Double Shot..my new project, tube preamp..layout & pcb

[d95err]

Agree, it does look cleaner. It's important not to have signal wires near AC heater wires, but twisted or not apparently doesn't change much.

Twisting AC wires helps to cancel out the magnetic field generated, but it works just as well to have the wires paralell and close together (e.g. like in a piece of lamp cord). The Soldano heater busses work because of the 90 degree angle to all crossing signal wires.

[dschwartz]

I took another look at the schematic (the first one you posted, don't know what you've changed since then). Some more notes:

The first three gain stages share the same power supply node. This could be very bad in a high-gain circuit, and could lead to oscillation and "motorboating". The best solution would be to add another RC node on the power supply for the first (and most sensitive) gain stage. Another solution would be to move the third gain stage to the "B" node. Adding more filter nodes seems like a good idea though, considering that a typical high-gain amp will have about 5 nodes including a choke.

Hmm.. another RC node will not fit in the box.. maybe i can redistribute the B+ supply.. since i have 2 RC nodes, could be 1 for stage one, and the other for the rest of the stages?

Some of the frequency shaping is odd. For instance, the 220p cap across the 50k Clean volume pot will only affect frequencies above 14kHz, so it will let through noise but do nothing with guitar frequencies. The 1n cap across the 1M Dist gain pot seems very big (470p/1M in a Plexi bright channel is considered extremely bright). The 2.2n cap across the 680k resistor before the Dist pot essentially takes the resistor out of the circuit, since everything above about 100Hz will pass through the cap.

Actually, clean volume is 500K now.. 50k is too low.. i want the clean channel to be very bright..will 680pF be better for this job?..anyway, i did not put those caps on the layout cause i want to test different values.. The 2.2n/680k is just like mesa preamps..Eh! and i changed the first coupling cap for 3.3nF (2KV) to clear the bass content...i went too far?..

The 3M3/50k voltage divider on the clean channel takes the signal down 36dB even at full clean volume. Are you sure you want to attenuate this much?

Ehmm nopes  ... i didnt put that part on the layout  either, cause i wasnt sure.. wich value do you recomment to get a bright clean, but not that attenuated?

EDIT: Another thing - how is the grounding layout? Tube stuff is very sensitive about grounding and it's important to isolate high-current ground returns from sensitive signal grounds. For your preamp, the main retifier ground, heater rectifier ground, heater regulator grounds and the relay ground are high current and should not mix with signal grounds. Try to use some kind of "star" grounding scheme. Getting the grounding right is critical in order to reduce hum in a tube circuit.

Hmm i tried to do the best with that.. i couldt separate heater and b+ grounds because of the way i did the box (separated transformers).. so.. i only could have 3 wires going to the enclosure..The filtering/regulation is done inside the main box and have only one ground output..
so i did this:
- led´s and relays ground goes directly to the PS ground
- B+, signal, and heaters share the same ground at the PCB. i think the PCB has plenty of ground and not looping..star ground? i dont know..you tell me....
- signal ground of wiring and jacks goes to chassis ground


thats it...Great Observations by the way

[d95err]

On second thoughts, I may have gotten some of the frequency response calculations wrong, so don't put to much faith in them. Comparing to existing values of well-known circuits should work well. As for the cap and resistor values , just experiment! It's not likely that you'll get the perfect sound the first time around anyway. Try to get something working as soon as possible and then tweak to perfection.

About the filtering, I saw that the Soldano Supercharger GTO used only 2 filter nodes, so I guess you should be allright with three. Try moving the third gain stage to the "B" node if you can. I think oscillations and motorboating will be more likely if you have two gain stages with the same phase on the same node, since there is a risk of positive feedback leaking through the power supply.

Keep going! I'm looking forward to hearing the results! If my posts were any help I'll take all the credit... 

[dschwartz]

yes, mi plan is to make it work, and make a lot of tweaking after..something like design "as it goes" (sorry for my english, i tend to make up  words and phrases sometimes)

i´m having a big PITA with the wiring now..it´s a LOT, even though i made a PCB for the tonestacks, i have to use really long cables (wich i hate)  coz i put the input too far from the bypass switch..and it´s starting to look scary now..if i pull up the lit, i can see a lot of cables..and there only for leds and switching..

