Jfet matching for phase 45 build

[Locrian99]

K I'm thinking I'm at least getting the gist of this.   Reading the electro smash article as well.   Plenty of it flies over my head, but getting an bit of understanding from it.   

[Rob Strand]

K I'm thinking I'm at least getting the gist of this.   Reading the electro smash article as well.   Plenty of it flies over my head, but getting an bit of understanding from it.   

Electrosmash have some very good articles.  There's a lot to take in.

The stuff I mentioned about varying voltages and adjustments is beyond that!  These are fine details that are never addressed.   I've never seen it mentioned in any articles.  The issues really only show up in phasers which use JFETs.  If you can get to the point where you can see the stuff I mentioned *could matter* then you have done extremely well.

[Locrian99]

So my diodes came in.  Set up the little breadboard test.   Basically just a 1k resistor between the power rail and the diode.   I'm getting 4.85v with the 4 I tested.   Should this be good?

[Rob Strand]

So my diodes came in.  Set up the little breadboard test.   Basically just a 1k resistor between the power rail and the diode.   I'm getting 4.85v with the 4 I tested.   Should this be good?

Yes it will work fine.

The down side of using 1k is there is 4.2mA going down the zener.   Whereas the original had 0.42mA, somewhat less battery drain.    With 2.2k you only use 1.9mA.  The voltage will be a little lower but it's good trade-off in terms of current.   Ideally you want at least enough current to keep the zener voltage in the 4.5V to 5.0V region.   With a 10k and 5.1V 1W you start to loose out as the voltage is 4.0V or even lower in some cases.

[Locrian99]

Cool I'll check out some slightly larger resistors see where it gets me.   

[Locrian99]

Hey if you have a second to take a look at this layout, is this circled 10k r15 from the schematic?   Only one I see coming off the 9V source so I figure it's got to be, just seems odd it's heading straight to the source strip on the jfets. 

[Rob Strand]

Hey if you have a second to take a look at this layout, is this circled 10k r15 from the schematic?   Only one I see coming off the 9V source so I figure it's got to be, just seems odd it's heading straight to the source strip on the jfets. 

Looks like the resistor to me.   All the sources wire back to the zener (Vref, 5.1V).  It's the only 10k resistor between +9V and Vref.

[Locrian99]

Cool found some .5 watts on eBay going to socket the diode and resistor.   Thanks again for answering my questions.  You've been extremely helpful as I try and figure this stuff out.   

[Locrian99]

Well it's built and working.  Biasing this thing is harder than in anticipated.   I'll have to try to dial it in tomorrow. 

[Phend]

Hi:
I completed this 45 and it works. Maybe it can work better.
Reading this interesting topic has lead me to a question(s).
First, I added a 1N4001 "protection" diode to the circuit, is this ok on this build ?
(See picture)
Second, What should the voltage be at D1 ?
My battery is 9.23 v and at D1 to ground is 3.51.
D1 is a 1 watt 1n4732A from Small Bear.  I have lots of 1n914.
Should I try that in series with D1 ?
Thanks in advance..

[PRR]

> D1 to ground is 3.51

Let's peek at 1N4732A specs.

This is a One Watt part, test for 4.7V at 53mA. Here it is used at 0.4mA, 0.02 Watts. We would expect voltage drop to be less. The other data is 500r dynamic resistance at 1mA. That would be a half-Volt different, and we are significantly under 1mA. And low-Volt "Zeners" are really soft knees. 3.5V is less than I expect but not unreasonable. Yes, if I thought I needed to be closer to 4.7V I'd try a couple '914s.

But does it have to be a specific voltage? I think it needs to be "mid-way", on a varying battery, and steady (actually tracking JFET Vgs). 3.5V may be a little tight on the downswing but if you are not clipping there may be no "better" here.

There is another plan which uses a JFET for bias, actually two, plus a trim. Looks better in theory but seems to be a pain in practice.

[PRR]

Ah, real-like curves.


This seems to be a smaller series than your 1N4732A; note that the "4.7" part hits 4.7V at like 7mA (blue dots). The "4.3" part seems closer to 1N4732A spec showing 4.7V (purple dots) @ ~~50mA (which may be out of bounds). It shows 3.3V in P45 circuit.


LEDs are good regulators. The several colors are several innate voltages, and also vary with current. Three IR LEDs at low current may be awful close to 4.5V, but you can't know if it is "on". (g)

[Rob Strand]

Second, What should the voltage be at D1 ?
My battery is 9.23 v and at D1 to ground is 3.51.
D1 is a 1 watt 1n4732A from Small Bear.

The correct voltage is about 4.8V with the correct zeners.

The correct zener is 1N5231B, 5.1V 500mW; see attached schem.
In the phase90/phase45 circuit, with a 9V supply and 10k, it produces about 4.8V.

Problem 1: What happens in the DIY world is someone measures the voltage at 4.8V, then they assume the
zener voltage is the measured voltage ie. 4.7V.  But it's not a 4.7V zener, it's a 5.1V operating at a low current.   

