Tonepad MXR Pahse 90. Not much mojo.

[Sea Chief]

My DIY phase 90 works, but effect is weak instead of a nice strong phase. The trimpot is at its best setting already. Quite a big vol drop once on. Also seems to drain battery very fast.

Twds middle it says CLICK HERE for the layout & schematic:

tonepad -- file information

Any ideas?

Thanks SC.

[Mark Hammer]

Are the JFETs you are using reasonably well-matched for Vgs, and is the Vgs in a usable range?

While they don't have to be exquisitely matched, JFETs closer to each other in properties tend to produce more pleasing-sounding phasing, simply because they are all responding to the LFO in similar fashion. When unmatched, it can be the case that one or more reaches the "end of travel" and stops while the others are still changing their drain-source resistance, whether at the top or bottom end of the sweep.

[Sea Chief]

Hi Mark- to be honest I dont know. How do I tell? I seem to remember choosing from 15 or so I bought.. but was a few yrs ago. I recently dug pedal out again.

[J M Fahey]

There is a method posted out there,*guess* it was by RG Keen, guess there's other by Rod Elliott, about matching Fets.
Hey, even I posted my own

That said, and to discard it in this case, mine was centered on matching them for repeatable consistent preamp designs (I basically matched bias at 1mA), while RG's was specifically centered on phase shifters (he matches fet equivalent resistance at a given bias, a different thing) so in this case his is the most approppriate by far.

Unmatched (or poorly done) FETs wreac havoc with the effect achieved.

I remember a Studio type one with individual trimpots, go figure.

Rod's was halfway between ours.

[Sea Chief]

Can you remind me how I would (or maybe I did/ maybe I didnt) match transistors? I mean if its a BYOC affair the builder surely isnt required to buy a FET matcher thing. Whatever that may be. Can they not be bought as a matched 4?

A good thing is I have the transistors in 'cradles' so able to remove simply.

[frus]

put just one FET in any socket
adjust the trim pot to a position where you hear the best phasing effect. note the position for that FET.
repeat the above for all the FETs you have
choose 4 for which the position was closest. put them in sockets and of course set the noted position on the trim pot

that's it

[frus]

as for the volume drop issue, there's a good advice right on the tonepad page:

"The resistor to change the level is the 150k which connects between the collector and base of the PNP transistor: 120k to reduce gain, 180k to increase it."

that's the (black) 150k resistor on the schematic, on the transistor closest to output. I remember having to increase the gain on mine also.
Maybe the output volume depends on how close the FETs are matched, so you need a little tweak if they are not a perfect match (just my assumption, may be totally wrong )

[Sea Chief]

put just one FET in any socket
adjust the trim pot to a position where you hear the best phasing effect. note the position for that FET.
repeat the above for all the FETs you have
choose 4 for which the position was closest. put them in sockets and of course set the noted position on the trim pot

that's it

Hi frus. Sorry Im not with you.. I have one trimpot and 4 FeTS. I dont have any others/ long since gone awol. The trimpot seems only to have any effect audible between just a very small incriment of 11.30 and 12.30am (looking at the trimpot mounted vertically as I have it, IE as a clockface).

Can you buy 4 matched FEts from anywhere? I seem to rememebr it was a massive pita to test & 'choose' 4. Did I have 15 or so? Id little idea what I was doing tbh so close & small all the tiny damn numbers were.

[J M Fahey]

Can you remind me how I would (or maybe I did/ maybe I didnt) match transistors? I mean if its a BYOC affair the builder surely isnt required to buy a FET matcher thing. Whatever that may be. Can they not be bought as a matched 4?

A good thing is I have the transistors in 'cradles' so able to remove simply.

No need to buy anything , just kludge this on ProtoBoard (if you don't have one, you can't realistically expect to experiment with pedals):
http://www.geofex.com/article_folder...h/fetmatch.htm


the Op Amp is *any* single Op Amp (LM741/LF351/TL071) ; if you have doubles (TL072/RC4558/etc.) just use one, and neuter the other by connecting its output to the "-" input and the "+" input to Vmid and ignoring it .

