Tube supply question

[rankot]

I want to build this supply for tube preamp / SS amp, but I can't buy transformer with additional 6.3V secondary for heater, so I thought that I could make that supply using two separate transformers, one for +/- voltage for pre/amp, and another one for heater supply, like this:



I have two questions for experienced builders:
1. Is it OK to use common ground, like this?
2. Is it better to put transformer primaries in parallel, as shown on picture, or I should put them in series? I presume parallel is better.

[merlinb]

1. Is it OK to use common ground, like this?

Yes. Ideally you should treat them separately before finally connecting the heater ground to the other ground, at one point.

2. Is it better to put transformer primaries in parallel, or in series? I presume parallel is better.

Yes, parallel is the only way this would work.

[rankot]

Thanks a lot, Merlin! 

[EBK]

Diy-tubes.com sells some transformers that would probably work for you if you want to go a little less roll-your-own.  I just bought some stuff from them to make my very first tube supply, so looks like I'm high voltage adventuring along with you. 

(Safety first, of course.  )

[antonis]

You also may try to use unrectified 6.3V for filament supply, connecting each side of it with 100R resistor to a point created by a resistive voltage divider from screen grid supply (if you use a pentode on output stage) or from preamp plate supply with a 22μF cap, to avoid hum or any other noise..
(I can't post a schematic for the moment but if you find a AX84 power supply diagram it will be all clear..)

You actually "off-set" ground reference point to a DC one, avoiding hum caused by filament "e-field" emission due to unrectified supply..

Just for educational/fun purpose.. 

[rankot]

I know all that, but I live in Serbia and it is not practical due to shipping time to order transformers from abroad, not to mention shipping costs

However, I am asking all this because I have built this preamp already and it works fine when using it's own, separate PS with step up from 9V to 50V for B+, but when I connect it to amplifiers PS (B+ directly to +50V rail, and heater to the same rail as shown here), I experience huge hum. I can't use it's separate PS inside main amp (no place to put that inside), so I need to use amp's PS.

So I presumed that the problem is with heater, it draws too much (and dropping resistor also waste too much power) and I wanted to provide separate power for it. And it is already connected to common ground, so I asked if it is OK, because if it is not, I can also fix it, but it means that I have to desolder tube socket and solder heater supply to appropriate pins using separate cable.

This is how it is connected right now:

[EBK]

I know all that, but I live in Serbia and it is not practical due to shipping time to order transformers from abroad, not to mention shipping costs

Got it.    I'm in the US, so I took a deep breath before clicking to see what the shipping costs from Russia were, and I feel your concerns.  I ended up paying $15USD on a $30USD order.

I'm mentally adding Serbia to my list of countries that I am aware of at least one forum member designating as their home.  To me, this long list is one of the very best things about this community.  I have to admit that I misread it as Siberia at first though.   

[rankot]

You are definitely not the only one person confusing Serbia and Siberia

It's nice meeting people from all around the globe here, cause music is the ultimate common language. I understood that for the first time while watching first videos of "Playing for change". It really touched me!

[rankot]

You also may try to use unrectified 6.3V for filament supply, connecting each side of it with 100R resistor to a point created by a resistive voltage divider from screen grid supply (if you use a pentode on output stage) or from preamp plate supply with a 22μF cap, to avoid hum or any other noise..
(I can't post a schematic for the moment but if you find a AX84 power supply diagram it will be all clear..)

You actually "off-set" ground reference point to a DC one, avoiding hum caused by filament "e-field" emission due to unrectified supply..

Just for educational/fun purpose.. 

There's one: http://ax84.rru.com/media/ax84_m116.gif

But I am not in the mood to use this, because I am not sure if using that kind of common ground will affect solid state amp behind my preamp

I'll try to use the 1st schematic I posted, with separate transformer and rectifier.

[rankot]

Well, I have posted this video if one has will to watch it - and maybe help find the solution for the problem.

I have noticed that turning volume potentiometer affects the hum, and this can be shown on video. This is the schematic I used to build this (heater supply not shown, since it is already posted):



I did two experiments: when I plug another preamp into override input (this preamp still running, but disconnected from amp), there is a hum, but lower in intensity. I have also tried to desolder heater supply and then there is no hum at all, and of course not audio through this preamp

[thermionix]

Irrelavent here, but C2 is unnecessary.

[rankot]

Indeed it is, but I usually put one high voltage capacitor there to block DC when working with high B+, so I don't have to use expensive high voltage caps for the tone stack. I forgot that this one works on 50V only

[Rob Strand]

This is how it is connected right now:

I'm assuming you have it like this right now.

You need to connect the ground side of the heater to the junction where all of C1, C2, C3, C4 join together.

It is normal for the a volume control at the end of the circuit to increase the hum.  The hum is a fixed level within the circuit and you are just controlling the volume of it like you would the level of the guitar.

