GTFO - Full tube high-gain pedal (2x 12AX7)

[Uriziel]

So far I have tried:

2ohm in series after the dc jack positive pin... this is as much as I can go because of the enormous voltage drop across a resistor in this circuit... calculated current usage is somewhere between 200-300 mA... the 470uf capacitor from +12V input to ground is suposed to act as a low pass filter even though it doesn't have a resistor before, but the traces going there act as a small resistor?

i also tried ground lifting... 10ohm between the effect ground and dc jack ground pin... ground is aprox at 3V atm

i cut the +12V trace going to the heaters and used an alternate power source for them

i removed the tube sockets and put shielded wire between the socket and the pcb

I tried sound probing... in bypass mode the tonestack (i know it's not in play when bypassed, but went over it anyway) components give out a high-pitched periodic sound... i guess the triger from 555 bleed through since even though bypassed from the first stage the others are still working and want something in their inputs... also the input resistors when touched with the capacitor give out a bangy noise, maybe it's just because of their sensitivity since they are there to prevent that noise when inserting a jack when the device is powered... some other components do that also, a few near V1 but component around V2 don't... i don't know if this is a indication of anything but i'm pretty much out of option short of redesigning everything and ordering a new pcb that might not improve anything anyway, so need to fix it, or if I can't then just find the cause so i can design a new pcb trying to avoid the mistakes i have made with this one...

maybe i'll try and replace the trimmer with fixed resistors since i already thought about that possibility and designed my layout accordingly when i find a good setting i like since everything i have gathered about trimmers from various sources is that they are noisy.

any help you can give my, your expertise, your experience?

Thanks

[gtudoran]

Hey Uriziel,

Hmm there is smth i would like to say: why  did you use SMD?? There is a reason why i didn't use SMD in the first place (i could do that but... is much easy to use THD components and also the space is bigger so the interference is smaller). I harder to use SMD components in this design. I saw a lot of vias and a lot of traces that goes under components... that i guess generates a lot of interference.

I just an idea and nothing more, but i would recommend to stick with the original THD design ... you could build the SMPS with SMD components but i would stick to THD on audio signal parts. There is also a reason why the smps is placed in that place. Keep in mind that the SMPS that is used is... a very rudimentary one and was chosen for the small parts count and for the size and so ... the equilibrium is very thin.

Best regards,
Gabriel Tudoran
Analog Sound

PS: You can design a proper fly-back smps and that would solve much more of the problems.

PPS: Don't try to find solutions where there is no problem - i'm referring here to the separate heater PSU, i don't believe that ... DC heaters are better then AC ones in therms of musical quality if everything is designed properly.

[Uriziel]

Hey. I think that THD vs. SMD is more an ideological problem. One isn't better then the other. This project in particular has to have components with parameters high enough that the size comparison between smd and thd is negligable. I simply like smd more. And for me it's easier. The choise bwtween smd or thd is pretty much how easily can you get certain components and at whay price.

I think i just made some mistakes designing i guess. There would be traces going under thd components aswell so i don't think that is the cause. Vias are there to provide the shortest path to ground although they are placed totally randomly whereever i could fit them... I didn't see an issue in that since there aren't any sensitive ic's in the schematic that can't handle even small interfetences. As far as i have seen, there are designs of this project that are much more compact than mine.

Since i don't know where is the actual issue, i'm just trying diferent things that come to mind. And it is my first experience with tubes and i have a simple learning process.

[gtudoran]

No offence intended and by far no argue about THD vs SMD (i'm using SMD all the time - more in the digital field but that is my preference) Just said that you are more prone to mistakes when using SMD (especially when working with tubes).
Ohh a question, where can i find SMD capacitors for high voltage?! i saw that you used some there and i would like to know where i can buy some.


Best regards,
Gabriel Tudoran
Analog Sound

[tubegeek]

any help you can give my, your expertise, your experience?

Here's something that I would try based on your troubles:

Hopefully you have another 100uH 3A part in hand, same as L1. If so, at the connection between D1 and R4, try breaking the connection from there to C4. Then connect that path back together using that spare 100uH inductor. You'd basically be making a two-pole LC filter instead of relying on an RC filter, which as you point out, is too wasteful of voltage in this circuit to be useful. The cutoff frequency of the LC filter would be 7.3KHz and drops 12 dB per octave so it would be very effective at 80 kHz, roughly 40 dB down or 1/100 of the noise voltages that you have now. It will filter the DC into your HV supply node quite well, and especially clean up high frequency hash. Not so perfect at 60 and 120 Hz for filtering mains frequencies, but those are already gone in the SMPS arrangement anyway. (That's why you typically see much, much larger inductor values in "old-school" linear tube power supply filters - they're there for filtering mains hum, not hash.)

