Interpretation Q about the Guytone FL3 schem.

[solderman]

Hi all
Found this and wanted to do a PCB. (Thank's Praying_V )I need some help to understand what is what here. As I see it the K30A FET just in front of the last OP-Amp stage before OUT is part of the original bypass setup. To me it looks like a millennium or RAT kind of bypass and the bypass is actually only killing the wet signal and letting the dry signal pass trough the U1a and U4b Op-amps no matter how the bypass switch is positioned.

If I have read the shem correct as explained above I can leave out the SK30 and all the components connected to it's gate ( D1 and all components starting with 0) Exept C04 witch I see as a filter Cap. for a true bypass setup.

The thing I am most uncertain of is R04 and R05 that to me looks like a Vb Bias net for the TL22 and  the C01 then being a filter cap for that. There is no value for the R5. Can some one guide me here? If some one forced me I would use a 4,5V Zener or a 4.7K resistor as R05

Am I totally lost here can some one pleas help me?



BTW I checked to se if everything was there and found that this is extremely close to the application example for the MN3207/3132 only Added an LFO and some tweeks.
http://www.datasheetcatalog.org/datasheet/panasonic/MN3207.pdf

[PRR]

> ...and all components starting with 0 ... Am I totally lost here...

Just a little... summa those zero parts stay needed.

Omit: Q4 (short), D1, C03, R010, R011, D6, R012, D5, SPDT (open).

"R5" appears to be a Zener, not a resistor.... ah, that's what the text says. It's just a reference for the LFO. I'd be inclined to drop another 4K7 in there; it should work. C01 should filter most wall-wart ripple, not that it matters much on an LFO. Maybe if the power supply has deep dips, the LFO burps: then maybe a 4V-5V Zener to hold the reference steady.

[solderman]

Thank's

I just found that R8 and R9 have no value but they seam to be part of a referens voltage net as well
What about R8=47K R9=47K?

[Mark Hammer]

Your inference about the "switching" is correct.  Indeed, the vast majority of chorus pedals, flangers, delays, and phasers - all of which are circuits that produce their "effect" by combining dry and wet signals - use the identical arrangement: a single JFET used to lift the wet signal, cancelling the "effect".

[solderman]

Some one have an idee about R8 and R9?

If the just are reference voltage why not use the the one produced by R32 R33 ?

[PRR]

> extremely close to the application example for the MN3207/3132
> idee about R8 and R9?

Biasing the MN3207? You have its datasheet handy; I have not seen one in decades. But IIRC, the distortion varies sharply with input bias. What do they suggest?

[solderman]

---BUMP--

And is there anyone out there who actually has built a working effect from the schematics above?

I have made a PCB layout and ordered some MN 3207 and MN3102. So it would be nice to know that I am using a working design.

About the R8 & R9

The datasheet does not suggest any input bias as i can find some I will omit then but leave the pads on the PCB there for them.

[doc_drop]

Well, as luck would have it, I recently bought some MN3207's and 3102's on E-Bay. I was looking for circuits to use them on, and this one popped up. So, I am going to try to breadboard this in the next few days. It looks like it is going to be a difficult breadboard to do, but I am going to challenge myself. I was planning on using a zeneer for R05 to start, and I am going to leave out R8 & R9 to start with as well.

I'll let you know when I have something to report.

Anyone know of another circuit these chips can be used for? Or is that another thread...

[Mark Hammer]

Some one have an idee about R8 and R9?

If the just are reference voltage why not use the the one produced by R32 R33 ?

The bias voltage virtually always needs to be set manually.  That can be done by ear or by scope.  ear is usuallygood enough, though I recommend use of a headphone amp if you are going to do that since you need to be able to hear when you reach the point of cleanest delay signal.  It will also be useful, if doing it by ear, to hear only the wet signal, so temporarily lift one end of that 100k resistor that brings dry signal to the output mixing stage so that you only hear the delay signal (when the switch/light is on).

As for what you use for biasing, seems to me the best place to look is at the Boss BF-2 schematic for how they do it.....which is pretty much how 90% of all flangers also do it.

[solderman]

Hi Thank's Mark

It is clear to me even if the schematic i found of the BF2 was blur

a trimpot with lug 1= GDN lug 3=+V and viper to pin 3 ot the MN3207 in series with a 100k Resistor and a 50nF cap in parallel to GDN.

By By R6 & R8 hello VR2

[Mark Hammer]

Yup.  You're just adding a DC voltage on top of the AC signal....or an AC signal on top of a DC voltage...which ever you prefer.

The smoothing electrolytic cap is not absolutely necessary, but does maintain some degree of stability in the bias voltage in the face of fluctuations in the actual supplythat feeds it, so it is strongly recommended.  1uf is not critical.  Feel free to use 2u2, 3u3, 4u7 10uf, etc.

[PRR]

> is there anyone out there who actually has built a working effect from the schematics above?

I oversaw EE students building BBD effects when they were so new the documentation was hand-sketches notated in Japanese.

> About the R8 & R9 ... The datasheet does not suggest any input bias

mostly-English Datasheet:
http://www.experimentalistsanonymous.com/diy/Datasheets/MN3207.pdf

Bottom of file page 3 (book page 82) has figure "THD-Vbias".

It shows that THD (distortion) is low for Vbias from 2.5V to 3.0V; past 2.25V or 3.25V THD gets significant and then horrible, like a 13 foot truck through a 7 foot door.

This plot is for 5V supply. Your plan works at nearly 9V. What changes? Well Page 4 (book pg 83) figure "Vbias-VDD" shows (optimum) bias against VDD. In fact the optimum bias is "a hair above half voltage". (Those students' thesis paper explored why this is.) "Half supply" will "work". For most useful supply voltages, 55% of supply will be a little cleaner. The exact best voltage varies with chip processing.



The Datasheet's "Application Circuit" shows a full plan. I see the biasing is not so obvious. Pin 3 gets biased audio through two stages of opamp. A DC analysis (or a lucky guess) says the pin 3 DC voltage is equal to IC1 pin 4 and the 47K+33K+"50K*" divider. The "50K" is an adjustment (and probably a 0-50K rheostat/pot). This divider will make 41% to 64% of supply voltage. If 50K pot is set to center, it makes 55% of supply, which is usually as good as it gets. I assume the wide range is so they can use low-tolerance parts.

> By By R6 & R8 hello VR2

That will "work". But over most of the pot turn you will get no sound at all. Below 50% or above 60%, even when you get sound, it will be disgustingly distorted. At least pre-set the pot for ~~55% of supply voltage before you expect sound to pass. Now that 2% parts are cheap, a 47K and a 56K will give a "good" bias right away, so you can debug the rest of the system (it will need bebugging). For final sweetening, you want a 50%-60% trim range. Trying to find the sweet-spot with a 0%-100% trim is very frustrating. 47K+47K+25K pot will spread-out the target bias zone.

[doc_drop]

O.K. I have this thing up on my breadboard. It isn't working, though. I finally got the LFO pulsing, but I only get clean signal at the output of the circuit. I did an audio probe, and I have clean signal at the input of the MN3207, but nothing at the output (7& pins.

A couple of things I discovered. I get a stronger LFO using a 4k7 resistor for R05 instead of the zeener I was using. I did the biasing technique Mark mentioned from the BF-2. I am going to see if biasing the chip as PRR mentioned helps.

The other thing is that I get a strong LFO at the depth pot wiper as measured by my DMM. But it is gone on the other side of the 330k (R23) resistor. Shouldn't some of the LFO make it past the resistor to the MN3102?

I will report my results once I have some.

Anybody else have any ideas on why the MN3207 isn't passing signal or what happened to the LFO after R23?

[solderman]

Hi
Very Interesting news all though not that good news. So cheer up, At least the LFO is working you have to appreciate the small things in life

According to PRR (thank's PRR for the in depth and very good explanation BTW) the correct bias is important.

> By By R6 & R8 hello VR2

That will "work". But over most of the pot turn you will get no sound at all. Below 50% or above 60%, even when you get sound, it will be disgustingly distorted. At least pre-set the pot for ~~55% of supply voltage before you expect sound to pass. Now that 2% parts are cheap, a 47K and a 56K will give a "good" bias right away, so you can debug the rest of the system (it will need bebugging). For final sweetening, you want a 50%-60% trim range. Trying to find the sweet-spot with a 0%-100% trim is very frustrating. 47K+47K+25K pot will spread-out the target bias zone.

I'm really curios about your progress since I'm going for a PCB at ones and skip the bread board. My Chip have not arrived jet. I am planning to etch a couple of boards to morrow. The extra trim pot for the bias has shaken my PCB layout. There is no place for it that will make it fit in a 1590A box.

[metallo]

If you are producing in limited numbers you could just use a pot when building and then measure and hard-wire the correct values in a way that would fit your enclosure.

[doc_drop]

Eurika! I got it flanging, and it sounds awesome.

This was definately a difficult breadboard, but the sweet smell of success is making me happy. 

I used 2 33k resistors in series with a 50k trim pot to bias. It worked great. I got a fixed flange sound by finally getting the MN3207 to have some signal on the output. (I'm afraid it was a wrong connection on U4 that was partly to blame as well.) Then I decided to eliminate R29 to see if I could get more LFO to the MN3102. That worked perfectly. There is the option to have too much depth at the extreme like this, but it works great.

So FYI, I used a TL072 for U5. And 2n5089's for the LFO trannies. I used a 100KB pot instead of the resonance switch. And RC4558's instead of the NJM's. I left out the bypass componants as suggested at the beginning of this thread.

Solderman, how long do you think it will be before you post a layout? Because then I would definitely be cheered up...

[solderman]

Hi
Congrat's I'm verry pleased to here that the "Eagel has landed" cred to you for bread boarding this one
About the PCB...Right now actually. The only drawback for you is that I have a habit of shrinking things to fit in a 1590A box. This one is no exception. I had to stick to the original 2 board design to make it fit. Of cause it's not verified yet. My IC:s are in the mail as we speak. So maybe next week. Specialy like the 100K pot in stead of switch mod. Good thinking
But if you care use it...


PCB
http://solderman.fatabur.se/FL3/PCB.pdf
Component Layout
http://solderman.fatabur.se/FL3/Component Layout.pdf
Silk
http://solderman.fatabur.se/FL3/Silk.pdf
BOM
http://solderman.fatabur.se/FL3/BOM.pdf

[doc_drop]

Thanks Solderman. That is perfect. I especially like the bypass switch pads you did.

I will probably use your layout as a guide and put this on perf. I kind of like doing that when I am still getting to know a circuit, even though Frequency Central makes fun of my perf skills. (See Causality phaser thread...) I will let you know once I get it together, and whether I think your layout is verified.

[Ummagumma]

Hi, any progress here?  Is it verified?

cheers

[doc_drop]

Umma,

Well, I started to try to imitate Solderman's PCB on perf. But, I was having a hard time. So, I scratched that idea and created my own sloppy perf. layout. I don't have much experience with layouts, and maybe a 5 IC layout was a bad way to get more, so it was a challenge. I will be testing it later today, but so far all the continuity is correct. I was planning on eventually posting my layout. But, I am a little afraid of all the "Did you have to run that many jumpers?" questions.

Anyway, I will report back when A) It works. And B) I get over my pride and post my perf. layout.

It sounded great on my breadboard, especially after a couple of tweaks I did to make it more extreme.