Phase Revolution - Perf & PCB layout for R.G.'s Phase 180 Plus

[1878]

I don't want to be pushy, but has the PCB been revised yet ?? I'm new to making PCB's and I don't have the confidence to run with it myself !! I've also managed to get myself 8 matched 2N5952's which has helped to treble my excitement/impatience.

Thanks.

[frequencycentral]

I don't want to be pushy, but has the PCB been revised yet ?? I'm new to making PCB's and I don't have the confidence to run with it myself !! I've also managed to get myself 8 matched 2N5952's which has helped to treble my excitement/impatience.

Thanks.

I haven't got round to it yet, I'll try to find time over the weekend, though I'm off abroad on Monday, so it's going to be a busy weekend preparing. I'll try to make time for a fellow scouser even if you do support the wrong team. 

[1878]

Ooohh... I suppose I'll have to take that one on the chin 

[frequencycentral]

Here's the revised PCB and PNP, which i now believe to be correct! I had to make the board a bit bigger to correct the errors. If anyone makes up any boards and would like to send me one that would be dandy!

[1878]

Cheers !!

I'll sort one for you. PM me your details.

[doc_drop]

Paging Mr. Central to the white courtesy telephone...

Hey Rick. I am just about ready to "smoke test" my build of your phaser based on the layout info in this thread. I am curious about the resistors you show in one of your diagrams around the bias pot. Did you leave those in when you made the corrections to the layouts? In other words, do I need those resistors to make the circuit work?

I have high hopes for this guy, so I hope I won't have too many debugging issues. I'll know after my Saturday morning pedal building session tomorrow.

Thanks

[frequencycentral]

Hey Doc! You know the perf layout has errors right? I revised the PCB layout but didn't get around to revising the perf yet. Though I did explain the 'fix' for the errors in the perf. If you use a 250K resistor for the Bias pot you should be fine, you don't need those other resistors. Good luck!

[doc_drop]

That was fast Rick. Thanks!

I do know about the errors. I actually used the corrected PCB layout you posted to figure out the connections. I have done a continuity check, and I appear to have everything connected per the PCB layout.

But, I thought the bias was a 100KB pot. It sounds like you are saying to use a 250KB instead?

Are the other pot values per your little "pots on perf" layout?

Chris

[frequencycentral]

If you look at the schematic it's a 250K Bias pot. From my experience with the P45 build I did, I know the useful range is mostly in the middle of the range of the pot, so my idea was to use a 100K with fixed resistors. In practice it's easier just to go with the schematic, after all, it's just a voltage divider whatever value you use. If in doubt consult the schematic. I added the resistor between the bias pot and the width pot on the advice of Mark Hammer (I think), to make the range of the width pot more useable.

[doc_drop]

O.K. a 250KB bias pot it is. (I definately made good use of the schematic while building this...But I really appreciate the work you put into the layouts and your generousity in sharing them.)

I'll post my results after this weekend.

[doc_drop]

O.K. After a bit of further debugging, and finding out that as usual I didn't have all the correct connections the first time, I got myself a working phaser this morning. 

I LOVE IT! Having all those parameters to tweak makes this guy really versitile. So much more fun than a one knob MXR Phase 90. I love how resonant the sweep can be. I have been getting all kinds of filtery sounds out of it. And being a synth guy, I like me some filta!

Thanks Rick/FrequencyCentral for the layouts. Thanks R.G. for the circuit!

If anyone builds this, even on perf like I did,  I recommend using the PCB to do your wiring. It is correct and easier to read.

Time do go sound like, "thrweeeeeshsssshhhhhhhhhzzzzzz......ChuggaChugga......................"

[frequencycentral]

Hey good for you Chris! Yeah it's so much more that a phaser isn't it? Extreeeeeme! I'm curious as to what FETs and opamps you used. Did you have to do loads of FET matching? Good to know the PCB layout is verified anyway, I'll have to get around to correcting the perf soon...........

[doc_drop]

I used 2N5485's, a BC560, LM324's and a 4558. Almost exactly per your parts list. (I had some LM324's from a value pack I got from Futurlec, so it was nice to finally use those.) I had 28 2N5485's that I took Hfe readings for using my DMM. They ranged from about 798 to 805. I took 8 that were right at 800. The phase sounds very symmetrical, so I guess it worked...

I have really been enjoying experimenting with the pot settings. The names don't do them justice. I did use a 250KB bias pot, which works like a sweep center frequency control. The Regen works like a Q control for the filter effect. I left the mix pot 100KB like your pot diagram even though the schematic uses a 200k pot. It seems to be a decent range like this. But it is backwards on your diagram so I switched lugs 1 and 3 so the "more effect" direction is clockwise.

I am putting this in a box with a Magnus Modulus I just built as my first introduction to etching my own boards. That too is a deep and extreme circuit. Kind of like a Tremulus Lune mixed with your own Clari-Cubed circuit without the fuzz. This will be my ultimate mod box!

