Modifying a 12ax7 HiFi valve preamp to be a guitar preamp

[hoho]

One thing I have noticed during the measurements is that there is some delta between my 12.6VDC and 300VDC grounds in my power supply. But not as big (like 0.2, 0.3 VDC). I'll first try to decrease it to the proper 6.3 (maybe the valve is acting that way taking that much of voltage for the heater), and I'll connect the 300VDC and 6.3VDC grounds together. And I'll surely measure the resistance, and will clean the board better, though the joints don't really look bad to me... Speaking of the soldering — what is the best compound for doing it? I use something which is called Sn99,3Cu0,7 PBFREE.

[Gnobuddy]

Originally Posted by hoho

One thing I have noticed during the measurements is that there is some delta between my 12.6VDC and 300VDC grounds in my power supply. But not as big (like 0.2, 0.3 VDC).

I'm not surprised, and this is exactly why I don't like the PCB design that grounds the heater to PCB signal ground.

For preamp valves, heater currents are huge - hundreds of times bigger than anode currents in these valves - and all that heater current flowing through thin copper traces will inevitably drop a little voltage across them (Ohm's law, V = I R).

This isn't much of a problem if the heater current was perfectly noiseless DC. But such a thing doesn't exist. In practice, there will be some ripple and noise in the heater current, even if you're running it from a DC power supply. And that noisy current running through the PCB traces will generate a noise voltage across the length of the PCB trace.

The input triode can't tell noise voltage in its ground from noise voltage at its control grid - it's the difference between control grid voltage and cathode voltage that's amplified.

This means that the noise voltage along the length of the ground PCB trace will be injected into the input triodes, and amplified by them, just like any other input signal. So the whole amp will be much noisier than it should be.

This problem would be much worse if you'd been feeding the heaters with 60-Hz (or 50 Hz) AC, as the designers of the PCB intended. The bad design would inject hum as well as noise into the amps input, where it would be amplified and appear at the output.

So don't do the same thing with your custom PCB! It will be much better to return both heater grounds to a "star ground" at the power supply ground (Google will tell you more about what a star ground is.)

Originally Posted by hoho

I'll first try to decrease it to the proper 6.3

Originally Posted by hoho

(maybe the valve is acting that way taking that much of voltage for the heater)

Your measured data showed more than one volt of DC on each of the four control grids. I can't think of a way that heater over-voltage can cause that symptom, but leaving the control grids floating (bad resistors or bad solder joints) could conceivably cause it (though I have never seen this before!)

Your ohmmeter will answer this quickly. Just measure resistance from each of the control grids (pins 2 & 7 of each tube) to ground. You should read the same, +/- 5% or so, as the corresponding grid bias resistor (1M or 470k if that's what you used.)

Originally Posted by hoho

...I'll connect the 300VDC and 6.3VDC grounds together...

The schematic provided by the Ebay vendor from whom you bought the power supply, shows the 300V and 6.3 V power supplies are already grounded together, on the power supply PCB.

If you add an additional wire connecting these same two points together, you will create a ground loop, which is a bad thing: WTF Are Ground Loops? | Hackaday

I've attached a much more informative write-up on ground-loops from an MIT professor.

Originally Posted by hoho

And I'll surely measure the resistance

Please do!

Originally Posted by hoho

Speaking of the soldering — what is the best compound for doing it? I use something which is called Sn99,3Cu0,7 PBFREE.

What you're using is "lead free" solder. It's mandated by law (ROHS regulations) in just about every commercially manufactured product these days, because lead is a poisonous metal.

However: lead-free solder is much harder to work with at home. Without getting too technical, the lead-free stuff is very unforgiving of the slightest movement between the parts you're soldering, it melts at much higher temperature (damaging components), and it is very easy to make bad (dry) solder joints with it.

Commercially, lead-free soldering is done with accurately computer-temperature-controlled soldering ovens, with a careful ramp-up to exactly the right temperature, and a careful ramp-down back to room temperature.

