Completely rebuilt my amp, strange sound when moving hand over chassis...?

[guitylerham]

Hey guys, so I posted a few weeks ago about my single channel AB763 w/o trem. Was using donor transformers, had some issues. Fixed some of them but just wasn't happy with how it turned out. Sooo, I ordered new correct transformers and completely gutted the amp and rebuilt it. This time, much more care was taken in layout and organization.

Results:

I have good voltages on the pins, decent signal to the speaker, all controls work. However, I am getting some ugly distortion at moderate volumes, even when I pull the reverb driver. I added a Mid pot so when it's cranked it defeats the tone stack. The ugly distortion rears its head more easily with the Mid cranked. Originally, I was getting some nasty squealing (kinda like a dj infinitely dragging a needle across the record) that would "wind up" by getting louder and higher pitch as I turned down the volume to zero, at that point the squeal was gone. I could recreate that effect over and again. I swapped the OT secondary leads and that seems to have cured that issue. But the distortion is still there. And what's strange is it seems to be related to me touching the chassis with my hands. So static interference? When I signal trace the circuit, I don't hear any distortion along the signal path up until the plates of the PI, from that point to the output tubes the signal is massively distorted at really any volume setting. This was the case before so I can only assume it's normal due to high voltage (50vdc about) into my listening headphones and I don't hear the same quality coming through the speaker. I've chopsticked all the wires and components to see if I can alter the fizzy distortion. If I boost the treble or turn up the volume a little and hover my hand over the preamp side of the chassis, in the right spot I can get a motor boating sound almost like a quick choppy tremelo. I've never had that happen before. Layout issue? Steel chassis. I've measured all bypass caps to ground, made sure my grounding buss reads continuous with chassis (is 0.1-0.3 ohms pretty close?). I've traced the signal with the schematic so I think everything is basically correct. I added a nfb defeat switch that breaks contact between the 820R and the junction of 22k/01.uf/47R components. I assume that's a legit way to do that. I don't hear much of a difference, to be honest though.

I'll keep picking away at it but I wanted to pose some of my observations to see if anything sounds textbook incorrect. It's just funny to me that having rebuilt everything with new resistors and caps too that I'm still dealing with inferior sound. One of the common between now and then are my preamp and output tubes. I've been swapping preamp tubes around to no avail. Grounding scheme is a floating buss bar, at one is the HT CT and at the other is the first catchode bypass cap. The preamp side of the buss bar is grounded at the input jack, so only two points attach to the chassis (earth ground is the other). Amp is pretty hissy though, unfortunately. Lots of noise from the reverb circtuit too. I added a Dwell control which replaces the 1M to ground right after the 500pf reverb cap, wiper goes to the input grid of the driver.













It ain't as pretty as some of you guys incredibly do but each time is getting better for me.

[guitylerham]

Doing some more research, I'm lead to believe that my PI outputs should NOT be distorted! If they are, what could that describe?

[eleventeen]

One thing I notice is that your reverb can connectors are right next to your tone stack components. That could be a problem. That's a high strength signal out of the reverb tranny. Your body could also be acting as an antenna if you have dimmers or fluorescent lights in your shop. Those Ajax and the other blue molded caps...are new?


The HV CT should be lashed up to a PT bolt, not running a "floating" ground bus running down the length of the amp, behind the tone controls, and connecting to the preamp cathodes (through their bypass parts) Same with the 6.3 CT. You know, you have ".1 ohm to .3 ohm to ground (whatever ground is) from some part of your ground bus, but the idea that your HV CT isn't BOLTED to ground right next to the tranny ...is not a happy picture in my book.


You have two light blue molded .056 caps. I don't see where in the schematic those are.


I think the thing you really need to establish is the brutal no-questions-asked ground right at the PT. It is more important to do that with a steel chassis than an aluminum one because steel is not that great a conductor. Indeed, the only bolt-to-the-chassis-with star-washer grounds I see are your AC line ground and the reverb ground. I'm not seeing (that does not mean it doesn't exist, it just means I am not seeing it) a "signal" ground chassis connection near where the preamp signal is. If the P/S ground is not established MUCH closer to the PT than whatever you're using for the signal ground, you could be in for a world of hurt, noise wise. I suspect that's at least 75% of your problems.

[Willabe]

I think you PI is wired up wrong and some wires coming from the PI right across the output speaker jack, that can't be good. Input right next to output, 180 degrees out of phase = oscillate. 

The output tube grid wires look pretty long, ~1'?, they should be shielded as well as a few other grid wires in the amp, long and unshielded = problems.

There's 2 caps that don't seem to be soldered up in the top turret holes.

And you have a LOT of wires bundled up and running across the top of the pots. I would have run most if not all under the pots in the amps chassis corner away form the pot connections.

I agree with 11teen about the ground.

[guitylerham]

I think you PI is wired up wrong and some wires coming from the PI right across the output speaker jack, that can't be good. Might ocliate. 

There's 2 caps that don't seem to be soldered up in the top turret holes.

