6J7 preamp

[Shoog]

You could try raising the bias of your heater supply to +20V. Take a voltage divider from your B+ and connect it to the midpoint of your heater resistors.
+B
470K
Heater reference
33K
Ground.

PS- It would be entirely expected that the hum would die away instantly if it is been caused by the PSU - once the mains is disconnected there is no more source of hum to bleed into your signal.

Shoog

 

[emk2]

OK, ripple is almost flat so that can be ruled out.
No microphonic tubes, so we can also rule out picking up of transformer resonance.
- Where is your transformer located? Is it radiating into a sensitive part of the amp?

Last option before I throw in the towel is to check for HF resonance. Did you put a resistor in series with your audio signal, close to the grid?

Transformer (red arrow):

https://dl.dropbox.com/u/425350/6J7%20Preamp/IMG_1634s.JPG

- it's a toroid with multiple secondaries (wires cut and isolated, but certainly "touch" the bolt)
- as you can see, diodes and chokes are not very far of it.
- top toroid cover home made with thin steel, not in contact with chassis

About the grid resistor, it is located near the pot.. I did a mistake, I'll solder it to the top cap. 2.4K better than 1K ?

I can try to remove the toroid and connect an EI transformer with 250V-0-250V secondary, too much voltage but I have a variac, although the heaters won't have the correct voltage..

 

[Shoog]

Toroidal radiates far less magnetics than an EI so is unlikely to be the cause - especially in its current location. Do not alter.
Since the grid wire is shielded the grid should not be picking up radio interference in its short run to the top cap. You can try introducing a 1K resistor at the top cap to see if it helps - but make this reversable as I doubt it will help. Do not disconnect the other resistor until you try this first.

Shoog

 

[emk2]

OK, I don't remove the toroidal.

- soldered a 1K resistor on each top cap: still hum.
- heater elevation (18V) with 30K & 420K: still hum.
- 3.3uF in parallel with ground and the 30K resistor: still hum.

I recorded the noise:

https://dl.dropbox.com/u/425350/6J7%20Preamp/Untitled.wav

After 5 seconds, the hum change, it's when I touch the ceramic top cap.. after 10 seconds I don't touch it anymore.

 

[Shoog]

Reinstate your output transformers and see if that reduces the hum to acceptable levels.

Shoog

 

[emk2]

hum level is acceptable with 13:1 reduction (enough gain) and with a small 15W amplifier, I wonder if it will be the case with the 807 PP.

I ordered Svetlana 6J7 and GE 6J7 to see if the problem comes from the tube itself..

I'll try again with the 10K pot, I don't understand why but sound is now flat, "dull", bass is strange. Source is a vdac (100 ohms impedance). Tried to invert speakers cables, no effect.

 

[Wavebourn]

I see from your picture you have ground loop problems.

You need to resolder your grounds. You have to remember that any wire has resistance and inductance. You have many capacitors, but take ground to power the amp from the first one, instead of from the last one. It must go from the rectifier to the first one, then to the next one, and so on, and the real clean "ground" should be at the last one in the chain. Here is my example: lots of thin wires, but absolutely no problems with hum and undesired feedbacks through ground or power. Think about currents! Pulses that charge caps create tiny voltage drops on wires, that being amplified cause loud hum!

http://wavebourn.com/forum/download.php?id=508&f=7

 

[emk2]

Thank you for your advice.

Maybe in ten years I'll be able to make a circuit like yours, without hum

Could you please tell me if I understood correctly the modifications ?

https://dl.dropbox.com/u/425350/6J7%20Preamp/IMG_1634s.JPG

Sorry, I am a beginner and not really gifted in electronics..

 

[dhtrob]

Correct me if I am mistaken, but if the hum is 50hz as 6J7 mentioned, wouldn't the ac-heating be a point of attention? Hum from powersupply or groundloops is often more like 100hz or like a buzz..
I used 6J7 as driver in a power-amp, no experiences in a preamp..

Rob

 

[emk2]

Hi,

in fact hum is not a "clean" 50 Hz, seems to be mixed with other frequencies, probably also 100 Hz..

