Question about electronic switch

[wcampagner]

Hello,

I'm trying to make an electronic switch.
I tried some FET's like the J111 and the BF245C and it didn't work.
I tried also with the 2N3819 and it worked fine...

There was a problem only when the stompbox volume was full i could hear a noise when switching On/OFF. What i want to know is what FET do you use in your bypass circuit and if someone can help me solve this noise problem...

I have another problem... when the stompbox is in bypass mode, i can hear the clean sound and a little bit of the stompbox sound... this sound is comming from the ground of the stompbox... so if anyone can help me solve this problem too... i'll be glad.

Thanks a lot in advance,
Wagner.

[gez]

There was a problem only when the stompbox volume was full i could hear a noise when switching On/OFF. What i want to know is what FET do you use in your bypass circuit and if someone can help me solve this noise problem...

I have another problem... when the stompbox is in bypass mode, i can hear the clean sound and a little bit of the stompbox sound... this sound is comming from the ground of the stompbox... so if anyone can help me solve this problem too... i'll be glad.

Is this on your breadboard?  The bleed-through you're getting when 'bypassed' could just be stray capacitance due to bad layout.

To reduce switching noise there should be no significant DC offset between either end of each FET's channel.  Usually, this means referencing both ends to half Vcc and AC coupling the signal.  I try to design my circuits so there's no significant offsets between stages and directly couple without all the bias crap.  It works fine, but often you need to turn on the FETs slower, something you might want to try anyway to see if it reduces the switching glitches.

How are you doing the momentary switching?  I recently had a problem as a result of dumping charge rapidly to a shared ground connection.  A little help from the wiser members of this forum solved this, just used a small value resistor to slow down the discharge of the cap and used separate ground paths (thanks Ton & RG).

I use MOSFETs for my switching, one less component per FET (don't need the reverse diodes to the gates).

[wcampagner]

Hello Gez,

First of all thanks for your reply.

Yes, it is in my breadboard. i always make my stompbox in the breadboard to test and make mods... So i have this problem of bleed-trhough... Do you think that the breadboard can be the cause of the problem? If i make a PCB do you think it will be solved? I've already made a GT-2 and a Tube Screamer in the breadboard and both had this problem so i didn't make the PCB yet...

About the switching problem... i am referencing both ends to half Vcc. About the AC coupling... Can you explain me what's that? i don't know...

Anyway, i'm using the GT-2 bypass system... you can see it in my page www.planetadosite.com.br/bypass.jpg (sorry for the bad quality, but i think you can understand it)... Can you please take a look and see if it is right?

About your problem of dumping charge rapidly... can you explain me how do you do the separate ground paths and the small value resistor?

I think the bypass system is good for me... no pop's and low noise switching... the real problem is happening just when the stompbox volume is full... it is not noise that i hear... it is a fast whistle when turning the stompbox ON/OFF.

Thanks a lot for your help,
Wagner.

[gez]

I'm a bit confused by your diagram.  The output should be taken from the junction of the two FETs and the 510k resistor, this resistor being used to bias an output buffer, which isn’t shown here.  The input buffer of the audio circuit will connect to one FET, and the output of the effect to the other FET.  The connections would be where the 'in' and 'out' on your schematic are drawn.

For AC coupling the points marked 'in' and 'out' would also be referenced to half Vcc by 510k resistors and AC coupled, i.e. DC blocking caps used.  However, as I mentioned previously, if there are no significant offsets between both ends of each FET's channel then you can usually get away DC coupling (though you might need to turn the FETs on more slowly - increase the value of the .047u caps - just to slow things down if there are very slight offsets).  If there are significant offsets, you need the caps; when the channel suddenly opens during switching there will be a flow of current due to the difference of potential between both ends, which can 'translate' as a pop.  

You say there are no switching glitches though (as in 'pops'), so I'm not sure what the problem is.  The 'whistling' you describe is odd.  See if AC coupling cures it.

looking at the schematic, you don't have to worry about the charge dumping thing, there are resistors in place to prevent too rapid a discharge of the caps when switching occurs.

I’d always recommend trying to get things to work on a breadboard as well as you can before committing to a PCB.  Many times, things will be improved once everything is in place on a PCB, but only if you're careful about layout.  To be on the safe side, use separate ground connections for all the switching stuff and join everything at the -ve power terminal.  Try this on your breadboard and keep wires from the switching circuit well away from the audio stuff.  Also, be careful with wiring etc between the clean and effect paths.  If you still get bleed-through it might be an idea to kill the signal going into the effect as well as killing its output when you switch to ‘bypass’ mode.  This will require an additional FET.

[wcampagner]

Hello Gez,

Thanks again for your reply...

Here is a link to the complete schematic of the stompbox i am working on: http://www.generalguitargadgets.com/diagrams/ts-808.jpg.

It's an original tube screamer... the only diference is that i've used the IC 4013 instead of the transistor Flip-Flop...

As you can see, the input and the junction of the two FET's are already AC coupled... the other FET going to the volume potentiometer is not AC coupled... do you think it is necessary to AC couple this part too??

