AC30 build lives - a bit of a hummmmm, though

[Willabe]

I've built a couple amps with the PT and OT oriented just like yours, got it from TUT 3. Their cores are 90 degrees from each other, just like yours. No noise problems.

Merc Mag. makes very good transformers, I doubt the transformers themselves are the problem. AND your PT/OT both have steel end bells that shield them.

I think it's either the Vox power supply or your layout and lead dress or a little of both.


 

   

[markmalin]

Did you tied to rotate the PT and OT of 90° eachother (in the other way) ?

K

I didn't.  I could get them both about 45 degrees, but they were offset (not lined up).  See pic's - that's all the room I had before the wires stopped me

[markmalin]

I built that cable front end for the listening amp that Doug has on his 'Tools' page.  I plugged it in to a little solid state marshall practice amp and started poking around searching for the hiss.  I guess this doesn't surprise me, but I hear it start at V2 where pins 2 and 6 tie together on my schematic (pins 1 and 7 on Doug's).  It's the plate of the A side to the grid of the B side of the tube.  Not very conclusive as I've already replaced the 100k (R12) with a Vishay/Dale metal film resistor in case the his was coming from the carbon comp's.  I don't hear hiss on the grid or cathode of the A side.  Anyhow, I did some reading and someone suggested rarely the 10K metal film resistors in the power supply dropping string can cause hiss, so I swapped out the 10K in the "D" supply.  No difference...still have a hiss.


Here's an observation, though, which may or may not mean anything.  When I touch the probe of my listening amp to the D supply resistor, there's a definite hum.  Poking a listening amp probe at the DC supply may be a no-no, I don't know.  Just thought it was interesting re the hum.


I think at this point I can live with the hum, but I really need to do something about this hiss.  (...I maybe should start a different thread on the hiss?)

[shooter]

Put a gator-clip from V2 grid to ground, I know the pot *should* do that but....  re-probe/listen to pin 1/7, hiss?
Haven't looked at the listnin amp but I'm guessing there is a cap between probe tip and amp, that "should" block DC from the power rail, but I wouldn't recommend headphones

[kagliostro]

I don't remember, did you tried other preamp tubes ?

A friend, some time ago, solved a hiss problem that way

Franco

[markmalin]

I don't remember, did you tried other preamp tubes ?

A friend, some time ago, solved a hiss problem that way

Franco

I swapped them out when I was working on the hum, but i can swap them out - I have a few sets.  I'll try that.

[markmalin]

Put a gator-clip from V2 grid to ground, I know the pot *should* do that but....  re-probe/listen to pin 1/7, hiss?
Haven't looked at the listnin amp but I'm guessing there is a cap between probe tip and amp, that "should" block DC from the power rail, but I wouldn't recommend headphones

Shooter - last night I swapped out some of the carbon comp resistors around V2 and actually this morning poked around with the listening amp and no longer hear the hiss prevalent in V2.  I grounded the grid of V2 with a clip lead, with or without the grid grounded now I don't hear anything worth mentioning.  Now I hear it just at the plate side of the PI, which seems peculiar.  Not on the grids of the PI, but on the plates.  I may try some different PI tubes to see if it changes anything.

[shooter]

I think you're in the home streach now, pretty soon you'll be building the 2nd one, bug free

[markmalin]

I think you're in the home streach now, pretty soon you'll be building the 2nd one, bug free

Thanks, shooter.  I sure hope so!  but I'm not out of the woods yet, unfortunately.


I thought I had narrowed it down to the PI, but I started tracing the hiss again with the gain way up on my listening amp.  I've traced it back quite a bit, actually.  Seems like the 500pf cap, C3, has his on the output side of the signal (toward the volume) but not so much on pin 1 of V1-B (or pin 6 if you use Doug's original schematic).  So it's like the plate is somewhat quiet, but after that 500pf cap there is more hiss.  The difference is faint (and there's enough 60 hz hum in my listening amp from stuff in my shop) that i wonder if that could be the issue, or perhaps V1 or the 33k resistor on the input...but then again, why fainter on the input side of the cap if it's not the cap??


