DIY boost problems - does not work right away

[powerslide]

Hi!

This weekend I had some time and decided to do a boost scratch build on a piece of stripboard.

Here are my references: http://diy.musikding.de/?p=357&lang=en
This is the schematic I have been working with.

Now I wired everything together. The bypassed mode works with no problems, of course. When I switch the effect on, nothing comes out unless I hit the strings real hard. Or just unplug my guitar cable, touch the tip a couple of times, put the jack back in and it works. Bear in mind this is just like jump starting a car. Once it gets a loud input signal it normalizes and stays that way. But as soon as I move the pot left or right, the sound stops again. For some reason, I can't "jump start" it when pot is set to noon (50%). Even if I hit the strings hard, all I get is a small fart.

If I leave the amp and the effect on (no guitar plugged, open cable on the input and pot set to max), after a minute or two I start hearing the usual hum and intereference I usually hear from my equipment and room. As if the pedal has charged up enough, like a battery, and starts working.

All in all it seems as if the guitar doesn't deliver enough output. Only when all the usual hisses, noises and interference are audible, it feels okay to play. I have old florescent lights in my shop with an EM ballast and it only picks it up when the pedal is "fully charged". Guitar problem here is out of the question, I even tried it with EMGs to get the most input to the board and it's the same.

I have a bunch of DIY LPB-1 clones, all on the stripboards, they all work okay. So this is not my first build.

All in all it feels to me like something is eating up the guitar input I am sending to the circuit.

I already replaced: electrolytic caps (except the input cap C1), pot and the transistor. I took care to cut the tracks carefully and thorougly cleaned any remains.

I used one of my used 3PDT switches in this build (as I always do before I paint the enclosure), could this be the culprit? Still waiting for new ones to arrive.

I might be better off just making a whole new circuit instead of debugging this one. But then again, I am not really sure the schematic is okay. I cross-checked it with the PCB that is also included in the document and the only error I have found (vs. the actual PCB) is the junction between input and D2/D3 doesn't have a dot.
I have put together a lot of Musikding's kits and they have all worked flawlessly. Maybe in this case someone was kinda lazy about the diagram? I hope it's me who is to blame

Any ideas or suggestions what to look at and try next?

Thanks!

[EBK]

Could you post a picture of your board (top and bottom)?

[GibsonGM]

Don't the mosfet heaters have to warm up? 

Just kidding!  Sounds like a build error....board pic, please....a bias resistor not connected properly could cause this, but let's have look first...

[powerslide]

Hi!

Wow, thanks for fast replies!

I hope these pictures will be of some help.

https://drive.google.com/drive/folders/0BzwEA6PcFDKlWWc0OGVmQ2NoX1U?usp=sharing

First off: some pics of the BOTTOM view (soldered connections) should be rotated 180° to get the right perspective - pot at the bottom, not at the top. The TOP views are okay and they go by the schematics - starting with the D1 and C3 on the top left. The wires from the TOP view are: 1 (red) - 9V power; 2 (white) - input; 3 - empty; 4 (red) - LED; 5 - empty; 6 (black) - ground. The white wire on the right is the output wire.

You can easily find the R4 on the TOP picture as it is the only one that is mounted left-to-right. Right next to it are two jumper wires - from left to right: first one is feeding the R3 (right at the transistor), and the other one is feeding the drain of the transistor.

Another jumpered section is on the lower side - right beside the C3 is one that brings the current from transistor source to the potentiometer. And on the right side, below the C2 there's another one that only bridges ground to an additional strip.

Signal goes from white wire to the green capacitor and then continues on the 4th row/strip. I was really confused where to put the D3 and R1.

I know this layout is cringe-worthy and very very crude. I worked completely on the fly and half way into my work I realized I should be careful how wide my PCB is going to, as I like to mount such small PCBs directly on the pot itself, so this helps eliminating hassles by mounting the board to the enclosure. As I only had the solder lugs pots, I had to improvise.

I will take some more pictures and upload them right away.

[GibsonGM]

It is not bad. I build things with far worse layout than that, ha ha.

Can you get us a voltage on the MOSFET pins?  They should be Drain, Gate, Source from top to bottom in the pic (flat side to the left).   Be careful not to hit more than 1 pin at a time with your meter probe (don't short the pins together).     

Everything that is supposed to go to ground does in fact go there, right?  D3, R1, etc...back to the batter "-" terminal...

[EBK]

I know this layout is cringe-worthy and very very crude.

Give yourself some credit: metal film resistors, no ceramic caps, socketed transistor, as small as it needs to be but not smaller. 

We've seen much worse.

[EBK]

Double check that you still have as many cuts as you planned.  There is a subtle crease in one spot of solder in the middle row.  (I haven't traced through anything yet -- just scrutinizing the solder work, which looks pretty good overall and it looks like you've scraped between rows where you needed).

[GibsonGM]

As always, check and recheck resistor values!  If the 1M resistors are not correct value, or the drain resistor, you may be misbiased, which COULD lead to symptoms you reported.
That's why I would like the voltages on the 3 mosfet pins...

[powerslide]

Here's the data:

1 - when pot is at 100% (fully CC)

• D - 4.30V
• G - 2.09V
• S - 0.00V

2 - pot at 0% (fully CCW)

• D - 7.45V
• G - 2.75V
• S - 1.69V

Tested with an older Boss PSA230 wallwart, 9.05V at the DC jack.

