S.E.X buzzing

[Paul Joppa]

In the SEX 2.1 (with PT-7 power transformer) the heater power is only connected to ground for two short periods during the 60Hz power line cycle - just when the diodes are conducting. Most of the time it floats. For that reason, if there is a little leakage between heater and cathode, the voltage will drift away from symmetrical. You can put a couple 100-ohm resistors in series across the DC power and ground the center point if you want to see matched voltages; you are not however going to hear a difference unless the leakage is so bad the tube should be replaced.

The center tap is grounded so that the winding itself is always grounded. This allows it to act as an electrostatic shield for other windings.

 

[mcandmar]

Paul,

I just tried this out of interest and it does make a difference, enough that i am going to have to wire it in permanently.

I initially tried it with just the scope connected and could see the sharp peaks were knocked off the noise with the resistor center point grounded, i then plugged in a pair of headphones (also before the 120r resistors at the headphone socket) and i could clearly hear a reduction in the higher pitched bzzzt noise any time i connected the ground. Voltages evened out a bit from +3.6/-2.4, to +2.9v/-3.1v, though i find it curious they swung the other way to be slightly more negative.

Before i warm up the soldering iron is it worth experimenting with small value caps across the heater circuit, or directly on the transformer secondary to help further reduce any switching noise?

 

[Paul Joppa]

Interesting - thanks so much for posting!

My best guess at this point is that it's common-mode noise. Capacitors aren't going to help unless they have lower impedance at 60Hz than the two 100 ohm resistors, which are effectively in parallel. 50 ohms at 60Hz is 53uF, that's getting pretty silly.

You can replace the resistors with a pot, probably anything from 50 to 500 ohms will do. You'll want a 3-watt or greater rating. And make sure the tubes are the same make - some have series heaters, so the best balance will be off center for those. (Also make sure they are the same phase - pin 7 to pin 7, 8 to 8.) Adjust it for the lowest hum.

The Schottky diodes don't have reverse recovery spikes, but they do still have switching noise. So it's possible that a common-mode choke will have an effect. If you put one between the transformer and the rectifiers, it will reduce coupling to the other windings (i.e. the HV winding). Putting one between the DC filter and the tubes will have more effect on what noise gets to the heaters. I have, at this point, no idea which would be more effective, or whether either would be audible at all. But it's a standard approach, so there's a good chance.

Before getting too carried away, do consider getting a second set of tubes. I imagine that different tubes, especially different makes, might have different sensitivity to heater noise coupling.

 

[mcandmar]

My best guess at this point is that it's common-mode noise. Capacitors aren't going to help unless they have lower impedance at 60Hz than the two 100 ohm resistors, which are effectively in parallel. 50 ohms at 60Hz is 53uF, that's getting pretty silly.

I was thinking smaller in the 10-100nf range directly on the secondary winding as i've seen it on various schematics with the explanation of preventing switching noise from propagating back through the transformer into adjacent windings.

You can replace the resistors with a pot, probably anything from 50 to 500 ohms will do. You'll want a 3-watt or greater rating.

I will look into that, from the bit of research i did i found this schematic you posted a few years back, http://usr.audioasylum.com/images/0/5625/FC-1_application_300B.GIF , i'm curious as to the purpose of the RC filter, whats the advantage to that vs directly connecting the center of the pot to ground?

And make sure the tubes are the same make - some have series heaters, so the best balance will be off center for those. (Also make sure they are the same phase - pin 7 to pin 7, 8 to 8.) Adjust it for the lowest hum.

Tubes are always identical pairs, i'm a little OCD about that.  I only have one series wired tube which is one of the Zeniths supplied with the kit, i never used that set as they were very noisy, but i'm now wondering if that is because one is series wired and the other is parallel wired.  I cant remember which of the two had the problem so i will need to retest them.

The Schottky diodes don't have reverse recovery spikes, but they do still have switching noise. So it's possible that a common-mode choke will have an effect. If you put one between the transformer and the rectifiers, it will reduce coupling to the other windings (i.e. the HV winding). Putting one between the DC filter and the tubes will have more effect on what noise gets to the heaters. I have, at this point, no idea which would be more effective, or whether either would be audible at all. But it's a standard approach, so there's a good chance.