At home, with a soldering iron in your hand, we cannot recreate such an exact soldering temperature and accurate temperature/time profile. And the result, very often, is poor-quality solder joints.

If you can get hold of it, 60/40 (tin/lead) solder is much, much easier to work with. Here in Canada, I can still buy this for hobby work. I don't know if you can still legally buy it in your country.

There is also 63/37 tin/lead solder, which is a little harder to work with than 60/40, but much easier to work with than the lead-free stuff.

Here is a You Tube video that will tell you more about the problems of using lead-free solder: YouTube


-Gnobuddy

[Gnobuddy]

The handout (pdf) on ground loops is rejected by the forum software as too big (file size). If you Google for "EST016_Ground_Loops_handout.pdf", you'll find it.

-Gnobuddy

[PRR]

"EST016_Ground_Loops_handout.pdf"
43 pages, 4MB
https://web.mit.edu/jhawk/tmp/p/EST0...ps_handout.pdf

[Gnobuddy]

That's it!


-Gnobuddy

[Gnobuddy]

By the way: I bypassed the grounded heater centre-tap problem entirely by deciding to use two Russian 6N2P valves in my PCB.

These valves are essentially identical to a 12AX7 as far as the guitar signal is concerned, but the heater is different. 6N2Ps have a 6.3V heater wired between pin 4 & 5. The heater has no connection to pin 9 - instead, pin 9 connects to an internal electrostatic shield, which is intended to be grounded.

This neatly solves the problem with grounded heaters in this particular PCB. Pin 9 is grounded, which is exactly what we want for a 6N2P. I supply 6.3V heater power (instead of 12.6 V) to the PCB, which routs it to pins 4 & 5, also exactly what we want!

I've added a pair of 1k resistors across the heater power, and their junction connects to power supply ground, to keep the heaters from floating off to some unwanted arbitrary voltage. The resistors are big enough to keep any substantial heater current from flowing through the PCB traces (there are only about 3 mA of current flowing through those 1k resistors.)

So far, I've only powered my board up long enough to test DC voltages on all pins. All of them seem to be okay, so the next step is to start adding the missing bits (input and output sockets, gain, volume, and tone controls, a channel-switching relay, et cetera.)


-Gnobuddy

[hoho]

I have just got it working!

I have dropped the voltage to 6.3. Connected the board's ground to my 12V supply's ground (so the high voltage supply's output ground is lifted, and the common ground is the one which comes from 220->12 supply). Resoldered the pots (the multiple holes under the caps which as I suppose are designed to accommodate various kinds of capacitors are the friends here — no need to drill the board, just cut the trace between the resistor and the capacitor and use one of those holes).

Now my main problem is that I've sent my Les Paul Junior for the maintenance and the luthier is really busy so that I will get it back in 5 weeks only... Tested with the harmonica microphone a bit, but everybody is sleeping here now so that I can't properly check the sound .

And there is another problem with the pots. I might have overheated them during the resoldering or something else, but they seem to be semibroken (like sometimes you have to hit it a little so that it starts working). I have these: 16mm Variable Resistor Potentiometer Pot Lin Linear Log Logarithmic | eBay (A500K), should they be sufficient? Maybe the wattage should be different.

As for the soldering compound — seems like there is Sn60Pb38Cu2 still available in some stores in Czech Republic. The stores say «for professional use only» — I will see if I will be able to get a little for the future work.

Anyway, I feel much more confident now that my custom board should actually work!

Gnobuddy, thank you so very much for all your help!

[Printer2]

Those are cheap pots, not that I do not have a few my self. I wonder if factory seconds get hawked on EBay, I have had a few where the wiper loses contact with the resistive track, broken connection between the lug and the track, something else that I can't remember. More mechanical issues.

[Gnobuddy]

Originally Posted by hoho

I have just got it working!

Congratulations!

Originally Posted by hoho

...the multiple holes under the caps...are the friends here — no need to drill the board, just cut the trace between the resistor and the capacitor and use one of those holes.