[rankot]

Thanks Rob, but what puzzles me is the fact that hum is almost equally strong on both sides of volume pot, while there is a certain decrease at approx. 10 o'clock.

This is how my PCB looks like:


Red GND is the place where I brought ground from junction of C1, C2, C3 and C4. Shall I add another wire (short and thick) from that place to heater ground?

[antonis]

You also may try to use unrectified 6.3V for filament supply, connecting each side of it with 100R resistor to a point created by a resistive voltage divider from screen grid supply (if you use a pentode on output stage) or from preamp plate supply with a 22μF cap, to avoid hum or any other noise..
(I can't post a schematic for the moment but if you find a AX84 power supply diagram it will be all clear..)

You actually "off-set" ground reference point to a DC one, avoiding hum caused by filament "e-field" emission due to unrectified supply..

Just for educational/fun purpose.. 

There's one: http://ax84.rru.com/media/ax84_m116.gif

That's the "wrong" one..
(most common "humming" fault in many guitar tube amps..)

I'm talking about the following way of connecting filament supply "GND"..
(you can see the difference between filament grounding through 100R resistors in your scheme and "offseting" AC ground to about 50VDC in the following scheme..)

[Rob Strand]

Thanks Rob, but what puzzles me is the fact that hum is almost equally strong on both sides of volume pot, while there is a certain decrease at approx. 10 o'clock.

Things like that are often a sign of multiple ground loop issues
It's probably cause by:
- the fact you have three ground points on your PCB
- the pcb connects to the power amplifier in your amp, and the preamp and the power amp share the same power supply.   
- Your filament current is also flowing through the same grounds as you signal.
- Possibly even more ground connections through sockets.

There's ground connections all over the place. Ideally you want to use a star ground connection.

http://www.geofex.com/Article_Folders/stargnd/stargnd.htm

You have to solve the hum with the volume pot at minimum first because that problem exists even without the preamp signals.

You might you need to complete separate the filament supply on your PCB (by cutting the grounds for those away from the ground plane)  then running them back to a star point at the junction of the big caps.    There's also a right way to connect to the caps:  draw a loop from the transformer through the rectifier diodes then through the big caps through some ground tracks then back to the transformer.    Any ground wires in that loop are considered noisy.  You need to connect to a ground point just outside of that loop.

Try to look how the amp handled the grounds and connections between the preamp an power amp on the existing preamp.

[Rob Strand]

Here's a good picture to follow:
https://media.tumblr.com/tumblr_m8scigy0fL1qascvc.png

Also note how the start point isn't right at the junction of the filter caps.

[rankot]

OK, Antonis, this is much clearer, but I can't use it here. I am upgrading existing solid state amp with tube preamp, so I must rely on it's existing PS.

What puzzles me is the fact that this preamp worked fine, no hum at all, with it's own PS, which was made of 9V/2A DC as the main PS, which later powered heater through 7805 (lifted to 6.3V with two diodes) and B+ was generated with 555 based boost circuit. And it was even built on a breadboard, just plugged to amplifier's direct input (original preamp overridden)! The ground was common for signal, power and heater, and also connected with amplifier's ground through the guitar cable. And it worked???

Now when I assembled it on PCB and connected to amplifier's PS, hum is horrific! I presume that + side of amp's PS is overloaded and that creates instability and oscillation, because when I measure voltage, +rail is much lower than -rail in absolute value, when I connect this preamp. They are practically the same when preamp is not connected. It seems that there's not enough power on positive rail to run all this.

I will try to make a separate, rectified heater supply, as shown on the first picture. It will share the ground with the others (so I don't have to alter PCB too much), and if there would be hum again, I will modify the PCB so to have separate ground for heater and everything else.

Regarding ground points (what Rob mentioned) - I actually use only one (marked red), and GP near input is for guitar jack connector, so it's not connected to the rest of the circuit in any other way than through the 'red' ground. No loop. I don't use ground point at output, but I had to make it because connectors I have are double, I don't have single.

[rankot]

This is how it is connected right now:



Wires on the left are original ones - red is positive rail, black is negative rail and blue is ground. This is the star point of the amp. Wires on the right are mine (OK, I know I connected colours incorrectly, but they go to right places on my PCB), and they go to the connector on PCB. Yellow wire goes from this star point to red marked ground point on preamp PCB.

[Rob Strand]

All the weird behaviour you are seeing is quite normal when you have grounding issues.   Separate supplies naturally decouple everything and avoid many issues.

I will try to make a separate, rectified heater supply, as shown on the first picture. It will share the ground with the others (so I don't have to alter PCB too much), and if there would be hum again, I will modify the PCB so to have separate ground for heater and everything else.

That's a good place to start.