Build Your Own Clone Message Board

has anyone else tried this? I just finished building a phase royal and noticed the same thing. Also I read elsewhere that changing one of the resistors helped increase the depth control. I think it was changing the 2M to a lower value. Can I do both mods or will it cause a problem?

In a previous post on this subject I couldn't be bothered to calculate the roll-off frequencies for those caps. Now I have.

C1, C11 and C12 are all inter-stage coupling caps and are all 47nF.

C1 is coupling in to a 470K so the low frequency -3dB point should be around 7.2Hz.

For C11 its 22.5Hz.

C12 sees R29, the 150K pull down in parallel with the input impedance of whatever you plug the phaser pedal in to. Lets say that is a pedal with an input impedance as low as 200K so C12 is driving roughly 86K- the -3dB point is still going to be about 39.7 Hz.

Now in theory you would think that all of those roll off frequencies are well below the lowest note a guitar puts out in standard tuning (bottom E = 82.407Hz) so increasing them should make no audible difference. In any case changing C12 should have the biggest effect (or you could just increase R29 to say 1M and make sure you run the phaser into something with a decent input impedance).

The modulation signal from R31 2.2M is being divided down by R30 and the bias trimmer. If the FETs are not being driven to the limits of their resistance change then reducing R31 to 1.5 or even 1M may increase maximum modulation depth and it can't do any harm.

Thanks for your reply and your help with this Tark. I didn't mess with the caps. I am just accepting the coloration for what it is.

I changed R31 to a 1M then to a 1.5M and settled with that. There are some great sounds in there but... Is their a way to keep it from clipping? If I'm understanding what you've told me, this is the sound of the FETs being overloaded? If that is the case, are there different FETs that could be used or is it not that simple?

Can you expand on what you mean by clipping? If you mean distorting then I don't think that would have anything to do with reducing that 2.2Meg.

If you mean that you can turn the modulation depth up past the point where there is a smooth change change in phasing and the LFO modulation starts to sound choppy then you have made that resistor too small and are driving the FETs past the point where their resistance continues to change.

Clipping distortion might be a problem if there were substantial gain anywhere in the circuit, but there is only one stage in the circuit that has a little gain and that is the output transistor stage.

Going back for a moment to the question of capacitors and roll off frequencies I notice the 10K Resonance control is also de-coupled by 47nF caps - C7 and C8. The roll-off frequency of a 47nf into 10K is approximately 338Hz ! Now with the resonance control turned down of course you wouldn't notice but I would think that would make quite a difference with it turned up.

The output transistor stage is almost always the culprit with the "inherent distortion" in the phase 90 circuit, and it seems to be more of a sensitivity thing (that trnsistor clips with hotter signals). In fact, in some percentage of builds, there is this weird output volume increase that changing the 150k feedback resistor to 120k cures (this goes back 10 or so years with cloning the phase 90 circuit, it's not unique to the BYOC). I've tried adjusting the feedback & bias resistors with no real luck. And I don't understand why they used a PNP stage there. When kit sellers do a phase 90 circuit without the PNP stage, the purists seem to go into a tiff, so almost all kits have this stage. I really prefer the circuits without it.

Any ideas on cleaning up that output stage Tark?

Thanks Morgan, that makes sense.

I guess what is happening there is that the simple base bias is not providing enough current to bias the transistor on far enough and the signal is clipping asymmetrically. A quick measurement of the collector voltage would confirm that was the problem. It should be around half the supply, so 4 to 4.5 or so. Variation in transistor gain would explain why it works in some pedals and not in others. And, as you say, decreasing the base resistor value a little would probably fix it in most cases.

Now I look at it again I realise that type of bias / feedback reduces the input impedance of that stage so quite possibly a bigger value for C11 might help the bass response.

And I agree - there doesn't seem to be any good reason to use a PNP transistor, an NPN version of that output stage should work just the same.

I'll give the 120k a try when I get a chance. I socketed it as you recomended.

So for now, I put the 2M back and switched the feedback resistor to 120k. I adjusted the trimmer to where it doesn't distort.

It sounds good overall it's just a bit brighter than the bypassed signal and doesn't have much low end.

When I adjust the trimmer to where the tone of the guitar and the phase shifting is deep, there is distortion when I strum the strings. Single notes sound good but chords distort.

Changing the resistor to a prescribed value of 120K won't necessarily bias that transistor output stage to its optimum point. The bias method used is dependant on transistor gain, which is probably why it works OK on some kit builds but not on others. The gain of the particular transistor will vary from kit to kit. The manufacturers data sheet for that transistor specifies that the gain may vary over a range of 3 to 1. If you can, measure the collector voltage.

To improve the low end I'd look at increasing C11 and maybe C7 and C8.

First of all thanks to both of you for your help, I do appreciate it.

Tark, can you recommend values for c11, c7 and c8. All I have on hand are 2 .22uf tantalum caps so I will probably have to order them.

I have now looked at the original MXR phaser circuits and run a SPICE simulation of the BYOC Phase Royal.

The original MXR circuit uses a different ref voltage of 3V rather than 5.1V in the BYOC so the bias point for all the op-amps and the emitter bias for the transistor is a bit different. The original circuit is DC coupled from the op-amps to the transistor and the use of a PNP in the old circuit makes a bit more sense. The 47nf coupling cap C11 is a BYOC change and it is chosen just because it is the same value as many of the other caps in the circuit. In the BYOC circuit a 47nF in that position results in a -3dB roll off at 70Hz for the simulation so that may explain the lack of bass. A 100nF (33Hz) or even a 220nF might be a better choice in that position, which would push the 3dB roll-off point down to about 15Hz.

