Excessive hum

[loudthud]

It is impossible to predict the results, but you should attempt to modify the PCB by cutting tracks and adding wires. Since I suspect English is not your native language, I will try to define some words. If you do not understand, please say so and I will try to make things clear for you.

PCB: An abreviation for Printed Circuit Board.

Crosstalk: Is an electrical phenomenon where a signal or noise from one wire leaks into another wire in close proximity even though there is no direct connection between them.

A PCB has two sides. One side is where you mount the components. This is called the Component Side. The other side is where the components are soldered to the tracks. It is called the Solder Side.

You will have to cut some of the copper tracks on the Solder side of the PCB. You will need to use a sharp knife. I have marked some of the places to make cuts on the attached photo. They are marked with a red X. You should actually make two parallel cuts about 1 millimeter apart and carefull remove the copper between the cuts to insure that there is no connection.

Because I cannot see all the tracks, you will have to determine where to make some of the cuts. There will be some copper tracks on the board that will not be connected to anything. This is ok. ("ok" means acceptable.)

The objective is to disconnect the heater voltage (6.3VAC) tracks around V1 and the replace the tracks with a twisted pair of wires from the connector block (where the heater voltage is connected to the transformer) to V1. Then another twisted pair of wires will connect the heater voltage from V1 to V2. It should not be necessary to replace the tracks between V2 and V3/V4. The circuits there are less sensitive to crosstalk.

Pins 4 and 5 of each 12AX7 should be connected together directly between the socket terminals. If there is a track routed between pins 4 and 5 that is connected to a resistor or capacitor, you should cut the tracks connected to pins 4 and 5 and connect them together with a short wire.

There are two green wires that I have marked to be removed. I think these are connected to the 6.3VAC. If they are not, then do not remove them.

You will need to determine what tracks are necessary to connect the 6.3VAC between V2 and V3/V4. Cut any tracks that may be connected to V2's heater pins but are branches that don't carry current to V3/V4.

Once the tracks are cut, solder a twisted pair of wires on the Solder side of the PCB from the connector block to V1, and then another twisted pair from V1 to V2. You do not need to observe polarity of the wires. The wires should not lay against the PCB but should be bent so they are suspended in air about 5mm away from the PCB until they are bent down and soldered to the socket pins of the tube socket.

Once the connections are complete and before you apply power, remove all the tubes, disconnect the 6.3VAC wires from the transformer and measure the resistance between the two screws at the connector block. You should measure an open circuit. If not, you have created a short that could damage your transformer if you apply power. You must eliminate the short before you apply power.

I have also marked a track to cut near R1. Solder a shielded coax cable on the solder side of the PCB and use that to connect to J1. Be sure not to create a short to other tracks on the PCB when you solder the coax. If you cannot obtain suitable coax cable, use a twisted pair but you will have to observe polarity when you solder to J1.

Grounds: The input jack in the picture looks like it is the type that connects the "sleeve" to the chassis. The speaker jack looks like the same type but may have a black insulating washer. If both of these jacks have the ground terminal connected to the chassis, a ground loop is created that may cause noise or oscillation. One of these jacks should be insulated from that chassis with plastic washers or you can use a plastic jack. There should be a connection to the chassis near the connector for the power cord for safety reasons.

Pilot light: It looks like you are using a 6.3VAC pilot lamp. Make sure that neither terminal of the lamp socket is connected to the chassis. Is there a connector on the PCB for connecting the 6.3VAC to the pilot lamp? Do not run wires from the left side of the PCB past the pot to connect to the pilot lamp. It would be better to solder a twisted pair of wires to the 6.3VAC at V3 of V4 to connect to the pilot lamp.

Power Switch: Your power switch does not have 4 terminals like the one on the schematic (S1A, S1B). It probably violates safety rules in your country to use a two terminal switch. Your instructor may deduct your grade if you do not comply with safety regulations. Route the wires between the power cord connector, the fuse, the power switch and the transformer along the right side of the chassis, and across the back side of the chassis. Do not route they past the input jack and the pots. Use a twisted pair of wires that are rated to safely handle the line voltage.

