mxr phase 90 zener diode

[domenico]

Hi! I have built my phase 90 clone with matched bf245.
I used 4,7 w zener diode but the phase effect was very weak then I tried a 5,1 w zener diode and now my phase "seems" to work , so my question is  how many values do exist for zener diods? should I try other values than 5,1w ?
please help me and thanks in advance!
domenico

[Boofhead]

The correct diode is 5.1V and 500mW.  Note if you use 1W then the actual voltage across the zener may be lower than desired.  In the actual circuit the voltage across a 5.1V zener will be around 4.8V (it's lower than the zener voltage because the zener is run at low currents, and it suppose to be like this).

Putting aside what the correct value is, with a given set of JFETs a higher or lower may sound better.   If you have to max-out bias trimpot to one then the JFET pinch-off voltage could be different to the standard JFETs.  The opamp choice can also have an effect too because if affects the VCO. If you don't have to max your trim pot out then it's a VCO issue.  I wouldn't be worried about using zeners upto 5.6V in that ckt.  If you can't decide which voltage  to use err on choosing zeners closer to 5.1V.   When changing the zener and making comparisons you *must* optimally adjust the trim pot in each case - if you don't you are wasting your time making the comparison.

[R.G.]

The correct diode is 5.1V

Depends on the definition of "correct". The original used a 3.9V zener and 2N5292 JFETs, which (JFETs) have a Vgsoff of less than 3V.

The reason the zener and JFET matter is that the zener simply has to be enough to turn the JFET almost totally off.  With a 9V battery that can get down to as low as 7V, you have to pick a voltage that's going to give you enough voltage room for the opamps to run as well as turning the JFET off, since the P90 opamps are biased from the same zener voltage. Most common opamps can only get within 2V of the + supply, so with a good 9V battery, you're OK at 5.6V for the zener, but really get into trouble when the battery gets down to 7.6V.

I don't know, but I believe from the voltage budget that this is the reason that MXR picked a sub-4V zener and used the somewhat hard to find 2N5292 with the low Vgsoff.

Another problem you can run into with higher Vgsoff zeners is that even if you get the opamps and the JFETs running correctly, you may not have enough voltage swing out of the LFO with the stock resistors. You'll have to diddle the LFO output resistors to get the output back up.

[Boofhead]

original used a 3.9V zener

I'm quite cetain the original was a 5.1V 500mW. This was confirmed by Steve Cerutti (thanks Steve).  It turned out this zener is also used in some other MXR effects.

The problem is the zener runs at a low current so the measured zener voltage is lower than the quoted part voltage.  I think this is where is discrepency lies between the figures people quote.  Steve also confirm the measured in-circuit voltage (which was 4.8V and that holds-up with the 5.1V part rating).

[R.G.]

I'm quite cetain the original was a 5.1V 500mW. This was confirmed by Steve Cerutti (thanks Steve). It turned out this zener is also used in some other MXR effects.

Perhaps you and I looked at two different models of the P90 - there have been at least three. But the ones I have traced were under 4V on the two earlier models. I have no doubt that 5.1V got used other places in the MXR line, and could well have been used in some versions.

The reason I'm so sure about this is that I spent a good, long time trying to find 2N5292 JFETs during a drought of them a few years back. I spent a long time pondering the design and why they used that @#$%# JFET in particular, and finally realized that the low VGSoff and zener value were a paired part of the design. I finally came up with the 2N5485 as an alternate part number, with its Vgsoff at about 3.5 max.

The comments about voltage headroom apply in any case.

[Steve C]

I'm quite cetain the original was a 5.1V 500mW. This was confirmed by Steve Cerutti (thanks Steve). It turned out this zener is also used in some other MXR effects.

Perhaps you and I looked at two different models of the P90 - there have been at least three. But the ones I have traced were under 4V on the two earlier models. I have no doubt that 5.1V got used other places in the MXR line, and could well have been used in some versions.

The one I have is the (just to be as specific as possible) script logo with a pot that dates 1975, and with the 2N5952's in it.  If it helps any.

[Boofhead]

But the ones I have traced were under 4V on the two earlier models.

Thanks RG.   It's damn hard to pin this stuff down sometimes (and arhh I though I put this one to bed).  Is that 4V figure measured?  I can see if a 5.1V 1W zener (as opposed to 500mW) is used it may end-up with sub 4V voltage - total speculation here of course.

Totally agree about the dynamic range issue - that's one reason why the voltage across the zener should be in the 4.5V region.

[R.G.]

I'd go look, but I no longer have the pedals I traced from. They were borrowed from a used equipment place. This was in about '95 or so.

It was both measured and looked up. The voltage was right for the zener, and the zener was marked with the 1N? part number, which was supposed to yield that voltage. I may have the original tracings in a box in the attic. It's been a while.

The knee of low voltage zeners is not too sharp, which is what I think you're meaning with the "measured" comment. If you starve them for current, they operate with a lot of noise in the rounded curve of the characteristic, and a lower average voltage. Avalance diodes (usually, everything over about 7V) have a much more abrupt knee.

No, it wasn't running starved, and was sitting where its part number said it should. I guess I never really thought about other options on that pedal, since the reverse engineering thinking I did linked the JFET choice to the zener voltage, and the zener to the power supply variation, and it all made sense.

Maybe EH wasn't the only outfit to build with available parts. Shoot, most of the worked, I bet.

[Boofhead]

It was both measured and looked up.

Very interesting, looks like we have a variant on our hands.  Now comes the messy business of working out which year (or maybe even batch!) the different zener voltages appeared in.  It was hard enough weeding this out last time.  Without Steve's help I wouldn't have been so confident about the value.  Steve's was an earlier model, and another guy's was an early but non-script version (IIRC).

If you are ever chasing spiders or squirrels in your attic, I'd be grateful if you could take a peek at your old tracings.  Don't go scrounging around up there just for this though.

knee of low voltage zeners is not too sharp

Yep.

. I guess I never really thought about other options on that pedal,

I wouldn't have questioned it either.  It's only over the past four years or so that I've realize just how many variations sneek in.  It's mind boggling trying to sort through the small variations and the sparsness of available units (particulary with details of components) makes matters worse.  In some cases even the manufacturers schematics haven't given the whole story.