BTW, for signal, i am using shielded cable...

debbuging and tweaking is going to be a real painstaking, horrible experience!!!!!

[dschwartz]

THE MOTHER &%&$$ IS ALIVE!!!!!!

i finished the wiring, checked connections and plug this thing on!!!!

first a pic:



well..i didnt posted a clip cause now starts the debbuging phase!!!! i sounds like crap!!!!! hahaha

first of all...it have a huge low freq hum, like a muffled hum...maybe that´s motor boating?
its a hum that affects the signal..making it sound fuzzy and octave-ish.

i have other bugs on the tonestacks..the pcbs were wrong, so bass and presence controls work backwards..but i expected those kind of errors.. and a pot seems faulty..affecting the overall response of the tonestack..

the clean gain pot works weird. (i havent put a cap there yet) about 10 o clock the sounds abruptly changes, loosing a lot of bass content..
it´s a mean clean channel!!! it gets to heavy overdrive.. but not a nice one yet..

the sound is fizzy and somewhat gated.. i think the loud hum produces interference distorting the wave in an ugly way, i hope that the sound improves a lot if i put a filter to stage 1 B+, and separate it from the other b+ (2nd and 3rd stages)..

i didnt noticed normal noises.. no hiss at all..and no typical magnetic hum...
the switching didnt pop at all, too. those are the good things..

so i hope with your help i can make this beast sound good!!
greeting for everybody!!!!

[DWBH]

dschwartz, can you take a full picture of the enclosure?

[d95err]

Could you post the latest version of the schematic and PCB layouts? It would make it easier to discuss the voicing and problems.

EDIT: What frequency is the hum? Can you tell if it's 50/60Hz (grounding problem or heater hum) or 100/120Hz (rectifier noise)?

[dschwartz]

the pcb posted above is the latest version..

the power supply is the same, but replaced the voltage divider resistors for a 7805...

the hum sounds like 50hz..feels like low freq oscilation to me as it distorts the signal..or probably is hum+oscilation...

tonight i´ll try to post some inside shots, and a clip..

[d95err]

It sounds like a grounding issue. The first thing I'd try would be to get rid of the heater rectification and run regular AC heaters. That would at least eliminate a possible error source. The second thing would be to get better filtering for the preamp gain stages. If that doesn't work, you may have to redesign the grounding scheme completely.

Since you're using a separate box for the power supply, a complete star ground may be impossible. One thing you could do though, is to move the last two filter caps to the effect board. Try to make sure the grounds of each gain stage goes directly to the negative pole of the corresponding filter cap using as short path as possible.

If you want to keep the DC heaters, it will be very important to separate the heater grond from signal ground. One thing you could do (also with AC heaters) is to use twisted wires for the heaters that are raised above the PCB, connecting directly to the tube socket heater pins.

And a question: How is the chassis connected to the PCB grounds? One single point? Multiple places? Are you using isolated jacks?

[dschwartz]

Hi D95! your'e really helping me out here..i´ll give you full credit when this thing gets done..
First of all i am redesigning the power supply..please look at this schem:



it has a lot of improvements over the current power supply:
- Groun loop elimination trick (some people has told me that actually works)
- Separated B+ for stages 1, 2 and 3, 4 and 5.
- Separated ground for B+ and heaters (is that separated?)

it looks nicer too hahaha!!!

well D95.. for grounding the chassis i soldered a wire between ground of output jack and master volume´s pot back cover..
since jacks are isolated from chassis i grounded all jacks directly to power supply at a single point..
Gain, tonestacks, and bypass caps switches use this ground...
pcb ground directly to power supply. too

all signal wires were shielded at one side (from the pcb side) except for the output jack that feeds ground to the chassis

did i mess it up?
and is my new power supply ok?

[d95err]

Hi, first of all, you're doing all the hard work, so you take the credit! I'm not a tube expert by any means, I've just built one tube amp (but I've spent far too many hours on tube web forums...) 
It's a lot of fun to assit you though!

