Help in Troubleshooting noise in audio Amplifier

[HLong]

Hi all, I just recently built my SE Tube amp with parafeed output transformer. Everything looked fine and as expected when I tested it with a 8 ohm resistor dummy load. However, as soon as I hook up to a speaker and crank up my signal generator, I heard a loud high pitch noise and the noise vary in pitch as I vary the input frequency. The outputs apprear normal on scope. I did not see any oscillation. I tested it on a different speaker and I have the same result.

I ran out of thought what would be wrong. Do you have any suggestion?

 

[felixc]

If you can hear it, you must be able to see it on the scope.
Does the noise add to the exepcted sound?
Does it only appear when there is another sound produced?
Does it appear at any volume?
Is it proportional to the volume of the input signal?

 

[cbarn24050]

Hi, thats the problem with you "Hi FI" guys, you hear things the rest of us can't. It must be harmonic distortion as changes with the input frequency, probably way below measurable levels, thats why you can't see it on your scope. Ask another family member if they ca hear it, if so then you might have a real fault.

 

[HLong]

The noise does not add to the output signal. Both input and output looked very clean and stable on the scope. With the dummy load I could hear the pitch also but it very low volume. It appears coming from the plate choke of the 300B tube, but not 100% sure.
Everyone obviously hear it when hook up to the speaker. It's darn..loud. As long as I crank up the volume on the signal generator, then you can hear it and the pitch varies with the frequency. It appears with any volume level. I have not actually tested it with a CD player yet because of this. I might try it to see if it makes any difference.

 

[felixc]

So the noise does not come from the speaker but from the amp itself?
Can it be magnetostriction noise from the tranformer?

 

[HLong]

Felixc, can you explain a little more in detail what you meant by magnetostriction noise from the transformer?

The plate choke mentioned is the load of the 300B tube and couple to the output transformer with an output cap.

I subsituted the plate choke with an identical one last night, but I still have the same problem. Changing different speaker would not help either. So I suspected that the noise came from somewhere within the amp. But why I could not see on the scope? and the sound appears in the entire audio frequency range. I could hear it up to 10Khz.

 

[felixc]

This is noise that coil windings and the metal plates in transformar generate when current passes through it. You can hear a 60Hz buzz in most transformers, this is magnetostriction. In a switching power supply, even if clocked at 500KHz had its regulation loop operating at some 10KHz, and I was able to hear a 10KHz pitch depending on the load at the power supply output.

It involves that there must be an oscillation somewhere in the circuit, or a ringing versus the signal input. But you should be able to see something on the scope.

 

[HLong]

Yesterday, I tested it with a portable CD player. The sound comes out very good. I did not hear any type of distortion. The signal for both input and output also looked very clean on the scope. During the silence passage, you do not see anything on the scope as expected.

I still do not understand why I have problems mentioned if the input came from the signal generator. I have tried two different signal generators with different brand and have the same results. It's still a mystery to me.

 

[Comcokid]

Does the power transformer on your amp have an internal electrostatic shield? This isolates noise on the primary from being capacitively coupled to the secondary.

Your portable CD player is isolated from AC power. Your signal generator most likely is AC powered through it's own power transformer.

Finally, your amp has several transformers and a plate choke. Make sure they are located far enough apart and oriented so that you avoid magnetic coupling between them.

 

[ScottyUK]

Is the signal generator output floating? Could you have introduced a ground loop which is provoking instability? The portable CD player obviously isn't ground-referenced.

Is it possible that you are over-driving the output transformer, i.e. the transformer is on the edge of saturation? This would give you a magnetising current comprising all sorts of harmonics, the lower ones falling within the audio band where you can hear them if they cause vibration of chokes or within the valve. If it is a single ended design there may well be even-order harmonics as well as odd. Similar wierd results may well occur if the transformer is carrying a DC bias, as this will also affect the magnetising curve and will generate harmonics. The closed feedback loop would act to minimise output distortion, so this problem would not necessarily manifest itself at the output.

I'm not suprised that you can't hear it with a musical passage - the effect might still be there but there is no pure tone against which you could identify any abnormal sound.



----------------------------------

If we learn from our mistakes,
I'm getting a great education!

 

[HLong]

Comcokid, yes I am aware of the maganetic couple between the choke, transformer and other components in the amps, especially the tubes. They located far apart and don't think it's a concern. Also the power tranny and power supply chokes are potted type, except for the plate choke.

I am not really sure if the power tranny have internal electrostactic shield (I will check on that it's been awhile I have not checked on that spec. I checked on it before though), but it's a military potted transformer. Since this is a prototype one, may be the wiring/grounding induced the noice into the system. I might try to put into the new chassis to see if it help.

 

[HLong]

Scotty, I don't think the output from signal generator is floating, but will double check on that tonight. Also I don't believe that output tranny is driven into saturation since the output tranny is the parafeed type; therefore there's no path for DC current to flow. Yes, it's a single end type power amp.

