sub bass odd harmonics?

[lowell]

this amp is sounding pretty good thus far, I've been tweaking things for 6 months now. however, there is this weird sub bass harmonic thing going on that makes certain intervals sound out of tune. 5ths and octaves sound good but anything else and it sounds awful. This is on both channels but only noticeable on the clean channel when it's almost cranked-cranked. I've thought about the power supply and wonder if the source is my decoupling b/t stage supplies.

below are the schem and power supply.

[loudthud]

You might try doubling up on the 47uF and 32uF series caps. What does the B+ run in the various modes?

[lowell]

not sure what you mean by modes LT...? so you think it's a PA decoupling/filter issue and not the preamp? Just curious what you're thinking behind this is.

[Tom Phillips]

Interesting amp.
Besides the heater CT, where is the 6V6 cathode circuit completed? (Point K)
If it's there in plain sight I missed it.
Tom

[lowell]

Hi yeah sorry I left that out of the schem here on ampage. it's to a Rk not shown, no bypass cap. forget what value Rk is at this moment as I don't have schem in front of me.

[loudthud]

Originally Posted by lowell

not sure what you mean by modes LT...? so you think it's a PA decoupling/filter issue and not the preamp? Just curious what you're thinking behind this is.

Well, there is what looks like a switch in the power supply schematic that bypasses the 5Y3. Another switch looks like it silences the 6L6s but that shouldn't affect the B+.

From your description I think you have a classic case of "ghost noting", usually caused by power supply ripple. In the 6V6 only mode, the 6L6s are still drawing current and causing more supply ripple than if they were turned off somehow. But this may be necessary to keep the B+ from going too high for the 6V6s.

[RudeBoy]

Maybe an intermod product. I wonder if you could be getting some blocking on the cathode biased output tubes? Those 470k are pretty large. You might try reducing C23 and C24 to say - .01u?

Is it actually wired like the schematic? G1 to shield, input sig to screen, supressor grid to what is normally the screen supply?

[guitician]

By "Cranked Cranked" do you mean distorted? Once your output tubes are clipping the harmonics generated are hard to control. Different tubes may have better transfer characteristics, and get less of the objectionable "sub odd harmonics".

[lowell]

interesting. Loudthud is there a way to see this on a scope or w/ a meter?

Sorry if the schem power tubes aren't wired right, they ARE wired correctly in the amp.

I can't image it's tube type that causes this. I've used many 6L6 6V6 amps that do not do this.

[guitician]

Exactly my point. With your particular amp layout/schem you've hit on just the right conditions to produce the "Odd sub bass". It may be that less filter on the PSU will fix it.

[loudthud]

Ghost noting is usually caused by ripple on the B+ that is rejected by a push-pull output stage until the amp starts to clip. There is a certain amount of this in most guitar amps but it's not to the point that it produces undesireable tones. The way to look at this with a scope is to simply look at the speaker output but sync on the power line frequency. The guitar signal will just look like a jumble of random lines but at the top and bottom of the waves you'll see these sawtooth divits. Use a sweep speed of 10 millseconds per division. If the peak to peak voltage of the sawtooth wave gets to be more than (I'm guessing) 20% of the peak to peak voltage of the guitar signal, it should be audible. So you are injecting a 2x line frequency buzz into the guitar signal and that can produce distortion products at frequencies that were not in the original guitar signal.

The first filter in your amp is two 47uF caps in series. Effectively 23.5uF. (You do have equalizing resistors don't you?) That's enough filtering to an amp like a 5E3 but maybe not enough for a 6L6 push pull amp. There are a lot of unknowns in your amp like what the static B+ current and voltage is so I just suggest an experiment to increase the filtering to see if that is what is causing the problem.

[lowell]

hmmm cool. ok so how do I "sync to line" w/ my scope?

Finally, this sawtooth is 120hz because it's been rectified at this point right? I want to understand all aspects of this as best I can.

thanks for your help w/ this!

[loudthud]

Just about any scope will have some way of triggering (starting the horizontal sweep) from an external source or possibly a second channel input.

If your scope has a "Trigger Select" switch with a position for "Line", use that. If there is no position for "Line", select "External" and connect a probe to one side of the heater winding from the "External Input" jack. Use the "Normal" Horizontal mode where the trace won't sweep unless it is triggered. Adjust the trigger level until you get a sweep. Connect the available Vertical input to the speaker output.

Yes, the sawtooth wave should be 2X the line frequency. There could be a bad diode or wiring error in the power supply that would cause the sawtooth to be the same as the line frequency. That may indeed be the problem with your amp. It would cause the ripple to be twice as big as normal. I'm just mentioning this because your amp isn't 100% yet.

[lowell]

ok i'll try that tomorrow.

I was thinking of using the screen supply for my PI plates. I'm wondering if this will cause any crosstalk issue in the filtering, or maybe it's fine? thought I'd pick your brain on that too.

[loudthud]

With a multi-Henry choke between the plate B+ and the screen B+, there should be enough filtering that the PI would not operate abnormally as long as the voltage is not too high.

