The Psychotenuse Tremolo... reverse sawtooth tremolo with very few parts

[moid]

I wanna be free, I wanna be free to do what I wanna do! And I wanna get (capacitance) loaded. And I wanna have a good time! ...and now the man is telling me that it doesn't matter what order I put my LEDs in? Heavy vibes man!

Ahem, thanks for that information, I didn't know that about series components - in this case I suspect I was wrong and just jumbled the notes up that I wrote as I was testing different LEDs. Now onto my botheration of your circuit. I took it apart, rebuilt it as you suggested and got nothing - I cannot get any LEDs to blink without using the values I used. I tried different Trig LEDs, putting in the ultra bright blue and white ones that didn't blink (I didn't test them for sound because I assumed no blink = no trem) but they don't blink and neither do the orange and red LEDs. I did get one combination of two ultra bright blue LEDs where one of them remained off and one did blink (just visibly) and they did trigger the red LEDs, but they ticked like crazy so I rejected them.

I even tried sticking the LDR onto the Trig LEDs instead of the Load LED - which was interesting, it produced a sawtooth trem (but not reversed) and one that had much longer duration (from sound off to sound on). The blue LEDS do fade up in brightness and take 200 - 300% longer than the Load LEDs do, so I got a better trem from them. The downside is the trem doesn't cut all volume (it's more subtle) and it's not reverse sawtooth but it's interesting in it's own right. It makes me wonder if I need the same transistor as you?

I did swap the 100uF cap (C1) for a 220uF - this did reduce the ticking a bit. I tried higher values (up to 1000uF but the amount of ticking remained the same as far as I could hear (I used moid jnr's better ears as well, he thought the same as me). Ticking is not much of an issue - it only occurs when moid jnr slams the circuit with one of his favourite powerchords, if he plays at a more sedate volume there is no click at all. I can remove the issue entirely by recording at less than maximum volume input in my DI box (which is what I normally do when I record) so I don't think we need to worry about this anymore

So all that probably means I screwed up somewhere on the breadboard so I made a nice clean layout of what is on my breadboard - would you mind telling me which part I stuck in the wrong place so we can all laugh at my inability to read schematics and I can fix it please? Thankee kindly

[duck_arse]

fly! be free! free from the burdons of waveshapes and response times and charge dumping. and suppliers that will cut all the legs off your parts. and then lose those legs.

I can only assume your Q1 pinout cause my datasheet is for smd. otherwise, your layout seems correct. (I jumped my LOAD and the scr across the bb channel to seperate, neaten.) I found some things. the freq pot value follows the R1 value in order to get any worthwhile range. and you need to tune R1 right to get the nice waveform - the stepped one I mentioned DUMPS off the leds, then they step on dim, then they sweep up to DUMP again ..... so the leds are much more on time then off, and the off is so abrupt the ldr(s) will bother to follow.

and even the goodest wave is shaped not so good for ldr messing, cause it dumps off so quick, and then is on for proportionally too long. but I tried it with a couple of real/own rolled ldr's and watched the voltage dividing in action. some strange results, with oddly over-the-hump then off waves. but - an NSL-32 and red as load and two tinted blue not-superbrites as TRIG gave a pretty good ramping sawtooth on the cro.

but if it's a ramping sawtooth you want, and you don't want to be messing with the low gainers in the other discrete thryistor mess (three transistors), then the dual opamp and diodes circuit is the easy reliable known way to go.

[moid]

Yes - the pinout should be fine - if I reverse the transistor on the board everything dies; it only works the one way round, and if I switch any of the legs, as it were, either nothing happens or the LEDs are constantly lit, so I'm confident I got that right.

I agree about R1 and the freq pot - I found that too when I was trying different values (which is why I got on better with the 20K pot and the 7K5 resistor).

I tried your Red LED /  NSL 32 combo suggestion, but it ticks much more and doesn't improve the length of the pulse at all... I even tried different LEDs in place of the red one and they all ticked, but some made the trem  softer (as I had found before). For me the amber LED plus Red LED / LDR combo works best. I can only assume either that I am doing something very odd or there is a genuine variation between your C106D and the BT169... maybe I should buy one?

