AC Voltage Measurement - Hum Reduction

[Diablo]

I'm try to do a little more hum reduction on my Magnatone M13, and I tried to measure the AC voltage at the speaker jack. I measure 6-21 millivolts, and the digital readout on the meter is jumping around. Can anyone tell me why the value isn't steady? I also get strange readings on the meter when I try to measure the AC voltage at the first filter cap connected to the rectifier. On the 700v and 200v meter scales, the AC voltage measures 6 volts. But, when I set the meter scale to 20v or 2v, the voltage reading goes much lower, like .6 volts. My meter reads fine on AC line voltage at 122 volts. So, is this a problem with the accuracy of my meter for AC voltage measurement, or some other phenomenon? I was expecting something like 40 volts AC ripple at the first filter node (40 mfd cap) with 429 DC plate volts. It's a cheap ($20) meter that's been doing fine for me for 10 years.

[Tom Phillips]

...I measure 6-21 millivolts, and the digital readout on the meter is jumping around. Can anyone tell me why the value isn't steady? ...

Your meter is measuring broad band noise. That, in conjunction with the sampling rate of the meter, will make the reading jump around

... I also get strange readings on the meter when I try to measure the AC voltage at the first filter cap connected to the rectifier. On the 700v and 200v meter scales, the AC voltage measures 6 volts. But, when I set the meter scale to 20v or 2v, the voltage reading goes much lower, like .6 volts. My meter reads fine on AC line voltage at 122 volts. So, is this a problem with the accuracy of my meter for AC voltage measurement, or some other phenomenon?

Inexpensive meters, such as your $20 unit, have a problem measuring a small AC voltage in the presence of a large DC voltage component. I suspect that is what’s happening to you.

...I was expecting something like 40 volts AC ripple at the first filter node (40 mfd cap) with 429 DC plate volts....

That would be unusually high ripple. (9.3%) Most amps measure 3% to 4% ripple at the first filter node when the amp is at idle. It would be higher when the amp is putting out full power. What are you measurement conditions? I suspect that your meter can’t make an accurate measurement of the ripple anyway. You could use a scope to make the measurement if you don’t have access to a better meter.

[Diablo]

Thanks Tom, you were right, it was my cheap meter. I borrowed a Fluke from work and took repeatable measurements.
With the amp idling and all pots set to zero, I measured 42 millivolts AC at the speaker.
6.04 volts AC at the plates of the EL34 power tubes, .92 volts AC on the screens, 200 millivolts at the next filter cap, 50 millivolts at the next filter cap, and 7 millivolts at the final filter cap, and 3 millivolts on the bias supply.
So, as far as ripple and filtering, it looks fairly normal.
However, the 42 millivolts at the speakers is a problem.
To check my methodology, I checked a 5E8A that I built, and it measures 4 millivolts at the speakers.

[soundguruman]

Hum reduction
Generally splitting the power supply ground from the preamp ground.
The preamp grounds on one side of the chassis, the power supply on the other.
This reduces noise in a significant and measurable way.

Twisting heater wires
This can cut hum by 50%
if it has not already been done.

finally if you are dealing with a really poor layout...
DC preamp filaments basically kills all remaining, after doing the above.

Just disconnect your AC filaments, and run it on a DC bench supply. That will tell you plenty.
If DC filaments are going to help, you can find out fairly quickly.

I am putting DC preamp filaments in a Vox AC30/6 TB.
because of the layout, this amp has one heck of a hum buzz problem.
(the AC runs right down the middle of the preamp)
After doing all "standard" hum reduction techniques, hum reduction was 60%.
After connecting the filaments to a DC bench supply it was 98%.

So if you need "studio" quiet, dead quiet, DC filaments might do the trick.
It's really easy to test this...

If you have a good layout, the AC filaments don't cause too much problem.
But when you have a real crap layout design, DC filaments can rescue the whole deal.
Example of crap layout:
This Vox had 3.65 volts PP hum buzz on the output, at fully cranked volume.
On DC heaters, it has almost none.
I'm not tearing the circuit board apart to cure the bad layout.
I'm just doing that simple mod to the preamp filaments.

[Diablo]

Hum reduction
Generally splitting the power supply ground from the preamp ground.
The preamp grounds on one side of the chassis, the power supply on the other.
This reduces noise in a significant and measurable way.

Twisting heater wires
This can cut hum by 50%
if it has not already been done.

finally if you are dealing with a really poor layout...
DC preamp filaments basically kills all remaining, after doing the above.