The schematic doesn't really tell if B+ and heater grounds are separated (a schematic wire is not the same as a physical wire), it depends on how you layout the PCB trace. If you want to make the separation really clear, you can use different symbols for heater ground and signal ground.

I'm a bit sceptical about the ground loop prevention circuit. As drawn in your schematic it isolates the rectifier ground from signal ground and heater ground, which I think may be a bad idea and could cause even more problems. I think it's better to try to get the basics right first, before experimenting with this kind of solution. You could try this later, but it should be between chassis ground and the whole circuit ground (signal, heaters, relays). That would help eliminate ground loop between your pedal and other devices it is connected to (e.g. an amp), but it won't fix the problems within the pedal. Hmm... it could perhaps be used to separate heater/relay ground and signal ground. Still, I'd try to keep things simple to begin with. The same goes for the 7812 regulator - I'd replace it with a suitable resistor for now and if things work OK, you could add it back later. You could still kee the same PCB layout for the regulator.

I recommend moving the first filter cap for each rectifier in the power supply box. The rectifier grounds are the most noisy and biggest potential problems. The first filter cap should be as close as possible (physically) to the rectifier. I.e. ideally, the ground end of the rectifier should be soldered directly connected to the negative side of the filter cap and the positive side directly to the positive side of the cap. That reduces the dirty ground currents through the power cable, and also makes the high voltage and heater voltages less noisy.

The heater currents are still a problem. Make sure they are separated from signal ground on the PCB and meet as close as possible to the jack for the power cable. If you could switch to 12V relays, they could use the same power supply as the heaters. That way, you could move the whole heater regulation in the power supply box and have only a well filtered 12V instead of the very noisy 17V in the power cable.

If it's possible, try to keep the gain stage grounds separate too. The grounds of each stage should meet at the negative leg of the corresponding filter cap. From there you can put a PCB trace to the common signal ground. By one stage, I mean from the cathode of a triode to every ground that follows until the grid of the next triode. So, don't take the grounds for volume and tonestacks to the shared ground, take them back to the filter cap ground of their corresponding B+ node.

Some of these tips may be overkill. I haven't built a tube pedal myself. I'm just telling you things I've tried in my tube amp (or should have tried), and stuff I've read about in books or on the web. It's a lot easier to do this stuff in a point-to-point built amp, because it's easier to move things around and add wires here and there. With a PCB I guess you need to plan ahead more...

BTW, please call me Erik! (d95err just happened to one of my first computer usernames 12 years ago...) 

[dschwartz]

i'm getting frustrated here...
took me all day to:
- Put a ground lift on the power supply...
- separated ground traces for heaters, relays and B+ at the power supply
- separated heaters ground from signal ground at the main board
- separated b+ for the first triode
- added a 47uF cap next to HV rectifier
- replaced the 12V reg for a 4ohm resistor (GAVE ME 12.4 VOLTS W/HEATERS ON)
- Changed coupling caps for 630V rating caps
- Added a 47uF filter for relays 5V
- Added a 10nF fiter cap for B+

and the big hum is still there!!! and hiss made appereance....

when i switched the ground lift at the supply (choosing the rectifier´s ground) this thing squealed like a pig
NOticed that even with gain at 0 there was the loud hum (cleaner, but loud..)
i think the ratio of noise is 1:1..

im losing my nerve here!!!!!

[stephanovitch]

Cut the wire of the actual 12 V supply
Try in first to supply the heater circuit with a 12V battery (car).

It's an easy mods to identify if your noise come from heater circuit.

[d95err]

Tube stuff can be tricky, especially high-gain...

I noticed one thing about your switching system. The input to the high gain stages (second and third) is not grounded when switched out. The unconnected input will work as an antennae and pickup lots of noise. Although the output of those stages is not connected to anything when switched out, there is a lot of amplified noise in there and that could affect other parts of the circuit through crosstalk between traces or leakage through the power supply. This is probably not the main problem, but it could make things worse.