You might (assuming the filter is not "cutting" it [pun intended]) parallel the 4.7 uF capacitor with a small ceramic cap of whatever value you happen to have around, say 1000 pF, because ceramic caps are typically able to maintain low ESR up into very high frequencies, where standard electrolytics can start developing a larger ESR component. In this way you'd have very effective filtering path as well as a large reservoir for transient demands in the 4.7 uF cap.

While you're breaking out the 1000 pF ceramics, connect each heater pin to ground with a cap to prevent coupling of HF through the heaters. This is usually only a problem with very high-gm tubes (which the 12AX7 really is not) so it's more of a belt & suspenders move than a necessary one.

High frequency oscillation problems often masquerade as other problems: excessive hum, excessive microphonics. If your tubes are oscillating at 80kHz, that may show up in the audible spectrum riding on small perturbations that would be negligible without the HF hash. The fact that you have a scope fast and sensitive enough to see your 80kHz problems means you have the tool to verify a kill once you take measure against them.

i removed the tube sockets and put shielded wire between the socket and the pcb

What was the shield on those wires connected to, and which connections did you shield? I'd say that's not going to really help matters any, I've never even seen that done - so it's certainly not a necessary effort.

[J0K3RX]

I like your layout, a lot! But, there maybe some geographical things going on here? I don't see why it would be that you used surface mount even though it's not used solely in very many tube amp/preamp designs that I am aware of. Even pedals like Blackstar use a mix of through hole and SMT so, I am not sure of the pros and cons or if there are any at all? We do know that Gabriels design works pretty much "hum free" at least the couple that I have built have been amazingly quiet! Assuming that all of your components are good the only inconsistency is the layout. Have you tried wiring the input and output jacks off-board from the closest points, not where the jacks are mounted from? Maybe it has something to do with the jacks being where they are?

[Uriziel]

No offence intended and by far no argue about THD vs SMD (i'm using SMD all the time - more in the digital field but that is my preference) Just said that you are more prone to mistakes when using SMD (especially when working with tubes).
Ohh a question, where can i find SMD capacitors for high voltage?! i saw that you used some there and i would like to know where i can buy some.


Best regards,
Gabriel Tudoran
Analog Sound

Hey,

I didn't take as offence in any way
I understand what you were implying and I am simply experimenting in my own way since i believe there isn't a right or wrong way of doing things. Even though in some cases there is proof for why some parts of the layout use certain type of parts... though i'm inclined to believe that mostly it would apply to the black magic of rf design and the digital realm... perhaps i'm wrong but that is why we all learn

Also all of the capacitors used I ordered from www.farnell.com They have distribution points in UK and USA (under the name of Newark or element14 in general)...

[Uriziel]

Here's something that I would try based on your troubles:

Hopefully you have another 100uH 3A part in hand, same as L1.

Sadly I don't... I don't have much experience with inductors but i will certainly try that... maybe i can reuse something from older piece of electronics or I will just order a new one...

Not so perfect at 60 and 120 Hz

In my part of the world its 50Hz and 100Hz for sine and rectified sine respectfully... Shouldn't affect much

What was the shield on those wires connected to...

I connected all the shields together at the tube end and i used shielded wire for all of the connections, besides pin 4 that is gnd and pin 9 that is nc... and I connected the shield to tube end gnd...

[Uriziel]

...there maybe some geographical things going on here?...Have you tried wiring the input and output jacks off-board from the closest points...

I can't think of a reason why it would make any difference but hey worth a try...

Can't do anything about the layout now but when i can't seem to find the source no matter what then i'm thinking about adjusting the layout to geographically look like the best noiseless version of it so far. Since I anyway need to fit it in the same box, the tubes and the jacks and the pots need to stay where they are, but anything else can move around... maybe that will help somewhat...

