can i use 2N5457 instead of 2N5952

[cameralover]

i wanna make MXR noisegate and MXR phase 90, however, i can't find 2n5952. i have only 2n5457 in my hand(that's because of i buy a ton of wrong number). could i use 2n5457 instead of 2n5952, and is it make some different for the output sound?

thank for your kindness help 

[MartyMart]

I used them in a Phase 45 , so it's possible ... BUT you have to build a Jfet matcher ( search ) which
is a simple way to get a set of matched ones for the P90 , without that your "phase" could be very
weak or ... not work at all.
You'll need at least couple of dozen to choose from also !

MM.

[Mark Hammer]

Many other FETs have been used to produce phase shifters besides the 2N5952.  It IS important to match them, no matter what type they are.  You may have to bias them differently than 2N5952s, and you may not be able to get as much sweep range out of them. 

Remember that the FET is being used as a voltage-dependent resistor.  What voltage must be applied to the gate to produce a change in resistance?  What is the minimum and what is the maximum?  How does that overlap with the voltage range of the LFO output?  What are the minimum and maximum possible resistance values that this results in?  Using the capacitor value in the phase-shfit stage, where will that situate the notches produced?

These are all questions to be considered when selecting a FET.  Keep in mind that some people LIKE a narrow sweep more than a wide sweep, and that placing the notches where you like them and where they show up best might be simply a matter of changing the cap values for something a little higher or a little lower.

Two things come out of that: 1) The circuit CAN work with other FETs, but not necessarily in the identical way, and 2) The circuit may be persuaded to work similarly to the original by changing some component values.

I guess you will see when you get there.  I will say that the 2N5457 is a good FET and I have seen it often used as a voltage-dependent resistor, most commonly in compressors and gates.

[cameralover]

thank you very much for your explanation, even though i'm a newbie and don't understand much about your information...i'll try and try

tomorrow i'll do both of it and wish that i'll fine

god bless me harrrrr

[Mark Hammer]

The FET has 3 pins: the drain, the source, and the gate.  When a voltage is applied at the gate, the path from drain to source (which is the pins connected between ground and the capacitor in each phase shift stage) changes its resistance.  If the FET is able to change that resistance over a very wide range of possible values, then the sweep of the phaser, from lowest to highest point, will be very wide.  If a voltage at the gate goes from, let us say, +3.5v to +5.6v (this is just an example, and not the actual voltages) produced a change from 3k to 1.4M in one type of FET, but only produce a change from 44k to 300k in another type of FET, the first one would give a wider sweep than the second one.

In the Phase 90 (which is what I guess you are building, since you mention 2N5952s), the combination of the FET and .05uf caps allows the notches produced to go as low as around 145hz.  Look at the schematic, and you will see that there is a 22k fixed resistor in parallel with the FET in each phase shift stage.  Remember that when two resistances are in parallel, their combination can never be higher than the smallest one.  So, even though the FET may be able to have a drain-to-source resistance of several megohms, when combined with a 22k fixed resistor, the combined parallel resistance will NEVER be any higher than 22k.  If the FET only had a narrow, and low, range of possible resistance values for different gate voltages (e.g., 100k to 300k), then the parallel resistance would not change much (from around 18k to around 20k).  For example, 100k to 300k, in parallel with 22k and using a .05uf cap, would mean that the notches produced would never go any lower than 155hz (not a huge change at the low end) but never really go above 353hz (for the upper notch at highest sweep point), only about 1.3 octaves.  Not much movement in the location of the notches at all.

Because it is the combined parallel resistance (the FET plus the fixed resistor) that sets the resistance range covered, in this case, that is good news for you. If you find 4 matched 2N5457 (or any other) FETs, and they don't sweep over a very wide range of resistance values, you can simply change the 22k resistors for a different value.  In our example of 100k-300k, if we changed the 22k resistors for 47k values, the resistance range covered would be 32k - 40k.  Use 100k instead, and the resistance change is now 50k-75k.  So the bigger the value of the fixed resistor, the more usable range can be gotten from even a FET with very limited resistance range.

One thing you can see is that every time I increased the parallel resistor to change the range, the actual resistances went up.  With 7k, the lowest notch would be at 42hz.  This is much too low.  Changing the cap to .015uf, though, moves that lowest notch up to 142hz, which is very close to stock.

Make sense?

[cameralover]

not god but you, Mark Hammer who bless me....it's clear. And as i search this forum for jfet match, i find a number of the same question as me and clarified me that i can use any number of j-fet as long as they are match, right?

so i need to build jfet matcher, i'll try ....

thank you verymuch again and again

[Mark Hammer]

You are most welcome.  EVERYTHING can be explained clearly... eventually...to anyone.  You just have to try.

You CAN successfully build a phaser without matching the JFETs, but the sweep will not be as good.  No. Let me correct that.  You will not be able to guarantee the sweep will be good unless you try to match the JFETs.  Sometimes it turns out well without matching, but usually it turns out better with matching.

One of the good things about phasers that use photocells instead of JFETs as the varying resistance is that the photocells do not have to be matched.  Unfortunately, it is more awkward, and sometimes more expensive, to make a photocell-based phaser.

Phasers that use OTA chips (CA3080, CA3094, LM13600, etc) such as the Electro-Harmonix Small Stone, the Moogerfooger Stage Phaser, the DOD FX20, and the old Ross phaser also do not need to be matched.

[calculating_infinity]

   Nice explanation, had to read it over a couple times to get a better understanding.  I am going to build a phaser sometime in the future so this information was helpful.  Thanks Mark! 

[cameralover]

i just finished DIY boss ce-2....sound so nice and i like it more than small stone which i made it one. Moreover, i finished MXR noisegate and MXR phase 90 too, but not wiring it anymore...(these 3 FX give some pain to my spinal bone herrrrr ). i plan to wiring it for tomorrow and hope that it'll ok

thank all