Measure and label all FETs you have for future use and disassemble the tester if you need to free the Protoboard for other use.

Then use FETs with close matching Vgs ; within millivolts if possible.

FWIW I used to buy 50 or 100 FETs (when they were cheap, now they became sort of an exotic component, speciallyn in T92 case) , then when I need any I pick whatever they need.

I hardly waste them (money is money ), because the ones with low Vgs (1.8V or less) are used as preamps (think Randall/Yamaha/Roland) and those with high Vgs (3 to 4.5V) get used as switches or in compressor-limiters so nothing goes to waste, so it pays to get a bunch of FETs once which will last years.

[frus]

Hi frus. Sorry Im not with you.. I have one trimpot and 4 FeTS. I dont have any others/ long since gone awol.

well if you need 4 and all you have is 4, then there's really nothing to match
but you can test if they are matched at all, using the above metod. just plug in one fet at the time and wiggle the trim pot until you hear the slight phasing effect. if the position of the trimpot for all the fets was different, you're out of luck. Btw, Juan's method is much better and more precis of course

I've found some listings on matched transistors on ebay, searching for "2n5952 matched"

[Sea Chief]

Hi- great info thanks. I'll read through later. Ive dug out 4 more of the same FETS so can try see if 4 of the 8 match good enough. I guess its a Q of whether the best 4 numbers are indeed a 'match' or not: without knowledge here Im blind. So maybe I'll post figures up & you could check for me? If a poor match then the ebay option link from frus- lucky there.

cheers excellent help as always on MEF. SC

[Sea Chief]

J M Fahey- I have only a TL072, and trying to understand even how to neuter one side of this is making my brain hurt (whats a Vmid?) let alone working out how to build the circuit on a breadboard without a layout and photos of both top and underside (connections). It would take me days to make this.

I wonder if its worth it anyway if Ive only got 8 of these 2N.... transistors to test, not 100 of them. (If Id known how complicated this JFET matching procedure is Id never have made the pedal tbh).

Does anyone have 4 matched they could sell me? surely the easiest way.

[J M Fahey]

J M Fahey- I have only a TL072, and trying to understand even how to neuter one side of this is making my brain hurt (whats a Vmid?)

Vmid which by the way is mentioned twice in the RG Keen schematic I posted is "the point which has half the battery voltage, so sits at the V(oltage) mid(dle) of 9V" , a pretty descriptive name for it.
Neutering refers to interconnecting 3 legs of the unused Op Amp so it does not disturb the other sharing the same room ... er ..... case.

let alone working out how to build the circuit on a breadboard without a layout and photos of both top and underside (connections). It would take me days to make this.

Sorry to hear that.
Ok, no problem, nobody is born knowing nor you are forced to do so, problem is that it limits your possibilities, such as being forced to "paint by the numbers" , meaning you need very clear and detiled instructions to build something electronic.
Nobody will hang you for that, far from it, problem is that for *some* projects it limits you.

You would really need a nice full kit, including 4 pre-matched Fets.

AFAIK, BYOC, GGG, DIY Stompboxes and most others only sell the board, maybe a specialized shop like Small Bear or similar ones offers a full kit like I mentioned, some even including a drilled and printed case..

Problem is it will probably cost as much or more than an Asian series made pedal

Less than a boutique pedal anyway

Maybe somebody sells matched Fet sets for Phasers , ior somebody here can offer you some, wouldn't hold my breath on that.

Personally I would offer you a set , problem is I'm 6000 miles away and much worse, 2 Customs barriers (plus a money exchange one) away .

I can only suggest that in the long run you practice reading schematics, transfering those symbols into physical parts layouts using paper, pencil and eraser until you get the hang of it, and then use that to build stuff on Protoboard, the experimenter's best friend.

Go at your own pace, practice makes perfect.




