Tips on building a Paia Phlanger

[kuttlefish]

I've had a Paia Phlanger PCB that I bought off ebay years ago and never got around to building and I recently acquired a SAD1024 for a "reasonable" price (under $100) so I think it's time to finally get this project going. I've built a few fuzzes from kits and a 4ms tremolo years ago, but this will be one of the more complicated projects I've ever tackled. Given the rarity of these parts, I want to take it slow and make sure I get everything right. I'm also open to making some improvements to the design if that's an option.

I've found the proper high-resolution version of the manual/build instructions, and Vol 3, issue 4a of Polyphony that shows the low noise mod (replace two of the opamps from 3900 with a 4739).
https://www.dropbox.com/s/ugqjt4xngjhjk96/1500.pdf?dl=1
https://www.dropbox.com/s/mf47zmdhukzn9ad/Polyphony-V3-4A.pdf?dl=1

Some initial questions I have before I get started:

   
• Is the transformer/diode network necessary? It looks like the power gets dropped down to 9V before touching anything important in the circuit. Can I just use a 2.1mm jack and wire it to where the positive side of c15 and the negative side of c16 would be?
• If the transformer can be removed, are there any advantages to keeping it? Increased headroom? secret mojo sauce? It seems like a clunky way to get 9V from the wall (since I imagine a 9V battery would last 2 seconds in this circuit) I have a Cioks power supply that seems like it would be more consistent and quieter power source if I can use it. It will give me 660mA at 9volts, 500 @12V, 400@15V or 330 @18V
• The Polyphony article mentions a 4739 dual low noise opamp. I think I saw another post in this forum that said there are better options that could do the same thing. Anyone have any recommendations? Or at least know what specs I should be looking for? The article also mentions removing R16 and replacing it with a .1 uF coupling cap. Would a different opamp require a different value here?
• Anyone have any ideas for a case? Like I said, I just got a spare PCB and it's 70's huge. 9.5" x 4.75" I have some decent woodworking tools, but no metal machining capabilities
• Any other ideas for improvements or modifications to this circuit? Any general tips for the build?

I'll probably have a million more questions as I work my way through this, but that's what I've got for the initial planning phase

[Govmnt_Lacky]

   

• Is the transformer/diode network necessary? It looks like the power gets dropped down to 9V before touching anything important in the circuit. Can I just use a 2.1mm jack and wire it to where the positive side of c15 and the negative side of c16 would be?
• If the transformer can be removed, are there any advantages to keeping it? Increased headroom? secret mojo sauce? It seems like a clunky way to get 9V from the wall (since I imagine a 9V battery would last 2 seconds in this circuit) I have a Cioks power supply that seems like it would be more consistent and quieter power source if I can use it. It will give me 660mA at 9volts, 500 @12V, 400@15V or 330 @18V

I cannot see the links however, I would be hard pressed to think that this circuit has the delay portion running at anything less than 12VDC. This means that somewhere in the circuit there is a regulator (7812, 7815, 78L12, or 78L15) which would be powered by the transformer/diode network you speak of.

[Fender3D]

According to schematic this Phlanger indeed works at 9V.
As Greg said above, BBD should work at 12 to 15 V for a better headroom.
I don't know if VFO/VCO works ok at such voltage supply, though...
About op-amps:
I don't know if you'll hear any significant difference swapping op-amps, since BBD will hiss far louder than any op-amp available around, expecially in this circuit with no input filter and with a well far from useful passive filter on output, anyway I suggest any modern op-amp such as 5532, TL072 or, for some more pennies, OPA series

[Mark Hammer]

The PAiA Phlanger was my first effect pedal in the late '70s.  It would take several more decades for the rest of the pedal world to catch up to what the Phlanger offered in 1976.  Before you tip your hat to me as a builder, note that I acquired it from a tech in the med school I was working at.  He's the one who actually built it.  I think I eventually sold it, but I still have a partly-populated board I etched.  It really spoiled me for subsequent flangers.

1) The 4739 was a pretty damn good op-amp at the time, and certainly well above a 3900 Norton in terms of audio performance.  These days, it has been eclipsed many times over by other op-amps, and is also hard to find and expensive.  Fender3D's recommendations for substitutes are spot on.

2) You may find Marvin Jones' project article in Radio Electronics a useful supplement to the PAiA-supplied build manual, though it looks like Jones may well have written both.  You can find it here:  https://worldradiohistory.com/Archive-Radio-Electronics/70s/1977/Radio-Electronics-1977-10.pdf

3) The circuit wants a bipolar supply, not simply +9V.  I can't seem to find the current requirements for the circuit, but it is hard to imagine it requiring more than maybe 40ma a side; certainly not for the negative supply.  There's a good chance you can get away with using an LT1054 charge pump to provide the requisite negative voltage from your CIOKS supply.  There's risks and benefits from this.  The benefit is that no transformer gets near the audio path, and you don't have to worry about bulky electrolytic caps for smoothing what is already smooth.  The risk is that both charge pumps and BBDs depend on clocks for their functioning, so I don't know what the potential risk of annoying heterodyning might be is a charge pump is used.