And you have a LOT of wires bundled up and running across the top of the pots.

The output tube grid wires look pretty long, they should be shielded as well as a few other grid wires in the amp, long and unshielded = problems.

Good eye! The top of that 0.047uf cap is actually soldered to the yellow lead that disappears behind the green ground buss wire. It's just not visible in the picture. And the other cap with the questionable solder connect at the bottom turret is definitely soldered and not lose. The two blue caps are definitely soldered to those top turrets, they just didn't fill with solder.

I do have lots of wires running up along those pots. I've seen it done before but perhaps it's just not good practice. Looks tidy tho!

And those blue output tube grid wires are unnecessarily long. I'll shorten those up. I was under the impression that an correctly built amp shouldn't require shielded wires anywhere so that was my first goal. I can always come back and add some though. Thanks, Willabe!

[guitylerham]

One thing I notice is that your reverb can connectors are right next to your tone stack components. That could be a problem. That's a high strength signal out of the reverb tranny. Your body could also be acting as an antenna if you have dimmers or fluorescent lights in your shop. Those Ajax and the other blue molded caps...are new?


The HV CT should be lashed up to a PT bolt, not running a "floating" ground bus running down the length of the amp, behind the tone controls, and connecting to the preamp cathodes (through their bypass parts) Same with the 6.3 CT. You know, you have ".1 ohm to .3 ohm to ground (whatever ground is) from some part of your ground bus, but the idea that your HV CT isn't BOLTED to ground right next to the tranny ...is not a happy picture in my book.


You have two light blue molded .056 caps. I don't see where in the schematic those are.


I think the thing you really need to establish is the brutal no-questions-asked ground right at the PT. It is more important to do that with a steel chassis than an aluminum one because steel is not that great a conductor. Indeed, the only bolt-to-the-chassis-with star-washer grounds I see are your AC line ground and the reverb ground. I'm not seeing (that does not mean it doesn't exist, it just means I am not seeing it) a "signal" ground chassis connection near where the preamp signal is. If the P/S ground is not established MUCH closer to the PT than whatever you're using for the signal ground, you could be in for a world of hurt, noise wise. I suspect that's at least 75% of your problems.

Yeah, I questioned my placement of those reverb jacks so close to the preamp tube.

Those Ajax caps are not new, I acquired a bunch and wanted to see if all the legend of their superiority is true. Very possible they are leaky.

As far as the CT, I can only defer to your expertise on that. I've read that ideally, you have one connection to chassis for Earth, and another for signal. So I have a buss bar that is grounded only at the preamp end to the input jack/chassis. The other end has the CT, resevour caps, power tube cathodes, 6.3 CT. I've grounded the circuit using the flow of the schematic as you travel from the input to the output. You think this is still inferior design? The HT CT needs to be physically close to the transformer or does it just need to be at the same potential as the circuit it's connected to?

Here's the end of my green "buss wire" and it's connection to chassis/input jack via the green cloth wire.


The 0.056uf caps were a choice I made in lowering those values from 0.1 as per the schematic because I've heard people say it tames some of the low end when you're pushing the amp hard. I will go ahead and change those back to stock to at least eliminate them as a variable. I was admittedly jumping the gun hoping I'd get different result than I did the last time I built the amp.

Thanks Eleventeen!

[Willabe]

I do have lots of wires running up along those pots. I've seen it done before but perhaps it's just not good practice.

We've all seen wires run along top of the pots but I've never seen that many, that long, traveling across that many pots, that I can recall.

The bottom side of the pot, opposite the solder lugs is shielded. If you run the wires down in the bend of the chassis then those wires would be away from the solder lugs. Certain wires running right across those lugs, their signal might mess with the signal going in and/or out of those pots.

I was under the impression that an correctly built amp shouldn't require shielded wires anywhere so that was my first goal. I can always come back and add some though.

Well define 'correctly built'. The way you have your layout, you ended up with a lot of extra wire length.

It helps greatly to keep all the capacitors (C's) and resistors (R's) of the tubes circuit they go with as close as possible to their tube, but you can't always do that. You want short grid wires, your output tube grid wires look to be at least ~1'.


                  Brad   

[guitylerham]

Ah, certainly the bottom of the pots are shielded! I'll rewire the pots now. Yeah, my layout was far from correct. My point was that I could only hope it was good enough and then come back and touch stuff up later if needbe. It needbe now.

The output grid wires are damn near 8". I more concerned with diverting them away and at right angles to other wires more so than the most direct route to the sockets. I'll redo all those long wires and see what happens. Live and learn! Thanks, Brad.

[Willabe]

I think you PI is wired up wrong and some wires coming from the PI right across the output speaker jack, that can't be good. Input right next to output, 180 degrees out of phase = oscillate.

You think your having distorted signal from the PI output, from what I think I see in your picture the PI looks like it's wired up wrong.

[guitylerham]

I think you PI is wired up wrong and some wires coming from the PI right across the output speaker jack, that can't be good. Input right next to output, 180 degrees out of phase = oscillate.

