I'm designing a guitar effect (simple booster) which uses a CD4053 cmos switch as outlined here (http://www.geofex.com/Article_Folders/cd4053/cd4053.htm)

As you can see from the second diagram, RG suggests using the "existing in and out capacitors" of the effect itself to properly ensure that the CD4053 does not experience any DC offset (which cause clicks).

Notice however that the link between 2 and 12 (the bypass link) does not have capacitors. So RG clearly assumes that there is no need for them in a simple link.

Now... This is the boost circuit I'm trying to implement (I like it because it also an attenuator, it has a volume knob, which is fundamentally a problem in this case)

Firstly, because RG has no caps on his bypass link, do we think it's safe here to omit the boost circuit input capacitor as long as I properly bias the input to 4.5v?

Secondly, what do we do about the output potentiometer, could I just use 4.5v instead of ground, then I don't need to re-bias the output to 4.5v?

In short what do we think about the booster circuit here :

enter image description here

LTSpice seems to indicate that the circuit works, but I have no idea what it would be like in an actual application.

s you can see here :

enter image description here

Whilst I have enough space on a PCB sized for the hammond 1590B to implement a Booster with a few more capacitors, I'm quickly running out of space so I couldn't use this design as a template for other effects. I'm pretty stuck on using the WIMA caps.. What is the point of making your own circuit if you don't use the best bits?

Does anyone have any suggestions as to how I can improve the booster circuit generally? And particularly how I can do that without using up much more PCB space?


See below same thing but with JFET switching as per the tube screamer :

enter image description here

Here's the swtich.. notice a nice ramp up in volume caused by the capacitors on the JFETS.. Much better than the rock hard (but instant) switching of the cmos chips.

enter image description here


1 Answer 1


Ok, firstly I think that switching circuit is going to click, no matter how many caps you scatter around, simple charge injection in the cmos switches will see to that at the impedance levels you are running there.

The trick with coupling (as opposed to filter) caps is to keep the AC voltage across em negligible, which is to say make them large compared to circuit impedance. So long as you keep the AC across the cap below 80mV or so the distortion is negligible, and as you know the DC conditions on your caps (4.5V give or take) there is no problem just going with 10V, 22uF electrolytics here, they will not be a problem and are far cheaper then messing around with PP or PPS film. A further note, your output cap is 2.2uF into 10k, which sets the LF rolloff, your input caps do not need to be sized to roll off at a frequency 1/30th of this, so maybe you can downsize them? Bit of luck, C0G ceramic may not be out of the question.

My first thought were I doing this thing would be lose the analog switch and go for a couple of jfets instead, use a dual opamp so you can buffer to get the impedance down to something sane, couple of jfets to do the audio switching (Probably need to pick parts with low pinchoff voltage as you only have 4.5V to play with).

The trouble with just floating the whole thing 4.5V above battery negative (which is sort of tempting) is that looking at that board you have an external DC input, and many pedalboard supplies make a nasty assumption that it is ok to connect all the negative rails together.

  • \$\begingroup\$ This is very interesting. In terms of the impedence, what if I place a buffer before the switch with a much lower output impedence. Would that improve anything? Why would jfets be better? This is kind of a warm up project for a bigger one featuring a 'radio button' concept where one of 8 effects can be selected by footswitches... so I'm really hoping to perfect this because that's based on a 4066. As for your cap size points, yes.. 2.2uF may be a bit large for some of these.. I just used the same values because it's easy to do that in LTSpice. Your point about earth loops etc.. noted. \$\endgroup\$
    – Richard
    Feb 18, 2020 at 17:20
  • \$\begingroup\$ Nice thing about jfets is that you can control the switching speed, so you can sort of ramp the resistance up and down over a few ms to reduce the tendency to click. There is a good chapter on audio switching in Doug Selfs "Small signal audio design", well worth picking up a copy. \$\endgroup\$
    – Dan Mills
    Feb 18, 2020 at 21:06
  • \$\begingroup\$ thank you. I have seen that book somewhere... I'll dig it out. \$\endgroup\$
    – Richard
    Feb 18, 2020 at 21:11
  • \$\begingroup\$ Dan.. what would be a suitable inexpensive fet... 2SK170? \$\endgroup\$
    – Richard
    Feb 18, 2020 at 23:46
  • 1
    \$\begingroup\$ 2SK170 is obsolete and basically unobtanium, Linear systems do an equivalent, but do check the price, cheap it isn't. 2SK3666-2-TB-E is available (SOT23, not hard to hand solder), is cheap and will cutoff with -4.5V on the gate. \$\endgroup\$
    – Dan Mills
    Feb 19, 2020 at 12:55

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.