The Raven - CKD

[Phil_S]

More power.

Sounds like Kirk talking to Scotty.

[xtian]

More power.

Sounds like Kirk talking to Scotty.

"I'm a doctor, not an amp technician!"

[Meat&Beer]

Two bulbs in series = Presto! Thanks shoggoth!

Here's to all of my questions from here on out being as easy as this one. I'll keep you guys up on the progress of this mammoth monstrosity.

Now where did I put my phaser....

[Colossal]

Great job! Quite an effort!

[Reeltarded]

You are a nut omg holy crap and wow what the hell

Neat!! lol

[Meat&Beer]

Ever work on something so long, and determined to get it working right, only to have it still uncooperative, that it becomes… not so fun anymore? I bet you have, and this is pretty much where I’m at with this. Hoping someone can shed some light…

It’s been decided that this needs a complete re-build, due to component location, channel proximity, yada yadda. For more than a month, I've been applying all of my free time on this, all of it. I systematically removed circuits and things not pertaining to just the “Kill” channel, which now has ONE remaining problem we can’t seem to shake.

Reminder; this amp is a two channel 50 watter, with a mixer at the end of the two preamp channels to be able to play through both channels simultaneously if desired.

The problem; There is what sounds like an electrostatic signal coupling to the actual signal. It’s treble-y, harsh, and way distorted. I recorded a clip, and looking at the waveform gave me a new clue. It’s unsymmetrical! In the past, this has meant some sort of bias issue. But, it could also be a parasitic oscillation somewhere?

Soundcloud, the host of this audio file, unfortunately compresses it, and it also does not make clear just how NOT symmetrical it is. So, you get a picture of the real waveform attached as well, in case you’d like a gander.

I strum the strings at zero volume, then just crack it open, strum, bring it up to 25%, strum, up to 50%, strum up to 75%, strum to 100%, with a different chord I descend in volume in the same manner. At less than 10% +/-, the waveform is at its peak in dB. Here it is:

https://soundcloud.com/agormortisrr1/tag-test

Sounds terrible, huh? This is through an Engl E212-Pro loaded with 2xV30’s, recorded with a Sennheizer 421-U5, (About 2.5’ away!) into a Presonus Audiobox, into Adobe Audition with ZERO effects. Just in case you were curious, it’s not a crappy phone recording this, it really does sound that bad. Photo of the amp settings, FFIW. These are the settings that are normally my favorite when playing it. It did behave, for the most part, and in this box, at one time! Both tone stacks engaged, ignore the knobs labeled "off".

The ONLY items left in the enclosure are the bare essentials, one preamp, the mixer, and the power amp.

Here’s what I’ve done in effort to mitigate;

- Tubes
- Changed out every preamp coupling cap
- Every stage is in its own bottle, uncluttered from others
- Tone stacks & volumes, moved and consolidated, given space from other components
- Re-built and checked each component as I did for each stage (Can’t check/verify most caps with the meter I have at this time, however.)
- Yelled profanities at it

Voltages;

Stg 1; Plate 203v - Supply 343v - Cathode - 1.38v (12AX7)
Stg 2; Plate 208v - Supply 346v - Cathode 1.37v (12AX7)
Stg 3; Plate 224v - Supply 360v - Cathode 1.34v (12AX7)
Stg 4; Plate 122v - Supply 360v - Cathode 7v (12AX7)
Stg 5; Plate 48v - Supply 126v - Cathode 2v (12AX7)
Mixer; Plate 227v - Supply 338v - Cathode 110v (12AU7)
P I ; Plate 177v - Supply 372v - Cathode 61v (12AU7)
Bias ; -42.1 (6L6GC)
5U4 Rectifier


I’ve attached the schematic for only the portions that have been described, and that are in question. It’s a LOT of amp in total, I don’t want to get anyone confused, yet… The whole schem will end up on here eventually as it moves along. But trust me, don’t build the whole thing unless you’re really good at this stuff. What a terrible idea for a first build!

This enclosure has become just another prototype. Please excuse the mess.

I have a scope, sig gen, dummy load, a couple meters, Mr. Manning's super fresh light bulb limiter, variac, etc..

Any ideas? I’d love to supply any additional info. Thanks so much for viewing!

[Meat&Beer]

And more for the viewing pleasure;

[xtian]

Since you have a scope and a signal generator, start from the input. Put 1kHz sinewave at 1v p-p at the input, and use the scope to view the sine wave at the plate of V1a. It should be close to 40v p-p, and it should be undistorted.