You think your having distorted signal from the PI output, from what I think I see in your picture the PI looks like it's wired up wrong.

I'm gonna go over it all again. Creating new accurate layouts on the fly often doesn't go so smoothly...

[Willabe]

I think you have the 2 PI plate resistors flipped with their power tube grid R's

It looks like your feeding the junction of the 2x220K power tube grid leak R's the B+. The junction of these 2x220K's  should be fed the -grid bias.

And it looks like the 100K and 82K PI plate R's junction, instead of being fed B+ are going to ground?   

[eleventeen]

" The HT CT needs to be physically close to the transformer or does it just need to be at the same potential as the circuit it's connected to?"


When you or I say "same potential"....long experience teaches us that we need to follow what we call "best practices" and not think we know better on certain things that have tried a zillion times. 


"So I have a buss bar that is grounded only at the preamp end to the input jack/chassis."


See, this is a formula for trouble. The CT of the HV winding is NEVER *exactly* at the center of the HV winding. It's nominally 5-10 volts off, perhaps more, perhaps considerably more. Go ahead and pull your GZ34 and measure OHMS from the HVCT to the plate pins of the GZ34. Oh, they are pretty close. But not exact. It's VERY common for there to be a few ohms difference. Is that a big deal? Well, as far the production of B+ through the rectifier, no, it's not a big deal. If I could instantaneously measure the peaks of the pulsating DC that are fed to the first filter cap, what would I find? [I *could* of course do this with a 'scope] I would find one phase is (let's say) 425 volts and the alternate phase is 420 volts or 430 volts. Maybe more of a difference. Well this gets evened out through the entire filter cap network into one pretty steady DC level *relative* to a ground established by lashing the HVCT to ground. *IF* that CT with its +/- 5 volts (and maybe more) is lashed up to ground as close to the rectifier as possible it will make no difference as far as our B+ is concerned. Our filter caps will just average out the two voltages into one steady voltage level and nobody will know. EVERY TRANSFORMER that has ever been wound by man has this kind of variation. But now, let us stand on the cathode of V1, the first preamp tube. From that standpoint, if the 5 or 10 volts of DC<---should be "AC" has not been killed or swamped within the rectifier section (by forcing it to ground ASAP) then that AC is being fed to the first preamp tube....because that tube isn't really referenced to "zero"...it's referenced to this shifty pseudo-zero "wire". That tube amplifies the raw guitar signal by 100x by the time it gets through the amp. Maybe all the filter caps you've installed along your bus wire whack that 5 volts of hum by 100, so it is only .05 volts. It still gets amplified by 100x by the action of the preamp....and 5 volts of hum through your amp is definitely not nothing.


The point is, we don't want to intellectualize this. The CTs have to be grounded near as possible to the PT and that's all you really need to know (and execute) It's urgent that whatever hum exists within the power supply is kept out signal_land as much as possible. It can cause nothing but trouble. One way we do this is to provide enough filtering with our e-caps. The other way is to FORCE the ground of our B+ to be "zero" BEFORE it gets to anything that handles signal(s). All of the signal section utilizes the same ground reference. And then, at ONE PLACE, we connect the signal ground bus to a place AS CLOSE AS POSSIBLE to that bolt-through HVCT ground. This is the way to avoid problems.

[guitylerham]

" The HT CT needs to be physically close to the transformer or does it just need to be at the same potential as the circuit it's connected to?"


When you or I say "same potential"....long experience teaches us that we need to follow what we call "best practices" and not think we know better on certain things that have tried a zillion times. 


"So I have a buss bar that is grounded only at the preamp end to the input jack/chassis."

With that explanation, I see no reason to question it! That makes much sense. I'll change that as soon as I can get back to it. I'll keep you all posted. Thanks a million for these pointed insights!

See, this is a formula for trouble. The CT of the HV winding is NEVER *exactly* at the center of the HV winding. It's nominally 5-10 volts off, perhaps more, perhaps considerably more. Go ahead and pull your GZ34 and measure OHMS from the HVCT to the plate pins of the GZ34. Oh, they are pretty close. But not exact. It's VERY common for there to be a few ohms difference. Is that a big deal? Well, as far the production of B+ through the rectifier, no, it's not a big deal. If I could instantaneously measure the peaks of the pulsating DC that are fed to the first filter cap, what would I find? I would find one phase is (let's say) 425 volts and the alternate phase is 420 volts or 430 volts. Maybe more of a difference. Well this gets evened out through the entire filter cap network into one pretty steady DC level *relative* to a ground established by lashing the HVCT to ground. *IF* that CT with its +/- 5 volts (and maybe more) is lashed up to ground as close to the rectifier as possible it will make no difference as far as our B+ is concerned. Our filter caps will just average out the two voltages into one steady voltage level and nobody will know. EVERY TRANSFORMER that has ever been wound by man has this kind of variation. But now, let us stand on the cathode of V1, the first preamp tube. From that standpoint, if the 5 or 10 volts of DC has not been killed or swamped within the rectifier section (by forcing it to ground ASAP) then that AC is being fed to the first preamp tube....because that tube isn't really referenced to "zero"...it's referenced to this shifty pseudo-zero "wire". That tube amplifies the raw guitar signal by 100x by the time it gets through the amp. Maybe all the filter caps you've installed along your bus wire whack that 5 volts of hum by 100, so it is only .05 volts. It still gets amplified by 100x by the action of the preamp....and 5 volts of hum through your amp is definitely not nothing.


