Maestro PS-1A layout...for your pleasure

[jdub]

I put mine on the trace side of the board, straddling the pads for the 100k resistors that run from pins 1 & 2 and 6 & 7 of the opamps. 

[tomas1808]

I resumed the work on this pedal and housed it in a 1590DD enclosure using a 12V AC adaptor to power it up. However I am now getting noise when the effect is on.

Since this needed a buffer on the front anyway I made this "buffered-bypass" instead of true bypass. When bypassed there is no noise, perfectly clean sound.

The noise is what seems to be 100Hz hum (double the mains freq?). And its not a clean, sine 100Hz either, its kind of "sharp" noise with lots of harmonics.

I believe I didn't have this problem before the switch to enclosure/new PSU so I suspect is has something to do with the new PSU maybe?

I used this design:



Any ideas??

jdub, you said you also used a AC adaptor. How is it working for you?

Thanks!!

[duck_arse]

what is the DC voltage you have at the INPUT of each regulator? it may be too low for headroom.

[jonnyeye]

The input buffer in the original is not a great design (especially for guitar purposes) - we're talking this part right here:

From the design (DC coupled right to the output!) it's clear that the emitter is supposed to sit at 0V, but that depends on the hfe of the transistor used (the 2N3638A datasheet I found gives a "typical" hfe of 140, which gets to about 100mV off), and likely the reason the bias resistor to V- is given as 470k or 680k is to trim for least error - whatever the error is will appear at the output, potentially messing with whatever comes next in the signal chain and (if true bypassing is being used) causing switching pops. The input impedance is at most 33k, which will load a pickup significantly (loss of both volume and treble).  And there is no RF rejection cap, so radio can get in.

What were the problems with this pedal again? 

The modern fix is to rip out the transistor buffer and replace it with an opamp buffer (Rpop is something 1M and up):

You may get away with leaving the output stock (with the 4.7k output resistor and no cap) with this one - the offset voltage should be small.

Or if you don't want to hack up your board too much and can get a PNP Darlington like a BC516, you can try something like this (adding the 68p cap from the base of the transistor to ground two traces over):

The input impedance is now about 400k. This one I would still change the 4.7k resistor on the output to a cap+resistor (220n+1M works) to ground.

[tomas1808]

what is the DC voltage you have at the INPUT of each regulator? it may be too low for headroom.

Actually I am getting 17.5VDC and -17.5VDC respectively on the input of each regulator. Is this ok?

The outputs are 11.60V and -11.60V.

Great info jonnyeye! What do you think about putting a opamp buffer on the front without removing the transistor buffer? This is what I did and seems to be working correctly.

BTW any idea why my build is noisy when the effect is on?

EDIT: Started tracing the signal from the input and the noise starts creeping right at the inputs of the phasing ICs. The noise get louder on each successive opamp output. Any ideas?

EDIT2: When I dump the FETs bias voltage to 0V the noise goes away. Not too sure what to make out of this.


Thanks!

[jonnyeye]

What do you think about putting a opamp buffer on the front without removing the transistor buffer? This is what I did and seems to be working correctly.

It works fine, but buffering a buffer offends my aesthetic sensibilities 

BTW any idea why my build is noisy when the effect is on?

EDIT: Started tracing the signal from the input and the noise starts creeping right at the inputs of the phasing ICs. The noise get louder on each successive opamp output. Any ideas?

My guess (given what other people have said in this thread) is RF interference - start by putting on the solder side of the board a 47-100pf cap from the base (centre leg) of the input transistor to ground (from the base, going towards the centre of the board, skip one trace over).  If that doesn't work add the other suggested 50pf caps (47pf works fine).

EDIT2: When I dump the FETs bias voltage to 0V the noise goes away. Not too sure what to make out of this.

When you dump the FET bias to 0V, you are turning them fully on - lowest resistance.  Basically you're shorting out the hiss, but you're shorting out the signal too.

[tomas1808]

Thanks

The 100pf cap didnt fix it, sadly.

The transistor buffer was clean when I probed it. Noise started right after

Started tinkering with the grounding and was able to remove the humming, however now I am getting quite severe clicking..

As I try different grounding schemes it seems to be a tradeoff between clicking noise and hum noise.. if I get rid of one I get the other..

I had already followed jdubs directions and placed the 100r resistor and the 22uf cap before all this. The only thing left to change would be the IC to a lower current equivalent.

When I close the lid whatever noise there is gets louder.

This is crazy  

[jdub]

Tomas, I'm using an 18vac adapter into the GGG bipolar power board for mine...initially I was running it at +/- 12v and did get some ticking, but then I switched out the regulators to run it at +/- 15v and oddly enough the ticking went away.  I recently switched them back to 12v, and sure enough the ticking came back.  TBH, I have no idea why    Haven't had any hum problems, though.

I'm not quite clear how you are getting 17 volts at the inputs of your regulators using a 12vac adapter...

I have been playing around with various 9v to +/- 12 and 15v power supplies (using two LT1054s, an LT3467 like here: http://obsoletetechnology.wordpress.com/projects/studio-electronics/dc-dc-bipolar-power-supply-for-effect-pedals/) but haven't been able to get them to work quite right (usually noise problems).  Juansolo did some great work getting his happening with a 9v supply, but I haven't tried it- the dc-dc converter he used isn't available in the US as far as I can tell.  However, I have requested a sample of a similar unit from Murata, but don't know if they'll actually send me one- it's a $17 part.  He also used the Klon buffer at the front end.
Surprisingly, considering the low input impedance, I haven't personally had much problem with high-end loss nor low output volume; it sounds pretty good to me.  This one does seem to be pretty touchy noise-wise though. 

