Oahu Tonemaster-Hum and 21khz distortion

[wsscott]

I've posted about this amp before.  All of the resistors and caps and tubes have been replaced.  The field coil speaker has been replaced with a permanent magnet speaker using a choke.  A schematic is attached.

It has an Instrument Channel with only 1 wired input jack.  The other jack was disconnected for future removal and use of the chassis hole for something else.  The other channel is the MIC channel.  Each channel has its own Volume Pot, but there is only 1 Tone pot.

After working on this amp for some time, it has a 60hz and 120hz hum even with nothing plugged into the input jacks, and even if the jacks are disconnected from the circuit.  If the 6SN7 tube is removed the hum goes away.  Other pre-amp and PI tubes have been tried with no effect on the hum.

The amp has very little treble produced from the Tone pot.

The Instrument channel reacts normally, just low Treble range.

The MIC channel is another story.  With the Volume pot at 2 (8 o'clock), and the Tone in any position, no problem except for the hum.  After turning up the Volume from  2 through 7 (2 o'clock), low level treble hiss begins.  At 7 thru 9, the signal is distorted greatly.  But just past 9 it clears up.  Still treble is minimal but it has increased.

I checked the signal frequency of the distortion and its around 21khz, and a high amplitude, so very noticeable.  When I adjust the Volume below 7 or above 9, that signal practically disappears.

I hope someone can help in solving this signal distortion.  Thanks.

[rumpus]

On the schematic, the input jacks don't shunt the input to ground when nothing is plugged in. This leaves the grid inputs floating, is also true if the input jacks are disconnected. What happens to the 60 Hz hum if you plug in a shorting plug on the input, or just short it to ground with a clip lead?

You didn't mention replacing the power tubes. Are they functional? How well balanced are they? If they are imbalanced, 120 Hz ripple from the B+ won't be fully canceled in the output stage. (But if removing the 6SN7 make the hum completely go away, this explanation is less likely.)

What is the signal you're using to measure the distortion? And how are you measuring it? I'll just note that "around 21 kHz" is just about the Nyquist frequency for the common 44.1 kHz sample rate. If you were, for example, sampling the signal with a computer sound card for distortion measurement, could that frequency be an artifact of the sampling?

[wsscott]

I replaced those jacks with shunting jacks.  So they are grounded.

The power tubes are new and matched.

I've measured the signal 2 ways. One with an App on my iPhone called FFT Plot.  I just open it up and put it in front of the speaker.  No input signal being used.  I also connected my scope to the Output jack with no signal injected and get the same results.  So its getting the hum and the distortion even if no signal is injected into the amp's circuit, and even if the input jacks are physically disconnected from the circuit connection to the grids.

[shooter]

If the 6SN7 tube is removed the hum goes away.

how much work do you want to do??


If it was a keeper, i'd make each triode it's own thing, no sharing cathodes or plates
I'd add grid-stop resistors at the tube socket.


did you increase PS filter capacitance?

[wsscott]

I also replaced the OT.  The PT is original.

[wsscott]

I replaced the filter caps with individual caps of 20/20/10.

All the tubes are new.

Good idea Shooter on re-wiring the sharing of cathodes and plates.  That shouldn't be too difficult.

I did add grid stoppers to the OT's but it didn't make any difference and I later removed them.  Did you mean add grid stoppers to the grids of the 6SN7 pre-amp tube?

[scstill]

For the 60hz hum, Have you tried elevating the 6v heaters?
If there is a separate wire from the 6v CT, remove it from chassis and jumper it to the 6v6 cathode

also without seeing the build are the heater wires away from signal path?

[scstill]

for the 120hz hum
without seeing your build have you grounded your PS nodes properly?
Preamp B+ node ground to the preamp ground
Power Tube B+ node grounded to the Power Stage
there is a great grounding diagram posted here
https://el34world.com/charts/grounds.htm

[wsscott]

I never heard the amp before my buddy started working on it.  So I don't know what it sounded like.  I just know what it sounds like now.

All the grounding is original.  Nothing has changed with that.  The preamp ground is separate from the power supply grounds.

The only wire that was weird was a bare wire from the PT.  I later learned that in these old amps, ie. 1947, they often grounded the PT chassis to a ground point with this wiring.

I haven't lifted the 6v grounding. It is original.  Its wires are original.  As far as separating the filament wires from the signal wiring, it is what it is because the space is so small and all the tubes are packed in like sardines.

[shooter]

Did you mean add grid stoppers to the grids of the 6SN7 pre-amp tube?

If the 6SN7 tube is removed the hum goes away.

yep

[scstill]

Sounds like you are making modern circuit changes from original.
As well modern ground techniques might help with vintage hums.
And the hum might be created with modern AC wall voltages that are higher than vintage AC wall voltages

[scstill]

Grid Stoppers are for RFI right? not for 60/120hz hum right?
Nevermind - after rereading the original, you are addressing the signal distortion

[uki]

Have any pictures of the amp ?! The guts of it.