The point is, we don't want to intellectualize this. The CTs have to be grounded near as possible to the PT and that's all you really need to know (and execute) It's urgent that whatever hum exists within the power supply is kept out signal_land as much as possible. It can cause nothing but trouble. One way we do this is to provide enough filtering with our e-caps. The other way is to FORCE the ground of our B+ to be "zero" BEFORE it gets to anything that handles signal(s). All of the signal section utilizes the same ground reference. And then, at ONE PLACE, we connect the signal ground bus to a place AS CLOSE AS POSSIBLE to that bolt-through HVCT ground. This is the way to avoid problems.

Odd, my reply didn't post!

What I said was thank you so much for explaining that to me in straight forward manner. Makes perfect sense. I'll switch that as soon as I get back from my gig. Thanks again everybody, we're on course for resolution!

[PRR]

> normal due to high voltage (50vdc about) into my listening headphones

Why do you have high DC in your listening headphones?

[guitylerham]

> normal due to high voltage (50vdc about) into my listening headphones

Why do you have high DC in your listening headphones?

No, that was just a dumb statement on my part. The capacitor in the audio probe filters all DC.

But my assumption was that perhaps the PI amplifies the signal so much that it over drives my solid state listening amp input but that was an uneducated guess and I might be wrong. It's just that the last time I built the amp, I had the exact same observation when listening to the PI output.

[guitylerham]

A diagram for your viewing convenience. This is what I have wired in my amp now pertaining to the phase inverter and proximate components. It's a bit clearer to see than the photographs.

So just to clarify:
-I am in fact feeding the 220k resistors negative bias voltage from the bias pot.
-The 470K screen resistors on the output tubes have been upped to 1k 3w and are designated in the diagram in the far left.
-The output tube grids have the 1k5 resistors soldered to the pins 5 and 6 (using 6 as an empty terminal)
-I hope someone spots an error in my PI wiring and the "cross" of resistors because I sure don't see one yet, even though it's confusing.
-The NFB switch just breaks contact between the 820R and the junction of the 22k and the 47R (100R on the schematic, fyi) to ground.
 


Tomorrow, I will finish adding shielded output tube input grid wires (grounding one end of the shield at the output tube cathodes, which are also tied to ground).

I will also look into tidying up more of the other wiring. Not sure what to do about the reverb jacks and transformer wires yet. But I'll see what happens after I fix some of these blatant errors.

I've grounded the HT CT to a chassis bolt and am also considering splitting the ground buss bar between C and D. Running A, B, and C to the transformer bolt ground, and then leaving D grounded at the input jack as per Hoffman's scheme. I'll do some more reading too. Now I sleep. Stay tuned.

[guitylerham]

Ok! The amp is sounding pretty good now. I did the following:

-kept the HT CT attached to the first A and B filter cap grounds and then ran a wire from there to the transformer bolt. I then seperated the ground buss at the C and D junction, running the D ground to the input jack.

-ran shielded wires to the grids of the power tubes
-ran shielded wire from the output of the first preamp tube to the volume pot (was a little noisy when I tapped it).

-ran the pot wires underneath the pots instead of across the tops.

The amp seems to be much quieter and doesn't have any of the gross distortion that was happening earlier. There is just a constant hiss happening that doesn't change with the volume. The amp actually sounds pretty quiet as far as hum goes when I turn the volume up and down, only constant hiss is apparent, no hum. The reverb circuit is really noisy. When I turn up the Reverb control it gets drastically noisy with pops and such, then when I turn up the Dwell, it gets even noisier with a "woooosh" sound. So that's unfortunate. I'll try and track down what could be causing hisses. I used CC resistors everywhere but the grids of the preamp tubes, that can't be the only cause of the hiss but maybe so. What causes random pops? Static?

Also, the amp is not as loud as I was expecting. It's basically feels like the same SPL as my 8-10w tweed Harvard that I built with a 10". Seem strange? The amp breaks up about halfway. I put a 12au7 in V1 to tame the breakup and when cranked, it's cleaner but just not very loud. My friend's DRRI sounds very loud if memory serves me correctly. I have a 12" 75w Celestian in there now. Even an old Jensen c12n sounded similar in volume. It's just not a punchy amp at all. Perhaps that's the nature of this amp.

Think it'd be worth throwing in another filter cap/smoothing resistor to feed just V1? As it stands, the whole preamp/reverb is just powered by one filter cap node (D). I wonder what the results would be.

So this is where I stand with the amp. I'm optimistic! Now it's just fine tuning all the crap that I don't like. Any thoughts come to mind with you guys?