The original circuits use 150K resistors for the mix values for unity gain and there is no mix balance control. BYOC have a 22K network with a 10K mix balance control. The mix balance control drops the signal level by around 1/5th and then feeding the final transistor with a 22Ks (R24, 26) instead of 150Ks adds gain of around 5 to bring the signal level up again. Changing that 150K feedback around the transistor to a 120K doesn't so much affect the bias point as it reduces the gain of that stage (and therefore through the pedal) a little bit.

Just saw your fresh post - I'd try 1uf electrolytics (+ ends to the pot) for the resonance (or you could go with 0.47uF plastic film) and a 0.22uF for C11.

I replaced c11 with a .1uF and now ther isnt a noticeable difference between the bypassed tone and the effect tone with the mix all the way clockwise. Sounds great thanks for your help.

Excellent! how's the distortion? I think I have a mod that would give you more clean headroom if you need it.

If you have an idea, I would definitely like to try it! The tone is great, very little coloration at this point. If you have a way to possibly get more headroom that would be much appreciated.

OK here goes -

The fix is this - A 68K resistor fitted between the base of the transistor and the 5.1V supply (the emitter) will re-bias the collector from +4.45 volts to around 3.1V allowing at least a 2V peak output swing. In fact a resistor of any value from 1 Meg down to 68K will move the collector bias point down and will improve the clean headroom. Ideally you will also need the original bias resistor (R27) value of 150K.

The explanation -

The PNP output stage supply voltage comes from the zener diode rather than being 9 volts directly from the pedals main supply. In the original MXR the zener was only 3 volts, in the BYOC it is 5.1 volts. In other words the BYOC Phase Royal output transistor is running on a 5.1 volt supply, so the output signals peak positive swing (with DC offset) cannot exceed + 5.1 volts.

The way that transistor is biased, with a 150K from collector to base, the collector sits at about 4.45 volts, roughly 0.6 volts below the zener voltage. That means the positve swing of the output signal cannot be greater than 0.6 volts peak before clipping occurs. So clipping distortion will occur with an input to the pedal of between 0.3 to 0.4 volts peak.

Re-biasing the transistor so that the collector voltage sits closer to half the supply will allow a greater signal swing before distortion occurs. it is possible to take this mod even further because, given that the BYOC circuit is AC coupled, it does not need to be supplied from the 5.1 volt zener diode. The emitter of the transistor Q7 could be connected to +9 volts and the extra bias resistor value adjusted to bias the collector to around 4.5 volts for an even bigger output swing before clipping.

Wow, amazing difference! I went with the 68k and it made a drastic improvement. I might try a higher value in a few days just to see if it can get even better. I put the 47n cap back in c11 because i had lost some of the highs. And switched R27 back to a 150k. So at the moment it's bone stock except for the 68k... And a few sockets

No don't try a larger resistor instead of the 68K, larger values will just move you back towards earlier clipping (when I said any value between 1M and 68K I meant that putting almost any value in there between those two would be an improvement, just in case you did not have a great choice of resistors. 68K should be the optimum value). Switching back to a 47nF will not boost your highs, it will just roll off the low end again.

Are you saying that adding the 68K lost you some highs ? Or that you only now feel you want more highs ? You could probably boost the clean or phase shifted highs a bit with a small cap, maybe a 2000pF, across R26 or R24.

Yes, I lost some highs when I added the 68k. I forget now which order I did it but between returning C11 and R27 back to stock, I got the highs back. Should that not have happened? Could it be because I have the trimmer set differently now?

If I were to change the voltage from 5.1 to 9, how would I go about it?

No I wouldn't have expected you to have lost highs simply through adding that 68K. I expect it had more to do with your trimmer adjustments.
With the stock 47nF for C11 you will be missing the low end.

If you have got rid of the distortion through adding the 68K I really doubt there would be any audible improvement if you disconnected the transistor emitter from the Zener 5.1 and moved it up to the 9V supply. If you did that you would need to change the 68K to something like a 100K.

I don't know yet if I'll try the 9v or not, that would allow me to get an even deeper phase effect since I could adjust the trimmer even more. Either way, the pedal sounds so much better already.

Thanks again for your time and helping me with this pedal.

"that would allow me to get an even deeper phase effect since I could adjust the trimmer even more"

No it wouldn't, just in case there has been some misunderstanding, the further mod I was suggesting only involved changing the supply rail for the single transistor output stage to improve output signal swing overhead. This would have no effect on the depth of the phase or its adjustment range. The Vref supply connection for the trimmer and FETs must remain the same. And in any case the phase adjustment trimmer actually adjusts the bias to the JFETS, which alters the centre point offset of the phase shift swing, not its depth.

Note: Split topic into a separate thread since it wandered wildly off the original thread (Moderator housekeeping)

Wow! I can't believe this was missed! The other clones that have the same clipping issue also have the same zener value. Wouldn't replacing that zener (D2) with a 3 volt fix it as well; instead of trying to find a way to hack in the resistor across the base & emitter?

Can I simply replace D2 with a 3 volt or do I need to change anything else? If so, Ill try this next time I order parts?

I had tried the resistor but didn't like the way it changed the sound... right now, I'm just using it with distortion

I would not recommend using a 3V zener. This would give you less clean headroom, not more.

The original MXR circuit seems to have been designed to use 3V because it gave a suitable bias for the FETs they were using at the time. The 5.1V zener in the BYOC allows a wider adjustment range for the FETs and greater headroom for that transistor stage. Re-biasing that stage with a resistor should have no effect on the sound other than reducing the possibility of clipping.