4) Do you have an LM566 for the clock?

5)  The Span, Center and Speed controls work the same way that Rate, Depth, and Manual work on a number of other modulation pedals, like the Boss BF-2 shown here.  "Span" mixes the relative amount of two sources of control voltage.  One is an LFO and the other is a fixed voltage.  Since they sum, you can't have too much.  So the Span control, like the Depth, takes from one as it gives to the other, in reciprocal fashion.

[kuttlefish]

Thank you to everyone so far! Here's the official schematic for easier reference:




For the power supply - what if I used two 2.1mm jacks connected to two separate, isolated outputs from the Cioks DC7? Something like this, but replace the batteries with power jacks:

The build sheet calls for a 300mA transformer, so amperage shouldn't be an issue.


I checked the datasheets and all the chips can handle 12-15V without a problem. Are there any other components I would need to check/adjust/worry about if I wanted to boost this up to 15V? I'm prepared to do some math, I just don't know enough about electrical engineering to know what to do.


I do not currently own an LM566. Does anyone have any good info on what to look for? Do I need to get an original or are the modern Chinese ones good enough? Or can I (should I) go with a completely different VCO?

[Fender3D]

2 DC jacks is easily prone to miswiring, especially on stage...
you'd better use 1 AC jack connected to transformer output as in this circuit



simply adjust transformer voltage to your need...

[kuttlefish]

Unless I'm misunderstanding you, I think that's basically what's in the original design. I was just seeing if there was a way to skip the transformer (and potential noise issues) if I already have a clean/smooth supply from the Cioks.

Or are you saying to put a 3 prong socket between the transformer and the diode network and move the transformer off the unit?

FWIW, i'm not worried about stage use or how others may plug it in wrong or anything like that. This is purely a personal project for my "home studio" to help me learn and experiment.

[Kevin Mitchell]

Mark had a good suggestion of trying an LT1054. The risk he had mentioned is viable but hasn't stopped folks from adapting it in other flanger designs which worked out perfectly fine. I'm not sure how efficient the layout's trace routing is to avoid said heterodyning. But I can't imagine it being bad enough to be an unresolvable issue. Even so I'd say it's a low risk factor and would for sure be worth a try.

[Mark Hammer]

For the power supply - what if I used two 2.1mm jacks connected to two separate, isolated outputs from the Cioks DC7? Something like this, but replace the batteries with power jacks:

The build sheet calls for a 300mA transformer, so amperage shouldn't be an issue.

I checked the datasheets and all the chips can handle 12-15V without a problem. Are there any other components I would need to check/adjust/worry about if I wanted to boost this up to 15V? I'm prepared to do some math, I just don't know enough about electrical engineering to know what to do.

I do not currently own an LM566. Does anyone have any good info on what to look for? Do I need to get an original or are the modern Chinese ones good enough? Or can I (should I) go with a completely different VCO?

The circuit, and calibration, are for +/-9V.  That doesn't mean it couldn't operate at a higher supply voltage, but based on the nature of your questions, I'm not confident you'd be able to make the adjustments needed.  In that case, best to stick with +/-9V.  I don't see increasing the supply voltage as contributing much to audio quality in this instance.  Keep in mind that the headroom on a BBD is quite limited, and the op-amps in the audio path will be powered by 18V (9+9).

Can your CIOKs provide a bipolar output?  I'm not familiar enough with the unit to know.  For instance, if one uses a pair of 9V outs to mimic a bipolar supply, can the + of one and - of the other form a ground?

[kuttlefish]

For increasing the voltage, if it's not plug-and-play, it's probably above my level. Maybe for another day...

I'll read up on the LT1054 - it looks like a lot of you have already answered my initial questions about that in other posts, so I'll just have to do some digging.

Is there an easy way to take an 18V output from the Cioks and split it to +-9? Even if it's not simpler, would it still be preferable to provide constant +9 and -9 than swapping back and forth?

The Cioks DC7 doesn't have a built-in bipolar option. It has 7 isolated outputs, each switchable between 9, 12, 15, or 18V (with decreasing amperage available as the voltage goes up, but way more than what this needs).  It's the same as the modern Eventide power supplies if that's helpful. Here's the two page manual if anyone wants to look at it: http://cioks.com/wp-content/uploads/2019/12/CIOKS-DC7-manual_PC_1.1_26.02.2019.pdf

[Mark Hammer]

18V can be transformed into a functional equivalent of +/-9v, the same way that +9V is regularly turned  into +/-4.5V.  Trouble is that it requires feeding a bias voltage to the inputs of the op-amps, once again upping the tech know-how requirements.

Using a charge pump is actually surprisingly easy to do, with the appnotes providing "textbook" illustrations.  Once you have the little daughter board wired up - quite feasible on a small piece of perfboard - It's a simple matter of connecting your CIOKS 9V to the charge pump daughter board and those three outputs +9V, gnd, -9V to the Phlanger board.