Jonnyeye, thx for the buffer tips.  If I get some time, I may try to adapt the layout for the addition of your buffer solution. 

[tomas1808]

Thanks, I will try going for a higher voltage.

The AC adapter actually outputs 14VAC between the two poles according to my cheapo DMM. This may account for the (actually) 18.5VDC at the input of the regulator.

BTW I am talking RMS here, so 14VRMS would be 19.8V peak.

Thanks

[tomas1808]

It seems the input voltage isnt enough for the 15V regulators as noise seems to have increased. I'll go back to 12V.

I just swapped to a TL062 and the noise is pretty much the same. Also when I put another cap in parallel to the 22uF to increase capacitance, the ticking doesnt change AT ALL. Even when putting a 1000uF!

The only thing that reduces ticking is placing a small cap between pins 5&6 but thats messes too much with the LFO.

jdub, how did you ground your maestro? I believe I have a grounding or PSU problem becuase I didnt have ticking before putting it inside the case & using a new PSU.

[jdub]

I just grounded to the input jack, as usual...that is, once the bipolar +v, -v and 0v were attached to the board, I connected the board ground to
the jack.

[jonnyeye]

A possible fix to reduce ticking that requires modifying the board:

Cut the trace from pin 4 of the LFO op-amp (the vertically oriented one at the top of the board, for reference).  Run a wire from that pin to as near the -12V input as possible. My guess is that the LFO op-amp sharing a current path with the signal op-amps is what is causing the ticking.

Note that I haven't build this, but I'm planning on it as soon as I get some of my other backlog projects completed - juansolo's sample is the best phaser I've heard so far. I think it's the shape of the LFO that really does it for me.  A friend of mine has a Leslie 130 which has a single bass rotor and a solid state amp that sounds great with a guitar through it (especially at low speed... subtly thick), and this gets real close to that tone while weighing about 85lbs less.  Once I get mine built I may be able to help more on identifying noise sources.

[tomas1808]

I have been testing for a while and the problem is indeed the wallwart PSU.

I managed to get NO ticking with the computer PSU.  

There was some ticking with the computer PSU but I managed to remove them by "filling" the ground traces with solder. I kinda created a little wall of solder between the audio and LFO traces. This reduced the ticking a great deal to my surprise. Then I placed a small sheet of aluminum on the other side of the board also between the LFO and audio components. The aluminum goes to ground. This killed the remaining ticking.

Everything is dandy when using the PC PSU, but when going back to the wallwart PSU the ticking is literally as loud as the audio signal.

What the problem with the wallwart psu??

Also,  the unit is quite hissy when the effect is on. Its white noise and seems to be constant as in not affected by changing the grounding at all. The white noise drops when bypassed. Is it supposed to be like this?

Sorry for flooding this thread with posts!

PS: jonnyeye, this is by far the best sounding phaser/stompbox I have ever built. Absolutely addicting. You are spot on with the shape of the LFO being what sets it apart. Its more like a throb than a predictable sine wave. You must build it

[johnk]

I just finished making, building, and testing the pcb for mine today. I use mine with a bass and mine has no ticking whatsoever using a walwart or my benchtop power supply. i'm running mine on +9V/-9V and it soun ds the same as running it with a bipolar 12V supply so i'm just using a a negative voltage converter to get the -9V. it has a very slight volume drop and definitely does cut the top end. I can adjust my sweep to be perfectly even with just about any Jfet. I first used 2N5485's and MPF102's but i'm now using 2N5457's for all of the jfets (since I have a TON of them). to get back the top end, i'm just going to use a tiny (and simple) Jfet buffer in front of the circuit since I already tried it that way and it sounds great with it. thanks for posting the PCB layout. this was the first one that I've ever made using the laser copier/gloss photo paper/etching method and I think that it came out great.
I slightly shrunk the width of the PCB by 'bending' some of the traces in photoshop so I can fit it into a 1590BB and it'll just fit in there now.
here's some pics of it as it is now:

the board (top):


the bottom:


the tested PCB:


showing how it will just fit in a 1590BB:

[hangingmonkey]

Looks great. Can I ask, how do you cut the pcb to size, looks like nice straight lines. I usually use a Stanley knife and a metal ruler but the end results aren't as nice as yours.  Especially on fr4

[duck_arse]

johnk - that's a nice looking etch for first go. do you have any smaller drill bits?

hanging - in all instances, when working w/ fibreglass pcbs, 400ish grit wet-or-dry (wet, on glass, with gloves) is your friend. it can turn the ugliest raggedy edge into a smooth straight, better than many fab houses.

[johnk]

johnk - that's a nice looking etch for first go. do you have any smaller drill bits?

hanging - in all instances, when working w/ fibreglass pcbs, 400ish grit wet-or-dry (wet, on glass, with gloves) is your friend. it can turn the ugliest raggedy edge into a smooth straight, better than many fab houses.

i wanted to get it built as quickly as i could so I just used the drill bits that I already had which are .038" (#62).

[johnk]

Looks great. Can I ask, how do you cut the pcb to size, looks like nice straight lines. I usually use a Stanley knife and a metal ruler but the end results aren't as nice as yours.  Especially on fr4

i first cut it with a dremel with #409 cutoff wheel, then i smooth the sides with a sharpening stone block.

[hangingmonkey]

Thanks for the tips guys! I'll try that on my next build.

[johnk]

She's done and it sounds awesome!
pics below: