proposed dr. boogey layout -- seeking comments

[John Lyons]

Victor
The ground jumper at the bottom doesn't do anything because the ground connects almost everywhere across the board already. Gaussmarkov made grounding points to form a star ground but your layout mostly does not use this approach.

The "crosses" that center each component should be turned off or switched "off" so they are not seen. They obscure the partes values and names.
Some of the parts names and values are overlapped as well.
There are also some missing values,  and you should probably use the transistor component outline as well to make sure you have the correct orientation.

I don't mean to pick apart your layout but it's hard to follow.

John

[Cursor]

Hi all,

I'm jumping on the Boogie bandwagon, albeit on vero, and doing F4T from the original Dual Rectifier schematic.

Just to check (there's a LOT of talent posting in this thread!) - I'm not using the tonestack, and I only have 4 J201s to hand (nicely matched, by coincidence). Since Q5 is a buffer, can I just drop in a 2N5088?

Many thanks for everyone's great work.

[Pushtone]

I'll try to add an external buffer to my Dr. Boogey (Buck layout), and see what happens with it on and off my signal chain...... maybe is worth the shot.... 

I just finished a GGG IC buffer on perf last night for just this purpose.
It was too late to test last night. Give it a go when I get home.

I was also thinking about the AMZ MosFET boost as an output buffer.
It's input impedance is very high, 10meg I think.

Back to work now.

[drizzt]

Great stuff. Every day is a new day!  and every day is a full day of improvements.
All I wish is that all of you electro gurus to keep up with this great project. This must be the greatest topic on this forum in terms of ideas and self-development. every morning i come here only to be amazed with your stuff 

I've learnt great things with you guys. Althought, i wish i could participate more in this, but i don't have the knowledge it demands.
I've built a Dr. boogie (failed! - too many osc), but now, i'm willing to build another.

Thank you very much.

[Victor]

John,

Thanks for the tips. There are some little things in Eagle that I don't know yet, and I'm glad I "found your help". 

I'll delete those two jumpers located at the bottom as you said, since there's room to connect gnd to it without the need for more wires. Just moving some components and it's done. A "jumperless" layout would be very nice. 

I'm also worried about pots connections in the board. They look nice, no-crossing lines, but "presence" wires will still pass close to "mid" wires. 

Pushtone,

In my opinion, I don't think you need to worry about AMZ Mosfet. Its input impedance is high, but we don't need that, I believe. And also it is much more components than another single fet source follower stage using another J201. But that just me..... 

Cursor,

You can't just swap a BJT (bipolar junction transistor) for a FET. Each one needs different polarisation techniques, so be careful. It's possible to use something similar to source follower FET - emitter follower - but, in my opinion, it's going to be some waste of board space to accomodate another components.

[John Lyons]

The main thing to worry about in crossing wires is that the connections do not cross the later connections off board.
This means that the gain control wires are short and stay clear, as far as posible from the tone controls.
The closer the in and out connetions the worse the oscillation will be.

A good way to look at it is:

Input on the left and the output on the right. Try and keep the parts placement starting on one side in and moving to the other side and then out. This gets a little tricky with the switch wiring but that's where shielded wiring helps.

Any buffer should work well I think. BJT or FET. They both should use about the same number of parts.
Using the Mosfet buffer isn't really the thing we need though. That's mainly an input circuit not necesarily on to tack on the end. The input impedance is too high I think. Victors look good as far as I can tell. A BJT should work as well but not a drop in for the fet circuit already posted. You will have to bias it a bit differently.

John

[gaussmarkov]

you all have been busy!!    it's nice to come back and read through all the posts.  and to find that pushtone has completed a successful build that has no oscillation problems, even without shielded connections.  i hope we hear from some other builders who were waiting for his report.  i am wondering whether abandoning the star grounding system is a good idea at this point.  it seems like that is where victor is headed.  or am i misinterpreting?

the output buffer seems like a good idea.  i'm with john on bjt or fet.  many, many circuits use an emitter follower in the same location, including the good old reliable ts808.  but let's go with the source follower that victor advocates.  pushtone's build suggests to me that we have room to reconfigure for either type of buffer and i'd be happy to adjust my layout for that.  given how tight it is at this point, i would probably end up moving things around a bit.  i don't think we need more room for the caps, WIMA box caps will fit fine, but i will keep my eyes open for that concern also.

it's good to be back, gm

[John Lyons]

Alright, here we go!
Glad to see your back now gauss.

I don't see the point of abandoning the star grounding. May not be crucial to the circuit but it's good practice to get into the habit of and you've already done the work! Can only help.

Adding a buffer seems like the way to go. High impedance input, low imp output is where we need to be. One FET, an output cap and a source resistor, no biggie! Does C22 on Victor's schematic need to be there? My understanding is that the Tone stack blocks DC and there isn't a need for a coupling cap before the volume pot. Right?

John

[gaussmarkov]

Adding a buffer seems like the way to go. High impedance input, low imp output is where we need to be. One FET, an output cap and a source resistor, no biggie! Does C22 on Victor's schematic need to be there? My understanding is that the Tone stack blocks DC and there isn't a need for a coupling cap before the volume pot. Right?

John

i think you are right about what is coming out of the tone stack, but the source follower -- even though it is near unity gain -- introduces a DC offset also.  so i think we also need C22.

[Victor]

gm,

Nice to see that you are back to work with us....... 

My idea on using some kind of approach that minimizes jumpers on PCB does not leaves off star grounding...... I believe we should maintain that idea, to star-ground the external connections such as jacks, 3dpt, and other things that are necessary......... but I'm working on a jumperless board, and with each pots wires going as straight out from the board as possible, not crossing any of them, if that is possible (I don't know if you guys can fully understand me, my english is kinda bad these days  ) 

By the way, my layout on output buffer of boogey has thinner copper traces, as you'll see below. I hope that does not look like a problem, maybe PNP or couche, nicely applied, can transfer that to the board with no problems. Here's the progression on it (without bottom GND jumpers):

[John Lyons]

Victor
Looking better with the cleaner layout.
Star grounding means that each sections ground connections go back to the star point without passing through the other sections.
If you can imagine all the different grounding points within each stage connected together, then a sible wire from each section connected at the point of the cleanest ground source which is the power supply capacitor and single ground point which connects to the battery or DC jack.

The Jumpers in Gauss's layout are the connections that connect each gain stages star point. with your board layout the grounds are all connected together to one big ground plane. As I mentioned above, this may not be too big of a problem, Gauss is just working with the theory of star grounding.

The thin buffer traces aren't a problem I don't think. Why not just fasten them up? There seems to be room.

Gauss I'm not sure about C22... Are you thinking DC is coming from the tone stack or are you thinking DC is coming from the Buffer back to the volume. My ignorance showing here...

John

[Victor]

John,

So you're saying that, in order to tame down noise and oscillations, we need to isolate every stage of the circuit and connect each ground path from them individually to the "lowest/better" ground point, to do the star-grounding? In my board, stages' grounds are all mixed up, so can we expect some kind of noise/osc coming from one stage from another, even if the're all connected to star-ground point?

In my head, I imaginne that your star-grounding technique could still cause some kind of noise/osc problems, mainly because all GND points are still connected together, similar as in my board, to the main star ground junction point. Is this makes any sense? Or am I just confused? 

I need my books! hehe     

[gaussmarkov]

Gauss I'm not sure about C22... Are you thinking DC is coming from the tone stack or are you thinking DC is coming from the Buffer back to the volume. My ignorance showing here...

sorry, i got my labels confused.  i think we need C23.  like you, i don't see a need for C22 except to change the highs with the volume level.