[Toxophilite]

Carbon comp resistors = hiss
They add a lot of hiss to the amps sound, in my opinion it's their major contribution sonically.
If you are set on using them you have to accept the hiss.
However you could have something else gong on but the carbon resistors won't help in that department

[guitylerham]

Carbon comp resistors = hiss
They add a lot of hiss to the amps sound, in my opinion it's their major contribution sonically.
If you are set on using them you have to accept the hiss.
However you could have something else gong on but the carbon resistors won't help in that department

I dig. So maybe try replacing the 100k plate for the preamp tubes first? What about the cathode bypass resistors, do those add noise?

[Willabe]

So maybe try replacing the 100k plate for the preamp tubes first?
What about the cathode bypass resistors, do those add noise?

Yes and yes.

The debate goes that some believe that a hi(ish) voltage (100v+?) across a carbon composition (CC) resistor causes a non-linear distortion that makes the amps tone warmer/sweeter. But even if true, then grid R's and most cathode (K) R's would not be affected, because they only normally have a few volts across them. (Cathode followers (CF) and split load PI's K R's do have a hi voltage across them.)

So the place to use CCR's (if you think they add to the amps tone) would be for a plate R and where there's a hi voltage on a K R. Using a CCR for a grid leak R or for the series input jack R's or most K R's will only add noise/hiss.

Remember the 1st stage preamp tubes plate/grid/K and input jack R's, the slightest noise introduced because of them will be amplified through each stage after. It's signal to noise ratio, each amplifier stage after the 1st will have a larger signal that will help swamp out the noise.

And the output from the reverb tank going to the verb recovery gain stage has even less signal than from a guitars pick ups (PUP's) so that's a gain stage to be careful with.
           


               Brad     

[eleventeen]

When CC's are new, they seldom introduce much hiss. It's when they are older they cause noise. I've found the 100K plate resistors are the worst offenders and in an old amp that hasn't been up on a bench for many years, I usually shotgun-replace all of them with 2 or 3 watt metal films. Of course in a new build those go right in, don't even think about it. Those plate resistors will fix 90% of hiss in a typical old Fender, there really isn't a cheaper, more dramatic fix you can do to an old amp. Grid leaks, too, while you're at it.


Now this is hiss....although really old plate R's will make rodent noises, squeals, and intermittent noises...they don't especially stop an amp from working. From your description you have other stuff going on. I myself didn't see anything wrong with your PI wiring.

[guitylerham]

So maybe try replacing the 100k plate for the preamp tubes first?
What about the cathode bypass resistors, do those add noise?

Yes and yes.

The debate goes that some believe that a hi(ish) voltage (100v+?) across a carbon composition (CC) resistor causes a non-linear distortion that makes the amps tone warmer/sweeter. But even if true, then grid R's and most cathode (K) R's would not be affected, because they only normally have a few volts across them. (Cathode followers (CF) and split load PI's K R's do have a hi voltage across them.)

So the place to use CCR's (if you think they add to the amps tone) would be for a plate R and where there's a hi voltage on a K R. Using a CCR for a grid leak R or for the series input jack R's or most K R's will only add noise/hiss.

Remember the 1st stage preamp tubes plate/grid/K and input jack R's, the slightest noise introduced because of them will be amplified through each stage after. It's signal to noise ratio, each amplifier stage after the 1st will have a larger signal that will help swamp out the noise.

And the output from the reverb tank going to the verb recovery gain stage has even less signal than from a guitars pick ups (PUP's) so that's a gain stage to be careful with.
           
Exactly the sort of things I've read about but haven't tested for yet. My only piece of observation is that the hiss doesn't change with the volume control. So my assumption is that whatever is adding hiss comes after the volume. There are a few gain stages following the volume that could be the culprit. I Think the strength of the reverb after the tank is especially strong reason to shield that wire!

               Brad     

[PRR]

> my assumption was that perhaps the PI amplifies the signal so much that it over drives my solid state listening amp input

Very possible. Signal levels run from 0.02V at the guitar jack to maybe 1V-2V in much of the amp to over 50V out of the driver to the power stage. Most of the things you would use for a Listening Amp might take 1V clean, but maybe not 50V clean.

This should be an imPRESSive amp. I wish I knew why it isn't.

[guitylerham]

Well, just got back from a gig with my amp (yeah, risky running a brand new amp on a gig but I brought a backup amp just in case). Besides the room with its mega interference, which left me unable to play my single coil guitar in any pickup position other than humbucking mode, the amp performed pretty well. It's incredible how your ears get used to sound pressure working on an amp in your shop and then at the gig on stage, the amp sounds so much louder. I must be careful with my ears in the future if cranking the amp didn't seem very loud. At the gig, I ran the volume at about a quarter plus for the first half of the night and bumped it to halfway for the rest. Was a nice mix of "clean enough" for the Motown/soul stuff and then grindier when I'd dig in. I didn't really notice the hiss at the gig but my next step might be to replace all the 100k plate resistors with metal film to further quiet it. The reverb circuit is still really noisy. I'll keep replacing wires with shielded ones until it quiets down a bit too.