What you have learned: Crosstalk on PCBs must be taken into account during layout. You can not simply place components at random and "connect the dots".

[Rebel]

sorry for may bad English but are more than 5 years that i don't speak or write in English.

Thank you for detailed instructions.I followed them but there is still noise.

I have cut and removed heaters circuits and green jumpres and i have replaced them with twisted wires. i have connected 6,3V to V1AB, from V1AB(4,5,9) to V2AB(4,5,9), from V2AB(4,5,9) to V3(4,5), and from V3(4,5) to V2(5,4).

I have not yet connected R1 and R2 with shielded coax cable as you described to me. I used shielded coax cable from the input jack to pcb input connectors. is it very important use your configuration? i will use it tomorrow.

Input jack is connected to ground and to the chassis and output jack is insulated.
I have connected to the ground also power and output transformers.

Pilot light is 6,3V lamp. neither terminal of the lamp socket is connected to the chassis. I connect it with twisted wires from the 6,3V connectors on the pcb (blue wires on the left) they not pass around pots.

My power switch does not have 4 terminals it has only 2 terminals because there was an error in the shipping of this pieces from Germany, soon i will receive 4 termnals switch.


this is the circuit now.


Tomorrow i will remove power circuit and mount it on another pcb , could it be a solution to reduce noise?

[loudthud]

At this time it is important to know where the hum is entering the signal path. My concern with cutting the track at R1 was that it runs next to one side of the heater track and underneath R36 and R35. This is a very sensitive track because it has the most Voltage Gain to the output.

As an experiment, install a short at the input connector. Does that eliminate the hum?

As an additional experiment, remove V1 from it's socket. Does that eliminate the hum?

Another experiment to try would be to connect an external 6.3V DC power supply to the heater connector in place of the transformer. The external power supply should be rated at 2.2 amps or higher. A battery could also be used but should be capable of supplying at least 2 amps. When power is applied to cold tubes, they will draw about 4 amps for a second or two until the tubes begin to warm up. This is normal. After about 15 seconds, turn on the switch to the transformer and listen to the amplifier. Is the hum now at an acceptable level?

[Rebel]

I tried to use DC voltage at the university with a dc power supply at 6V 2.5A but the noise was more loud.

I cut the track from J1 to R1 and i connect a shield coax cable as you suggest to me (i didn't connect the ground from R2 to J1) and the noise is decreased, but is still here.

when the jack is not conected i have a low noise on the speaker, when i put a short at the input connector to the ground or i attack the jack of the guitar the buzz increase
i think it is not normal. it is possible that i have a ground loop?
I take the ground from the power line and put it on the iron chassy on a screw of power transfoermer, the circuit get ground from this point; transformers get ground from the chassy; power switch, volume and tone pots, input jack take ground from their bodies connected to the chassy; output jack is not connected to ground.

now i have moved the power out from the case, i used a 1000uf 450V to filter ripple but the noise is still here so i think the hum does not depend on the power. Now i have no idea about what i have to do.

I removed V1 but i can hear the noise.

[loudthud]

I tried to use DC voltage at the university with a dc power supply at 6V 2.5A but the noise was more loud.

I removed V1 but i can hear the noise.

These things mean that the problem is more serious than just heater hum leaking into the signal path due to crosstalk on the PCB. Try the amp with V1 and V2 removed. Please measure the hum at the speaker jack with a voltmeter and post the readings.

Are you sure it's 50Hz and not 100Hz?

You never answered that question. If you don't know how to make that measurement, ask one of your instructors. This is important information!

output jack is not connected to ground.

The sleeve of the output jack should be connected to ground on the PCB. Verify by measuring the resistance from the jack to the chassis. It should not be grounded by touching the chassis.

it is possible that i have a ground loop?

Yes, you have a ground loop but most of the hum should be reduced when V1 is removed. There may be components on the PCB that are not connected to ground. You should verify all the points that should be grounded measure a low resistance to the chassis.

I cut the track from J1 to R1 and i connect a shield coax cable as you suggest to me (i didn't connect the ground from R2 to J1) and the noise is decreased, but is still here.

That is one of your worst ground loops. You are connecting the noisest ground to the most sensitive circuit.