So you're saying that, in order to tame down noise and oscillations, we need to isolate every stage of the circuit and connect each ground path from them individually to the "lowest/better" ground point, to do the star-grounding? In my board, stages' grounds are all mixed up, so can we expect some kind of noise/osc coming from one stage from another, even if the're all connected to star-ground point?

exactly.  and our collective experience seems to show this.  pushtone built both versions and his star-grounded version is tamer.  here's the an example of how the experts explain this:  you don't want the ground of a sensitive part of the circuit (like the input) passing by (or through) the ground of a high gain section.  there is a more detailed explanation on Aiken Amps that is commonly referenced: Star Grounding.

cheers, gm

[gaussmarkov]

I'm working on a jumperless board, and with each pots wires going as straight out from the board as possible, not crossing any of them, if that is possible (I don't know if you guys can fully understand me, my english is kinda bad these days  ) 

victor, that's fine.  lots of people like to avoid jumpers.  in some cases, jumpers are like crossed wires that lead to noise/oscillation problems and should be avoided.  in other cases, jumpers are just resistors with no resistance.   we run traces under resistors all of the time.   so why not under resistors with no resistance? 

in the case of my layout of dr. boogey, the jumpers are for ground and they are crossing supply rails.  no traces carrying signal are involved.  the supply rails and ground rails are next to each other on purpose because our experts (e.g., R.G.) tell us this is usually a good thing to do.  it's like twisting power supply wires around each other, a common stompbox building practice.  passing a jumper for one rail over the another rail seems to me like it is in the same spirit.  but i could be wrong.  maybe keeping things parallel is critical.  i just haven't seen such advice yet.

By the way, my layout on output buffer of boogey has thinner copper traces, as you'll see below. I hope that does not look like a problem, maybe PNP or couche, nicely applied, can transfer that to the board with no problems. Here's the progression on it (without bottom GND jumpers):

for a long time, i also used thinner traces, too.  some of the older layouts on gaussmarkov.net still have 16mil traces.  but several folks on the forum have commented that fat traces and pads are preferable and that makes sense to me.  so i am switching over.

thanks for your input!  gm

[Pushtone]

for a long time, i also used thinner traces, too.  some of the older layouts on gaussmarkov.net still have 16mil traces.  but several folks on the forum have commented that fat traces and pads are preferable and that makes sense to me.  so i am switching over.

thanks for your input!  gm

For me fat pads are more important than fat traces.
The home etched boards lift pads real easy after the second or third soldering.
I've never lifted a trace, just the pads.

Basicaudio's suggestion about the PhotoShop filter "Filter ->Other ->Minimize"
works great for fattening up bitmap traces before printing.
Great tip, thanks Basic! I'll be using that one fer sure.




Comparing my two builds the biggest advantage this new layout provides is the single edge of off-board connections.

Both my builds work but the Buck version squeals when nothing is connected to the input jack. I can also make it squeal by pushing the MID pot wires closer to other ones. Both builds produce the same amount of buzz and hum.  If the star grounding was making a difference wouldn't it produces less hum and buzz? Can the lack of oscillation in the "Gauss" layout be attributed to the star grounding or is it solely the off-board connections?

I still have yet to try MPF102 trannies which might go a long way to lower the hiss.
I'm still biasing both units at around 7.5V to get lowest amount of hum and buzz.

This weekend I'm going to swap out the 100K VOLUME pot I used in
the "Buck" build with the stock 1meg to see if that makes to two units sound identical or not.

Tried a second time to get a good sound out of the DB with a 2005 Fender Twin.
Still yucky, harsh, high-end heavy, not enough bass.
Anybody getting good tones with this setup cause I'm not.
Other amps sound OK to very good with the DB. 

Glad to see your back on this thread Gauss.

[John Lyons]

Trying to clarify.
Grounding has to do with Buzz/noise/a little to do with hiss and general cleanliness of the circuit as far as grounding noise.
In amps the grounding is crucial to the noise floor of the amp. In pedals it's to a much lesser degree but still a good thing to look at in making layouts.

Oscillation comes from sensitive high impedance points of the circuit. gate inputs of fets and to a lesser degree the resistive dividers at the inputs/gates. The more resistance in line the higher the noise. Make the the resistor a larger suface area (although not at all not practical for PCB pedals) and the noise will go down. Using metal film resistors will make a big difference as well.
Keeping the snsitive ares away from the high gain areas is the thing we want to do. 
Setting up the board so you don't have to cross sensitive areas when wiring up the pots iis a good thing.