I'm curious so i'll keep an eye on ebay for something common-mode and capable of handling a few amps to experiment with.  Unless something salvaged from a computer ATX power supply would be suitable, i have a box of those toroidal style chokes here that are rated for a few amps at 240v.

Before getting too carried away, do consider getting a second set of tubes. I imagine that different tubes, especially different makes, might have different sensitivity to heater noise coupling.

I have a bit of a collection at this stage of various different brands, interestingly one of my tubes started making an intermittent russling sound last night, i'm not sure if its just coincidence or if i finished them off while testing yesterday as i had the amp powered on/off a few times during the process.  I went back to an older set that i had it my bad pile marked as "slight hum" and so far they are silent, i need to go back and re-evaluate my collection of bad sets to see if they will work fine now.

 

[mauroj]

Guys, I am still having this problem. It is a loud buzzing sound more in the left channel than the right but the volume doesn't change with the volume control. I have tried 3 different sets of tubes and my voltages are -2.5, 3.7, -2.5, 3.7 at C1,2,4,5 and A7,8, B7, 8. All other voltages and impedances are perfect. I have gone over every terminal and resoldered any that had even a hint of error. I don't know what the problem could be. Please help. Thanks in advance.

 

[Paul Joppa]

Sorry for the confusion, there are two thread in this thread and I do get them mixed up.

Can you quantify "loud"? Like, by measuring the AC voltage at the speaker terminals? I understand that not everyone has a meter that will read a millivolt or less - I'm just trying to guess what candidate problems to look for.  If your meter will do this, it could help.

If you have some clip leads (cheap 10-pack at Radio Shack; very handy!) you can clip the power section grid (pin 1) to ground. If it still hums the same amount, it's in the power section, if not then it's coming into the driver section. The 450v caps are the respective filter caps to look at - 47uF for the power section and 22uF for the driver. Check for orientation, possible bulging, and good solder joints.

 

[mauroj]

From the left channel terminal I am getting 144 mV and the right channel terminal I am getting 5 mV. This seems high so I may have done it incorrectly. Black lead on black terminal and red lead on red terminal? Once it warmed up it went to 39 mV and 6.6 mV respectively. This is so discouraging. Thanks again.

 

[Paul Birkeland]

Guys, I am still having this problem. It is a loud buzzing sound more in the left channel than the right but the volume doesn't change with the volume control.

This is likely a loose solder joint in the power supply, probably around the 22uF capacitor on the left channel.

 

[mauroj]

Ok, I am still at it but the problem seems to have shifted slightly. Now when I power it up there is a faint hum. After a few seconds it starts to get louder until it gets just as loud as it was before but now the sound is approximately equal in both channels with headphones. Without the headphones plugged in, I measure 6.6 mV on the left channel and 5.4 mV on the right. I have gone through and resoldered every terminal that looks less than perfect, especially those around the power supply capacitors and it is still buzzing. Please give me more suggestions. Tell me what else I can do. I am dying to hear this thing playing music. Thanks again in advance.

 

[Paul Joppa]

OK then, back to my previous suggestion to ground the grid of the power section of the tubes and see how much noise is left.

In addition, you have used the terms "hum" and "buzz" in this thread. Hum means it's a nearly pure low frequency tone, usually 120Hz - not deep bass but unquestionably in the bass. "Buzz" means the tone has a lot of harmonics and a raspy nature; what we used to call buzz-saw back when there were such things. This is a useful diagnostic - power supplies with some filtering but not enough will give a hum, while capacitive coupling such as heater to cathode coupling is biased towards the higher harmonics and has more of a buzz character. Can you sort out what you are hearing?

