FET's for Phase 45

[Steben]

I had a bunch of J201's lying around.
For the rest I've "stock-" built it all, but I only get clean sound on the total range of the bias with ALL J201's! Sticking them in out only resulted in variatons of tone controls as you would expected, but without any sweep. Are them J201 prone to be distortion queens and suck at the rest or what? Anyone built the the 45 with these?
What kind of fets did you guys use?
Maybe I'ld better check first for AC at the oscillator?

[A.S.P.]

if it were possible, I`d post the URL of what the search-function, everybody is so fond about in this new forum, had to offer...

[MartyMart]

J201's wont work I'm afraid !
You need a "matched" pair of 2N5457's/5458's
I built the "Jfet matcher" to help with my 45
which is Torchy's version & Univibe mods.

Marty.

[dj_death]

Actually J201's work very well. I used RG's jfet matcher and i found 2 couples from a bunch of 15. Both couples work very well in MXR Phase 45.

John

[spudulike]

... if you have a 5.1V zener and the JFETs seem to work in the matcher at 5.9V, you can't expect them to work for phasing. You need JFETs with Vgsoff at less than the zener voltage. Your numbers suggest that your JFETs have Vgsoff too high to work with that zener.

You need to build RGs FET matcher and check the Vgsoff like the great man said 

[slacker]

I'm not an expert at this but I don't think the type of JFET matters. The important thing is that they work within the voltage range set by the bias pot. I built mine with 2N5457s and out of 10 that I bought only 3 worked at all and only 2 gave a decent sweep.
I haven't tried this, but you might get your J201s to work by tweaking the bias setup a bit. As standard the bias pot only gives a bias range of between about 1volt and 3.5volts, if your JFETs need a voltage outside this range they won't work. You could try jumpering R25 and R26 which would give the bias pot a range from 0 volts up to whatever Vref is. If you're lucky this might bias your J201s.

[Steben]

Can I change the Zener diode to green or yellow LED? This will lower the 4.7V voltage drop to 3V a 2V?

[Paul Marossy]

I used J201s in mine and it works great. And on top of that, it was the first two I randomly picked out of a pile of them! 

[MartyMart]

That's odd, i was specificly told by Torchy that J201's would not work in a 45 !!

oh well

Marty.

[Paul Marossy]

That's odd, i was specificly told by Torchy that J201's would not work in a 45 !!

Yes, it is odd. But they do work for me! 

[R.G.]

Guys, this is not an unknowable thing, nor is rumor a good way to figure it out. Consult the datasheet.

(1) JFETs have a resistive region of their operation if the voltage from drain to source is about +/-0.1 to 0.5V. There are ways to make this bigger, but the raw device does this.
(2) The JFET goes from on (that is, resistance between source and drain is Rdson, a few ohms to few hundred, depending on device) to fully off (many megohms) with a voltage between its source and gate of 0 (fully on) to -Vgsoff. For the J201, the Rdson is 750 typically.
(3) Vgsoff is unique per JFET type, and per JFET. JFET types specify the Vgsoff range. For the J201 this is -0.3 to -1.5V.
(4) For a phaser, you want the phasing resistors to be quite similar in resistance.
(5) In JFETs, very close Vgsoff is associated with close Rdson over the range of 0 to Vgsoff. Not perfect, but close enough for phasers.
(6) In a P90 style phaser, the source of the N-channel JFET is hung at a reference voltage in the middle of the "9v" power supply; the P90 was 3.0V, the tonepad version upped that to 5.1V so that more JFETs would work.
(7) You can only pull the gate down to ground in a P90 style phaser, so the Vgsoff has to be less than the reference voltage, which is the zener voltage in this circuit, so any JFET you use MUST have Vgsoff of less than the zener voltage to have a chance to work. It is possible that some JFET with a datasheet range above the zener voltage MAY work if you select specific JFETs that have Vgsoff lower than the zener voltage.
( The LFO voltage in the phaser must be able to drive the JFET VGS from nearly off to nearly on. Outside these ranges of voltage, nothing happens. So a J201 needs a small (for the 1.5V Vgsoff devices) to miniscule (for the 0.3V Vgsoff devices) LFO voltage. The P90 LFO may need tinkering to do a sweep voltage this small. The stock P90 used a 2N5292 (Vgsoff  = -1 to -3.0V).
(9) There are J201's that will work, according to the datasheet. These will be the ones in the upper end of the Vgsoff range. There are J201's that will not work. These will be the ones from the lower end of the Vgsoff range.
(10) J201 is chancy. However, so are all specific JFETs. I like either the 2N5292 or the 2N5484 or 2N5485. However there are many that can work.
(11) Your best chance is matching to a voltage within the 1V to <zener voltage> range. For tonepad's 5.1V zener, get a JFET with a specified Vgsoff of -1 to -5V, and match to about -2.5. I use the 3.0V zener and match to -1.5 to -1.8V. They have always worked right when matched that way.