So overall, I think the amp is in pretty decent shape. I think my earlier grounding scheme was just not correct. The signal at the PI is now clear when before it was saturated. I noticed the amp seemed to lose some of it's definition or punch as the night went on but I think that's normal behavior for tube amps when heat builds up. I might install a DC fan inside the cab. I'll post some more pics of the amp tomorrow so you can see what it looks like. Thank you for all your very informative and explicit directions everyone. Your input made it possible to get a working amp, finally!

[guitylerham]

I wanted to thank everyone for helping me get my amp up and running. It's doing so much better than before.

The thing I've learned with this build: Make sure you have a solid and functioning ground scheme otherwise you'll spend all your efforts trying to chase symptoms and infer what is causing them when actually one of the most fundamental aspects isn't correct in the first place. Amps are so complex and the behavior of electrons can appear to be so random, you have to do your best to eliminate the most common mistakes. Anyway, I can always depend on everyone here to offer up quick advice when I'm in a pinch. Thank you!

[sluckey]

That looks good! Kinda begs for a Cannabis leaf, some mushrooms and a peace symbol on the grill cloth.   

[tubenit]

Very nice paisley covering!  That is a very sharp looking amp cab.  Thanks for sharing the pics.

with respect, Tubenit

[eleventeen]

Very nice looking amp! Glad you got it ironed out.

[sluckey]

Looking back thru your gut shots I see at least two pieces of loose wire laying on the board. They are long enough to roll around and put a short between terminals. I also see several solder droppings and smaller pieces of wire inside the chassis. I hope you will take the time to get all that stuff out of the amp.

[guitylerham]

Looking back thru your gut shots I see at least two pieces of loose wire laying on the board. They are long enough to roll around and put a short between terminals. I also see several solder droppings and smaller pieces of wire inside the chassis. I hope you will take the time to get all that stuff out of the amp.

Thanks, Sluckey. I did run a can of compressed air throughout the chassis to remove all the loose stuff. I'm in the process of tracking down the hiss now. The hum stays present as I remove the first two preamp tubes but once I remove the third reverb recovery tube, the hiss is gone completely (when that happens, there is an obvious 60hz hum as well). I've replaced the plate load resistors with metal film but the hiss remains. So I'll probably come back with some metal film 1k5 cathode bypass resistors next. I've been swapping tubes as well to no avail. Pretty much all of the wires coming off the reverb recovery tube are microphonic regardless of the tube I have in there. Not sure there's much I can do about that though. I am so impressed when you guys can build your amps and have them turn out silent without noise and gremlins. That's my goal someday!

[sluckey]

I did run a can of compressed air throughout the chassis to remove all the loose stuff.

That works sometimes. Other times it just hides the loose stuff even better. I'm a firm believer in picking that stuff out as soon as I see it. Needle nose and a toothpick with a bit of paste flux on the end will pick up even tiny stubborn pieces.

The hum stays present as I remove the first two preamp tubes but once I remove the third reverb recovery tube, the hiss is gone completely (when that happens, there is an obvious 60hz hum as well).

That kinda narrows it down to the third tube circuit. Remember, there are two triodes in that tube. One is the reverb recovery and the other is the third preamp gain stage.

I think your layout  and wiring are major contributing factors in your noise and hum, especially in the very sensitive reverb recovery circuit. I would move that reverb return jack over next to V3, mount the 220K directly on the RCA jack, and run a short wire directly to the grid of V3.

You also have a lot of very long unshielded wires that are carrying low level signals. Talking about wires from your Dwell and Reverb pots, volume pot, and input jacks. Plenty of opportunity for noise and hum to sneak in on those long wires.

Those are just a few of the issues I see. There are plenty of other examples of wiring that were not well planned. Take a look at the Fender layout for the Deluxe Reverb or study Hoffman's AB763 layout. I think you have a big challenge with the layout as is.
 

[guitylerham]

I did run a can of compressed air throughout the chassis to remove all the loose stuff.

That works sometimes. Other times it just hides the loose stuff even better. I'm a firm believer in picking that stuff out as soon as I see it. Needle nose and a toothpick with a bit of paste flux on the end will pick up even tiny stubborn pieces.

The hum stays present as I remove the first two preamp tubes but once I remove the third reverb recovery tube, the hiss is gone completely (when that happens, there is an obvious 60hz hum as well).

That kinda narrows it down to the third tube circuit. Remember, there are two triodes in that tube. One is the reverb recovery and the other is the third preamp gain stage.

I think your layout  and wiring are major contributing factors in your noise and hum, especially in the very sensitive reverb recovery circuit. I would move that reverb return jack over next to V3, mount the 220K directly on the RCA jack, and run a short wire directly to the grid of V3.

You also have a lot of very long unshielded wires that are carrying low level signals. Talking about wires from your Dwell and Reverb pots, volume pot, and input jacks. Plenty of opportunity for noise and hum to sneak in on those long wires.

