Anharmonic Tremolo - yes another harmonic tremolo, but with op-amps

[pinkjimiphoton]

can't wait to see it... i heard a demo of baja's vibrotrem on youtube i think, but this one sounds like, REAL close to the "real deal"....

anxiously awaiting!!

[samhay]

Not being entirely confident with my layout, I thought I would verify (or verofy?) it first. The LFO works, but there is a bug in the audio section, so I need to spend a little more time with the fine-tooth comb.

I did find another error in the schematic, which crept in when I re-drew it - C5 should be 10n not 22n. Not a major problem, but it will have a flatter (almost Butterworth) response this way. The schem in the first post and the Eagle file have both been updated.

[samhay]

Here's a vero layout - it works, but I haven't squeezed it into a 1590B yet. Sorry about all the jumpers and the odd vertical resistor - I had a hard time laying this out how I wanted it. Anybody and everybody is encouraged to better my efforts.
The LFO section is designed to be pot-mounted and sit at the non-foot switch end of the box. I know this will fit. It might also just fit in a more traditional orientation.
The audio board should float over the volume and frequency pots and hopefully will fit if you put the jacks or foot switch near the bottom - sorry, there is not likely to be room for a battery.
I set the space between the LFO pots to be quite wide so that I can fit a switch between each pot. You may need to use smallish switches or lay it out differently.

The circuit is reasonably easy to get working. Setting the trimmer is quite easy to do by ear - you want the volume to be similar when the frequency pot is at each extreme of its rotation - at one end, you will have volume with the LED on, and the other end when it is off. You may have to play with R28 and R29 (LED resistors for the vactrol) to get a pleasant duty cycle.
I currently have a home-rolled LED/LDR working with 1k resistors at both R28 and R29 (as marked on the vero layout). I may replace this with a Silonex NSL-23SR3 'vactrol' and will probably have to bump R28 and R29 up a bit if I do.

The frequency pot works by decreasing the crossover frequency as you turn the pot clockwise. If you use a C100k pot, then wire it backwards (pin 1 <-> pin 3).

Edit - you may have to increase the value of R23 (1k restor to pin 3 of the rate pot) as the LFO may lock up near max speed (unless you see this as a feature). I need to investigate this a little further.

Edit - I left room in the top right corner for a trimmer, so if you don't want a frequency pot, you can set-and-forget the volume. There is a volume boost when using the 'treble' function, so I think a volume pot is probably a good idea. YMMV.

Edit - oh and there is one significant change from the schematic. I changed the order of the audio op-amps. The input and output op-amps are now one dual package (a TL072), while U1b and U2a are now the other dual package - which I used an NE5532 for. I found the layout easier this way and it gives you the best of both worlds - the TL072's fet input impedance and the low(er) noise of the NE5532 for the filtering. However, you don't need to use an NE5532 - anything with a TL072 pinout should work.

[pinkjimiphoton]

 

looks sweet!! gonna try and give 'er a go!!

[samhay]

Hope you like it Jimi.

I gave the LFO values a tweak last night - it seems that the dodgy connections on the breadboard were helping with its mojo.
1k R to pin 3 of the rate pot -> 1.8k (stopped the LFO locking up)
220k R -> 680k R (gave a better range of depth)
22u C -> 10u (gave a better range of LFO speeds.

Should get it somewhat boxed up this evening, so I guess we shall see if it all fits.

[samhay]

Here are a couple of simulations of the frequency response of the circuit with the frequency pot set to 10k. Both assume that the LDR can swing an order of magnitude either side of it's resistance when the depth is 0. This is probably a little over-optimistic, but you get the idea.
Of note - and in response to a thread by tca (http://www.diystompboxes.com/smfforum/index.php?topic=103114.0) - is that in this circuit the crossover frequency moves around as it pans. The simulations look pretty ugly, but it sounds pretty good...

The first sim is 'anharmonic' mode and the second 'treble' mode.

Edit - I should probably stress that this circuits works very differently than the original Fender design, which may well have a much cleaner crossover frequency.

[drolo]

Thanks for posting this, Sam. The sims are very interesting.

I had thought that the circuit would pan between a fixed HP and a fixed LP filter.
But in your sim, if I understand it right, it is varying the amount of filtering of both passes out of phase from each other, but not panning between them. (is that correct?)

[samhay]

No problem. The crossover frequency of the HP and LP filters is fixed (by the frequency dual-gang pot and the 10n and 22n caps). However, when we pan between them, we are using a voltage divider consisting of an LDR and a trimmer. This is the least parts-count approach (for other options, it might help to take a look at RG's 'panning for fun' article), but it has some drawbacks, one of which is that the amplitude of both signals change during panning. In this case, this causes the apparent crossover frequency to move all over the place. I thought long and hard about whether to do the panning in a 'better' way, but decided against it for the time being as it would double the circuit size.

[samhay]

Well I finally got it boxed up. It is a bit of a mess, as I have been tweaking as I go, but it fits in 1590B (with a few trick borrowed from my 1590A builds) and works as intended. .The wiring is all a lot longer than I would generally like, but I kept it like this so I could readily pop the boards in and out. Despite the rat's nest, it is pretty quiet - the LFO noise is all but absent.
The audio board is identical to that shown in the schematic and vero layout, but a few of the LFO values have changed again, and I decided to drop the square wave switch as it wasn't adding much. These are all essentially 'tweaking to taste' changes, but I will update the schem etc shortly.

For anybody interested in building it - be warned. As-is, it is a squeeze and you will probably have a lot more fun nicely laying it out in a larger enclosure. That said, a better (PCB) layout would help a lot. I could do with some help on this front. I will probably get round to doing this one day, but I will next be trying to get it to play nice with the TAPFLO chip.

[jtn191]

I'm interested in building either this or Jon's Xardinal Trem...any way to get rid of LFO noise or switch to normal tremolo? Your demo sounds great

[samhay]

I realised after building this and playing with it for a while that my direct recording setup doesn't have a very good earth - USB powered audio interface to a laptop - so it can get noisy with LFOs that throw hash onto the ground. If I play my Anharmonic tremolo through my amp, which is grounded via a 3 prong power plug, the LFO noise is very difficult to hear - certainly no worse than a Tremulus Lune and far better than an EA tremolo.

I would suggest that if you are interested, you put the circuit together on a breadboard and see how it sounds. The LFO noise will likely be at its worst in this case, so if it is ok like that, it should be fine in a build.

[blackieNYC]

What is the p-p voltage of your LFO? I'm trying to get a square/triangle option on another trem, and the square is much hotter. If i voltage divide it down it crashes after about 5 seconds.  Not sure I understand what your switch is doing.  Do you attenuate the square to match the triangle? Thanks

[samhay]

^What is the p-p voltage of your LFO?
The square and triangle LFO voltage swings should be similar and you can change the square wave amplitude by varying the value of R22.

However, a square and triangle LFO with the same amplitude will probably sound like they have different depths when used in e.g. a tremolo.

[Chugs]

Anyone have the schematic for this? Seems to have vanished.

[samhay]

Sorry - this is on my list of stuff to migrate from my dropbox public folder.

Schematic:
https://www.dropbox.com/s/43wbr6cu6ejobwr/Anharmonic_Tremolo.jpg?dl=0

Vero layout:
https://www.dropbox.com/s/1drqi9af31eoa5l/Anharmonic_Tremolo_Vero.jpg?dl=0

(this is a pretty tight fit in a 1590B):
https://www.dropbox.com/s/rpetyli12usrqaq/Anharmonic_Tremolo_guts.jpg?dl=0

[Chugs]

Thanks! Much appreciated.

You posted some interesting ideas for LED drivers in this thread. Any chance of reattaching those?

http://www.diystompboxes.com/smfforum/index.php?topic=102551.0

[deadastronaut]

''Sorry - this is on my list of stuff to migrate from my dropbox public folder.''


yep, know the feeling..

[samhay]

Thanks! Much appreciated.

You posted some interesting ideas for LED drivers in this thread. Any chance of reattaching those?

http://www.diystompboxes.com/smfforum/index.php?topic=102551.0

No problem.

For this post:
http://www.diystompboxes.com/smfforum/index.php?topic=102551.msg909902#msg909902

I think this is the relevant picture:
https://www.dropbox.com/s/lib4ge1mvqjjbyy/LuneLFOLEDDrivers.jpg?dl=0

I'm not sure how well 'B' will work, but 'C' is how it works in the Anharmonic tremolo.
However, to make 'C' work well, you have to make R3 large enough to not load down the LFO circuitry too much - i.e. probably 100's k or more.