[Meat&Beer]

I'll admit, I'm not the most savvy 'scopist on the block. The DMM tells me .75v AC, the loudest the sig gen goes. Here are a couple pictures of the display and settings. (Hope I'm not embarrassing myself too much here!)

Edit: Is that actually 1.5V entering the amp says the scope?

Of note; there are no cathode bypass caps in this preamplifier circuit.

I forgot to mention, and I think it's a good clue;

If a 100k is placed from the 270k feeding the grid/Zener diode node to ground at the entry to stage 5, it greatly reduces the fizzy added distortion sound. Stage 4 grid to ground also. More R than that, the good signal really deteriorates in volume. Less, it doesn't have much affect on it the bad signal. I didn't notice much action like this from other stages.

Overall volume goes down a bit when these r's are implemented, but the undesirable signal goes down MUCH more in comparison.

More poking around after work...

[xtian]

Your scope shows the signal looking perfect after the first gain stage. So now move your probe to the plate of the next stage, etc.

If a 100k is placed from the 270k feeding the grid/Zener diode node to ground at the entry to stage 5, it greatly reduces the fizzy added distortion sound.

By adding that 100K (making a voltage divider) you're reducing the level of the signal. So it sounds like you're driving your cascaded gain stages WAY to hard, causing all kinds of ugly distortion.

You're on the right track.

[Meat&Beer]

Some light is being shone, I do thinketh.

The 1k signal can be followed with the scope all the way through the preamp until the front of stage 5 with great results visually. Adjusting each "volume" knob along the way to deliver a beautifully clean sine wave along the path, backing it off to just before clipping.

Attached is a closeup of the stage five schemtastic.

At the node of the two 270k's and diodes, the signal is smashed in distortion, top and bottom of the wave FLATTENED. Before the first 270k, a beautiful princess wave shape exists.

Disconnected the diodes, the bottom of the wave looks great, the top still severely distorted. (At the junction of the two 270k's again) I tried doin' this stuff that didn't work;

- Disconnect voltage regulator, made it its own "normal" supply of +/- 120v
- Checked the 270k connected to grid (Lifted from grid, newly opened leg looked perfect on the scope)
- Tube. Again. Nope.
- Swapped out 100k plate resistor
- 4.7M and then a 1M grid to ground (Is this called a grid leak? Geepers.)
- Already swapped out coupling cap and cathode resistor beforehand.

We know it is not the mixer, which resides directly after this stage. (I tried jumping over it from stage 5 exit to PI) It is not in the power amp, when I had an FX loop installed it sounded perfect with guitar into the return jack. (I also have lots more evidence why it's not PA induced.)

It's not anything connected after this stage, (off the plate) because I unhooked that wire and nothing cleaned up at the grid junction.

After these tests, settings were for "clean" volume, I plugged a guitar in. I could of course still hear the crunchy crap signal, but if plucked very lightly, guitar was clean... Probably doesn't amount to anything though. Does it? (Two guitars actually, one with actives, one without. Same resuslts!)

Does anyone see something wrong with this stage?

[shoggoth]

Try reducing the 10k cathode resistor on the 4th stage, it may be biased too negative. Or maybe reducing the 330k plate resistor.

If that's not it, maybe the coupling cap between 3rd and 4th is leaking DC, causing the 4th stage to be biased funny.

[shoggoth]

Oh, I didn't see the bit where you said the wave was fine after stage 4 but before the zeners.

If that's still the case (signal fine before the 270k resistors), I don't see anything that would clip a signal at that point in the schematic. Maybe you're driving the grid positive on the 5th stage? Grid-to-cathode resistance becomes very low when that happens.

If that's so, try adding a resistor between the coupling cap and 100k/bright cap junction to attenuate the signal before it hits the grid.

The zeners you removed may have been doing more than clipping, they may also have been keeping the signal voltage within the headroom of the 5th stage.

[ampdoc1]

I don't mean to be a buzzkill, but I'd bet very few builders on this forum (probably none), started out like this. You weren't satisfied to build a Model T to learn the basics,...you went straight to the Maserati!

And, not to criticize, but with your wiring going in ALL directions, crossing over one another, etc...my money would be on "this will never sound good".

Hone your skills by starting simple.

But, I do wish you good luck!



a'doc1

[xtian]

I don't mean to be a buzzkill

We already did that; see top of thread.

Now OP is troubleshooting like a boss.