[lowell]

Ok so I scoped it and I'm not seeing the sawtooth, at least I think. There is definitely clipping and some angled peaks though. I changed the B+ supply to series 100uf and 220uf w/ 220k balancing resistors. Seems that problem persists. Should I up the screen supply cap values? How much current do screens draw?

[loudthud]

The screen supply could do it but the choke would have to be bad. Is the choke something new or from a working amp or something unknown that's been laying around for 20 years?

You still haven't talked about the two switches for the power tubes and the rectifier. Does the amp have to be in one particular mode to make the sub-harmonics? Will it do it when either the 6V6s or 6L6s are removed from the amp?

Have you tried a different speaker? Can you rule out a bad connection in the amp or a bad tube that is affected by vibration? Can you extend the speaker connection so the speaker is across the room and still create the sub harmonics?

[lowell]

I beefed up both supplies, choke is brand new Mercury Magnetics. The tube rectifier switch hasn't been installed yet so just solid state rectifier as of now. The output switch just puts the 20uf cap across the 6L6 grids thus canceling the signal. Could be a bad tube I guess, I've had these tubes in here for the 6 months I've been tweaking this, and I've cycled the power on/off many many times. maybe I've stressed the tubes over time. I'll try new ones. Oh and I also reduced the 6v6 coupling caps to .022uf and their grid leaks to 220k.

[pdf64]

I'd measure the Vac riding on the B+ at the OT and Screen grid supply nodes, at no signal, max sine wave, and max square wave. Best to use a true rms meter, but any should give a reasonable indication.
Also think about the ground return paths for each power supply cap, the power tubes, phase splitter and pre-amp. Best to keep the path for the power tubes and reservoir cap (which are going to carry large ripple currents) seperate from the others. Peter.

[guitician]

A fresh matched set of tubes my changes things indeed.

[lowell]

Originally Posted by pdf64

I'd measure the Vac riding on the B+ at the OT and Screen grid supply nodes, at no signal, max sine wave, and max square wave. Best to use a true rms meter, but any should give a reasonable indication.
Also think about the ground return paths for each power supply cap, the power tubes, phase splitter and pre-amp. Best to keep the path for the power tubes and reservoir cap (which are going to carry large ripple currents) seperate from the others. Peter.

Ok, what am I looking for in the results though? Increased AC on B+ at higher levels?

I originally laid this amp out quite well concerning grounds so pretty sure that's not the issue.

[loudthud]

There has got to be something in your amp causing the problem. Other than the extra gain stage, the amp is pretty close to a 59 Bassman. Everything looks fine except there is that output stage with the 6L6s and the 6V6s. The switch is a little unusual. We are kind of on a treasure hunt here. None of us can see or hear what's going on. Could the output stage be oscillating during parts of the input signal?

Does the amp create sub-harmonics with just the 6V6s alone or with both 6V6s and 6L6s enabled?

Does it do it with either set of tubes removed?

Does it do it with any of the preamp tubes removed? Some if the V numbers in the preamp don't make sense. Are they correct on the schematic? In a preamp with lots of gain you really don't want two tubes in the same bottle if one is too far down the signal chain.

What is the B+ voltage? Is it too much for the 6V6s to handle?

[lowell]

let me know post a correct schem w/ correct tube numbers. i'll look into those other things.

[RudeBoy]

Originally Posted by lowell

Ok so I scoped it and I'm not seeing the sawtooth, at least I think. There is definitely clipping and some angled peaks though. I changed the B+ supply to series 100uf and 220uf w/ 220k balancing resistors. Seems that problem persists.

Do you hear the problem when playing only one note or does it require two or more?

A spectrum analyzer plot of the problem would help. For a two-tone input, if the sub-bass harmonics are related the intervals by f_lo = 2*f1 - f2 and f_hi = 2*f2 - f1 then they are normal distortion products. And if they are prominent and depending on the frequencies of the two fundamentals they can be non harmonically related to the two fundamentals (would sound bad). What you are hearing may not be related to the power supply ripple but might be normal products of a badly distorting amplifier.

Just thought I'd float the idea...

[lowell]

it's only when minor 3rds and 6ths are played. i don't have a spectrum analyser. are they expensive?

[loudthud]

What are the voltages on V6? One side is marked 12AX7 and the other 12AU7. What type tube is in that socket?

[RudeBoy]

Originally Posted by lowell

it's only when minor 3rds and 6ths are played. i don't have a spectrum analyser. are they expensive?

There are a number of FFT type analyzers available that use a PC sound card. A google search will produce several.

[Steve Conner]

My personal favourite is Spectrum Lab. You could easily see 60 or 120Hz sidebands with it. The only drawback is that it has so many features, you could spend a day figuring out how to make it work.

Minor 3rds and 6ths are kind of marginal with distortion, anyway. Church music never changes key because of the long reverberation times of cathedrals, and metal ended up all 5ths and octaves because they had to be intelligible through all that distortion.

[lowell]

very good reminder of some things I forgot Steve.