So close and yet so far.... The "Heavy" preamp mess continues

When it comes to vinyl playback crosstalk isnt all that good ,but Id say it adds to the atmosphere on certain program material .Its one of the reasons a cd copy of the Beatles just doesnt sound the same as vinyl.

Maybe a good plan is to double up on the decoupling ,in other words whatever your dropper resistor is, divide it by 2 then add an extra RC stage ,same cap, half the resistance value ,so the voltage drop remains the same. Even in mono high gain systems this strategy decreases any tendency to motorboating due to feedback via the power supply ,capacitive coupling is trickier to deal with ,but less of an issue due to it tending to increase with frequency. 

Heres a pic from Pj Baxandalls 'Low cost high quality amplifier' to illustrate.
I might scan and post the entire document later as its a great primer for anyone getting into building and designing their own equipment .

Also note the table in this pic relating to the high pass section in the mic pre,

Didn't get to work on this at all last night, so no chance to come up with any actual crosstalk numbers.

What I do know is that it's worse than just your garden variety "inside the tube," or acceptable level of crosstalk.  And the fact that moving a couple wires made it almost disappear tells us it's a layout issue.

Then I got thinking more about using the 12DW7 for each channel.  While thinking about how to keep things far enough away from each other to prevent problems, it hit me.  All this time I've been assuming a single circuit board with both channels on it.  But, if I use 12DW7's, I don't have to have both channels on one board.  I could give each channel its own smaller board and place them on opposite sides of the case.  Two self-contained, mono preamps powered from a common supply.  (Just have to be careful about grounding...)  I already know the basic design works properly.  It's just when I get both channels going that I run into problems.  OK, so both channels don't get along living on the same board?  Separate 'em!

Sounds like a plan.

Cheers

Ian

ruffrecords said:

Sounds like a plan.

Cheers

Ian

Click to expand...

That's what Guy Fawkes said!  ;D

I'll start breadboarding this idea tonight.  Stay tuned!  In the meantime, a grounding question....
Assuming each channel's circuit board will have its own signal ground, would grounding both of them to the power supply ground work?  Individual ground wire from each board to the power supply ground?

CurtZHP said:

That's what Guy Fawkes said!  ;D

I'll start breadboarding this idea tonight.  Stay tuned!  In the meantime, a grounding question....
Assuming each channel's circuit board will have its own signal ground, would grounding both of them to the power supply ground work?  Individual ground wire from each board to the power supply ground?

Click to expand...

That should be fine. Make sure it is a nice thick one. Assuming the ins and outs are both balanced, that will be the only 0V wire from each board.

Cheers

ian

Got one channel of the new circuit perf-boarded up, just to test the signal flow.  If that works, then we're off in a whole new direction.

Might actually finish this Frankenstein's monster by summer!

Told my wife, if I ever do this again, she's to divorce me with extreme prejudice, until I'm picking through dumpsters for my daily bread.

Well, the perf-boarded prototype for the new audio circuit (12DW7) went very well.  Grounding and shielding could have been better, but overall it works!  (How the heck did that happen??) 

Just for giggles, I tried a different wiring idea with the existing circuit, and that made the crosstalk so bad you would have though I plugged the tone generator into the same channel I was listening to!

So that layout has been consigned to the scrap heap.  Now to clean up the new circuit and move forward.

Spent the weekend testing the new audio circuit with the 12DW7 tube.  Much better!

Started working on some potential PCB layouts for it, a couple based on the perf board layout.  Posting them here just for  ;D  ...

The layouts are based on putting two separate channels on a 3-inch by 5-inch PCB and cutting it down the middle.
Sample layouts 1 and 2 are set up using a tube socket soldered to the board.  3 and 4 have the tube attached using a 90-degree angle bracket at one end of the board.  (Gives a little more room and a little more pinout flexibility...)

I'm trying to avoid using jumpers, except maybe for ground runs.  Also, on 3 and 4, the filament wires go directly to the tube, so you won't see traces for them on the board.

I am guessing you plan to etch this boards yourself which is why it is single sided and both  channels are on one PCB but will be separately after the board is made.