I'm curious how the Slo century at OLCircuits layout looks. To sell a pedal without all this fiddly osc issues is a feat. Maybe they used all MPF102s or something.

________________

Larger pads are very much nicer where you can fit them. I try to use .100 pads as much as possible if not bigger even. Soldering them is simpler and getting a good shiney connetion is easier because you have a bigger metal surface and it gets hotter and works with the tip of the iron better at leasty to me. I use a 1/16" screwdriver tip in mr iron.

Bigger traces are nice for Photo paper etchers (me included) .040 or .050 is what I use.

John

[Victor]

gm,

Thanks so much for your help (and thank others too). I guess I'll leave aside my board and work on a new version based on yours. Since Pushtone had built it successfully, and now I have a clue about star-grounding, I think the winner for me is still the stock version of the board, made by you...

So, maybe star-grounding is not the main issue on gm's board, as we can see.

I was wondering if maybe using some more theory to determine an optimum operacional point for all FETs would be nice. I still have some doubts regarding the fact of drain's voltage being half of supply voltage, if it's worth to maintain that (I've tested some trimmers tweaking on my Buck's Boogey to see and there are not so much sonic differences between some close values of drain voltages to my ears)  or just replace all trimmers (I read somewhere that they're "noisy") by fixed resistors........ to do that, we need to determine Vp and Idss of all FETs, and calculate a closer drain's voltage assuming comercial resistor values...... JFET matching is the solution for this, and some paper and pen, maybe.....

Vp in J201 is around -0.3V to -1.5V.
Idss in J201 is around 0.2 to 1mA.

I'm working here with Vp= -0,9V and Idss=0,6mA, so far I found Q1's drain resistor = 22Kohms, for Vdrain=1/2 of Vsupply..... I'll keep working on that to see the possibilities..... 

And I'm worried about those floating wires crossing in tone-stack pots....... would it be nicer if we made them on the board?

So, the ouput buffer, no tone-stack scaling, maybe some new copper traces to give tone-stack a "neater floating look" and the possibility of working into FETs matching plus resistor on drains....... does it sounds good (at least on paper)?

[gaussmarkov]

So, maybe star-grounding is not the main issue on gm's board, as we can see.
...
And I'm worried about those floating wires crossing in tone-stack pots....... would it be nicer if we made them on the board?

i guess that's right.  the star-grounding does not seem to matter much after all, according to pushtone's latest report.  too bad.  nor does all the care to get a simplified signal path.  it seems like this supports a hypothesis that i have been wondering about for quite some time:  the care used in building amps often does not make any difference in stompbox builds.  but it is also consistent with another hypothesis:  that i have not implemented star-grounding and signal path protection sufficiently. 

if neither star-grounding nor signal path protection matter here, then we should just follow victor's suggestion and focus on the pot connections.

[John Lyons]

I think all the details to groyunding and placement does matter maybe not as much in an amp but were going down the right theory path at any rate and that can't be a bad thing.
If we can get the pots pacement and wiring worked out then that's as good as it's going to get.

See if you can work in a buffer gauss and we'll move on to something else.
Just like trainwreck amps, the wiring layout is key.
It would be good to have a wiring layout to go along with the board layout.

John

[gaussmarkov]

on second thought, maybe the signal path routing has made a difference.  as john points out, we can protect ourselves from oscillation by careful routing of the signal path.  also, pushtone said that he didn't use any shielded wires for his hookups.  so doesn't that suggest that we should stop worrying about pot wiring? 

I think all the details to groyunding and placement does matter maybe not as much in an amp but were going down the right  theory path at any rate and that can't be a bad thing.  If we can get the pots pacement and wiring worked out then that's as good as it's going to get.

See if you can work in a buffer gauss and we'll move on to something else.  Just like trainwreck amps, the wiring layout is key.  It would be good to have a wiring layout to go along with the board layout.

John

o.k.  i started looking at it last night and it's going to be a bit of work.  even though the buffer is really simple, it still takes up too much room to just squeeze it in there.

pushtone, i hope that you get a chance to try that buffer you were going to insert in your build.

cheers, gm