I've made a search with the word "bleed-throught" and find some topic about... most of them i've also tried all the things they mention... the only good solution i saw was someone that told about adding another FET to "open" the signal path when bypassed: http://forum.musikding.de/attachements/Boss_SD-1_mod_c.gif as you also told me in your previous message...

What i'll try to do now is follow your advice about separate grounds... i'll start wiring the circuit again but now separating the modules as my breadboard has 8 slots i think i'll do this way:

first slot: bypass module
second slot: input and output buffer module
third slot: effect module
fouth slot: power supply module

Each module have it's own ground all conected to the ground of the power supply.
Do you think it's a good separation? Maybe this way the bleed-through can be avoided...

About the whistling i'll try to test it again when the modules are separated... maybe this will not happen... maybe it is the FET i am using... I'm using the 2n3819 FET... i think it is a good one... because it is the only one that worked in the bypass circuit...

About your idea of using another FET to kill the signal i think it's a good idea, but i want to use it only if it is the last opition because it does not solve the problem... it is only a way to bypass the problem... i think the original tube screamer does not have this FET so i'll try not to use it...

Thanks agains for your help and i'll post the results as soon as i get the modules separated, ok?

Wagner.

[gez]

Unfortunately the link doesn't seem to be working.  You're right though, this should work with just two FETs because the original did.

The 'whistling' effect might be caused by the circuit oscillating somehow when disengaged.  Seems odd though.

Also, do you notice the bleed-through only if you play hard?  If you turn down the guitar a bit does it go away?  Try changing your layout on the breadboard and see if this cures the problem.

[wcampagner]

Hello Gez,

The link had a period at the end... that's why it didn't open... the correct link is here http://www.generalguitargadgets.com/diagrams/ts-808.jpg

There is a trick that i do to hear the bleed-through... on the output jack i leave the tip "open"... so the jack is only conected to the ground... this way i can hear only the "bleed-through" signal... One thing i am sure is that the bleed-through signal is comming from the ground... and if i diconect the IC ground pin (4) the bleed-through is over... that's why i want to try to separate the modules... i think it should work...

The bleed-through signal on the tube screamer is not very noticeable... just when i turn the amp. volume i can hear it (using that trick above)... I think that if i didn't do that trick i would never notice it... but i have this problem... i want everything to be perfect... so i'll try to solve it too... on the GT-2 this is a real problem, because i think is a hi-gain stompbox... the bleed-through signal very noticeable.... even with low amp. volume.

When i finish and test it i'll post the results here...

Thanks again for all your help,
Wagner.

[wcampagner]

Hello Gez,

Thanks a lot for your help.

I tested the new layout now... and the "Bleed-Through" signal was almost gone... almost untoticeable...

So i tested the trick explained in http://forum.musikding.de/attachements/Boss_SD-1_mod_c.gif adding another FET... and it worked perfectly... no more "Bleed-Through"... so i think i'll use this trick... because i don't need to worry about designing the ground routing and i thought that adding another FET (another point of opening and closing path) would make the bypass POP more... but it didn't... the bypass kept on very silent....

About the "whistling"... i took out both .047u caps from the bypass circuit and it was perfect... but don't tell me why.... i just tried it and it worked...

Now, i want to add an additional "temporary" switch in the bypass circuit to have two options... one to switch between ON/OFF and the other to switch "momentanly" ON/OFF. So i'll need to remember all those things about Flip-Flop that i've seen at University... if you know how to do it can you tell me how...

Another thing is about the buffer transistor's... i'm using the 2N3904 on them... but i don't know if there are better transistor for this purpose... maybe the 2N5089 that is low noise is better.... do you know what transistor is good for buffers?

I'm happy now with this stompbox... it was a lot of month for me to get it working the way i wanted...

Thanks again for all your help it was really important to me,
Wagner.

[gez]

I tested the new layout now... and the "Bleed-Through" signal was almost gone... almost untoticeable....

Most of the time, when you make a PCB and pay attention to its layout all trace of bleed-through will disappear.  A lot of the time all you can do with a breadboard is minimise noise issues, then hope for the best when you move to PCB.  So far I've had a 100% success rate with my PCBs when I've had similar issues, though it's always nerve-racking!  

About the "whistling"... i took out both .047u caps from the bypass circuit and it was perfect... but don't tell me why.... i just tried it and it worked...

OK, the whistling you describe might be the sound of the cap ramping up during switching, this gets interpreted as a signal and is coupled through via the FET's gate capacitance.  You did use the correct values didn't you?  Pulling the caps isn't a good idea as you're likely to get switching clicks.  Using a FET with low gate capacitance would be the best thing to do and reinstate the caps.  Making the caps larger might be another option, if you can slow down the time it takes for the cap to ramp up so that you have a slowly rising DC level (below audio frequency), this might cure the problem.  Again, you did use the correct values didn't you?

Now, i want to add an additional "temporary" switch in the bypass circuit to have two options... one to switch between ON/OFF and the other to switch "momentanly" ON/OFF. So i'll need to remember all those things about Flip-Flop that i've seen at University... if you know how to do it can you tell me how...

