Fet array for a phaser??

[emstin1]

I was cruising through on online electronics supplier's catalog and saw a bunch of FET arrays.  Now, my knowledge of what goes into a phaser are pretty limited, but would it be possible or desirable to use one of these arrays in a phaser?  I'm thinking no, because all the ingenious people on here would probably have jumped on that right quick, but I thought I would ask anyways and maybe learn something new and interesting.  Thanks in advance. 

[Mark Hammer]

Well they certainly have been used in past.  See, for example, the early orange Ross phaser, shown here.

It is not the fact that they do or don't come in an array as mush as the arrangement of the array.  It is common for transistor arrays to either share, or not share connections to certain pins.  For a FET array to do the job, the right pins have to be available to you.  In the Ross, above, you can see that two pins of each FET are made available to the user, and the third is common across all the FETs.  Of course, once you have that particular aspect tackled, the FETs still need to have the appropriate range of control and resistance change.

[edvard]

I've asked this question before because I've found numerous "cookbook" phaser circuits that used a fet array.
Here's the circuit in question:
http://www.aronnelson.com/gallery/main.php/v/edvard/phaser.jpg.html
The Ross phaser is fairly close.

What website did you find the FET arrays at?
Since this circuit has intrigued me for some time, I'd be willing to buy a few just to have the satisfaction of building the thing...

[emstin1]

It was some Asian supplier.  Hong Kong I think.  Here's where I found them http://www.alibaba.com/product-gs/225641170/MOSFET_Array_IC_Integrated_Circuit_STA509A.html

A lot of it seemed to be SMT.  Maybe I was a little misleading calling it a catalog though...

[Mark Hammer]

The datasheet is here: http://www.alldatasheet.com/datasheet-pdf/pdf/38436/SANKEN/STA509.html

I have no idea of their applicability.

[edvard]

Aha... Alibaba is one of many sites for finding wholesale quantities of inexpensive asian goods.
If you request a quote, you'll most likely find they won't talk to you for quantities less than a few thousand.
Still, good to know the stuff still exists somewhere...

[emstin1]

All the Sanken arrays (like the one I posted) seem to be motor drivers or something.

[Taylor]

I suspect that the PWM'ed resistor is a more sensible direction in which to go if we want to eliminate FET matching and all that from phasers. Or the Bad Stone inverter trick. I could be wrong, but I don't think we really gain any magic with FETs - they do distort, and there are probably people who like that, but you can always get distortion elsewhere.

[R.G.]

Some perspective.

I believe the circuit you found was in an internet copy of a book of a collection of circuits, and that the original circuit was in the National Semiconductor Audio Applications handbook from 1973... 

Back in that day, National Semiconductor and some other companies made JFET arrays for analog switching purposes. At that time, CMOS transmission gates which are the basis for all CMOS analog switches, had not yet been made available as CMOS logic. In fact, I'm not sure that there was ANY CMOS available commercially yet. The JFET array was typically used to switch the currents into the inverting input of an opamp for signal switching.

The JFET arrays were (as I remember, it's been a while ) never specified as linear, only for switching. But they did work as linear if you perverted them that way. As such, the linear matching was not specified or tested, but was generally good if the JFETs were on a common substrate.

They largely vanished when cheaper, better performance CMOS analog switches made their debut. I don't think there are any JFET arrays in current production, although I have not looked. I think they were last made in the 70s, maybe 80s. I think I still have a tube of the AM? arrays I picked up in a surplus store and never used.

It's useful to differentiate between making one or two pedals and making quite a number, especially to sell. If the design depends on a magic part, then I would not start building until I had one of these magic items in my hand. If I had to make a dozen, I'd locate a dozen first. If I was doing a design for publication or production, I would not design in a magic part at all. I've handled too many questions about "where can I find X?"

[Mark Hammer]

Though I haven't spoken to him in a while (I do owe that boy a phone call), phaser genius Mike Irwin had explored use of invertors as control elements for phasers, starting with the 4049 used in the 6-stage ETI Phaser ( http://synthdiy.com/show/?id=1069 ).  He found it serviceable, but lacking.  The chief limitation was signal handling, which necessitated use of a pre-attenuation/post-boost strategy.  That, in turn, made the S/N ratio less than stellar, and also made use of feedback (noticeably absent from the ETI design) contra-indicated (you really don't want to tax the invertor sections with feedback if they can't even handle an attenuated signal very well).

All of that being said, with 2N5952 and other acceptable FETs, being as plentiful and inexpensive as they are, one is always free to use the NatSem design with discrete JFETs.  All those AM9709/9711s were is a design/layout convenience.  You can certainly match JFETs yourself.  And if the goal is to dicker with a one-off, or maybe a personal unit and build for a friend, it's likely easier to just get the JFETs, match 'em and be done with it, than drive yourself crazy looking for arrays in a chip that meet some unspecified specification.

[edvard]

Some perspective.

I believe the circuit you found was in an internet copy of a book of a collection of circuits, and that the original circuit was in the National Semiconductor Audio Applications handbook from 1973...  

Actually, the circuit I have is a scan from a book that I have a hard copy of.
Before the internet, those kinds of books were my Boy Scout Handbook for pedal building, and I scoured the second hand shops for them.

You're absolutely right about the "magic part" thing; although I did try to build that very circuit with discretes and it just didn't work, so I always wondered if there weren't some "magic" to the nigh-unobtainable part after all.
Then again, that was ~1994 or so, maybe I should try again...

[deadastronaut]

rather than start a new thread,

i'm looking for a fet array to replace the 4 fets in a phaser

any recommendations...?.

just thought i'd give it a go as i have a phaser on bread at the mo, might be interesting. ...

[Hatredman]

I know you want the atray for fun, but why not go the PWM way?

[deadastronaut]

hmmm...i would, but i'd want to try the fet array for other stuff too...

[anotherjim]

Rare as Rocking Horse sheet!
There are JFET analog switches, but I doubt those will work linear.
Even BJT arrays like the good old CA3046 are uncommon now.
Besides, JFET's being as awkward as they are, they might not match how you like even when they are formed on the same die.

[deadastronaut]

yeah i thought they might be a good way to go 'matching' wise....hmmmmm...

maybe not then..