[kuttlefish]

I just found this:
https://www.goldpt.com/virtual_ground_circuit.html
It's a $5 preprinted circuit board that creates virtual ground circuit using an LM317 and LM337. Can I use this to replace the whole power section at the bottom of the schematic?

[Mark Hammer]

An LT1054 would be easier.  Seriously.  One 8-pin chip and three capacitors.

[Kevin Mitchell]

In case you're interested in the LT1054, here's some perfboard layouts for the charge pump. Small enough to snug into most pedals. Bottom layout is the one you'd like.
Charge Pump Perf Layouts

It really is the best solution. Using the charge pump allows the use of a standard 9V pedal power supply.

Is there an easy way to take an 18V output from the Cioks and split it to +-9? Even if it's not simpler, would it still be preferable to provide constant +9 and -9 than swapping back and forth?

You can divide the voltage in half with series resistors, buffer it via an opamp and use the output voltage as your ground reference. Your power supply's ground would be the -9V reference and the positive voltage (+18V) would be your +9V reference.

[kuttlefish]

Thank you both. I'll keep reading about the LT1054 before making a decision. I'm hesitating because you both mentioned heterodyning could be an issue and if that pops up I don't know how difficult that would be to fix. But it sounds like supplying a constant +9 and -9 would eliminate that problem. Each of the 7 outputs on CIOKS has 2 dip switches that allow you to choose between 9, 12, 15 and, 18v, and I wasn't planning to daisy chain anything else the output for this pedal since I'm assuming it's going to be noisy as hell in the best of circumstances. So the difference between a standard 9v supply vs 18V doesn't matter to me. Also, I'm ashamed to admit I don't currently have any perfboard - so the price of the goldpoint pcb + components vs the cost of perfboard + LT1054 ends up being about the same.

So if the difference is between spending 1-2 extra dollars and soldering half a dozen components now vs spending an unknown amount of time later tracking down a problem I don't really understand - option 1 sounds better.

Here are photos of the PCB (still in the shrinkwrap). I mirrored the picture of the backside so the two will line up better. The 566 is IC4. If I completely bypass the current PCB's power layout, I was thinking the +9 would go to the positive hole of c15 would be, -9 would go to the negative hole of c16, and ground can be anywhere in the giant glob.





Also, now that I'm taking a closer look at the circuit (and have a slightly better idea of what I'm looking at) - it looks like the -9 is only for the 566 VCO and the rest of the circuit is only seeing the +9.  This is probably above my current level, but I think the goldpoint board can be set up for an asymmetric split (+12 or +15 / -9 - I can combine two cioks outputs for 24V or 30V if necessary) Could that boost the headroom of the BBD and audio path opamps? I imagine the + and - going to the 566 would need to be the same for timing, and I'm sure I'm getting way ahead of myself and talking nonsense. If the goldpoint board allows for future improvements I might want to keep that door open.

And to be clear, I'm not trying to question anyone's expertise. Y'all are clearly far beyond my skill and knowledge levels. I'm just trying to wrap my head around all the potential options before messing with anything that will be difficult to replace. And if I'm being annoying and you know of any books/resources I should check out before asking more dumb questions, please point me in the right direction.

[Kevin Mitchell]

I've used the LT1054 design in a handful of BBD based pedals (flangers & delays mostly) and plenty of non-BBD effects with no problem at all. Usually if there's any noise it's only natural of the pedal at hand and simply increasing the value of the power filter capacitors rids the issue. I've only had to do that for a Univibe but that's because the LFO is a noisy, current hungry design to begin with.

Let's talk about what heterodyning is and it's part in designing a PCB layout. It's when two oscillating signals crosstalk by unavoidable capacitance between traces through the PCB material and interfere with each other. If the BBD's clock signal and power supply (or any other signal really) are crossing over or running long in parallel close together the risk of crosstalk noise increases. That's why it's a popular practice to isolate clock signals with a ground plane or trace as these oscillating signals can pour over into anything if they're strong(current) & fast(frequency) enough. If these precautions were kept in mind the pedal would be healthy in regards to noise to begin with which in turn will lessen the risk of heterodyning when using any sort of switching power supply. In this case with the PAIA, it's a minimalist circuit and the risk is far, FAR lessened.

I feel that you're overthinking the mentioned risks. With sufficient filter caps on the power rails it will be a constant voltage. Heterodyning is more prevalent in inferior charge pumps like the LT1044 or the 7660 where you'd have to jump through more hoops to get rid of it.

Hope that info helps in the least bit. But do what you feel is best for you!

[Fender3D]

Unless I'm misunderstanding you, I think that's basically what's in the original design.

It's not what's in the original schematic. My suggested schematic shows single voltage transformer instead of dual as in stock schematic

Or are you saying to put a 3 prong socket between the transformer and the diode network and move the transformer off the unit?

If you use the circuit I linked, you can use a normal 2 prong socket.
It will carry AC instead of DC.
Transformer out of the unit, of course...