[scstill]

sorry of you already addressed it and I missed
but is that .006 cap in the mic channel still good?

[wsscott]

It’s a .005, the reading isn’t the best to see. Yes it’s new and good.

[AlNewman]

Do you have an as built schematic?

[wsscott]

The only real change was adding the choke.  Everything else is wired as it was, just new parts. And the schematic attached is what I’m working with, and the wiring follows the schematic. I’ve checked it 5-6 times looking for a mistake.

[wsscott]

Shooter-Where should I install the grid stopper resistors for the 6SN7?  One end at Pin 1/Pin 4 for the 2 grids, and then where does the other end of the resistors connect?

Is a 68K resistor the appropriate size?

Thanks.

[wsscott]

Should I disconnect the other components that are currently connecting to each of the grids, and install the resistor on each of those socket pins (ie. 1 and 4), and then connect the other end of the resistor to the components that were disconnected?

This amp is point-to-point wired, with very few terminal strips, so it gets tricky--at least for me- ie. adding a wire with no solid connection point.

[shooter]

this is where I'd start, where you end is up to Hoyle, Roberts, or Murfree   

[wsscott]

Shooter-Great!  Thanks for your help.  Fingers crossed this will get rid of the oscillation, and maybe the hum too.

[wsscott]

Shooter-No luck with this mod.  I still have 60 hz and 180 hz hum.  The 120 hz seems gone based on the frequency analyzer app.  But I still have the same issue with the 21khz oscillation/distortion.  It's only in the MIC channel.  It seems to begin appearing when the Tone pot begins to add some Treble, and if it's Vol. pot is above 2.  Its really noticeable when the Vol pot gets up to 7, regardless of where the Tone pot is set as long as its adding some Treble to the circuit, BUT it cuts out when the Vol gets up to 9.

Could this be some kind of parasitic oscillation that's maybe coming from the PT?  But only the MIC channel seems to be affected.

[shooter]

120hz is a good indication of poor filtering on the DC, bad/sketchy grounds, sloppy sockets n connectors etal.
the 21K is a good indication of parasitic osc. or that channel picking up stray RF from the airwaves


try grounding the mic-channels grid to see if the 21K goes away.  is the tone pot original??  you could temp-in a new TS/vol or bypass for testing


you can also "walk" a solid chassis ground wire to each "ground point" and see if the 120hz "changes" as you ground grounds

[wsscott]

Shooter-all of the pots have been replaced with new ones.

The 120 doesn't seem to be showing up much on the analyzer, but it showing 60 and 180 hz noise, and the 180's amplitude is much higher than the 60hz.

The amp pics up this 21khz wherever it's plugged in.  It's not just one location but at locations miles apart.  I don't think its RF from the airwaves for that reason, plus its "tunable" with the Vol/Tone pots but ONLY on the MIC channel where maybe the impedance is different than the Instrument channel.

Is there a way to test the Power Transformer to see if its the source of parasitic oscillation?

I'll try grounding the grid and let you know.

[wsscott]

YES, grounding the Grid eliminates the 21khz noise!!  So what do I do?

But there is a 60hz hum at -73db, then a slightly larger 120hz hum, but a very prominent 240hz at -45db.

The Tone control brings that in when it moves past 2 on the pot, just starting to let the Treble through.

[shooter]

YES, grounding the Grid eliminates the 21khz noise!!  So what do I do?

1. verify your new shorting jack is actually shorting (you can also just temp-short ring to sleeve)
1a. verify the ground is solid at the 1 meg grid-leak


2. "bypass" the pot n .006uF with a jumper, you want the 1MEG N grid-stop "in-circuit"  BUT!!! don't add signal, just see if the 21Khz comes back.

[shooter]

there is a 60hz hum at -73db, then a slightly larger 120hz hum, but a very prominent 240hz at -45db.

this sounds like your measuring device isn't very good a(t) discernment 
a scope with FFT/spectrum feature comes in real handy

[wsscott]

Shooter--#1 and #2 checked out fine.  Ground continuity.

With no inputs in the Input Jack or any kind of input signal, and the Grid still jumped to a ground, I jumped the cap connecting the wiper on the Mic Vol pot to the 1 Meg resistor.

The 21khz remained gone.

[wsscott]

FYI-I grounded the Grid by connecting a jumper at Pin 4 Grid with the other end of the jumper to a ground point on the chassis. 

Everything else remained connected in the circuit, but no Plug was inserted into the Input jack, and no Signal was introduced.

[tubeswell]

Re-tension the faulty jack switch tip.

[wsscott]

The jack tip switch works fine.  It goes to ground when nothing is inserted, and goes to circuit when a plug is inserted.

[shooter]

and the Grid still jumped to a ground, I jumped the cap connecting the wiper on the Mic Vol pot to the 1 Meg resistor.

that proves nothing


once you proved the 21khz was "left" of the grid, then it's time to eliminate what's there, so UN-GROUND the grid, jump the cap, test, if still there, jump input jack-TIP to grid-stop/grid leak junction, "bypassing  both cap and pot.


