Saturday 11 August 2012

MXR Phase 45 modded

Now Miro has verified the original, here's a modded version of the Phase 45 and offering a few useful addition such as external Bias and a Mix control, Phase 45 / Univibe 10:1 Ratio switch and a Rate LED.  The switch is 2 position 4PDT but you could also split it and use 2 x 2 position DPDT which will then allow you to select the caps individually offering a little bit more versatility.

I actually managed to get this 2 columns smaller because of the removal of some components and others going on the switch instead of the board, so it will easily fit in a 1590B.



52 comments:

  1. Any particular reason for the change of jfets from 2n5952's to 2n5854's? I presume 2n5952's will be ok?

    ReplyDelete
    Replies
    1. Most JFETs should work here. Don't know why the modded schem suggests 2N5854s though.

      I think this is the file that someone pointed to on the original 45 post:
      http://www.lynx.bc.ca/~jc/phase45modded.gif

      I'd be interested to hear how the bias control works, if someone builds this.. :)
      +m

      Delete
  2. External bias control? Have to admit that i didn't read the docs on these mods, but i can't understand why anyone would want to take that control out of the board. The original has 250K trimmer, and here the sweep is limited to 100K, so it would be more useful..

    As it's really about biasing the FETs to nearly open and close with suitable voltage, the external control would still mean that there is just one spot where the phase could work? Or am i missing something?
    +m

    ReplyDelete
    Replies
    1. I thought the same but I suppose it depends how any of the other mods affect the bias like setting the mix control which replaces 2 x 10K resistors at the output. I don't know if varying that or any of the other mods could mean the bias needs tweaking at some different settings. The Bias pot is 100K but with 2 x 47K's either side of it, so it probably keeps the rotation in the more useful central area giving you greater resolution for fine tuning. I still suspect it could be a "set and forget" though! :o)

      Delete
    2. We'll just have to wait for someone to finish it :)

      If it still acts the same way that the trimmer in the original does, it's really bad control to have on the box. The "wrongly" biased JFETs do not sound good, as they may take the phase effect away altogether.

      As you probably noticed, i'm somewhat against having unuseful controls on my boxes :)
      +m

      Delete
    3. That said, i don't feel comfortable "wasting" my matched JFETs on a build that may have useless/harmful control on the face of the box :)
      +m

      Delete
    4. If I'd paid for matched JFETs I'd socket everything and just put them in whichever phaser I was using on that particular day! :o)

      Delete
    5. This was what the guy who did the mod schematic had to say about the Bias control: "the external bias control came later after hearing the light/dark effect it had on the phasor, similar to the 'offset control' mod on the bulb driver in my Univibe".

      So there you go! :o)

      Delete
    6. im about to build a phaser for a guy, which one should i go for, this or the causality 4 mkii?

      Delete
    7. Both in one box! :o)

      It depends mate, even with these tweaks the Causality is a bit more extreme but some people really like the Phase 45 sound and the extras will make this more versatile.

      Why don't you link him to vids of both on youtube and let him decide.

      Delete
  3. If you dont have a zener on hand a 10k resistor will work fine in that position.

    ReplyDelete
  4. Is this it ?

    http://www.youtube.com/watch?v=BpzAJP29Iq8

    ReplyDelete
    Replies
    1. Yes it must be, it has exactly the same pots and a single switch, and they all seem to be acting as expected.

      Delete
    2. So the bias pot could be useful after all. May be that the value misguided me to a bad first impression..
      +m

      Delete
  5. Works. But isn't as good as the original, unmodded 45. On the demo above, the dude must have used smaller value pot for bias. As it works just like i feared. There is only 1-3% of usable surface on that taper. I'll have to play more with this on next friday (when i have the time) before i can say more. But with this testing, it's not as good as the original.
    +m

    ReplyDelete
    Replies
    1. Oh. And i used 2N5259s. Might try on 5458s on friday too..
      +m

      Delete
    2. Hmm strange, I'll have to hunt around the forums about this mod and see if others have used different values to those listed. If you get the sweet spot of the bias isn't is as good? Because when the bias is set right most of the rest is as per the stock with a couple of pots replacing resistors, and in one switch position it is stock too. Anyway thanks for verifying mate, this really is the last tag because I'm going to bed! :o)

      Delete
    3. http://www.freestompboxes.org/viewtopic.php?f=10&t=14449&p=166123

      This guy says that it's better with 100K pot in place of 500K and 10K in place of 100K.. Might try those once i get to it. Fender3D also suggests that lowering 22K resistor will be good workaround if it's under unity. Might even try to replace that with something like 20K pot and one, possibly 5K1 resistor. This will need some more tinkering before i can call it a success.
      +m

      Delete
    4. First of all. Mine isn't below unity. It's rather just a bit over, but we're talking about 1-2dB here, so gain is not a problem. (cuthbertg's problem may be with his choice of pulldown resistor)

      100K for speed works good, much better than 500K. But.