I am completely unsure about what I did at R1/D3/D2/R3 section. The way I understand the diagram, all of them should be on the signal path from the capacitor before the signal hits the gate of the transistor, right? If you take a look at the PCB on the original Musikding's document, the D3 and R1. Maybe I am just not getting it. Are D3 and R1 only needed for the initial signal path before it gets to the gate of the transistor or after it?

I have checked all of the resistor values. Heck I even matched the two 1Meg so that both are 998KOhms. I also tested all three diodes for blocking and passing.

[powerslide]

Double check that you still have as many cuts as you planned.  There is a subtle crease in one spot of solder in the middle row.  (I haven't traced through anything yet -- just scrutinizing the solder work, which looks pretty good overall and it looks like you've scraped between rows where you needed).

I scraped all of the cuts and everything between the strips themselves, to remove any solder or copper residue. I used an xacto knife for cuting and lifting the copper layer, then followed with a micro brass wire brush to get all the crud out. Wiped whole bottom side with 99% isopropyl. Some cuts might appear lighter, that's just scratched fiberboard.

[EBK]

Hmm.  As far as I can tell, the gate should always be at half the drain voltage (there should be 0 steady state current into the gate or through that input cap)....  The diodes should not be conducting either...

I find myself wanting to blame c1, but not until I've traced your board, which I might try to do later.

[EBK]

Regarding that before/after question, are you referring to the PCB layout?
It really doesn't matter which side of the transistor those parts are physically placed.

[PRR]

> matched the two 1Meg

Not needed. They are equal only as an arbitrary convenience. If not equal, it only fouls EBK's "half the drain voltage" assumption. 900K and 1.1Meg would be 10% off of "half".

However there's another foul, meter loading. The gate impedance here is 500K. A 10Meg meter throws a 5% change on the true voltage. I accept that for your 4.30V-2.09V diff.... well, that implies a 17Meg(?) meter which is an odd spec.

> Once it gets a loud input signal it normalizes

I would bet on a bad contact. Tarnish does this. BIG inputs zap through and it may work a whole gig.

Pot-change should not matter unless the tarnish is in the pot.

[GibsonGM]

D2 and D3 aren't connected to C1, are they? 

There isn't a real node there, at the junction of all 3...C1 'jumps' and goes to the junction of the 1M resistors, where there IS a node...just a hunch that you might have them connected where they shouldn't be.   If they are intended to be connected that way, the schematic is wrong, and I would have NO idea what the h3ll they are for...

D3 anode goes to ground...D3 cathode to D2 anode...D2 cathode to R5 and C3, 9V.   C1 isn't part of it.

[EBK]

D2 and D3 aren't connected to C1, are they? 

There isn't a real node there, at the junction of all 3...C1 'jumps' and goes to the junction of the 1M resistors, where there IS a node...just a hunch that you might have them connected where they shouldn't be.   If they are intended to be connected that way, the schematic is wrong, and I would have NO idea what the h3ll they are for...

D3 anode goes to ground...D3 cathode to D2 anode...D2 cathode to R5 and C3, 9V.   C1 isn't part of it.

There IS a node there on the PCB.  It's an error in the schematic spotted by powerslide in the first post.  I think they are there because they resulting clipping would sound better than allowing the MOSFET to clip (just a guess without thinking it through).

[EBK]

Pot-change should not matter unless the tarnish is in the pot.

Adjusting the pot immediately makes it quit working.  Maybe it is the pot.  Wouldn't hurt to try a few fixed resistor values for that.  The 0-ohm voltages looked OK.

[GibsonGM]

D2 and D3 aren't connected to C1, are they? 

There isn't a real node there, at the junction of all 3...C1 'jumps' and goes to the junction of the 1M resistors, where there IS a node...just a hunch that you might have them connected where they shouldn't be.   If they are intended to be connected that way, the schematic is wrong, and I would have NO idea what the h3ll they are for...

D3 anode goes to ground...D3 cathode to D2 anode...D2 cathode to R5 and C3, 9V.   C1 isn't part of it.

There IS a node there on the PCB.  It's an error in the schematic spotted by powerslide in the first post.  I think they are there because they resulting clipping would sound better than allowing the MOSFET to clip (just a guess without thinking it through).

Yeah, I dunno why they are there unless it's that.  The '170 clipping would sound better. It won't, anyway, unless you put a preamp in front of it...

I build circuits like this where stages cascade and I think they sound a lot better than diode clipping, but that's just me...

Good catch, I often skim over 1st posts if they're long!  Can't hurt to take the pot out of the equation.

[PRR]

The diodes are over-voltage protection for the Gate. They MUST connect to Gate. (Especially if he is gonna "touch the tip a couple of times".) There should be a DOT there. (And otherwise: what are these diodes for?)

But that's not the problem. It will work without the diodes UNTIL static shocks blow-up the Gate and it stops working forever.

[EBK]

...UNTIL static shocks blow-up the Gate and it stops working forever.

"Something touched me deep inside
The day the music died"

Thank you, Paul, for once again turning an "I think" into actual wisdom.  You are an awesome resource to have around here. 

[duck_arse]

I have two questions - why not change the input cap? and why the hell are they using a 100uF output cap? if the 100k to ground following were missing, that cap might act funnee for a while, like until it charged up, or something.