I think I can see what you mean about using LEDs and LDRs - the way the LED charges up (both in this circuit and in others where I've made LED LDR combos) is not in a linear sort of fashion, and this does affect the sound of the audio a lot... I will look into a dual op amp / diodes idea (I will try that circuit that had variable rise/fall first, it looks interesting)

record scratch! we interrupt this broadcast to bring you the following news:

OK I did something clever / stupid / electrically inadvisable this evening. I was musing about how cool it would be to have a random stutter to the tremolo and was wondering how on earth could I do that, and then my mind goes a wandering, as it does and I thought maybe the audio signal of the guitar itself could disrupt the flicker pattern of the Trig LEDs? So I did an experiment. I plugged my guitar into a splitter box, took one of the outs and plugged that into the tremolo audio in. Then I took the other out from the splitter and plugged it into a booster pedal and then (with the help of crocodile clips and breadboard cables) connected the Tip to the +leg of the first Trig LED and the Sleeve to the ground of the board, strummed the guitar and..... nothing happened

But I don't give up as easy at that see? Especially when I've got a great idea and I decide it has to work, so I start plugging the tip of the boosted signal into different parts of the circuit but nothing does anything until I connect the signal with lug 2 of the freq pot... and oh boy, then I created a kind of envelope controlled stutter trem! The harder you strum (or the louder - to a point*) the signal is, the faster the tremolo goes - as a note decays the trem slows down. Did I walk around the room feeling like a god? Did I proclaim my intellect to the world? Did I even consider contacting ACME to get a business card with Wile E Coyote (Genius) printed on it? Well maybe some of those things, but either way I made something cool

* point being that with experimentation of my boost pedal, when the audio from the boost was too loud, the LEDs stayed flickering fast but dimly too much and the resulting output audio was very low and the trem effect disappeared (I think it went too fast to be heard). Adding distortion made things worse if it was too high and the volume was too high as well, and adding extra bass made things worse as well - the best setting was about a quarter turn of the volume pot and some gain or no gain.

Now because doing this is almost certainly a bad idea circuit wise, there are a few issues that may or may not be fixable... the circuit ticks spectacularly now It maybe that that is one of the issues with this sort of circuit and I have to live with it (because the pedal stutters slightly randomly it's not so bad, but if there is a clever way to control that (maybe some sort of diode or a cap to ground somewhere between the boost pedal and the freq pot?) that would be awesome. I wonder if this sort of idea would work on any optical trem? My brain thinks yes! So maybe it would work with a different tremolo that doesn't tick?

Also there is a pronounced hum (pronouced HHUUUUMMMMM very loudly) that is way too loud. The audio coming out of the pedal is a lot quieter.

Then I tried something else - I found that a larger pot, which normally would have a chunk of it's sweep do nothing to the blink effect worked for mostly the entire sweep with the extra boost audio plugged into it, and with a slower (anti clockwise) pot setting in the fist place the strumming could go from slow to fast with more of a noticeable range of speeds. Occasionally the LEDs would lock up though - twiddling the pot reset them back to blinking, but that might be an issue (or just a mass of cables shorting against something).

Then I got weird and wondered what would happen it I stuck a second BAT69 transistor on the board right behind the first one? Well a second one modulates the sound a bit - it adds a narrow pulse to the tremolo that runs about 2-3 times faster than the main trem speed - it's not massively loud, so I thought hmmm - try other transistors instead? So a BC337 and a 2N3904 did the next audition and both broke up the trem more, with the 2n3904 doing the best job (maybe other transistors would be better?). This part isn't that important but I had just reached the peak of my 'infinite number of monkeys with an infinite number of breadboards' behaviour and was having fun. Not sure if that will go in the final circuit or not. So I'm thinking I need to add a buffer/splitter to the front of the circuit, with one part going to the circuit as usual and the other part going to a simple loud boost (probably that 386 circuit I've used before), and then some sort of clean boost afterwards...

You know I bet you want a schematic of this madness don't you? Well I'll have to keep you in suspense because I doubt I'll get to it before the weekend sadly.

Anyway we now return you to your schedule programme




 

[duck_arse]

yes. what?

I must point out again that I havee not had any this connected to audio, so I dunno nuffing baht no ticks or cliks. also - they-who-shall-remain-nameless-ikey show C106 in several variants in their catledog. and they aren't 2 for 10$ like on the bay that shall remain nameless.

[moid]

Well here I goes and breaks open my wallet (sudden cloud of rising moths) let's go wild and buy a few! I've found the C106D for sale at a good seller for cheap (£0.38 each) so I'll order some and we'll see what happens. I just need to make an order of things I don't have / haven't got enough of (that shouldn't be hard!)

[duck_arse]

... what madness is this? one magik ctrl key mistake and the post dissappears? bloody thing.

as I was saying. while you are messing with the diode at the gate doing blocking and stuff, you could try replace that diode with whatever first transistor is laying on the bench in front of you. put new Q C to old diode A, new Q E to scr gate, and add a 1k resistor to new Q B. then do whatever silly you want into the transistor base, see what happens. obvs, without any/enough silly at the new base, the old scr won't trigger.

and if it doesn't werk - you ain't seen me, right?

[moid]

Well that made my scratch my head...and I thought of a couple of different ways of interpreting your wisdom o oracle! Does you mean do dis? Cos the old diode didn't connect to an anode; it went to the gate of the BT169.


Or somefink else?
First transistor is a BC337

The pinout of the BT169 is Cathode at the top, Gate in the middle and Anode at the bottom of this layout.

edit: I tried this (in the name of science!) and none of the LEDs light up. Nice true bypass signal though... but no trem. So probably not what you said?

[duck_arse]

put new Q C to old diode A, new Q E to scr gate, and add a 1k resistor to new Q B.

Does you mean do dis? Cos the old diode didn't connect to an anode; it went to the gate of the BT169.

hurrah. I knew I could confond you with the worded connections. now read on!



I meant like shown. but and, I'm about 100% sure it won't work like that, cause the npn won't leak any current for charging things. so maybe a pnp, reversed, as also shown, might have a better chance of doing something. mind, any time I mess with the other [pnp] polarity, things overheat, sometimes smoke.

not in a good way.

[moid]

Now before you go a-looking at the below pickchurr... I wants you to remember that I am a bear of very little brain, and to be honest your schematic (thanks for drawing it) confused me as much as the written explanation of it. I made an assumption that the resistor that goes to the X symbol means power. I wasn't sure how the BT169 was connected at all, and it seemed that you kept the diode in that I thought you removed in the written description... and then I tried NPN and PNP transistors (the BC337 and a 2N3702) in the below configuration (I tried reversing them as well). Most of the time the LEDs would light up but not blink, and some of the time they did not light up at all... and if I was a betting man, I would be betting on me making a complete pig's ear of your drawing...



No it's OK, you can stop laughing now Or crying...

But at least there was no smoke! And the C106Ds are on order (hopefully they will arrive around Wednesday)

[duck_arse]

Then I got weird and wondered what would happen it I stuck a second BAT69 transistor on the board right behind the first one? Well a second one modulates the sound a bit - it adds a narrow pulse to the tremolo that runs about 2-3 times faster than the main trem speed - it's not massively loud, so I thought hmmm - try other transistors instead? So a BC337 and a 2N3904 did the next audition and both broke up the trem more, with the 2n3904 doing the best job (maybe other transistors would be better?). This part isn't that important but I had just reached the peak of my 'infinite number of monkeys with an infinite number of breadboards' behaviour and was having fun.

can I speak with one of the monkeys, please?

see that paragraph, you were putting signal of some sort into a transistor which was modulating etc etc? well, the X in my dia is/are indicating wirepads, as a shorthand, to other parts of the circuit [which I can't see - don't really care about] which would be some sort of modulating signal [like you had before], not supply or ground. AND - in my diagram I have an npn connected as per my words description, but as I say, I think it won't work. the pnp shown floating in that diagram is indicating how I would replace the npn - so as to avoid prevent any confusion.

oh, what's that you say?

the diode in the trigger string is removed - the pnp transistor collector goes where the diodes Anode was, the pnp's emitter goes where the diodes Kathode was.

[moid]

can I speak with one of the monkeys, please?

Oook? ook. ook. ooOOOK!!  *Sounds of a banana phone ringing, faint jungle cries, a voice far away is singing "I'm the king of the swingers..."

Hello, Jungle VIP here, how about I lay a tall tale of woe on you mancub? I wanna be like you (oo-oo!) shoobidoobidoobidoo! I wanna walk like you, solder like you (oo-oo!)... Well today I finally had a few moments to not actually do any work and so tried to rebuild the entire circuit using your original schematic and the C106D transistor... with the results that the trig LEDs light up nice and bright, but the Load LEDs do nothing... I am using all the same values as you had in your pretty picture. I tried changing R1 from 10K to 100K (and every resistor I had inbetween I think... hmmm maybe I didn't try a 22K?  I tried 10k, 15K, 39K, 43K, 47K, 51K, 68K and 100K. And the trig LEDs stay nice and blue (they vary in intensity if I twiddle the pot) but they do not flicker and the Load LEDs, zey do nuzzing! Occasionally I accidentally touched some legs of one of the components together when I was sticking new resistors in while pulling old ones out and the circuit would flash crazily, then as soon as I stopped holding things everything went back to normal.... Anyway I wil look at this again, but I need some sleep now...

I'll be back!

[moid]

Scientific Progress goes boink! I finally had an afternoon to play with the circuit instead of doing work and lots of stuff happened.


I rebuilt the circuit on breadboard using Duck's layout with the C106D and we have success! I had to use 15K as the R1 resistor to get the blue TRIG LEDs to light up and also light up the Red LOAD LEDs. The pot has become a 25K lin (question about that below). I tried a bigger cap than 47uF for C1, but that just stopped the LEDs from doing anything... I didn't try sticking two caps in series to make something bigger than 47uF though... I tried a smaller cap (22uF) and that gives me a faster tremolo (as expected, so that's good). If I try a 10uF the TRIG LEDs stay on all the time and the red LOAD ones do nothing... so I might make a SPDT switch to flip between 22uF and 47uF.

22uF - speed of trem is 200msec to 550msec*
47uF - speed of trem is 450msec to 1700msec*

* question about the pot here - if I use a 25K pot a small area near the end of the sweep (as the speed goes slower) will set the TRIG LEDs on all the time with no trem sound at all. If I use a 47uF I could get to about 23K on the pot (using DMM to measure between the wiper and the other lug that connects to the circuit) before this would happen. If i switch to 22uF the trem dies at 21K... so my first thought was aha, just use a 20K pot, but that drops the longest trem pulse to 930msec... is there a way to make a 25K pot actually only travel from 0 - 21K resistance? If not I'll just use a 20K pot and live with it I guess. 21K gets me 1050msec pulse time.

Other problem - the audio doesn't fade up for the length of the trem - it does it quickly, just like using the BT169 transistor did. The good part is there is no clicking or ticking I guess what looks good on the oscilloscope doesn't always translate to audio? I did notice that the TRIG LEDs do light up more slowly, so I stuck an LDR on one of them and played through that, but because they don't get as bright as the red LEDs, the trem effect is softer, and also is not a reverse trem, but does take longer to fade out (about double the time of the red LEDs). I might stick two LDRs in the circuit if I can figure a way to disable one while using the other (having both on at the same time just seems to make the circuit use the red LED/LDR combo)... hmmm I bet all I have to do is stick an SPDT switch in, connect the two LDR legs that should go to the audio signal out to the outside lugs, then feed the central lug to the audio out... sounds good?

Other stuff I tried - I tried adding my weird modulation trem effect from a separate pedal plugged into the wiper of the pot, but that does nothing with the new transistor. I then re-read Duck's explanation of using an additional transistor to replace the diode and tried NPN and PNP transistors as he suggested but they either did nothing or turned the trem off... luckily nothing started smoking! Thanks for the suggestion though.