      10K pot doesn't bias the JFETs at all, so i sticked 100K trimmer to a 3x4 board and it biases the JFETs at ~43K. That could be made to sweep the sweet spot with 10K by swapping or adding resistors, but as cap and mix mods don't make that great of a difference, i don't see any reason to do these mods. Board works, but i'm in favour of the original unmodded board. Tonal depth differences come with cap switches, but..

      This was a nice experiment, but it won't get boxed. In my opinion, the bottom line would be something like this: Good intentions to get more out of 45, but the mods give harmfull external controls and couple controls that are too subtle to be worth while. These mods don't add that much to it that it would be good to have both. So i'm keeping (and boxing) the original clone and scrapping this one.

      If someone is looking to build a phaser, do either this or the original. If you do both, you're bound to be disappointed in either one :)
      +m

      Delete
  6. Hey Mark, could you do a vero for a Maestro PS-1A ?? those rule but, ya know, they're huge... i was thinking just one stomp switch and a 3way toggle for speed selection. no knobs needed. there must be a way to keep it "ON" and bypass the on/off switch. the original colors the signal A LOT but making it a true bypass like that would solve the issue. just an idea...

    ReplyDelete
    Replies
    1. Oh yes, a vero for the Maestro PS-1A would be the best Christmas present!

      Delete
  7. so I built this up without reading the comments like an idiot (i read the comments for the unmodded phase 45)... So I'm getting good solid dry signal through with the mix knob down, but as I turn up the mix knob I introduce more and more static'y hiss. until it's all hiss when the mix knob is all the way up.

    Could this be a bias issue? Could it be a matched jfet issue? I'm also selecting between 2n5457's and 2n5952's as i dont have any 2n5458's.

    Any Help would be greatly appreciated!! I love this blog and have had only success until this build!

    ReplyDelete
  8. by the way- i'm not using matched jfets, as I didnt have a 9v battery for my multimeter this morning (i've built the jfet matcher vero from mark's layout) but yeh I was under the impression that matching jfets just made it sound better.

    could it be the reason for me not getting any phasing? Also, I'm not getting any phasing, just hiss that is mixable with my dry signal

    ReplyDelete
    Replies
    1. To get phasing, they need to be matched.
      +m

      Delete
    2. would that explain the hiss instead of phasing?

      Delete
  9. by the way the pinout on the layout is opposite from the 2n5458 pinout i found online

    ReplyDelete
  10. so i got it working- it was either a bad ne5532 or a cold joint. the bias knob is pointless- i dont know who thought that would be a good idea to have external. Also the "vibrato" switch isnt really doin much for me.

    my first thing is- I'm getting some distortion out of this- i think i'm getting quite a bit of a volume boost and it's distorting my amp. could this be due to using 2n5952's or is it something else?

    I'd like to remove the toggle switch- i assume I'd just wire the 2 capacitors across sw 8 and sw11--- and sw 5 and sw 2 respectively... are the phase 45 values the 47n's?

    What about replacing the bias pot with a trimpot? Is that gonna be easily done on the board or should I just make a daughter board and wire it up?

    Would a 100kA master volume be the best way to limit the volume? or is the clipping coming from the circuit itself?

    Thanks so much- anything will help! mirosol and mark! you guys rock!

    ReplyDelete
    Replies
    1. I have let out some steam about the bias pot earlier.. If you don't want the switch and you do want the bias as a trimmer - you should probably build the unmodded one. Only difference after those is the mix..
      +m

      Delete
    2. not exactly a board I wanna rebuild. it cant be that difficult could it? how about doing a perf daughterboard for the trimmer- that should work fine right?

      and what cap values are the standard phase 45? would that be the 47n's? or the 10n 100n combo?

      I'll stick with the mix knob...

      100kA for a master volume sound okay? I'm getting a little bit of a boost- the distortion was me being a dumbass and having the gain channel selected on my blackstar ht1.

      I'm not too thrilled with the taper of the speed knob... at the end it goes from pretty slow to super fast... would a reverse log pot be a better idea?

      I know that's a lot of questions- but they're pretty simple. I dont need anyone to have experience with these mods to answer- just a more educated opinion than my own would be wonderful!

      Thanks again!

      Delete
    3. anyone- can someone just tell me what the stock phase 45 caps from the switch are? the 47n pair or the 10n 100n pair?

      any help will earn you double points in double bonus cool point competition!

      Delete
  11. Comparing the two layouts (unmodded and modded) I can see the BIAS trimpot in the stock one is 250K, while the BIAS pot in the modded one is 100K.
    Is there any evident reason in that?