So I ended up finding a 500pf silver mica cap, swapped it in and still have the hiss.  I swapped in a different 12AX7 for V1, granted just one to see, but still same hiss.


Which makes me wonder...the 33k (R4) is a carbon film, but I do have a Vishay/Dale metal film I could put in there so I'll try that.


Is it possible the tube socket could cause noise?  I'm using these porcelain style...bad idea??

[shooter]

Is it possible the tube socket could cause noise?  I'm using these porcelain style...bad idea

I have had issues with the ceramics, you should be able to "roll the tube around" and hear something.  I "rock" the tube in a circular motion while monitoring, be carful!! the tubes are hot and ceramic isn't very flexable.

[markmalin]

I actually swapped out ALL the remaining carbon comp resistors, and the 33k carbon film on the input of V1 pin 2 with Dale metal films.  Still hissing.  I experimented with more pre-amp tubes, still have the hiss. 


I've ordered some non ceramic tube sockets and plan to swap those in tonight.  I hope that does it because I"m running out of ideas....(and have pulled most of my remaining hair out!)  Wonder if it's worth posting this on Ampgarage too, just for the heck of it?

[shooter]

Wonder if it's worth posting this on Ampgarage too

As long as you warn them ahead of time, your "problems" have infected my build 
I might have to report you to the CDC

Although to be fair , I only caught the hummmm

[markmalin]

Wonder if it's worth posting this on Ampgarage too

As long as you warn them ahead of time, your "problems" have infected my build 
I might have to report you to the CDC

Although to be fair , I only caught the hummmm

No way!  Your build is humming?   


I'll gladly give you some hiss if you like :P  I have little hope at this point that new sockets are going to do the trick...

[John]

Ya know, I was looking at your pictures again. The only thing I really see is the B+ wires running pretty close to your pots. Maybe if you routed them up along the top side of the board instead?

[markmalin]

It's worth a try, I guess.  Would that induce hiss?

I guess this begs the bigger question.  What exactly is hiss and what can cause it?  From what I've been seeing on the interwebs (as Strongbad would say), it looks like biggest causes are (from this link http://www.geofex.com/ampdbug/hiss.htm  ):

1) tubes  (I've tried enough pre-amp tubes to satisfy me that it's not a tube, unless they ALL hiss)
2) carbon comp plate resistors (I've replace literally all the carbon comp's with metal film (but left 2 carbon film cap in the bass and on intput 1 which is not hissing))
3) noisy resistors in the "B+ decoupling string, often around 10k in value"  (as a shotgun approach, I swapped out the 10K metal film with another)

Those are some "common causes", but what exactly is it?  Why would it happen when there is no input signal (i.e. the input jack is grounded and nothing is plugged in)?  Some phenomena is "generating" it?  Is it that whatever "noise" is being picked up in the circuit, this "signal" passes through a component that somehow induces the hiss?

I really thought a systematic approach with a listening amplifier would allow me to pinpoint it.  But it's basically very faint but present at the plate of V1 (the half that has the tone stack) and gets amplified greatly at the cathode follower.  But where is it COMING from?!   

But alas...I digress....

John, I'll try moving those lines when I swap the sockets.  I stuck them down there because the wire I had has such silly big insulation.  I'll see if I can scare up some other wire.  On the D'clone builds I do I run them along the top of the main board...but on AB763 style builds I run them under the main board and along the front edge...but not near the pots...

[John]

Mark, *as far as I know* those wires wouldn't induce hiss, just hum. I've never had to chase a hiss, myself.... yet! I'm sure I will though. I only mentioned those wires because I had a similar hum I was trying to track down on my push pull amp, and it turned out I had run the B+ wire too close to a grid resistor in the preamp. What makes trouble shooting so tough (especially for me) is it can be so many things.

