PCB layout criterias??

[Learner]

Hi guys,
I know I really should be diggin into the goodol'area 621 for this but I would really like to get some thoughts on this here as well, I am going to try to finish up my La2a PCB layout with regulated PSU in this term break holiday(next 2 weeks) picking up where I left off end of last holidays. My concern is that I remember reading or hearing from somewhere/someone that it is not good to have tracks on the PCB parallel againest each other, because of the inductance that might interfere with its parallel tracks?

But I guess the inductance is depending on how much DC current is running on the track parallel to each other, is there anything in particular that people should watch out for when doing the PCB layout?

I would not want to complete building the circuit and find out the problem the hard way, it would be a total waste of time finding out it sounds "sh*t" because of the layout is crap.......

I am also very intersted to try the lo/hi pass filter side chain thingy, on the feedback and insert source. Might also implant the vocoder concept using may be 10-11 bandpass filter on the side chain insert to trigger/shape the envelope of the VCA, what about rectifying the 6AQ5A output to DC and incorporate attack and release control to the rectified DC to drive the El panel?

But first I will sort of the regulated PSU and take it from there, might not be a LA2A anymore by the time I am done.........:shock:

Thoughts!?

 

[gyraf]

Parallel PCB tracks is a minor problem in layouts, untill you reach frequencies much higher than what interests us here. Or if you possibly work with extremely high impedances and/or very high currents.

Layouting is more or less an empirical science - at least for me - I redesign my boards on a regular basis :razz:

There always is a detail that you want to change..

So go ahead, do a layout, build it, and see if you like the outcome..

For HT PSU regulation I'd recommend the TL783 circuit that I use in the G9. Very low noise, very stable. Adjustable. Only drawback is the max 120V input/output differential (this is the maximum voltage it will "sink")

Jakob E.

 

[Learner]

[quote author="gyraf"]

For HT PSU regulation I'd recommend the TL783 circuit that I use in the G9. Very low noise, very stable. Adjustable. Only drawback is the max 120V input/output differential (this is the maximum voltage it will "sink")

Jakob E.[/quote]

Hey thanks Jakob! :thumb:
That is what I'll be using, just a qeustion about the regulator have you ever tried to push and see what is the max output voltage this thing can handle at say max current output 700mA?

I am thinking of building a bench HT PSU with this, just by placing a potentiometer to replace R2 like say a 1 meg linear pot as on Fig 20 on the data sheet. Is there anything I should be watching out for?

http://focus.ti.com/lit/ds/symlink/tl783.pdf

Thanks!

:guinness:

 

[sismofyt]

I can't remember how much effect the TL783C can sink, but it's probably around 15W. So you cannot get much current out at low voltages if you're using, say 115V unregulated. Integrated Vregs are not the best for lab supplies, you need to go discreet with probably a MOSFET follower.

Also, I'm not sure that using the Gyraf trick to pump up the output voltage and making it adjustable at the same time would be smart. Thirdly, don't use a cap on the regulator leg or the TL783C will act funny. And lastly, it must see a minimum 15mA load to work properly.

But tube circuits seldom need precise voltages and if it's just for testing you could skimp a bit on the quality and use a 12 pole switch, some suitable zeners and a emitter follower.

Or you could use a multitaped trannie and/or dropping resistors.

Personally, I use a Variac plus a rectifier and some caps.

PCB design is very critical if you're designing a huge monster like an SSL. I think the ground in scheme in general is quite important. PSU distribution is more dependant on good decoupling than PCB design in my experience.

It's almost impossible to predict how a drwaing will actually work in real life. I too have had to redesign what I thought were well done circuits.

I'm a big fan of point 2 point wiring. I also do it with transistors.

 

[CJ]

Grid wires high, plate wires low....

 

[Learner]

[quote author="gyraf"]Parallel PCB tracks is a minor problem in layouts, untill you reach frequencies much higher than what interests us here. Or if you possibly work with extremely high impedances and/or very high currents.

Jakob E.[/quote]


hey Jakob,
Just thought of 1 more thing about PCB layouts, since I am doing the LALA2 and it has got 4 tubes(heaters) which draws a bit more current than the usual DC bias. Could this possibly be a source of noise creepin into the circuit due to inductance generated by the current? OK, I guess it all depends on how much current and how close to the adjacent tracks. Therefore leads to the question, has anyone actually done any experiment or know what is the minimum distance and current should be when parallel to each other?

I hope I am not goin super anal about this am I? Just tryin to cut out some noise if possible.......

 

[Learner]

[quote author="cjenrick"]Grid wires high, plate wires low.... [/quote]


HA100X gimme gimme gimme................A10 gimme gimme gimme....................A24 gimme gimme gimme.....................................

 

[gyraf]

Still, it's an empirical science.

What works fine in one design may not work in another. I would recommend:

Keep heater tracks wide (1/20 to 1/10")

Keep heater tracks away from signal tracks - specially high-impedance points.

Use star ground on your layout, and attach heater chain here only.

You will have much less problems running DC heater than running AC heater if you build on a PCB. In my opinion, AC heating should be reserved for point-to-point assembled units.

Jakob E.

 

[BradAvenson]

If you use DC heaters let them have their own returns to the main ground input on the board. That keeps the heater current off the rest of your grounds. (ie signal grounds)

You also might think about wiring the heaters with a twisted pair just under the board. (still think DC would be best.)

 

[rafafredd]

Should we twist the wires when using DC for heaters?
:?:

 

[BradAvenson]

When you twist the wires the field generated from the current will cancel out better. You wouldn't have to, but it keeps the wiring neater.

One project I worked on had a low magnetic signature requirement, so all the power wiring had to be coaxial to cancel the magnetic field generated by the current.