Exactly! (I mentioned this back in post #19, page 2.)

You do have to remember to cut the trace on both sides of the board, since the holes are plated-through-hole. I just use a suitably sized drill-bit held between finger and thumb to cut away the copper around one of the through-holes.

Originally Posted by hoho

...my Les Paul Junior...5 weeks...

Aw, man! That's a long time to have to wait!

Originally Posted by hoho

I have these: 16mm Variable Resistor Potentiometer Pot Lin Linear Log Logarithmic | eBay (A500K), should they be sufficient?

I have used pots just like that in some projects, and haven't had any problems with them. They don't feel wonderful to turn, though, so I prefer Alpha pots (from Taiwan, but carried by many electronics suppliers, such as Mouser): alpha potentiometers Alpha (Taiwan) Passive Components | Mouser Canada

In my experience, Alpha pots have a very nice smooth, precise feel when you turn them, and the cost is not much more than the very cheapest pots, so I use them when I can.

Originally Posted by hoho

Maybe the wattage should be different.

The wattage is probably fine. You may have up to 200 volts peak-to-peak (about 70 volts RMS) from your preamp, but the pots are 470,000 ohms. Power dissipated in the pot (P = V*V/R) is therefore a maximum of 0.010 watts - barely one-hundredth of one watt!

For our purposes (valve preamps), the voltage rating of the pot is more likely to be an issue. The pot could have up to 100 volts (peak) across it at any one instant, and it would be best to use bigger pots that can handle this much voltage with some safety margin.

As an example, the first Alpha pot in that list (Alpha part number RV16AF-10-20R1-B25K-LA) is rated for up to 500 volts.

Originally Posted by hoho

...Sn60Pb38Cu2...

Two percent copper, huh? I'm not familiar with this type of solder. I've never seen it before. But I think the copper is there to prolong the life of the (copper) soldering iron tip. Other than that, I believe it works pretty much exactly like 60/40 solder (Sn 60 / Pb 40).

When I began tinkering with electronics as a small boy, my cheap soldering iron tips used to be made of pure copper, and this material would actually dissolve slowly into the hot solder as you used the iron, so the tip would develop a pit. Every so often, you had to file the tip of the iron to remove the pit, and eventually, the filing would shrink the tip to the point where you had to buy a new tip (or a new soldering iron, because the old tip was corroded solid, and would not come out of the iron!)

In more recent years, all the soldering irons I've used have a copper tip that's iron-plated. The iron plating keeps the tip from dissolving into the solder. As long as you clean the tip gently (and do not sandpaper through the thin iron plating!), the tip doesn't pit.

I think if you have an iron-plated tip, you don't need the 2% copper in the solder. But if that's the only kind of leaded solder you can easily get, go for it. There's no real downside.

Originally Posted by hoho

Gnobuddy, thank you so very much for all your help!

You're very welcome! Congratulations on sticking with it until you got it to work, great job!


-Gnobuddy

[hoho]

Thanks for the pots suggestions! I've done some searching and have ordered a few Bourns which seem to be the opposite of cheap, but not that expensive given that I only need a few. I've favoured Bourns before Alpha because I will also be trying to add that overdrive circuit I've mentioned before, it requires A50K and B100K, and Bourns were the ones I've found A500K, A50K and B100K in the same form-factor (for the consistency of the future boxing).

Next step is to add the Zeners/Resistors part to my current board to see if it works properly with my ADC/DAC part. Right now, when the volume is bigger than something, I get the noise in both of my ADC channels (regardless of the fact that I currently have only one channel connected). Which is of course likely because the level my ADC is capable of taking is less than it takes.

If everything goes well and nothing is lost during the delivery, I will report my progress in a few months (sooner or later if something goes wrong ).

And once again — thanks to everybody who has commented on this thread!

One more observation: making a thing yourself gives an interesting imaginary feeling — apart from feeling the instrument, you start to feel it going through the resistors, the heaters and the glass of the valve . The level of intimacy between you and the sound you make increases .