Don't really understand what you mean Wagner.  You mean you want true-bypass too?

Another thing is about the buffer transistor's... i'm using the 2N3904 on them... but i don't know if there are better transistor for this purpose... maybe the 2N5089 that is low noise is better.... do you know what transistor is good for buffers?

Check data sheets for whatever you can get hold of and use one with low noise figures, the higher the gain the better.

[wcampagner]

Hello again Gez,

Yes, i'm sure i am using 0.047uF caps. on the FET bypass system... i am also using the 2N3819 FET and i don't know if they have low gate capacitance... anyway... when i took out the caps. i did not get switching clicks... i also tried other caps. values too... the higher the cap. more time it took to the "whistling" stop when switching...

I am satisfied with this configuration as it did not increase any switching clicks... do you think if i leave the bypass system this way may be a problem? Do you thing that without the caps. it can damage other components? Because if there is no problem i'll leave this way... as i said before... there is no click...

About the momentary switch... let me explain you... i want two switches... the first is going to work the normal way... each time you press it... it'll change the effect ON/OFF... the second switch i want it to be a temporary switch... it is going to change the effect WHILE pressed... understand? When this switch is pressed it'll change the effect....

For example, if i want to use this effect for about 2 seconds... i won't need to press two time the normal switch... i'll have to press the momentary switch for 2 seconds...

Thanks a lot again,
Wagner.

[gez]

i am also using the 2N3819 FET and i don't know if they have low gate capacitance... anyway... when i took out the caps. i did not get switching clicks... i also tried other caps. values too... the higher the cap. more time it took to the "whistling" stop when switching...

The FETs you're using are audio, you'd be better off finding ones designed for switching.  It's not a question of leaving components open to damage if you exclude the caps, just that you may (will probably) get a problem with clicking once you commit to PCB.  If I were you I'd leave a space on the PCB to put them in, better still use a different type of FET.

if i want to use this effect for about 2 seconds... i won't need to press two time the normal switch... i'll have to press the momentary switch for 2 seconds...

You need to give more info.  If you want a different setting for distortion or something within this circuit, you can simply configure the momentary switch to turn on a CMOS/JFET switch when engaged and turn it off when not.

First thing that springs to mind would be to connect a cap and resistor in series from V+ to ground, the junction of which connects to your switch (need reverse bias diode if using a JFET).  Connect one lug of the momentary switch to V+ and the other to a smaller value resistor which connects to the junction of the series cap, resistor and FET.  When the momentary switch is closed the resistor connected to it pulls up the gate of the FET and turns it on.  When the switch is disengaged the gate ramps down again, turning the FET/CMOS switch off.

If you don't mind switching noise, or if it's for something that won't affect the audio path, just use the momentary switch to connect up a resitor/cap/whatever in parallel.  Like I said, you need to be more specific.

I'm sure there's a better way, I'll post tonight if something springs to mind.  

[wcampagner]

Hello again,

As i told you... i've also tested with other FET's too.

I tried the BF245C and the J111... i think the J111 is a switch FET, but it also didn't work... the clean and dirty signal where mixed...

Maybe my pair of J111 is broken... because i think it is a really good FET for switching... used in the DOD stompboxes....

So, i'll try to test others FET... do you have any suggestion which FET should i use? I'll try to find... but i'm from Brazil and it is not easy to find parts here... but i can try...

About the momentary switch i'll try to put it in the bypass circuit... maybe using the SET/PRESET pins of the Flip-Flop... well i need to think about it... if i find something i'll tell you...

Thanks agains,
Wagner.

[gez]

I tried the BF245C and the J111... i think the J111 is a switch FET, but it also didn't work... the clean and dirty signal where mixed...

You need FETs with a low pinch-off/cut-off voltage/threshold.  If it's high then the FET won't necessarily be turned off when the gate is pulled to ground, hence the combination of effect and clean when bypassed - this isn't an indication that the FET is faulty, just that this individual one isn't  suitable for the job (though others with the same code may be).

All I can say is look through your suppliers' catalogues and check data sheets for whatever they've got.  Find a FET with reasonably low cut-off (Vgs off I think it's called - my brain is a bit fried at the moment - or Vp).  The figure is usually quoted as a range.  Even if you do find something there's no guarantee you won't have a problem.  Sometimes it requires bying a batch and testing them (or just sticking them in a circuit to see which work and which don't).  Obviously anything with a range that extends above a few volts is going to require testing (assuming that at least some devices will be within the range we require)

[wcampagner]

Hello Gez,

On the weekend i bought 3 2SK30A and replaced the 2N3819...
It worked fine... i also put back the 0.047uF Caps. and no whistling...
I'll be using those FET for my bypass circuit...

I also bought a pair of 2N5089 (low-noise) to replace the 2N3904 buffers... but i didn't notice any diference... so i'll be using the 2N3904 that are cheaper than the 2N5088...

Thanks for all your help again,
Wagner.

[gez]

i also put back the 0.047uF Caps. and no whistling

Good, better to leave them in, they're there for a reason!  

Glad it's all going well for you now.