You're probably right the problem's probably still there, but I can't hear it.

 

[ScottyUK]

HLong,

Could you expand a little on the term 'parafeed' w.r.t transformer windings? It's not a term I'm familiar with in the UK.

The above not withstanding, the only sure-fire ways I can think of avoiding DC through the winding would seem to be 1) capacitively couple it; 2) use a DC servo loop to drive the DC to zero. I think the latter would be tricky with a valve amp (not impossible though).

If the signal generator output is not floating, can you make it so by using a coupling transformer between it and the amplifier input before repeating your test.



----------------------------------

If we learn from our mistakes,
I'm getting a great education!

 

[HLong]

Scott, you're thinking right in term of avoiding DC through out put tranny. The parafeed term has been a popular term in audio cirle. The topology and the term is not new. It has been described in Radiotron Designer's Handbook by F. Langford Smith I think even before I was born. The output transformer was designed with no DC current allowed in mind to optimize for audio purpose. Some people like it. Some do not. Yes, it's couple via a capacitor.

I did try to feed the signal generator via my tube preamp, but got the same result.

 

[felixc]

So the problem only shows with your signal generator, and if you put audio components at the input, everything works fine?
Is there a switcher power supply in the signal generator, or a linear one?

 

[logbook]

Scotty,

Parafeed is apparently an abbreviation of "parallel feed". Parafeed is not indexed in Langford-Smith, but parallel feed is. The interstage coupling transformer in this valve based system is capacitively coupled so that no DC passes through the high inductance transformer.

Hlong,
I suppose your scope is earthed. Applying the scope therefore changes the conditions. Did I understand you correctly, did you say that you applied the scope to the loudspeaker terminals (not just the dummy load) and that no noise was visible with a pure sine input signal form the signal generator?

Since the output is clean with a CD input, you would think the problem was feedback from the loudspeaker to the signal generator. The fact that you have tried two entirely different types of signal generator negates that theory. Is there anything else different between the CD set-up and the signal generator set-up? (eg a signal cable routed over the output cable in both signal generator configurations but not with the CD system.).

 

[BigTom2005]

Check the quality of the cable between fun gen and your amplifier. Also, is there any DC coming into the amplifier from the generator? Most noise problems come from the source, and amplification just makes it much worse. Also, whats the output impedance of the function generators that you are using? Its usually 50-75 ohms, and that is what you want to transfer the maximum amount of signal into the amp. Check the cabling between the source and the amp, and check for DC components in the signal b4 the amp.

 

[BigTom2005]

Whats the output impedance of your amplifier? Could it be that its much greater than the 8-ohm dummy load? Perhaps the load is loading down the output of the amplifier, attenuating the useful signal. The scope has a mega-ohmic input impedance, so it would not load down your amp. I suggest that you find out what the amps output impedance is by placing different loads at the output, and using the voltage divider rule to solve for the voltage across the unknown impedance which is your output impedance. Measure the no-load voltage, then measure a load voltage. From there, you can find the output impedance. If its much greater than 8 ohms, then you have a problem--one of the fundamental concepts in communications is the issue of impedance matching.

Thomas

 

[BigTom2005]

Also, make sure that the input impedance to your amplifier is equal or greater than the output impedance of the function generator. You might need a impedance transformer circuit b4 the amp in case the source (FUN GEN) is getting loaded down. Good luck, and keep me posted.

Thomas

 

[HLong]

Felixc, yes I beleived it only happend with the function generator, but not 100% sure since the sound may mask the noise. I think one of the function generator might use switching power supply, except for the HP generator I am not sure if it has linear or switching. I note the HP generator caused humm whenever I plug it in the outlet even without turning power switch on. But the other one does not. I am still thinking about magnetostriction noise you mentioned in your earlier post. I believed the high pitch I heard probably came from the 300B loading choke when I tested it with dummy load.

Felixc, can you explain when/what case magnetostriction noise happen? Does it happen when you pass too much DC current through it? I am very sure that this loading choke can handle much more DC current than the current condition. I am passing from 60 to 75mA DC through it. And it rated at 50H @100 mA.

Logbook, yes scope is earth. I don't want to mess with this amp unless everything is reference back to earth since I am dealing with 800VDC power supply here. The routing cables are exactly the same as the function generator.

I did suspect some type of feedback, but why can't see on the scope??

BigTom, the output impedance of this tranny is specially designed for 300B output tube with a reflected impendance of 3K with 8ohm loaded. I do not think loading is the problem. The output from the function generator is 50 ohms and it feeds into the 6SN7 driver tube with 150K input loaded.

Thanks for all suggestions.

Just a note: I have been burning this amp for 8 hrs continously yesterday with my portable CD player. Sometime I did hear some distortion, but not really sure if it's from the actual burned CD or from the amp itself. I might have to repeat the CD to see if I hear the same thing. I play many different CDs though when sitting in my lab room doing other thing.