Those are just a few of the issues I see. There are plenty of other examples of wiring that were not well planned. Take a look at the Fender layout for the Deluxe Reverb or study Hoffman's AB763 layout. I think you have a big challenge with the layout as is.

I'm afraid you're right, Sluckey. My layout is and will be an uphill battle. That's unfortunate. I'll do my best to shield what I can and after that I'll consider my options as far as layout. I'll apply your suggestions and see what that yields.

[eleventeen]

One thing you can do, if we now believe one or the other of V3s' triodes are at fault, is to ground the grids of one tube-half at a time. A clip-lead is probably the best way. But the suspicion has to lean towards the rev recovery tube.

[Willabe]

That works sometimes. Other times it just hides the loose stuff even better. I'm a firm believer in picking that stuff out as soon as I see it. Needle nose and a toothpick with a bit of paste flux on the end will pick up even tiny stubborn pieces.

Agree. I do the same so I don't forget. I also use a small, cheap paint brush that I keep on my work bench. I also actually will pick up the chassis and hold it up in the air upside down and shake it out.

Needle nose and a toothpick with a bit of paste flux on the end will pick up even tiny stubborn pieces.

Great idea.

You also have a lot of very long unshielded wires that are carrying low level signals. Talking about wires from your Dwell and Reverb pots, volume pot, and input jacks. Plenty of opportunity for noise and hum to sneak in on those long wires.

Once the plate wire signal goes through the plate/grid coupling cap (also called blocking cap, because it blocks the plate dcv) it is now a grid wire. Grid wires act like an antenna. They take in noise and can cause instability.

The grid wires often need to be shielded wire, with their shield only connected at 1 end to ground. A few inches in most places won't be a problem but making long(er) runs from the tube socket at the back of the chassis to controls on the front of the chassis then back to the eyelet board or tube socket and from the PI coupling caps to the output tube grid stoppers, that's where you need shielded wire. 

   
               Brad   

[guitylerham]

One thing you can do, if we now believe one or the other of V3s' triodes are at fault, is to ground the grids of one tube-half at a time. A clip-lead is probably the best way. But the suspicion has to lean towards the rev recovery tube.

Using a clip to ground, when I short V1-A pin 7 (input grid after volume pot) much of the hiss goes away, only a tiny but remains. When I ground V3-A (input grid after reverb tank), a little bit of the his goes away but a large amount of perhaps 120hz goes away, nearly all of it. And when I ground V3-B pin 7, everything goes away except low 60hz hum. I like these clues!

[guitylerham]

Once the plate wire signal goes through the plate/grid coupling cap (also called blocking cap, because it blocks the plate dcv) it is now a grid wire. Grid wires act like an antenna. They take in noise and can cause instability.

The grid wires often need to be shielded wire, with their shield only connected at 1 end to ground. A few inches in most places won't be a problem but making long(er) runs from the tube socket at the back of the chassis to controls on the front of the chassis then back to the eyelet board or tube socket and from the PI coupling caps to the output tube grid stoppers, that's where you need shielded wire. 

   
               Brad   

Ah! That makes sense. I'm always curious where one part of a circuit begins and another ends. I hadn't thought that way before about the grid/plate divide. I think I should just go back and shield everything going to the pots on the front panel. I just have a hesitation to use shielded wire everywhere for the treble sucking that I've read can happen due to capacitance. Lesser of two evils?

[Willabe]

I'd try using shielded wire for the 3 grid wire runs you just grounded, that seems to be where the noise is getting in.

They look to be pretty long going from the tube sockets/eyelet board then across the pots.

Do the longest runs 1st and see if those are enough.

And yes, shielded wire is the lesser of 2 evils but you won't lose enough hi end to notice it.   
 

[eleventeen]

"Lesser of two evils?"


By far.


Willabe gave an excellent explanation of how this works. This is one reason why, despite our general wish to have neat-looking layouts on parts boards, there is something to be said for point-to-point wiring. Short wires. Often a tad easier to troubleshoot, too.


Incidentally, that 220K R on the reverb recovery jack is a known culprit.

[guitylerham]

Incidentally, that 220K R on the reverb recovery jack is a known culprit.

Speaking of which, I run a wire from the recovery jack to the board, then the resistor to ground. Should I move that resistor around or switch up to metal film or something?

Willabe, I'll try replacing those longs runs first and report back.

[guitylerham]

FWIW, I hear two frequencies of hum: the lower is a B note (60hz), the higher one is F# (not 120hz!) I thought that was interesting since I thought you generally have 60 or 120hz.

[eleventeen]

I am going from memory here, but if I recall, there have been a few forum members who decided to mount that 220K on their board or maybe it was close to the receiving end of the rev rec signal and they had noise problems. Fender mounts it right on the jack. Which is kind of odd, since that signal wire is held off ground by the 470K + 220K >---getting close to 1 MB. You can also see that that signal wire is on the "cold" side of the parallel combo of the 3.3M and 10 pf. So the observation is that that wire is a very high impedance feed into the grid of V4A (the "mix" triode) high-impedance implies high susceptibility to noise. While I just said that Fender mounts that resistor right on the reverb can jack, if you have another 220K resistor lying around, I would at least try tacking it in closer to V4 and grabbing a better, closer ground, and see what happens noisewise. That could be either a harebrained idea or it could help, but it's easy enough to try.


