MXR Phase 45 DIY volume drop problem.

[Rethfing]

Hi everyone. This is my first post and first topic at once and its about a problem, so sorry about that. Today I played my fresh-built Phase 45 (schematics from Tonepad). Its sounds great, but I hear a big volume drop. I read on the internet that I need to change value of certain resistor to increase the volume while the effect is on. Could you tell me which one do I need to replace? It also add very high tones, but Im aware that it is fault of ceramic caps.

Thanks in advance!

Edit: And also is there possibility to increase the volume of Tube Screamer (tonepad)? I built it for a friend but he would like to use it as booster, but it doesnt increase volume at all.

[Rethfing]

I try to answer myself cause I see there are no replies. Does lowering resistor value that is first in the signal chain will cause the effect to be louder?

Edit: I used 500k trimpot but I found out that it may cause signal loss so now Im going to replace it with 250k.

Edit2: No difference..

[R O Tiree]

The gain of each individual frequency should be a minimum of not very much at all (above about 150Hz) up to a maximum of about 1.25.  This according to my SPICE simulator.  The input buffer gives a gain of about 1.4, not 1.5 as you might expect, because of the 10nF input cap and the 10k input resistor which attenuate the input signal just a tad.  By the time the signal has got through the 2 phase-shift stages, the gain has only dropped the tiniest amount.  However, this thing is designed to shift the phase, so the "wet" and "dry" signals are never really going to be exactly in phase... so, even at their closest phase angles, there's still a little bit of cancellation, hence my figure of "about" 1.25 max for any one frequency you choose.

The LFO sweeps up and down, applying a variable voltage to the gates of the 2 FETs, altering their resistance, and hence the behaviour of the 2 phasing stages, so the phase angle of the "wet" signal shifts back and forth, and it's frequency-dependent.  At the same gate voltage, different frequencies are shifted by different amounts.  Sometimes they will almost be in phase with the input signal, sometimes completely out of phase, so the frequency "notch", where input and output cancel each other out, sweeps back and forth, giving rise to the swirly sounds.

So, given that the notch where a particular frequency is cancelled out, briefly, when it's mixed back with the "dry" signal out of the input buffer, will drop the volume quite a bit, the fact that all the other frequencies are boosted just a touch means that, overall, the average volume should just about be the same as the input volume.

Altering the 20k resistor going into pin 6 of IC1 on the Tonepad layout you're working from should alter the gain nicely.  Drop it to 15k as a first go...

The thing is, you shouldn't need to do this, I'm thinking... If you're getting a huge volume drop, then there's something else going on.  Confirm that you have a matched pair of 2N5952 FETs?  Check and re-check the value of the output resistor - 150k is brown/green/yellow/gold (for a 4-band resistor) or brown/green/black/orange/brown (for a 5-band resistor).  Also check those two 10k resistors that connect to the 47nF output cap->150k->GND.  Check that you didn't use 100k by mistake.  Also check the values of the input and output caps - the input cap is 10nF and the output one is 47nF.  If either or both of those is/are low (1nF and 4.7nF, perhaps?) then the output will be very low.

Hope this helps.

[Rethfing]

Thanks a lot for your reply! I was waiting so long for someone to reply here that I managed to built another Phase 45 (I had PCB already soldered just needed to put it all in the enclosure). What surprised me was that newly built Phaser sounds just fine. I checked if I didnt messed parts in the previous Phaser, but it was all OK. Then my dad told me that I should check transistors because they are responsible for amplifying the signal. I never thought of that because after I matched two 2N5952 I thought that they were all fine. How surprised I was when replacing them with another pair fixed the problem! I do think that the broken one sounded a just bit better, I remember that the voltage of matched 2N5952 was higher, this may be the cause. But anyway im very happy with results.

Thanks again for your reply R O Tiree, your post is full of useful info I might use in future

Edit: While the Phase is in effects loop it lowers the volume when engaged, but when its between guitar and amp everything is fine. Why is that?

[R O Tiree]

There's an impedance mis-match, somewhere... either the output impedance of your FX Send is too high, or the input impedance at the FX Return is too low for this pedal.  It's a real cheap and cheerful mixer on this thing - simply 2 resistors feeding that 47nF cap, then a 150k to GND.  Not an output buffer in sight.  47nF is quite a low value for an output cap, so its impedance will be high.  The amp's input impedance will be about 1M or better, so it will happily play with that 47nF, but who knows what the FX loop Return is like?  In your shoes, I'd get hold of a 1µF metal-film box cap, if you have one, or an Aluminium Electrolytic (negative terminal connected to the top of the 150k, positive to the two 10k resistors), which would be about as good, and slap it in there in place of the 47nF output cap and see what happens.  Also, the 10nF is pretty small... I'd slap a 1µF metal-film in there as well.

You're kind of stuck with those 10k resistors just before the output cap, but lowering them only gets you a few more mV - not enough to matter...  My SPICE simulator says this would be fine, but it's only a simulator.  And you certainly don't want to raise their values...

I'd also strap a 1M resistor between the input cap and GND - place the "live" end between circuit input and the cap.  This will ensure that the FX Send is presented with a large value.  I can't find the reference, but I think I read somewhere that the non-inverting input can have a lower impedance to the inverting one.  Whether it's going to be significant in this case...  It can't hurt, though.