[Paul Perry (Frostwave)]

That's a 'keeper'., R.G. And, I'm keeping it to remind me why I never build anything with fets (except buffers).

[nelson]

The matched fets have always put me off building an MXR phaser.

[R.G.]

The matched fets have always put me off building an MXR phaser.

Don't do that. Knowledge is the lever that lets you move large objects.

You only need an opamp, a battery, some resistors and your DVM to match JFETs. Get two to three times as many JFETs as you need. Order a dozen. They're under US$0.25 each. You'll use the others in something. Get 2N5292 or 2N5484, or 2N5485. Mouser has them all. Match away!

[nelson]

Sound advise.

I have always tended more towards Photocoupler based phasers, MXR phase 100 for example. I am currently working on a layout for the EH polyphase too.

I really should give a phase 90/45 a bash.

[RDV]

My Phase45 has been making me crazy for about a year now. It seemed like I had it happening, then it goes away. Tonight I tried J201s and found a couple that worked right in the middle of the bias pot. It's more subtle than before, but it's not distorting as bad either.

Here's a Sample From 2night

RDV

[Paul Marossy]

I made one of the FET matching circuits from GEO. It's a handy dandy circuit! 

[Fret Wire]

I have always tended more towards Photocoupler based phasers, MXR phase 100 for example. I really should give a phase 90/45 a bash.

The Phase 100 requires matched vactrols for best phasing sound. The factory calibration procedure calls for re-matching the vactrols (VTL5C3/2) when the relevant voltage test points aren't up to spec. If you want to build an MXR phaser (45, 90, 100), there's no escaping having to match components.

Question for RG: Didn't the original script 90 use a 3.9v zener?

[R.G.]

Didn't the original script 90 use a 3.9v zener?

Maybe. The few I've measured are about 3.0 to 3.2V. I don't know the original type number.

I suspect that the actual value doesn't matter as long as the JFET thresholds are smaller. A 4 or 5V zener should be fine. Much above 5 and you run out of linear room for the opamps, especially under low battery conditions. The opamps have a linear region that's smaller than the power supply by a volt or two on each side. Lower a battery to 7V, use a 5V zener, and have an opamp with an output voltage range of Vsupply - 1V and you can only swing 1V of signal without distortion. Ideally the threshold ought to be half the power supply, but the LFO is referenced to the negative supply - ground that is in this application. So they picked something that worked OK down to say 6.5-7V, and gave up signal range on the negative side, but not a huge amount.

[Fret Wire]

Makes sense, thank you. :icon_smile:  Maybe my thinking is flawed, but it sounds like zeners in the 3v range would be more battery friendly, and the mid 4v-5v zeners better suited for power supply operation. The P-90 also can distort with higher output pickups.....again, would I be off in thinking that jfets matched to a lower voltage zener would help to alleviate the distortion?

Any tips on testing the 5-lead VTL5C3/2 other than trial and error matching?