What gathers up noise is naked wires running over other stuff and running into a naked grid. If you cannot avoid this, shielded wire is your best defense.

[sluckey]

Speaking of which, I run a wire from the recovery jack to the board, then the resistor to ground. Should I move that resistor around or switch up to metal film or something?

Read reply #30 again.

I am going from memory here, but if I recall, there have been a few forum members who decided to mount that 220K

You're getting two different 220K mixed up.

[guitylerham]

Read reply #30 again.

Got it ;)

[guitylerham]

I moved the reverb return jack in between tubes 2 and 3, mounted the 220k right to the ISOLATED RCA and then to the RCA ground tab, ran that to ground. Got to run a 1.5" wire to the grid of V3-A. I think that helped some. Still getting decent 120hz when turning up the reverb tho.

There are two levels of hiss. A constant hiss that does disappear once you ground the grid of V3-B.

There is also another level of hiss that only appears when the volume is turned up and then disappears when the volume is on max. This hiss also disappears when you ground the grid of V1-B, and also when you ground the junction of the 0.022, 10pf/3.3m, and 500pf. Still working my way through shielding some wires.

[guitylerham]

I just spent six hours trying to solve this and I'm not making any progress.

I disconnected the plate of V1-A from the 100k to shut it off and the volume related hiss went away. So I shielded the input wires to no effect. That must mean a hissy tube then?

Then there's the constant hiss, somewhere between V1-A and V3-B (likely the reverb recovery stage). So I disconnected the plate of V3-A and the constant hiss was still there. I then disconnected the plate of V3-B and the hiss was gone. I have shielded wire running from the output of that tube to the reverb pot and then from there to the input of of V3-B.

I'm just about out of ideas and done. I now understand why the later Fender amps are considered more difficult to build than the tweed era.

[guitylerham]

You know what doesn't help?? Grounding the tip lug of your isolated speaker jack. No wonder my nfb switch was having zero effect. Well now the amp hisses a lot less. Still a little 120hz perhaps. My reverb doesn't feedback now either.

I am relieved but actually I'm more pissed off with myself. I don't how many times I look at each section of the amp. Man.

[Willabe]

You know what doesn't help?? Grounding the tip lug of your isolated speaker jack. No wonder my nfb switch was having zero effect. Well now the amp hisses a lot less. Still a little 120hz perhaps. My reverb doesn't feedback now either.

I am relieved but actually I'm more pissed off with myself. I don't how many times I look at each section of the amp. Man.

Doh!         (You'll laugh too in time.  )

It happens, this too will pass. BUT, YOU found it! We ALL make mistakes, as long as we learn from them and as long as we don't hurt anyone or ourselves, it's OK!   

We don't know how many amp's you've built/re built but it's always best to jump into the shallow end of the pool 1st and then go to the deeper end after we can swim. 

If you buy a pre-made clone chassis and use an already established, proven layout, like Doug's or Sluckey's or Tubenit's or many others here and from other sites, then things will/should go much easier.

We learn 'good lead dress' and good soldering skills this way, then we can stretch out a bit.

The 1st amp I built, like many was a Fender tweed Deluxe, was with a clone fender chassis and a clone Fender eyelet board, worked fine form 1st start up and I still have it.   
 

[guitylerham]

I couldn't have found out and resolved those issues quite so directly without all of your suggestions. No amount of reading can replace knowledge and wisdom above your own.

This is my third amp, none of which were from pre-fabbed boards or anything. Just a basic fendery layout and lots of schematic study. I built a 5F10 Harvard, which is one of my favorite amps, it has just the right amount of volume for a lot of the smaller gigs we play. I start the night with the volume on half and then by the end, it's just dimed and it works well. A little more volume couldn't but I've yet to play an amp that fits so nicely.

My second amp was a 5F4 tweed Super in a V-front cab with two 10" speakers. A pretty cool amp but doesn't have a lot of low end. Good for more of a stinging blues tone when pushed.

This Deluxe Reverb type amp is the first of that style that I've built. It is a different animal and I only appreciate more how much of a feat it is to make one function 100%. When building an amp, it's all about establishing your checklist and order of operations. That way, you'll more likely get the amp working at first power up. I've yet to get that aspect ingrained yet. I feel like I really learned a lot with this build. I actually learned what all the pins of a preamp/power tube are called! It used to be just a big blurry mystery on the schematic page. Amps really aren't that complicated, it's just that they're immensely complex in operation when you consider that we're getting ELECTRONS to do the work for us. Pretty cool. I think my next build will try and follow a proven layout more closely than I have done in the past. Thanks again to you and everyone else who kept coming back with new ideas. This forum is a beautiful thing!