    I'm about to build the modded version, as I'm looking for a "vibey" version of the Phase 45, and I'd try replacing the external 100K BIAS pot with an internal 250K trimmer... maybe a more efficient jfets biasing?

    ReplyDelete
  12. Hi,

    I plan to make a Phase 45 only with vibe mod. switch. Studing this option... I have some questions:

    Why the 4k1 resistor instead of 7k5? is necessary for the vibe mod? Thanks!

    ReplyDelete
  13. can't get my phaser to phase! this is the second attempt i've made at this, the first was the original torchys layout a few years back. i've used matched transistors, and NE5532s from two different sources and both results are the same. it passes unphased signal, rate LED works, and that's about it. can't believe that two separate builds came out the same way

    ReplyDelete
  14. exactly, mine doesn't work too,.....but the signal passes unphased, how do I have to set the bias?

    ReplyDelete
    Replies
    1. Did you match the JFETs?

      Delete
    2. mhmmm...no!
      without matched jfets doesn't it work?

      Delete
    3. It will work if you're lucky and picked out a matched pair, but it's unlikely you did. JFET phasers are temperamental and well known for not giving the phase effect unless you match the pair. Check out the JFET tested post, it's a simple circuit to test them.

      Delete
    4. I have found 2 jfet with I=5mA e V=2,08 and 2,03....can I consider them "matched"??? why my phaser doesn't work the same? :(
      the signal is always dry... damned phaser!!!!!!

      Delete
    5. If you're getting no phasing at all then it points to a problem elsewhere. Measure all the IC and JFET voltages and see if that suggests a problem anywhere

      Delete
    6. so jfets are both:
      D= 3,62
      S= 3,71
      G= 3,71

      IC1:
      1= 3,82
      2= 3,70
      3= 3,63
      4= 0
      5= 3,63
      6= 3,70
      7= 3,59
      8= 8,95

      IC2:
      1= 3,59
      2= 3,62
      3= 2,44
      4= 0
      5= 4,22
      6= 4,65
      7= 4,60
      8= 8,95

      Values on pin 5,6, and 7 of IC2 are not fixed (regular) but the tester reads variable values (even in a small range)

      Delete
    7. ok!!!! you changed the verse of jfet....now it works! :D sorry

      Delete
  15. Hello!!! I'm after the univibe mood, but I can't find a layout here. How close could be this mxr45 with the univibe mode on?

    ReplyDelete
  16. Hello again. Built. Works great, it sounds great. I would add some interesting notes:

    1 I recommend using 500 reverselog for speed, much better control.

    2 Mirosol said it did not sound as good as the original, and have been researching. I think the biggest difference is in the rate led, as it affects the LFO somehow. The rest can be adjusted just like the original. By doing tests I found that without the rate led sounded deeper and better. Finally I removed the rate led. I have also preferred to use the original transistors(matched quartet) 2n5952, and TL072 for ic's. 8k2 resistor in the LFO.

    3 Looking for better Univibe sound: I added two extra stages, using a mini toogle, and installed the four capacitors in the style of Univibe, but attention !!! adapting values to the resistor values ​​of this circuit, not the original LDR´s. People out there uses the original values( 200n, 15n, 4.7n, 470 pf) but doesn´t take into consideration that LDR´s swing values are more or less 200k to 20k. The equivalent swing here is 20k to 2k, more or less. I put 2uF, 150n, 47n, and 4,7n, since the resistance swing value of the phase 45 is ten times lower. So the frequency swing is the same, more or less. In this way the original´s characteristic throb appears, besides being a denser sound. I recommend it.

    I would like to know how to put a control to adjust the depth of the LFO swing, similarly to Univibe.

    ReplyDelete
    Replies
    1. Hey thanks for the,tip. i did a mod on a phase 90 and was like you said a thin throb . I will go back and try this thank you

      Delete
    2. This comment has been removed by the author.

      Delete
    3. Hi! i've also notice that led mod is changing LFO - without the led LFO is sine wave, and with added led it starts to sound like one half of sine is become square. Maybe if we add another led on second part of wave it will sound more chopier and tremolo like

      Delete
  17. Hi , Are the Caps on the 4pdt switch Ceramic or Polybox.

    ReplyDelete
  18. Where negative pole of rate LED should be connected? To the 3pdt where on/off LED is or?

    ReplyDelete
    Replies
    1. Yes, then the rate LED will only pulse when the effect is switched on. If you'd prefer it to pulse permanently so you can see the tempo when the effect is off then just connect it to ground.

      Delete
  19. Just below the Gate electrodes of the JFETs there is a full row that can be removed.

    ReplyDelete
  20. Would it be possible to make these same modifications in phase 90 with option 45?

    ReplyDelete