I wonder if its worth it anyway if Ive only got 8 of these 2N.... transistors to test, not 100 of them. (If Id known how complicated this JFET matching procedure is Id never have made the pedal tbh).

Does anyone have 4 matched they could sell me? surely the easiest way.

[Mark Hammer]

The matching thing is widely misunderstood, and overinterpreted.

All JFETs will have a range of values applied to the gate that will result in a change in the drain-source resistance. Too high, or too low, and you won't get the resistance changes appropriate to produce audible phasing.

The JFET gates in the P90, and all workalikes, receives a bias voltage from the trimpot, which summed with the LFO voltage. That combination can't be too low or you won't hear a nice sweep, and it can't be too high either.

The matching circuits are intended to identify a set of JFETs that have very similar on/off ranges, such that when the bias and LFO are adjusted to the right sum, ALL the JFETs will sweep at the same time, and none of them will reach a point at either the top or bottom of the sweep cycle where they simply give up and sit idle, waiting for the sweep to reach the end and come back again, where the JFET can rejoin the sweep.

Matching for resistance DOES NOT MATTER. I repeat, it does not matter. The 22k parallel resistors set the max resistance, anyway, and none of the JFETs will ever reach their minimum resistance. Besides, the resistance simply determines how much phase shift is produced, at Frequency X, in that stage, at that point in the sweep. The sound you hear is the cumulative phase acorss the four stages, which will vary by the tolerances of the 10k resistors, the caps, and the 22k resistors. So, unit-to-unit differences in resistance of the JFETs specifically will have negligible impact. What will matter most is that all 4 have their oars in the water at the same time, and paddle together.

[J M Fahey]

Sorry Mark but in this single post you contradict yourself, almost line by line

The matching thing is widely misunderstood, and overinterpreted.

So it does not matter.

All JFETs will have a range of values applied to the gate that will result in a change in the drain-source resistance. Too high, or too low, and you won't get the resistance changes appropriate to produce audible phasing.

So it does.

The JFET gates in the P90, and all workalikes, receives a bias voltage from the trimpot, which summed with the LFO voltage. That combination can't be too low or you won't hear a nice sweep, and it can't be too high either.

So it does.

The matching circuits are intended to identify a set of JFETs that have very similar on/off ranges, such that when the bias and LFO are adjusted to the right sum, ALL the JFETs will sweep at the same time, and none of them will reach a point at either the top or bottom of the sweep cycle where they simply give up and sit idle, waiting for the sweep to reach the end and come back again, where the JFET can rejoin the sweep.

So it does.

Matching for resistance DOES NOT MATTER. I repeat, it does not matter.

So it does not. ... So it does not. (sorry, couldn't resist )

The 22k parallel resistors set the max resistance, anyway, and none of the JFETs will ever reach their minimum resistance.

True, but within that range lies the phasing effect, the RC cells were dimensioned according to that.
Sweeping between 22k (in parallel with the fixed 22k will give you around 10k effective resistance) and, say, 1k , easy to reqch for any Fet, gives you a 10:1 sweep range , how's that?
Add that you have at least 2 , typically 4 and sometimes 6 cascaded stages and it's the effect we all know.
Now if one is 22k , other 10k , other 5k and other 2k , all at the same sweep+bias voltage and that Phaser is a mess.

Besides, the resistance simply determines how much phase shift is produced, at Frequency X, in that stage, at that point in the sweep. The sound you hear is the cumulative phase acorss the four stages,

"simply?" ... that's the point of the whole circuit.
So I guess it counts as: So it does

which will vary by the tolerances of the 10k resistors, the caps, and the 22k resistors.

Resistors and even caps are much more consistent than Fets, and that is an understatement.

So, unit-to-unit differences in resistance of the JFETs specifically will have negligible impact.

So it does not matter?

What will matter most is that all 4 have their oars in the water at the same time, and paddle together.

So it does.

Not sure about how to read this post: do I take those opinions as votes and count them to see which wins, do I average them? (almost the same) ... or consider the last one the definitive conclusion?
Oh well.