I cannot see all the tracks on the PCB from your pictures. Do you have any kind of graphic that shows the tracks and the component locations with the circuit designators (R1, C1, et al)? I would need this to show you how to modify the grounds.

On my profile, http://music-electronics-forum.com/users/150/ there are links to threads where I have posted schematics and a layout of how ground should be connected. The general idea is to follow the signal flow of the schematic from the input back to the power supply. On your PCB I can see where the grounds of C15 and C16 are connected almost directly to the input connector. This is not good.

now i have moved the power out from the case, i used a 1000uf 450V to filter ripple but the noise is still here

This is very dangerous. Serious injury or death could result. Do not do this!

Edit: Since there might be some build error that is causing the hum, please do the following:

Remove all the tubes and turn the power on. Measure the DC voltage to ground at each tube socket pin and post the results. On the heater voltage pins (4, 5 and 9 of V1 and V2),(4 and 5 of V3 and V4) measure the AC voltage. Also list the voltages at TP24, TP25 and TP26. This will help diagnose any as yet undiscovered problem.

[Rebel]

These things mean that the problem is more serious than just heater hum leaking into the signal path due to crosstalk on the PCB. Try the amp with V1 and V2 removed. Please measure the hum at the speaker jack with a voltmeter and post the readings.

I have removed V1 and V2 the noise is very very low on the scope i have this signal :

the scope was set on : 20mV per division and 5ms per division

when i use V1 and V2 at 0 volume and 0 tone i have this

the scope wa set on : 100mv per division and 5ms per division
i think the noise i of 50Hz becuase the signal is not definite but has peaks on 50hz.

if i connect the ground to the output jack i have 100hz noise (i think)

the scope is set on 100mv per division and 5ms per division

this is the origianl schematic of the pcb


i have modified it removing the tone control (because it make a ground loop) and i put it on pots:


now when i put a short on input jack the noise go down (a little).

the circuit now is it on the top


and this is on the bottom(R33 and r34 are R36 and R35)


this is noise at 0Volume and 0 tone at 20cm far the speaker:

http://www.easy-share.com/1918205060...at 0volume.WAV

[loudthud]

Note: When I refer to a component, I am using the name (that is C3, R4 etc.) from the Fender schematic. Some of the names on the layout do not agree with that schematic.

There is a note on the schematic that says to change R29 to 18K for 50Hz operation. Make sure that TP29 measures about -10.4VDC.

C5 and R27 (using the designators from the Fender schematic) seem to be missing from the board and the layout. When R27 is connected, the power amplifier may oscillate at an ultrasonic frequency. You can see it on the oscilloscope but not be able to hear it from the speaker. To fix this you must exchange the wires from the output transformer that go to pin 7 of V3 and V4. Remember to ground the sleeve of the speaker jack.

Remove those 47uF capacitors from the back of the board. They are not helping.

Is R32 burned? Check it's resistance.

Below is a picture of cuts and wires to add to fix the ground loop. Be sure to remove the wire near the minus end of C18.

[Rebel]

yes i'm using the fender schematic too, but the lay out was made in spice and power components have different names but when i talk with you i use the original schematic. The error on the top left on the layout (C12 connected to DC) was cecked before the make of pcb and i put a wire on your blu line.
Tp29 is -10.45V
C5 and R7 are not in the schematic layout because who make it forgot them but i have installed them on pcb.
On oscilloscope i can see the oscillation you mean, i exhanged blue with borown and it go better.

R32 is not broken.

now i'm cutting the power circuit from the pcb and make it in another part of the case so i put the power circui far from the preamp. i'm doing well?
and now im'using your ground wire. give me some minutes and i write you what appen.

now at 13:35 i have finished to change my layout.
I followed your ground wire and i have made a separate power circuit this is the amp now :



green wire = ground
red and black= heaters

from the power circuit:
blue=294Vdc =Y
red=232Vdc =Z
Brown=-10.4vdc =X

this is the power cicuit:


before this modification i had 200mv of noise when volume was 0 and tone was0
now i have 20mv of noise in the same conditions.
But when i increase the volume the noise increases too, is it normal? or there is another problem?

now i'm tring to reduce all wiring on the pcb with cable, do you have any suggestions?