About the blackstar, I think that the mix of SMT and thru-hole is somewhat a price considerations because when you look the prices of SMT's, specially HV ones then they greatly surpass the prices of THD components... not a considerable amount if you buy small quantities but at larger quantities you may start to notice the difference... Alse certain techologies eg. Silver Mica is not available in SMT and electrolytics usually have shorter lifetimes unless you are willing to spend a fortune... The resistors I use a MELF so they are pretty much the same as 6mm resistors w/o the leads... again pretty much most of the techologies used for THD have equivalents for SMD aswell and with the same level of quality... nothing else is changed but the enclosure for better us in reflow or wave soldering applications... imho...

[Uriziel]

I tried all of the suggestions above... ceramic cap, parallel to 4.7u electrolytic... decoupling caps to the heaters... two pole LC filter... on the oscilloscope that showed promise but testing through an amplifier it was clear that nothing had really changed... the noise level was the same as before... still need to try lifting the jacks of and maybe i'll even leave out the input resistors... the input seems too sensitive to me and i believe the noise comes from there and is amplified so it's quite impossible to filter it out afterwards w/o significant loss to the sound... it effectively acts as an antenna but instead of radiowaves it picks up the ground level noise... maybe if i try decoupling input to ground?

i ordered some larger caps and i'm thinking of replacing the 470u cap before the 555 to something like 1000uf and the 4.7u cap to something like 22u or 33u

i also tried reheating every component in case of cold solder spots but nothing... i think the fault is somewhat more fundamental than mere fiddling with values and filtering out unwanted waveforms since by now there should have been an audible difference to the changes i have already made, but everything is the same

[tubegeek]

You must be coupling the noise in via the ground traces if the scope looked good but the result was bad. Hmm. Thinking....

[vigilante397]

So I'm still kinda new to this, but I'm COMPLETELY IN LOVE with this pedal. Is there by chance a vero layout floating around for this? Because that would make a lot of beginners REALLY happy 

[Uriziel]

Trace the schematic and try to make a vero layout yourself. Then you can post it here and you'll get feedback from the community about if, what and how to change. It's a pretty good way to learn for possible future projects.

[vigilante397]

Trace the schematic and try to make a vero layout yourself. Then you can post it here and you'll get feedback from the community about if, what and how to change. It's a pretty good way to learn for possible future projects.

Started on the layout last night and fell asleep doing it It would be easier if someone already had one, but doing it myself will be a great learning experience I'll probably finish it up and get it posted tonight.

[vigilante397]

Okay, so I just finished my vero layout on paper (I've been doing it on graph paper during downtime at work ), and hopefully tonight I'll get it all put into the layout creator when I get home.

Keep in mind this is the first vero layout I've ever done, so I'm expecting there to be mistakes and I'm expecting some constructive feedback

[vigilante397]

One question I came across as I was doing it though, where's pot #2? I see pot 1, 3, 4, and 5 on the layout, but I'm not seeing #2. Help?

[vigilante397]

Okay, so I just finished my vero layout for this. I know there's tons of people on here that know a lot more about this than I do, so PLEASE share you expertise by taking a look at this and letting me know what's wrong and what I can do to fix it

I'm uploading the image file so you can take a quick look, and I'm also uploading the DIY layout file, so you can take a better look and maybe even fix some things for me

https://www.dropbox.com/s/3vpedakzcmwd46o/gtfo%20vero%20v01.png

https://www.dropbox.com/s/6ozunrutw3wj3a3/gtfo%20vero.diy

You'll notice I set a lot of things up off the board to just do point-to-point, because my soldering skills are significantly better than my board layout skills

[hopkinWFG]

Hi vigilante ...

very nice layout you posted ! I am also in the process of populating the parts for this project... do you mind take a pic of your done velro board? does it fit into a 6500 enclosure?

mind to ask how do you intend to sort out the tube sockets? do you intend to mount it on a Velro board as well?

[J0K3RX]

If you guys are gonna do this on vero with the tubes not connected to the main board you might consider mounting the tubes so the are parallel to the enclosure. That way you can enclose them inside the box so they are not protruding out of the top of the pedal? I know the tubes look cool sticking outa the top like a blower sitting ontop of the hood of a 69 camero... but even with the roll bars it scares me and I could still damage them easily..

I know these are not vero but just to give you some idea of what I am talking about..

[hopkinWFG]

I still trying to getta answer if I could getta conventional PCB tube socket type to mount it on a velro board?
I  still like the idea of the two 12ax7 half way protruding out from the pedal box and with handle bars sticking out...

while waiting for my parts to come ! i'll maybe try to draw out a layout plan ?  i'll already spent somewhat 100 USD on the kit lol...

btw that's a massive tone controls you did !