I tried different LDRs - some from one of my bags of random Chinese LDRs gave me much louder output than the 8K - 24K LDRs I usually use, so I'll probably use those (about time I found a use for them!)

I think this is about as far as I can take this now (it's a very simple circuit) so I reckon I'll try to draw up a layout for it and make it - it's nothing special, but it's simple and might be interesting for someone else, especially if I can make it small enough to squeeze into a old sweet tin I have here

[duck_arse]

your pot travel problem - add a stop resistor, like usual. looks like 2k2 or 3k3-ish from your measures? then the pot can sweep full rotation, will have a slightly slower slowest, too. ahh, a reread ...... a 25k pot with a 100k resistor across the ends won't go more than 21k. anotherjim used to say those sort of trem rate numbers were his favourites.

yer not supposed to be looking at the trig leds, they were just my replacement for a low-volt zener. that they provide some circuit operation indication is a bonus only. and as for the two leds as load, I knew you'd be wanting a rate indicator, so one led is meant for vactrol and one for front panel. but you know all this - it's so 3 pages ago now.

and as for the shape of the trem, whelp, that's what it is, that's how this circuit works. and doesn't tik you say? worst come to worst, you could buid a bank of basic oscillator with smaller caps, as audio oscs, and build instead a noise maker.

[moid]

your pot travel problem - a 25k pot with a 100k resistor across the ends won't go more than 21k. anotherjim used to say those sort of trem rate numbers were his favourites.

Thanks! I finally got part of an afternoon to get back to this, sorry I didn't get a chance to say thank you before. Just out of interest, if I wanted the pot to go  to 23K what value resistors would I use? Is there a calculator online I can check for this instead of annoying you? And by both ends, do you mean lugs 2 and 3 (the two that are connected to the circuit). I'm trying to draw a new schematic and layout of this at the moment.

yer not supposed to be looking at the trig leds, they were just my replacement for a low-volt zener. that they provide some circuit operation indication is a bonus only. and as for the two leds as load, I knew you'd be wanting a rate indicator, so one led is meant for vactrol and one for front panel. but you know all this - it's so 3 pages ago now.

and as for the shape of the trem, whelp, that's what it is, that's how this circuit works. and doesn't tik you say? worst come to worst, you could buid a bank of basic oscillator with smaller caps, as audio oscs, and build instead a noise maker.

I think I'll still use one of the TRIG leds though - with an LDR on it and a SPDT I can switch between reverse trem and forwards trem - never hurts to have options and it's easy to do. I'll use the spare LEDs as rate indicates on the box this will go in (thanks for thinking of that!)

regarding the use of smaller caps to make audio oscillators, I presume you meant a switch to substitute C1 for a small value so the trem would go into audio frequency range? At the moment that isn't possible - any cap smaller than 22uF at C1 and the trem stops working entirely (red load LEDs don't light up at all)... or did you mean something else?

EDIT: I am drawing the revised schematic at the moment and noticed that I had missed one part of Duck's design - lugs 1 and 2 of the pot are connected together, lug 3 conncts to the LEDs... at the moment I have been using lugs 2 and 3 and did not connect lug 1... is this important? Oooops...

EDIT2: panic over, it doesn't seem to make a difference whether lug 1 is connected or not. Or at least I can't hear a difference.  Although the circuit is now gently ticking, which it didn't do before - hopefully that is just because the components are a bit loose due to the breadboard getting knocked about a bit over the last month and doing all this with solder will fix it... hopefully.

[moid]

Picture time! Re drawn schematic with a switch to go between fast and slow tremolo, and another switch to go between forwards or reverse sawtooth. Can anyone see anything wrong with this? I want to draw a vero layout next. Thanks!

