOs, I keep wondering why he doesn't do a signal generator.

I did have to come up with a quick and not-too-dirty sine for a tuner recently, and fell back onto a walking ring counter ala Don Lancaster, which just worked. I keep getting tempted to just sit down and write out the code for an NCO=>CWG oscillator controlled by an analog pot-voltage read in on the A-D input on a PIC, expecially now that NCOs and CWGs are inside most PICs. Then I remember how much else I'm not getting done that I really do have to get done.

An interesting variant would be to run the PIC from a crystal and have the frequencies be only lowest-cents-error musical note frequencies, again read from an A-D pot input.

I think it would be relatively straighforward, given how much work Tom has already put into waveform lookups.

[merlinb]

I just built one of these, working great, but I found one small error on the DIY PCB layout:   the frequency pot is wired backwards. 

"All this does have a down side; the tuning becomes roughly exponential. In other words, the frequency does not vary evenly as you vary the tuning pot. Ideally we would need a pot with an anti-exponential taper (good luck). As a compromise I used a log pot backwards, so turning it anticlockwise increases the frequency."

I'm not sure which PCB layout you're referring to. On mine, the pots are mounted on the opposite side from the other components.
https://valvewizard.co.uk/siggen.html

[charbot]

I'm not sure which PCB layout you're referring to. On mine, the pots are mounted on the opposite side from the other components.
https://valvewizard.co.uk/siggen.html

Mine is built exactly like it is shown on your site (the sine gen pdf)  with the potentiometers and rotary switch mounted to the solder side of the board, in the same orientation(freq pot is upside down).   Weird.   ... maybe I reversed something else?   it not a big problem, its working fine w/ an audio taper pot after that minor modification..

Merlin, While I have your ear, I just want to say thanks for all the great projects and for continuing to maintain your site... lots of valuable info there.   I just finished a few of your other projects and they all work great:  Glass Blower(amazing!),  the Audio Burst generator on a DIY dual-sided PCB  (this one was a little gnarly to build and troubleshoot, but its now working!...,also FYI, to anyone who wants to build one: one of the layout sides in the PDF (topside, I think)  needs to be mirrored.)   Coming soon... an Engineers thumb.