Just disconnect your AC filaments, and run it on a DC bench supply. That will tell you plenty.
If DC filaments are going to help, you can find out fairly quickly.

I am putting DC preamp filaments in a Vox AC30/6 TB.
because of the layout, this amp has one heck of a hum buzz problem.
(the AC runs right down the middle of the preamp)
After doing all "standard" hum reduction techniques, hum reduction was 60%.
After connecting the filaments to a DC bench supply it was 98%.

So if you need "studio" quiet, dead quiet, DC filaments might do the trick.
It's really easy to test this...

If you have a good layout, the AC filaments don't cause too much problem.
But when you have a real crap layout design, DC filaments can rescue the whole deal.
Example of crap layout:
This Vox had 3.65 volts PP hum buzz on the output, at fully cranked volume.
On DC heaters, it has almost none.
I'm not tearing the circuit board apart to cure the bad layout.
I'm just doing that simple mod to the preamp filaments.

Yeah, I stare at this thing and scratch my head on the layout. It has 10 tubes on the top chassis with lots of local grounds. It has AC line voltage delivered to the bottom chassis, then routed up to the top chassis with a wiring loom to the power switch. The main filter caps are a cap can in the bottom chassis, so the preamp and power amp share the same filter grounds down below. I've tried to move the heater wires to minimize hum, and have done the same thing with the AC line. I added a more solid ground wire between the bottom and top chassis after I installed a grounded IEC power jack for safety and removed the pair of death caps. I think the next experiment is to run the heaters off a battery as you suggest.

[Tom Phillips]

Yeah, I stare at this thing and scratch my head on the layout...

I have seen small Magnatone combo amps with the same problem. It appears that they just built them good enough to sell in their day and because of the poor layout and grounding many of them are very noisy even after the traditional filter cap job. A problem that many people have is the belief that a component must have gone bad or a mistake made during a previous repair. However, I believe that many Magnatone amps left the factory with designed in excess noise. They can be fixed but it requires careful rework above and beyond common parts replacement. Unfortunately, I have never seen step-by-step instructions for such rework.

[Diablo]

I have seen small Magnatone combo amps with the same problem. It appears that they just built them good enough to sell in their day and because of the poor layout and grounding many of them are very noisy even after the traditional filter cap job. A problem that many people have is the belief that a component must have gone bad or a mistake made during a previous repair. However, I believe that many Magnatone amps left the factory with designed in excess noise. They can be fixed but it requires careful rework above and beyond common parts replacement. Unfortunately, I have never seen step-by-step instructions for such rework.

Thanks for letting me know this problem is common to the brand and may be characteristic of the layout and design. When I got the amp, it was 100% original, so I know that the noise isn't caused by some earlier botched repair. I was hoping that I could find a simple cure, but that seems unlikely now.

[soundguruman]

Yeah, I stare at this thing and scratch my head on the layout. It has 10 tubes on the top chassis with lots of local grounds. It has AC line voltage delivered to the bottom chassis, then routed up to the top chassis with a wiring loom to the power switch. The main filter caps are a cap can in the bottom chassis, so the preamp and power amp share the same filter grounds down below. I've tried to move the heater wires to minimize hum, and have done the same thing with the AC line. I added a more solid ground wire between the bottom and top chassis after I installed a grounded IEC power jack for safety and removed the pair of death caps. I think the next experiment is to run the heaters off a battery as you suggest.

Just running the preamp with DC filaments will tell you a whole bunch, really fast.
This is why I have a small DC lab power supply - a great tool to have around.
But that is correct. You are trying to achieve "battery" quiet DC, with no ripple at all.

The McIntosh tube hi fi preamp is a good circuit to look at for a DC filament supply.
Usually, only the first stages, and not all of them, need to be DC filaments.
The rest are usually OK on AC.

It's THAT, and splitting the power supply ground from the preamp ground.
Separating these two, on opposite sides of the chassis.

Using JUST those two, you can pretty much get a VERY acceptable result.

It does help to measure with a scope.
You can see if you are fighting AC sine wave - from improper power supply ground
OR spurious noise from heater buzz.
The scope helps you to see which it is, or both.

BACK in the old days - hum buzz was considered "acceptable." we did not have high gain amps.
And manufacturers did not pay a lot of attention to noise.
But now, as we crank these amps to 10, it becomes a problem.

HINT- some components - such as ceramic disk capacitors...
located close to AC heater wires...
will allow the AC to easily enter the audio path. And then, it is amplified along with the audio.