If all else fails, I'd install a little switch to put a 20k in parallel with the one going to pin 5 of IC1 (on the Tonepad layout you're using).  That will drop the total value to 10k and will increase the input buffer gain to 2.  If that's not enough, try a 10k in parallel, which will increase the gain to 2.5.  If it's too much increase the parallel resistor to 24k/27/30/33k... you'll find a good value.  The switch will allow you to disengage it when you're playing into the amp and re-engage it when playing it through the FX loop.

This problem has me interested, so I'm building my own - thanks

[Rethfing]

Thats very interesting what you said. I think Im gonna investigate this problem on my own, cause I own a chorus pedal made by Polish manufacture EXAR and Im getting even signal wherever I put the pedal, so they might know what is the issue. Or maybe it depends on the pedal, and P45 DIY just cause this problem.

This switch you are suggesting to add is also good idea for a mod - for instance when you want to play some solo's on clean channel - you hit phaser and you get phasing sound + some boost. Thanks again for reply!

[R O Tiree]

Just had a thought... your amp doesn't have a parallel FX loop, does it?  That can really mess with effects that mix "wet" and "dry" signals.  This pedal does not invert the output, but it's going to be at best a little bit out of phase with the FX Send.  That could account for the volume loss in this case.

Marshall amps tend to have parallel FX loops, and I've no doubt many other brands as well.  Some have a switch to get rid of the parallel path.  If yours does have a parallel loop and it doesn't have a switch to defeat it, then you're either going to have to mod your amp or mod the pedal with (yet) another switch to lift one of the ends of the 10k running from pin 7 of IC1 to the output cap.  Again, when the pedal is in front of the amp, that switch will need to be made (so the mixing goes on in the pedal) and, when it's in the FX loop, the switch will need to be "off" and then the mixing can go on inside the amp... the only signal coming out of the pedal in this case is the phase-shifted one.

[Rethfing]

My amp is solid-state Peavey Bandit 112. The manual says this about FX Loop:

Effects Send:
Load Impedance: 1 k Ω or greater
Nominal Output Level:
-5.4
dBV, 535 mV RMS

Effects Return:
Load Impedance: High Z, 22 k Ω
Designed Input Level:
-5.4
dBV, 535 mV RMS

I cant find whether it is parralel or not.

[R O Tiree]

OK, found the service manual with a schematic... It's NOT a parallel loop... which is the good news...  The bad news is that I can't find any reason why there would be a volume drop.

Here's the link to the service manual: http://dl.dropbox.com/u/91341319/peavey_bandit_112_service_manual.pdf

Perhaps someone else might suggest why this isn't working correctly?  The FX Send/Return circuitry is in the top-left corner of Page 5.

[Rethfing]

Does anyone have any idea what to do with it?

[R O Tiree]

Paging PRR... He knows his onions, shallots, scallions... anything even vaguely onion-y... and amp-y

[PRR]

> input impedance for the FX Return looks as though it would be ~470k

FX Return inpedance is in fact 20.56K (similar to the claimed 22K).

There's 1Meg and 470K bleeders. Then it goes to U1A, an inverter, with 22K input resistor. The "-" terminal of the opamp follows the "+" input, which is grounded, so the far end of the 22K is "grounded", near-enough for practical purpose. (The virtual earth point may be 10-100 ohms at midband, zero-enuff.)

That's quite low for guitar-cord effects. I'm sure they expected you to use rack-effects here; g-cord effects go in the guitar cord *before* the amp.

You "could" hit 100K by raising all impedances around U1A 5X. But that makes C6 a mere 20pFd. That seems tiny and susceptable to sneakage from other signals inside the amp.

I'd buffer the P45's output.

[PRR]

The P-45 opamps have some grunt to spare. Change some 10K to 1K-2K. This gives a good near-600-ohm output. Change-up the output cap to be sure of driving heavy loads well.

[R O Tiree]

Thanks, Paul...

@ Rethfing - looks like my gut reaction of reducing the output resistors and beefing up the output cap has been proven to be in the ball-park.  Go with Paul's values... I will be doing the same for my build.

You often see threads in forums all over the place complaining about the volume drop in this pedal, going back years and years... such a simple fix, and yet never implemented by MXR?

[Rethfing]

I never thought that I get this problem solved But would that solution increase the volume while using it between amp and guitar or everything will be OK no matter how I connect it?

[R O Tiree]

It should be OK wherever you put it in the signal path.

[deadastronaut]

You often see threads in forums all over the place complaining about the volume drop in this pedal, going back years and years... such a simple fix, and yet never implemented by MXR?

very true, i had the same issue myself...i'm amazed how a company like MXR could let this slip by them...its a pretty obvious drop...

mine was ok-ish on clean,(still a noticable drop though)  but as soon as i put a distortion into it, it sucked/dropped big time.......

[Rethfing]

You guys are geniuses, gonna try that thing tomorrow! Thanks for finally solving this problem, Im gonna post build report on tonepad with link to this forum and solution

[R O Tiree]

Well, we have to find out if it works, first...

[R O Tiree]

Wow!... what a lovely sound... no volume drop at all, anywhere in the chain.  Rethfing, let me know if yours is OK in the FX Loop?

Here's my board:



And the top-side:



Time to put it in a box...