[doc_drop]

A sad story, and a couple of questions:

Rick,

Sad story:

So, there I was loving this circuit so much...But when I installed it in a box next to my Magnus Modulus, I must have screwed something up because it stoped making any sound at all. I tried debugging it, but I decided that this was a good time to try my new PCB etching skils, so I etched the board and rebuilt it by pulling the parts off my perf version. I was very careful, testing every part before I reused it. But, the circuit only works halfway. It is filtering, but there is no sweep. If I adjust the bias knob, width or regen knob, I hear the filter being affected. But, there is no pulsing sweep to it. It is also self-oscillating a lot.

Questions:

I see that the schematic calls out for the Vref to be 3v. Right now mine is at 4.36v to 4.45v as I rotate the bias pot. Is there a reason you didn't try to get the 3v Vref? Would changing the zeneer to a higher voltage lower the Vref?

But more importantly, I believe there is an error in the PCB. Look at the attached image.



Shouldn't the tops of R12 and R13 be connected to the bases of Q4 and Q5? That is what is happening with R6, R19 and Q2,Q7. I started wondering when I noticed the pad for R13 is free standing and not connected to anything but one resistor lead. I guess I missed this when I was using the PCB diagram to debug my perfed version.

Thanks!

[frequencycentral]

Shouldn't the tops of R12 and R13 be connected to the bases of Q4 and Q5?

Ooops - you are correct sir! I'll edit and repost when I get the chance. An easy fix for you though......

I see that the schematic calls out for the Vref to be 3v. Right now mine is at 4.36v to 4.45v as I rotate the bias pot. Is there a reason you didn't try to get the 3v Vref? Would changing the zeneer to a higher voltage lower the Vref?

I figured it worked at what it was, so why sweat? But I'll do some experiments on mine and report back.

[doc_drop]

You sir are the promptest responder on the whole internet! I wanted to have you look at it before I made my "lil' bridges". But since you agree, I will make the change tonight, and report back.

I don't know if I really care about the Vref thing if I can get it sounding as good as it did at first by making the little correction above. But, I am all ears if you want to experiment a bit with that.

Thanks again for such a quick response, as well as posting this layout. I really like this circuit...at least I will when I get it up and running again...

[frequencycentral]

I've corrected the mistake and reloaded the PCB and PnP, so the images above are now correct, you may have to refresh you browser for the corrections to show. I really hate being human and therefore prone to error - why can't I be perfect like Mrs FC?

[doc_drop]

Well, something tells me Mrs FC couldn't whip together a perf board like you if she wanted to. Or maybe she can, in which case, does she have any sisters?...

My last 3 circuits have all been debugging nightmares, so I know how imperfection feels. It always cracks me up how happy I am to find my own stupid mistakes when I am debugging something. And I usually hate making stupid mistakes!

[doc_drop]

That did the trick! I created my 2 intentional solder bridges, and whammo, sweet phase - totally adjustable phase!

Guys, if you want to buid a phaser, do this one. It is verified and amazing.

[Mark Hammer]

Although it didn't come up yet, it may at some point.

One of the things often experienced as one adds more and more phase-shift stages is that noise accumulates.  With the traditional 4-stagers, it isn't much of an issue, but once you get into 8 and more stages, particularly if you have intentions to use regeneration/feedback around those 8 stages, hiss can get aggressive. 

There are a few potential solutions to this.  One is to stick a suitable cap in the feedback loop of every 4th phase-shft stage.  So, if one was using a P90 configuration with a 10k feedback resistor, stick a 1nf cap in parallel with that resistor on every 4th stage.  That would get you a rolloff starting just under 16khz.  If you don't ever plan to use it with acoustic guitar, you can up the cap to 1n5 and bring that rolloff down to around 10khz.

Alternatively, if you would like to keep as much bandwidth as possible, consider using a lowpass filter configuration in the regen path itself such that the hiss-reduction only really kicks in at higher regen settings.  So, as an illustration, let's use the Ropez/Ross circuit.  You can see a 27k fixed resistor leading to the "regen" point.  Split that 27k up into a 12k and 15k resistor in series, with the 15k first in line.  Now, run a 1nf (1000pf) cap from the junction of those two series resistors to ground.  Voila, a 10khz lowpass.

In the case of the Ropez, I've found that the unit tends to break out into low-frequency oscillation when more than 4 stages are used (i.e., it starts to "moo").  You can fix that by reducing the value of the 1uf cap after the Regen pot to a smaller value, like .1 or .047uf.

R.G.,

Although I haven't spoken to him lately, Mike Irwin told me that although the drain-gate R-C network around the FET does manage to reduce distortion, in reality it functions more to stave off distortion to a higher threshold, rather than reduce distortion at all signal levels.  His sentiment was that, although he held out high hopes for the R-C network to solve a S/N problem that was fundamental to FET-based phasers, the sound of the distortion that came into the picture at clipping point was simply godawful and an unintended penalty.  In his mind, in retrospect it was worth dealing with the more modest distortion at lower signal levels to NOT to have to deal with the type of distortion introduced at higher levels.  He went back to the no-network configuration.

Now, having said that, it bears noting that Mike is principally a synth guy, so the sorts of signal levels he might view as common enough to be higher risk might be generally way out of our league as guitar strummers.  In which case, we would never, or only VERY rarely, exceed the threshold set by the addition of the R-C network.

An empirical question I suppose, resovable with a bit of signal-generator and scope work.