For a mic pre, where noise and hum are especially important. I would recommend you consider having the PCBs made professionally. This is incredibly cheap. You can have five 3 by 3 inch boards made for just $18 including shipping and most importantly it means you can have a double sided board made. The reason this is important is it means you can have one entire side as a 0V analogue ground plane. I would also recommend you mount the mic transformer on the PCB so you can keep the wires from the secondary as short as possible. using a PCB solves a huge number of implementation problems in one go.

Cheers

Ian

ruffrecords said:

I would also recommend you mount the mic transformer on the PCB so you can keep the wires from the secondary as short as possible.

Click to expand...

The only problem there is that the transformer is a UTC LS-Series, so it's too big and heavy to mount to the PCB.  I could, however, probably get away with mounting the PCB to the transformer!  As for having them made, wouldn't I still have to provide some sort of layout?

CurtZHP said:

As for having them made, wouldn't I still have to provide some sort of layout?

Click to expand...

I must admit I assumed you were already using some sort of PCB layout program but maybe you are using a regular graphics program.

For PCBs you definitely need the right software but good programs for simple stuff are still available for free. They all produce manufacturing output in Gerber file format which you just send to the manufacturer in a zip file.

There is a learning curve of course and you will need to create footprints for unusual parts like tube sockets but it has got to be easier than doing it in Corel Draw or something similar.

Cheers

Ian

ruffrecords said:

I must admit I assumed you were already using some sort of PCB layout program but maybe you are using a regular graphics program.

For PCBs you definitely need the right software but good programs for simple stuff are still available for free. They all produce manufacturing output in Gerber file format which you just send to the manufacturer in a zip file.

There is a learning curve of course and you will need to create footprints for unusual parts like tube sockets but it has got to be easier than doing it in Corel Draw or something similar.

Cheers

Ian

Click to expand...

I'm using a program called ExpressPCB.  Free download.  I do my schematics in it too.  A little challenging when working with things like tube gear and other oddball components.  Had to do exactly as you said and make a few custom components along the way, but it's pretty slick just the same.  It does allow for up to four layers on a board, so doing double sided should be doable.  As for their cost to make them, I'd have to look into it.

I tried express PCB many moons ago. I found it quite easy to use. I did not like the idea of being tied to one manufacturer though. I cannot remember if it has a Gerber export capability. I then moved on to free PCB which I have been using ever since. Does not do schematics but the circuits are simple enough not to need to be able to import schematics which means you also do not have to create schematic symbols as well as PCB footprints for new components.

Cheers

Ian

What I like about ExpressPCB is that you can link the schematic to the PCB layout, so it sort of checks your work.  But you're right; it doesn't export Gerber files.

Meanwhile, back at the workbench....
I got the shielding issues sorted out with the new prototype.  No hum, no buzz, no squealing in my ear(!).  This thing sounds good!  Completely disassembled the original audio circuit and saved the parts for the final build.  Giant step forward! 

CurtZHP said:

What I like about ExpressPCB is that you can link the schematic to the PCB layout, so it sort of checks your work.  But you're right; it doesn't export Gerber files.

Click to expand...

I checked their web site. They will produce Gerbers for you for $60 !!!!!!!!

Cheers

Ian

ruffrecords said:

I checked their web site. They will produce Gerbers for you for $60 !!!!!!!!

Cheers

Ian

Click to expand...

Is that all?? 

Well,  it's been a while.....

But I've been very busy with this thing, and it's really coming together now.

Completely redesigned the audio channels, and created two completely separate circuits  for them.
Redid the power supply as well to improve ripple filtering all around and clean up the layout.  Also changed the physical layout of the whole thing to better accommodate its chassis (or, at least, what I hope to be its chassis).

Reassembled it on the bench this weekend and gave it the once-over.

It works great!  No hum, no buzz, no grunge.  No crosstalk! 

The only thing that seemed out of place, and this might not even be a problem....
With  no mic plugged in, the phantom power voltage reading from the XLR input is about 53-54 VDC.  With a mic plugged in, it drops to about 39VDC.  Tried this with two different condenser mics with the same result.  Both mics performed just fine otherwise.  I certainly would expect a voltage drop under load, but that seemed to me to be kind of too low.  Cause for concern, or is it a case of "if it ain't broke, don't fix it?"

I think the spec for phantom power is 44V to 52V so 39 is definitely out of spec. Is the supply regulated?

Cheers

Ian