 

[mauroj]

Ok, first things first. I did not have clip leads so I temporarily soldered in some hookup wire from each of the tubes pin 1 to ground. With that in place, it is virtually silent, only the faintest of hums. So I removed the temporary leads and resoldered the 22uf cap terminals as well as every other spot around the tubes sockets. I then put the tubes back in and tried it. Same sound, same volume, 5.4 mV left and 6.6 mV right which is what I meant to put in my previous post. Anyway, I tried a second set of tubes and the sound was still there, but lower at 3.7 mV left and 3.8 mV right. Then I tried the tubes that came with the kit, coin base Sylvanias and the sound went to 6.3 mV left and 3.7 mV right. As to the sound, it sounds to me as if it is a buzzing sound, not a hum. There are definitely variations in the harmonics. At this point, I think the tubes may be playing a role since the noise has changed in volume with each set of tubes. 
Also, the tubes that measure 3.7 mV really are fairly quiet and loud sound could be due to the sensitive headphones I have 105 db at 44 Ohm Shure SRH440 for testing anyway. It worries me though because I have Denon D5000 headphones and Klipsch Heresy 1.5 speakers and I don't want that sound to color my listening. What is normal hum/buzz with the SEX? Will it decrease significantly with the C4S upgrade which I have yet to install? Thanks again.

 

[Paul Joppa]

THIS POST HAS BEEN MODIFIED TO CORRECT AN ERROR IN THE SECOND PARAGRAPH!!

Great data! That really helps; it's not the high voltage power supply (except possibly the last stage with the 22uF capacitor) and it's not the power stage. So it's coming in at the first stage. And it's buzzy, so there is probably a capacitive coupling going on. And it's in both channels - it's always possible, but very unlikely, that there is the identical fault in both channels.

My best guess at this point is that the 6.3v winding, which is common to both channels, is not properly grounded. You can check the DC resistance from power transformer terminal 10 (the 6v winding centertap) to the chassis. It's in the middle of page 22 in the manual. I'd check the resistance from the G terminal of the IEC inlet to the chassis as well, since it's right there. In both cases, check the transformer or socket terminal to the actual chassis, which tests all the mechanical and soldered connections. Check the resistance from power transformer terminal 10 to power transformer terminals 1 and 2 4 and 5, to be certain that terminal 10 is connected internally in the correct way - I'm trying to cover all the bases here!

If that does not make a difference, then it's back to that 22uF capacitor. The lead that attaches to 12U is grounded by the wire from 12L to 13L, and 13 is bolted to the chassis as the single point connection of signal ground to chassis ground. So check the resistance of that capacitor lead to the chassis itself.

11L and 11U are presumably good connections since the high voltage power does get to the tube, and you are certain about 11U. The only way this could fail is if at 11L the two resistors are connected but the joint to the lug itself has failed. You can check the resistance from the positive lead of the 22uF to either resistor (8.2K to T6 or 150K to A5).

 

[mauroj]

Paul, thank you so much for the help. I checked the resistances around terminal 11 and the 22 uf cap and all checked out. I resoldered the ground connection from terminal 10 on the power transformer to the terminal strip 3. However, I do not know what the different resistances within the power transformer should be so I couldn't check that yet. Can you clarify that second paragraph for me? And can you clarify what an acceptable level of hum is for the SeX without C4S? Thanks again.

 

[mauroj]

Paul and others, thank you very much for your help so far. I am however still in the thick of it. I am hoping to get that clarification about what those resistances should be around the PT-7 as well as a good explanation and a picture of the 120 Ohm resistors on the headphone jack. I would love to know some more details about why one would want them, whether they will be ok if I use speakers more than headphones, and exactly how to install them. Thanks again.

 

[Paul Joppa]

Paul, thank you so much for the help. I checked the resistances around terminal 11 and the 22 uf cap and all checked out. I resoldered the ground connection from terminal 10 on the power transformer to the terminal strip 3. However, I do not know what the different resistances within the power transformer should be so I couldn't check that yet. Can you clarify that second paragraph for me? And can you clarify what an acceptable level of hum is for the SeX without C4S? Thanks again.

THIS POST HAS BEEN MODIFIED TO CORRECT AN ERROR IN NOTE 4)!!

Normal noise for a SEX amp is 0.3mV at the speaker terminals.

Second paragraph:

1) My best guess at this point is that the 6.3v winding, which is common to both channels, is not properly grounded. You can check the DC resistance from power transformer terminal 10 (the 6v winding centertap) to the chassis. It's in the middle of page 22 in the manual.