EDIT  DO NOT use "continuity" for testing, use OHMS, continuity might show a "short" when in fact it could be 10-50 OHMS of resistance

[wsscott]

With the jumper to ground removed from the Grid Pin4, The 21khz signal is NOT present when either the cap is jumped or the Input tip to grid/stop/leak is jumped.

[wsscott]

When I jump the Tip of the Input Jack to the lead going into the 1 Meg resistor, not only is the 21khz signal NOT present, but the hum is much quieter.

[tubeswell]

The jack tip switch works fine.  It goes to ground when nothing is inserted, and goes to circuit when a plug is inserted.

How are you so certain? It might appear okay from a quick glance, but the contact could be slightly loose, causing microphonic vibration.


I suggest using a chopstick to mechanically press the tip against the tipswsitch to see if that cancels the noise

[wsscott]

A week ago I checked the jack just as you said, and it functioned properly.  I just checked it again, as you suggested, and it is still functioning as designed.

[shooter]

When I jump the Tip of the Input Jack to the lead going into the 1 Meg resistor, not only is the 21khz signal NOT present, but the hum is much quieter.

ok, jump or cut out the cap, still  ^^^^^^^^^^^^^^^^^^?

[wsscott]

Shooter-As I said in #32, when I jump the cap there’s no 21khz noise.

Seems to me that maybe the 1 Meg resistor is the problem since a direct ground eliminates the oscillation, or whatever is going on.  Is that resistor too big and blocking the path to ground? Should it be smaller or eliminated?

[wsscott]

I decided to check the voltage on the Plates of the 6SN7.  It had previously read an actual voltage on each plate of 89VDC, with the schematic showing 75VDC.  It now reads 31VDC!  Now that's strange.

So I checked its B+ coming from the 10uf cap and it is now reading 190VDC at the cap, and it previously had an actual reading of 215VDC, with the schematic showing a value of 175 VDC.  So that's probably ok.

Then I checked the voltage going into the 100K resistor between that B+ source and the Plates.  It reads 190VDC going into the resistor, BUT 31VDC coming out and connecting to the Plates!  Quite a voltage drop!  I checked the 100K resistors value in circuit and it shows 99K, so its value is okay, but why the big voltage drop?

[shooter]

E (volts) =  I (current) *  R (resistance)


so if E is low and R is constant.............

[wsscott]

Then current is very low.

So the tube is not working properly and not drawing a load?

Is that because the tube has a bias issue?

The caps I used at the cathodes are 25uf/50V vs. 10uf.  I didn’t have 10’s.

Sorry but that’s the best I’ve got as a response without a solution.

[shooter]

did you split the cathodes?
IF yes, what is the VDC reading at each cathode?


did you use 2 1.5k vs original 3K?
IF yes, did you verify they where actually 1.5k n not 15K or 150K or...

[wsscott]

Yes, I split the cathodes.

Cathode Pin 3 = 1.195VDC
Cathode Pin 6 = 1. 188VDC

I used two 1.5 K resistors which read 1.49K and 1.48K

[tubeswell]

And you tried another tube in that input slot?

[wsscott]

At your suggestion, I put in another tube and the readings are the same, just around 30VDC.  A voltage drop through the 100K resistor, which reads 100k in circuit, of about 160VDC.

The voltage is fine going into the resistor but drops dramatically for some reason when it comes out to the plates.  Obviously the tube is not the problem.

Can the resistor read fine like this one does, but is defective in some other way that would cause this?  The voltage read fine after I installed the new resistor a couple weeks ago.

[wsscott]

I went ahead and installed a new 100K resistor, and checked its value before installing, and the results are the same as with the old resistor.  190VDC voltage drop and 31VDC at the plates.  The old resistor also tested fine out of circuit.

[wsscott]

The 2 plates at pins 2 & 6 are jumpered with a single source of voltage being supplied by the 100K resistor.  Since the cathodes are now separated and no longer jumped and sharing a single resistor and cap, would that affect the voltage?

[SEL49]

Since the cathodes are now separated and no longer jumped and sharing a single resistor and cap, would that affect the voltage?

Absolutely! Your original circuit shows a 3K resistor shared by both cathodes. This resulted in 3V bias on the cathode. In reply 19 shooter suggested splitting the cathodes and using two 1.5K cathode resistors. The smaller resistors biased the tubes much hotter and this caused more tube current which caused a bigger voltage drop across your plate resistor. This is wrong! The correct thing to do when splitting shared cathode resistor is to DOUBLE the size of the individual resistors in order to maintain the same bias. So, use two 6.2K resistors and the bias will return to about 3V on each cathode and your plate voltages should increase to a more likeable number.

[wsscott]

Thanks.  I'll give it a try.

[wsscott]

The closest I've got is 6.8K, so I'll give that a try.