Problem 2:  People don't realize 1W zeners and 500mW zeners are different.
A 5.1V 1W zener will produce a lower voltage than a 5.1V 500mW zener for the same current.

What's happened with some of the Phase 45/Phase 90  builds is an accumulation of errors
from these two problems:  The first means sometimes the zener voltage is incorrectly specified
at 4.7V zener.   The second means some people buy a 1W zener and not 500mW and net
result is the 4.7V 1W zener in the phase45/Phase90 circuit ends up way down at 3.6V!

What the means is the pedal has lost 1V of headroom because the wrong zener is used,
not great for a 9V pedal.  If you can bias Vref for 4.5V to 5.0V the headroom will return.
The exact voltage is not important as the trimpot has enough adjustment to set the
JFET bias.

PRR' first graph shows the first problem, however it doesn't show the second problem.

You will also notice PRR's curves doesn't quite match-up with the 1N5231B, 5.1V 500mW
zener in the Phase 90/Phase45 since it's showing about 4.4V at 400uA whereas
the voltage should be about 4.8V.   Those curves only apply to a specific model of
zener.

Each model zener has it's own set of curves, even if it has the same power rating.
Unfortunately those detailed curves are rarely given, so you
have to eyeball what's going on for a specific zener by making it pass through
a known point (from the single point in the datasheet for *your* zener.)

For problem 2 the curves for a 1W zener will have lower voltages for the *same
current*.  It's sort of like the curves for the 1W zeners are like a 500mW zener with
a lower voltage rating *but* that's technically not true - since it's the current axis which is
changed.  When we want to compare a 1W and 500mW at the same current we are forced to
compare voltages.




When you can only get 1W zener's you can either use a higher voltage and/or use a higher current,

https://www.diystompboxes.com/smfforum/index.php?topic=129436.msg1251001#msg1251001
https://www.diystompboxes.com/smfforum/index.php?topic=129436.msg1251103#msg1251103

However, if you raise the current on your 1W 4.7V zener it will help but since the voltage is wrong
the voltage is always going to be low - you would be better off changing it to a 5.1V 500mW.

typos fixed!

[Phend]

Thanks Paul and Rob. Fortunately resistors don't follow the same quirks as zeners, lol. Valuable info from both, thanks. Not knowing I guess I thought a Dollar General 5.1v zener was the same as a Dollar Store one. Thanks for the help and support. Chalk this up to lessons learned. Shall find correct zener.

[Locrian99]

Hmm.  Shouldn't the lfo be causing the voltage of the fete to be jumping around?

[PRR]

Just get a good voltage, not a magic Zener. 1N914 shims are perfectly valid. The exact voltage does not have to be as exact as the line down the center of the road. (You know what I mean.)

[Rob Strand]

Hmm.  Shouldn't the lfo be causing the voltage of the fete to be jumping around?

The opamps are biased to Vref, which is set by the zener.  That is a fix voltage.  It is desirable to set the voltage so the opamps can swing symmetrically for maximum headroom.  That occurs when Vref is about half the supply voltage.

The zener is also used to set the bias point of the JFETs.   The bias voltage is *between* the source and gate.   That's set by the trimpot, so as far as the JFETS go the exact Vref voltage isn't important (although you want it to be fairly stable and the adjustment must match the JFETs).

The LFO modulates the gate voltage.   The actual gate voltage is a combination of the bias trimpot and the LFO voltage.  The two are mixed together where the 470k resistors and 3M9 (or 3M3) resistors meet.

[Locrian99]

I'll go through it again I've got audio fine and when I play and Kobe the trimmer in what I hear a phase kikeneffect.   Leave the trimmer be there is none.   And I don't see any voltage movement at the fets or anywhere else.   after the kids go to bed I'll play with it some more.  I accidentally plugged it into my test box backwards for power (reversed polarity) wondering if I fried one of the op amps.   I'll play around with some more, only thing that struck me as odd was pins 5/6 of the bottom op amp were 0 when I did my quick check through.   

[Locrian99]

Tiny little solder bridge where the 15u cap was connecting to ground was grounding out pin 6 and the LFO.   Works great now, now to to play with the bias and resistors feeding that diode.   Tomorrow I've got some coming which will have a 5.6v 1w I'll try.   Had to get the 5.1 1/2 watt from eBay so hopefully sometime next week.   Will wait to box this up and call done until then.  Now to remember the chords to lightning crashes lol. 

[Rob Strand]

Tiny little solder bridge where the 15u cap was connecting to ground was grounding out pin 6 and the LFO.   Works great now, now to to play with the bias and resistors feeding that diode.   Tomorrow I've got some coming which will have a 5.6v 1w I'll try.   Had to get the 5.1 1/2 watt from eBay so hopefully sometime next week.   Will wait to box this up and call done until then.  Now to remember the chords to lightning crashes lol.

Good one!