[m4268588]

Just out of interest, if I wanted the pot to go  to 23K what value resistors would I use? Is there a calculator online I can check for this instead of annoying you?

https://www.google.co.jp/search?q=23*100%2F%28100-23%29

[duck_arse]

when you use a pot as a potentiometer, a voltage dividing device, you use all three lugs, going to different places. when you use a pot as a variable resistor, you only use wiper and one end. the free end/lug is then ALWAYS always always always ALWAYS don't make me say it again always connected [a "short"] to the wiper. this is what we do, no need for argument. then, the wiper and the lug it is shorted to are the same thing.

you end up with a pot with only two connections, which can then go in circuit this way, or that way, because it makes no difference - as you turn the shaft, the resistance changes, easy.


about audio - it's not fast trem I was thinking of, it's an actual signal source, audio oscillator, type affair. that I was thinking of. I don't know if these parts [scr] in this circuit are speed limited in some way or not, I was just drawing the conclusion from smaller cap equals faster oscillations. the smaller cap may well crap out the circuit, I'd have to have a go on it to see if it werks or not. [I did make an audio noise oscillator mess from discrete transistors connected as thryistors though, just to make a complication for my waking hours.]

your circuit diagram - you have a 100k then a pot then a 100k, all in series. why?

[moid]

Just out of interest, if I wanted the pot to go  to 23K what value resistors would I use? Is there a calculator online I can check for this instead of annoying you?

https://www.google.co.jp/search?q=23*100%2F%28100-23%29

Thanks for that, sorry about the slow reply, work is super busy at the moment - so I need to put a 29K resistor in place of the 100K resistor that Duck suggested I would need for setting the pot to 21K? Do I not need to put the 25K of the pot into the equation?

when you use a pot as a potentiometer, a voltage dividing device, you use all three lugs, going to different places. when you use a pot as a variable resistor, you only use wiper and one end. the free end/lug is then ALWAYS always always always ALWAYS don't make me say it again always connected [a "short"] to the wiper. this is what we do, no need for argument. then, the wiper and the lug it is shorted to are the same thing.

you end up with a pot with only two connections, which can then go in circuit this way, or that way, because it makes no difference - as you turn the shaft, the resistance changes, easy.

about audio - it's not fast trem I was thinking of, it's an actual signal source, audio oscillator, type affair. that I was thinking of. I don't know if these parts [scr] in this circuit are speed limited in some way or not, I was just drawing the conclusion from smaller cap equals faster oscillations. the smaller cap may well crap out the circuit, I'd have to have a go on it to see if it werks or not. [I did make an audio noise oscillator mess from discrete transistors connected as thryistors though, just to make a complication for my waking hours.]

Thanks very much, that explains why I see lugs 1 and 2 tied together on some circuits. Sadly the SCR seems to be speed limited - it doesn't go above or below certain speeds (the circuit stops working), so I guess that's a limitation of a part that was probably never designed to be part of a tremolo!

your circuit diagram - you have a 100k then a pot then a 100k, all in series. why?

First Racoon: Shit, we did something stupid again! He's onto us - quick type something smart so he doesn't realise we're really three racoons in a trenchcoat!

Second Racoon: I'm trying, but with you on my shoulders it's hard to see out of the trenchcoat - I'll draw something quick, let's stick a resistor near the pot, connecting lugs 1 and 3? Maybe that's right?

Third Racoon: Hey guys how come I never get to do any soldering and I just stand here holding you guys up so you can see the computer?

First Racoon: That's 'cause you're the strong silent type, dude! Girl's love the strong silent guys!
 
Second Racoon: Yeah us weak nerdy types do all the talking while you just stand there like a total babe magnet!

Third Racoon: Really? Oh OK guys, strong and silent it is... you guys are the best!

Second Racoon: Phew. I'm uploading a drawing now...

First Racoon: Just in time, I can hear him coming now! Quick everyone, look normal!

Moid: So is the below what you meant? Or something else?



First Racoon: He's looking at it, he's thinking - I reckon playing dumb means we got away with it, well done gang!

[duck_arse]

talking racoons? the possums are quite interested now. they didn't much care for the talking monkeys.



yes. like that.

[bluelagoon]

Love the Racoon skit, very funny! LOL