PS: thank you very much you saved my life
can i ask you why you know so much on ground and tube amps? is it your job or is a passion? you are very good teacher.
Now i'm readin all posts on your link to know more on ground thank you again.

[loudthud]

now i'm cutting the power circuit from the pcb and make it in another part of the case so i put the power circui far from the preamp. i'm doing well?
and now im'using your ground wire. give me some minutes and i write you what appen.

Every wire has resistace and when current flows through that wire, a small voltage is generated. If that voltage appears at the input of an amplifier, it will be amplified along with the signal we wish to amplify. In a guitar amplifier, or any amplifier that employs a power supply that converts AC (50Hz or 50KHz) to DC, care must be taken to keep any AC currents that flow to ground from generating small voltages between different points in the ground that are sensitive to noise pickup. Proper connection of ground requires you to understand where current flows to minimize the hum or noise.

A common mistake is to connect all the filter capacitor grounds together at one point on the PCB. Notice how I disconnected the ground for C18 and ran a wire to the preamp and grounded it there. Although there is not much hum current in the ground at the point where C18 was grounded on the PCB, there are currents from V3 and V4 in that ground. Those currents can cause distortion or oscillation when their small voltages are fed to the preamp. Note that any hum on the ground of a filter cap is conducted to the positive side of the cap with very little attenuation.

When large AC currents flow through a filter capacitor, there will be an AC voltage across the capacitor. C15 and C16 are in parallel so they tend to share current but C15 is closer to the source of current (the transformer) so it gets slightly more current because of the resistance of the tracks connecting C16. R32 and C17 form an attenuator for the AC voltage that is on C16 and there is also a DC voltage loss across R32. Note that the point B+ and point Y both go to V3 and V4. The ground currents for V3 and V4 should return to ground near those capacitors. The ground currents for the bias supply (C13 and C14) are very small but you would like to not have any small AC difference in voltage to be seen by V3 and V4 so the ground for C13 and C14 should be close to the ground return for V3 and V4.

R33 and C18 attenuate any hum present on C17. The voltage on point Z feeds the preamp which is more sensitive to hum so C17 needs to be connected to a quiet ground of the preamp.

PS: thank you very much you saved my life
can i ask you why you know so much on ground and tube amps? is it your job or is a passion? you are very good teacher.
Now i'm readin all posts on your link to know more on ground thank you again.

When I first started building audio projects, they all had hum unless they were battery powered. Many years later, after I had obtained some education, I was able to eliminate most of the hum. I got a job soldering circuit boards and took classes on electronics that were offered by the company after normal work hours. I only needed to take three classes to become qualified as a PCB designer so I did that and got into one of the engineering groups. It was around that time that I bought my first oscilloscope.

Having my own test equipment helped me figure out how and why things work. I also developed an intuitive feeling for how analog circuits work, not just a math based knowledge. Both approaches have merit. One place I worked designed and manufactured ICs. Whenever someone had an idea, they would want to sit down and run a simulation. I could usually get things working in the lab before they were done. Many areas of the electronics industry have contributed to my knowledge of high performance analog circuits.

Kevin O'Connor (sometimes called KOC on this board) wrote an excellent series of books about tube amps. He actually visits and posts here from time to time. His books are good but he needs to hire me as his editor! He doesn't get it right every time. Many of the members on this forum are quite knowledgeable, but sometimes opinions can differ. It is good to get all sides and make your own conclusion.

Here is a little tip that will ruduce some of the 50Hz hum you still have. Replace R31 with two resistors in series. A 220K 1W to B+ and a 47K to ground. Connect a 10uF 100V capacitor across the 47K. Now disconnect the junction of R35 and R36 from ground and connect that point to the junction of the 220K and the 47K. If you want to hear how it works, short the junction of the 220K and 47K to ground with an insulated clip lead. Power up the amp and when you can hear the hum (turn up the volume control) disconnect the clip lead and the hum will decrease in volume.