Try to disconnect as much of the circuit as possible to find the problems. Temporarily hardwire the whole thing for clean channel only. Remove the relays, MOSFET source follower, tonestacks and DC heaters. Cut power to the high-gain channel stages. Once you get that working, try adding features one by one.

Another thing - make sure you try different preamp tubes. You could have had bad luck and ended up with a microphonic or damaged tube.

Keep working! I'm sure you'll get it working eventually!

[dschwartz]

hmm... that battery idea seems good...
but i dont have one.. maybe if i use a separated 12 v adapter?

should i change the hole PCB???
anybody sees any suspicious traces or something....

[d95err]

Are you still using the same PCBs that you posted, or have you made any changes?

[dschwartz]

yes i´m using the same pcb's.. but cut some ground traces...

by the way i missed grounding the chassis..and that lowered the hum a lot!!!..but it´s still there..it let´s you play, but annoying anyway..
i´ll try to put the 12v regulator again....and maybe redo all the pcb..makinmg one big board including the power filters next toi the cathodes..and appliyng all i´ve learn thanks to you all!!!

[dschwartz]

more news...
i did these things:
- put back the 7812, with a diode for 12,6V
- star grounded to the chassis the b+ ground, main board ground, AC Main ground, heater filters ground.
- Added a 100 Ohm 2W resistor in series with the heaters ground (the ground going to the main board, i read somewhere that that will bias the heater ground more positive, reducing noise, although i have a DC regulated supply, i think that it works, or doesnt hurt at least)
- Isolated jacks from chassis
- Wired input jack ground to the board input ground, output and tonestacks to the board output ground

results:
- Goodbye oscilations!!!!
- less louder hum and cleaner..BUT..the hum wasn´t affected by the gain or volume settings, at min or max vol/gain, the hum was there.. my instinct (and empiric knowledge) tells me that it means that i still have a ground loop at the output wiring, somewhere between the master volume and the out jack (i have a passive fx loop between both)..

so i have the  SOB against the wall!!!!! i know where yua´re you bastard!!!

are my thoughts OK?
by the way.. it´s starting to sound REALLY good

[chillhuman]

more news...
i did these things:
- put back the 7812, with a diode for 12,6V
- star grounded to the chassis the b+ ground, main board ground, AC Main ground, heater filters ground.
- Added a 100 Ohm 2W resistor in series with the heaters ground (the ground going to the main board, i read somewhere that that will bias the heater ground more positive, reducing noise, although i have a DC regulated supply, i think that it works, or doesnt hurt at least)
- Isolated jacks from chassis
- Wired input jack ground to the board input ground, output and tonestacks to the board output ground

results:
- Goodbye oscilations!!!!
- less louder hum and cleaner..BUT..the hum wasn´t affected by the gain or volume settings, at min or max vol/gain, the hum was there.. my instinct (and empiric knowledge) tells me that it means that i still have a ground loop at the output wiring, somewhere between the master volume and the out jack (i have a passive fx loop between both)..

so i have the  SOB against the wall!!!!! i know where yua´re you bastard!!!

are my thoughts OK?
by the way.. it´s starting to sound REALLY good

Another thing you may want to try is unscrew the transformers, and physically move them around to see if the PT is causing you troubles. It's better (from a safety POV) to fire up the pre, listen, shut it off, discharge the caps, move the PT's, and repeat. While I'm thinking about safety, what I do when I'm working on an amp is keep one hand tucked away in my back pocket, or into the waistband of my pants. This greatly reduces the chances of me being grounded with my left hand, touching B+ with my right hand, and having the current travel through my heart. It takes longer to get the work done, but being dead REALLY pushes back the time of completion.

Also, I'm assuming you poked around in there with a chopstick to make sure your lead dress is up to par ? If not, you could try moving some wires around and see if it produces an audible difference one way or the other.

[dschwartz]

the PT is outside the box..in other enclosure.. y moved it around, and did´nt change at all..it´s shielded.. so, no hum from that...

moved the wires inside the box, too.. no change.. except for one that was "sensitive" to the touch,,

i´ll redo the wiring anyway..
} so.. i´ll keep you all informed..