--- Properly grounded means zero resistance between terminal 10 of the power transformer and the chassis plate. See note number three below. No transformer resistance is involved. Note that you will not be able to measure with precision fractions of an ohm.

2) I'd check the resistance from the G terminal of the IEC inlet to the chassis as well, since it's right there.

--- Sorry, on the current IEC the terminal is labelled E (for Earth) rather than G (for Ground). My bad, I didn't look it up. There are three terminals on the IEC socket, labelled E, L, and N (for Earth, Line, and Neutral). There is a wire from terminal E to the chassis ground lug, as described on page 20 of the manual. check the resistance between the terminal E and the  chassis; it should be zero. See note number three below. No transformer resistance is involved. Note that you will not be able to measure with precision fractions of an ohm.

3)  In both cases, check the transformer or socket terminal to the actual chassis, which tests all the mechanical and soldered connections.

--- Do not re-solder. Do not connect the meter to the solder. Do not connect the meter to the wire. Connect one meter probe to the actual terminal itself. Connect the other meter probe to the chassis plate.

4) Check the resistance from power transformer terminal 10 to power transformer terminals 1 and 2 4 and 5, to be certain that terminal 10 is connected internally in the correct way - I'm trying to cover all the bases here!

--- The actual transformer resistance of this winding is too small to measure. Meters that can distinguish such small resistances from zero cost $2000 and up. So just look for as close to zero ohms as you can measure!  Note that you will not be able to measure with precision fractions of an ohm.

Clear now?

 

[mauroj]

Yes, thank you very much Paul. That clarifies things nicely for me. I will check that later tonight and report back with my findings. Thanks again.

 

[mauroj]

Ok, I finally got those new measurements done. The resistance between terminal 10 and the chassis plate started at 0.1 Ohms and climbed up to 1.3 Ohm before dropping down and hovering between 0.4 and 0.9 Ohms. When I checked again after a couple minutes it hovered between 0.3 and 0.7 Ohms.
The E terminal on the IEC resistance to the chassis plate hovered between 0.2 and 0.5 Ohms. Finally, the resistance between terminal 1 and 10, 2 and 10 on the PT did not even register on my meter, just showed O.L. I am assuming that the fluctuating resistances are a bad thing but I don't know why. Is it a bad connection? Solder joint? PT? Thanks again for the help.

 

[Paul Joppa]

THIS POST WAS BASED ON AN ERROR OF MINE AND IS CANCELLED!! See later apology.

Bingo!  I think, anyhow. Looks like the terminal 10 is not connected to the center tap of the 6.3v winding.

First off, all resistances less than 1 ohm are probably zero, it's just hard to get an accurate measurement. Those are probably not the problem.

You can fake a good grounding, or ask Eileen (email the spare parts address) fora replacement transformer.

If you fake the grounding, it will confirm this to be the problem, so it might be interesting whether or not you ask for a replacement transformer. The easy way is to connect T1 (or T2, your choice) to T10 of the power transformer. Leave the connection of T10 to the chassis.

A better way is to use two 47 ohm resistors, T1 to T10 and T2 to T10. But the easy way will at least answer the question definitively, and will probably be good enough that no further effort is needed.

I should note two things:

1) If the easy way causes the fuse to blow, then your measurements of T1 to T10 and T2 to T10 were incorrect and we're back to Square Zero.

2) If either method works and gives low hum (finally!!) then you deserve a new power transformer if you want it.

If you can get a good high resolution close-up photo of the transformer T10, showing any internal transformer wires, that would be informative for us - but no sweat if that's not easy to do.

 

[Paul Birkeland]

I would also mention that it's possible that you accidentally snipped off the wires that connect to T10 on the power transformer.  If this is the case, you can reconnect them with some solder. 

 

[mauroj]

Ok, I will try to get a good shot to show T10 to you guys. I am reasonably certain I did not snip those wires but I will definitely check. Barring that, I will try the 47 Ohm resistor way tomorrow and see what I come up with. Thanks a lot gentlemen. Have a great night.