Real (high quality) silver mica capacitors (probably not Chinese ones)
will stop this ingress from occurring.
However, when I say "silver mica" I am talking USA Cornell Dubilier capacitors.
If you are looking at Chinese capacitors (labeled "SM"), I question if these are really silver mica, or just fake ceramic "look a likes." Which are sold with a misleading label "silver mica." The Chinese one may not be real silver mica cap at all.

[Diablo]

The hum I'm dealing with is not effected by any of the three volume controls. I'm going to test the heaters on DC with a battery. Yesterday, I tried two other experiments which didn't change the hum. I elevated the heater wires with +25VDC offset, by building a voltage divider network off the bias tap on the PT, a diode and a couple of 220K resistors and a 100mfd/100V cap and a 25mfd/50V cap. I also added snubber caps (.01/1000V) across each of the four rectifier diodes supplying the B+.

[Diablo]

I didn't pick up a lantern battery yet to power the heaters, so I tried another experiment. I bypassed the switch in the top amp chassis by soldering a wire across the terminal lugs in the bottom chassis near the PT. By doing this, the wire loom from the bottom to top chassis no longer carries AC line voltage, and can't induce any AC into the B+ wire in the loom that runs parallel. What did I learn? The hum didn't improve, so the AC line voltage in the loom isn't a problem.

Would an audio probe be of any use for troubleshooting hum, or is it a waste of time? No O-scope, so I'm trying to get creative on the cheap.

[g1]

I bypassed the switch in the top amp chassis by soldering a wire across the terminal lugs in the bottom chassis near the PT. By doing this, the wire loom from the bottom to top chassis no longer carries AC line voltage, and can't induce any AC into the B+ wire in the loom that runs parallel. What did I learn? The hum didn't improve, so the AC line voltage in the loom isn't a problem.

This is only valid if you disconnected the wires in the loom. Otherwise, even though they are not carrying current, the voltage is still in there radiating hum. I'm assuming you did disconnect the wires but thought I'd mention it just in case.

[Diablo]

This is only valid if you disconnected the wires in the loom. Otherwise, even though they are not carrying current, the voltage is still in there radiating hum. I'm assuming you did disconnect the wires but thought I'd mention it just in case.

You caught my whoops. No, I didn't disconnect the wires in the loom as I was thinking that the current needed to flow in those wires in order to generate an alternating magnetic field. So I didn't get any magnetic coupling, but I could get capacitive coupling? I also tried to connect a 6 volt lantern battery to the heaters, but that experiment failed miserably. The battery didn't put out enough current to heat the tubes enough to make any noise whatsoever. I did get a faint glow from the pilot lamp. I'll have to charge up some NiMH batteries and have another go at it tomorrow.

[Diablo]

Finally got around to using batteries for the filament heaters. Disconnected the two leads for the heaters that plug into the loom connector for the top chassis. Hooked up a 5400 mAH - 5-cell NiMH pack with alligator clips on the socket pins of one of the EL34 tubes. Amp powered up fine, but it's just as hummy on DC heaters as it was on AC heaters.

[soundguruman]

Finally got around to using batteries for the filament heaters. Disconnected the two leads for the heaters that plug into the loom connector for the top chassis. Hooked up a 5400 mAH - 5-cell NiMH pack with alligator clips on the socket pins of one of the EL34 tubes. Amp powered up fine, but it's just as hummy on DC heaters as it was on AC heaters.

Having eliminated that heater hum, the noise might just be coming from the B+ supply.
Then the filter caps are opening up? From Age? Or, a bad rectifier?
Seems like you're narrowing that down.

[Diablo]

Having eliminated that heater hum, the noise might just be coming from the B+ supply.
Then the filter caps are opening up? From Age? Or, a bad rectifier?
Seems like you're narrowing that down.

I recapped the entire amp, all new e-caps. I also built a second fixed bias circuit with pots so I could bias and match each of the EL34s and cancel more of the hum. I posted the ripple voltage in post #3 above on each filter stage, and they all look pretty normal. I put in new four new diode rectifiers (1n5408) for the B+, and I bridged across each one with a snubber cap. The only diode that's original is the diode for the bias supply, and since I am getting good bias voltage with low bias ripple, the diode has to be OK. So now, I'm thinking the hum has to originate from ground currents interacting. Too many local grounds on the top chassis, and a huge current loop between the B+ rectifier in the bottom chassis, up to the top chassis in the loom to the on/off switch, back down the loom to the B+ filter caps, then back up the loom to supply the filtered B+ to each of the stages, then a ground return at one point on the top chassis back through the loom to the bottom chassis.