I am waiting for other 6J7's to see if some are quieter.

Here is a modification:

https://dl.dropbox.com/u/425350/6J7%20Preamp/IMG_1635.JPG

- ground bar starts only from the first C1 near rectifier
- C2 (4 x 51uF in //) with the first cap ground connected to C1
- last C3 cap ground goes to the starground

Small improvement.. but still hum.

- something wrong with this modification ?
- wrong heater wiring implementation ?

I tried to add capacity to C2 (350uF instead of 200uF) with electrolytics, not better.

If you still have the patience to help me.. it's very appreciated.

 

[dhtrob]

I do agree on Wavebourn: You have an interesting grounding topology..

It could be that you have a ground loop over the mains. Disconnecting the mains earth from the chassis, does that make any difference? (for testing only! I'd like to point out that you should connect your metal chassis to mains-earth.. lifting your signal earth from chassis is better definitive option).

Rob

 

[emk2]

maybe entirely remove the bus bar and connect all grounds to starground ?

create a second starground for C1,C2,C3 and a straight wire between the two ?

I'll try to remove temporarily earth connection.

lifting your signal earth from chassis is better definitive option

?

not using the chassis as ground ?

 

[Wavebourn]

I do agree on Wavebourn: You have an interesting grounding topology..

It could be that you have a ground loop over the mains. Disconnecting the mains earth from the chassis, does that make any difference? (for testing only! I'd like to point out that you should connect your metal chassis to mains-earth.. lifting your signal earth from chassis is better definitive option).

I did not mean mains grounding, I meant B+ filtering.

 

[Shoog]

Personally I would bring all the power supply grounds to one central point and then take the small signal ground to that point before onward connection to safety earth.

Shoog

 

[Shoog]

maybe entirely remove the bus bar and connect all grounds to starground ?

create a second starground for C1,C2,C3 and a straight wire between the two ?

I'll try to remove temporarily earth connection.



?

not using the chassis as ground ?

Exactly how I would do it. At this stage you have nothing to lose.

Shoog

 

[Wavebourn]

Personally I would bring ground from rectifier to first filtering capacitor. From first filtering capacitor to second filtering capacitor, From second to the third one. And the ground near the last one would use as a "star ground" for everything else. No matter how many rectifiers I have, their own grounds go by a single wire from first to last filtering cap, and grounds of last filtering caps of all different voltage sources are interconnected by shortest wires, in the "star" ground that is close to all last capacitors. Here will go "grounds" from tube stages, from chassis, from mains socket.

 

[DF96]

Yes, always ground a PSU at the clean end not the dirty end. People often get this wrong then wonder why they have some buzz.

 

[Wavebourn]

Yes, always ground a PSU at the clean end not the dirty end. People often get this wrong then wonder why they have some buzz.

Because they assume that "ground" wire has zero resistance and don't understand impact of voltage drops on wires caused by currents that flow through them.

 

[iko]

Personally I would bring ground from rectifier to first filtering capacitor. From first filtering capacitor to second filtering capacitor, From second to the third one. And the ground near the last one would use as a "star ground" for everything else. No matter how many rectifiers I have, their own grounds go by a single wire from first to last filtering cap, and grounds of last filtering caps of all different voltage sources are interconnected by shortest wires, in the "star" ground that is close to all last capacitors. Here will go "grounds" from tube stages, from chassis, from mains socket.

Anatoliy, it's a bit hard to imagine. Can you show us a simple drawing of how you do this?

 

[Wavebourn]

Anatoliy, it's a bit hard to imagine. Can you show us a simple drawing of how you do this?

Follow how it is drawn on schematic. Wires from transformer go to rectifier, from rectifier to the first filter cap, from filter cap to the next filter cap, and when everything is already filtered, on the last cap, is the cleanest point to use. I.e. cleaner and cleaner, with each filtering cap.

Then take clean, filtered voltage, from the last cap. Both ground and B+

If you have more than one rectifier, do the same with them, and interconnect grounds of each cleanest end.