[Rebel]

you have a lot of experience, I'm glad to have met someone like you.

is this what you have told to me in your post?


what appen when i use it? i create a filter ( like low pass) that attenuates 50hz buzz?or what?

on monday i will use it, and i will tell you what appen.

[Bruce / Mission Amps]

Not sure what if any effect your LED lamp circuit would have on this but, I'd move that little circuit to be in parallel with your 100K B+ draining resistor... and then just eliminate the 100K resistor all together, as I doubt you need it anyhow.

My limited understanding of what this is doing is putting a DC voltage "bias" on the filament supply through the 47-100 ohm "voltage balancing" resistors.
The filament supply is running at 50Hz-60Hz 6.3VAC and the new DC voltage, fed at the zero voltage point of the AC cycle in the filament string, moves the reference of 0v (not ground) to a positive voltage... which in turn makes the actual 1vdc to 3vdc (etc.) bias voltage on the cathodes of your preamp tubes look like they are much further away from the 6.3vac filament voltage by the new +DC biasing supply.
So they are less likely to pick up and AC hum from the internal filaments in the tubes with respect to their cathodes....
yikes that was wordy! ha ha...

[loudthud]

Your schematic is correct. The 100K R32 is not needed unless you want to lower the voltage at C4 that feeds the preamp.

The red glowing heater inside a tube will emit electrons. It is not as efficient as the cathode, (which has a special chemical coating) but it still emits a number of electrons. Some tubes do not have a cathode, they use the heater as a source of free electrons. During part of the AC cycle, these electrons are attracted to the cathode because the cathode is positive with respect to heater. Because there is no capacitor from the cathode to ground, a small voltage is created at 50Hz that becomes part of the signal being amplified. By biasing the heater with a positive voltage (50V), now the cathode (1.5V) is always negative with respect to the heater so the cathode repels any electrons that might be floating around. This is discussed in the Radiotron Designer's Handbook.

The Fender Pro Jr did not include this circuit to keep the cost low. Fender used the LED as a pilot light. Since you have used a 6.3V incandescent bulb, you do not really need the LED, D5 or R34.

[Rebel]

I put R32 100k only for simulation of power (as load resistence) circuit and i forgot to delete it before the print of the screen, sorry.
I use a 6.3V lamp and when i connect it the hum decrease.
Tomorrow I will use your schematic and I will write you what appen. if i can i will make a video or sound recording to show you the result.

Thanks for the time you spent with me and for all intelligent solutions you have given me
I understand that this is a forum full of people educated and prepared.
I hope one day I will give an answer like yours to some one

[Rebel]

i used your configuration of 220k and 47k with capacitor but nothing changes.
i noted that if i touch the body of the lamp the noise decreases a lot! why?
i think that with my hand i put a capacity or an inductance on 6.3Vac and reduce thenoise.... I don't know what i have to do

[loudthud]

i used your configuration of 220k and 47k with capacitor but nothing changes.

The junction of the two resistors should be about +50VDC. Verify this with your Meter. Did you try temporarily shorting that point to ground and listen for a change when you remove the short. Make sure the volume control is not all the way down.

i noted that if i touch the body of the lamp the noise decreases a lot! why?
i think that with my hand i put a capacity or an inductance on 6.3Vac and reduce thenoise.... I don't know what i have to do

Hard to say what is causing this. Could be a shielding issue where your body injects enough hum of the correct phase to cancel some of the hum already present from other sources. Does it matter which side of the heater supply you touch? Try reversing the heater supply wires where they connect to the PCB.

Are there flourescent lights in the room where you did this? They can be a source of hum. For this and other reasons, guitar amps are usually enclosed in a metal chassis and metal shields are put around preamp tubes.

Here is another trick from the Radiotron Handbook. Remove R35 and R36. Replace them with a 100 to 250 ohm pot. The center terminal of the pot should go to ground or the junction of the 220K and 47K. Check the rating of the pot to make sure it is within it's power rating. Now adjust the pot for minimum hum. This is called a hum balance control. It is usually only accessable from inside the chassis. It will reduce 50Hz hum, but not the 100Hz hum. 100Hz hum comes from grounding issues in the power supply. You could try the DC power supply for the heater supply again. That should remove all the 50Hz hum.