[shooter]

What exactly is hiss

I think it's something like describing color to a blind man

When I look at "hiss" on a scope I see a complex set of frequencies like looking at white noise.  I'm not gonna re-read everything, sooo, have you tried clipping caps from plate to ground, or plate to cathode?
If not start around 500pF and move up to maybe .01uf.  These WILL have effect on amp tone.

[markmalin]

What exactly is hiss

I think it's something like describing color to a blind man

When I look at "hiss" on a scope I see a complex set of frequencies like looking at white noise.  I'm not gonna re-read everything, sooo, have you tried clipping caps from plate to ground, or plate to cathode?
If not start around 500pF and move up to maybe .01uf.  These WILL have effect on amp tone.

I haven't tried the cap idea, but I want to do that here soon.  I know in the D-clone style amps, there is often a 330 pf to 220pf cap across the second gain stage plate and cathode.  Might be worth playing around with that.

I've replaced 2 of the 3 ceramic sockets with plastic style (yeesh...these things are getting more cheaply made as the years go on!) and have re-routed the B+ supply wires to the top of the board, but it still hisses like a wild banshee...so tomorrow it's 500pf caps from plate to cathode!

Questions: if I want to temporarily bypass parts of the tone stack, would I ground the input to the cathode follower and jumper the wiper on the Volume pot to R20?
Or could I just jumper C6 to R20 and leave the cathode follower in, but bypass the tone stack?
Mark.

[shooter]

I would try it this way and see

[markmalin]

I would try it this way and see

Ok - so actually desolder where the X's are, or wouldn't a jumper from X to X cause the signal to take the path of "least resistance" being the jumper?  I'm just curious - desoldering is no problem, just trying to learn here ;)

[shooter]

I'm just curious

You're right in a shorted wire "should be" fine, I personally would un-solder, just so when I fall asleep I can be certain the TS was physically removed.  AC doesn't play as well as DC, I have found it wandering on grounds that measure fine in a DC world.

[markmalin]

Well, not much luck.  Here's what I've tried.


- I bypassed the tone stack (I left it in place, I just wanted to try it quick) with a clip lead.  Hiss worse, I suspect because of the additional gain due to no tone controls
- experimented with a 500pf cap across V1 cathode and plate (the section feeding the channel with the tone stack).  Still hisses the same though I didn't measure it, just listening
- same cap across V3 ( the PI ) no difference
(I need to mess around with more caps, just haven't had a lot of time)
- I replaced the 4 1.5K carbon comp's on the EL84's with metal film.  No difference in the hiss
- I replaced all 3 pre-amp tube sockets with non ceramic.  No difference in the hiss
- I re-routed some wires and did some lead dress improvements.... No difference
- I noticed V1 was pretty microphonic, so I grabbed a JJ Tesla and put that in.  No difference in hiss, but less microphoncs

So I'm stumpped guys.  This is getting discouraging.

I was wondering about the El84's bias setting, just wondering if they're running too hot.  I split the pairs out and am using 120 Ohm resistors, one on each pair.  I measure ~11 VDC across the resistor.  Is that about right?  That's like 95 ma/pair.  This is idling.
Plates on the EL84's are at 355 VDC.
FWIW - The Plate voltages on the other tubes are sitting at 155 V for V1, 192 on pins 2&6 of V2 (the cathode follower), 302V for V3 the PI. 

[shooter]

That's like 95 ma/pair

I'd call that good, I've built 4 amps with 84's, self biased and I shoot for 45-50mA/tube, with plate V's around 250.

Did you try and set the amp on a steel plate, surface?  When I did that, my hiss went from bad to completely acceptable.  I tried an aluminum plate and that worked also, so I cut one out, hope to install it soon.

[markmalin]

That's like 95 ma/pair

I'd call that good, I've built 4 amps with 84's, self biased and I shoot for 45-50mA/tube, with plate V's around 250.

Did you try and set the amp on a steel plate, surface?  When I did that, my hiss went from bad to completely acceptable.  I tried an aluminum plate and that worked also, so I cut one out, hope to install it soon.

So you set the amp on a steel plate??  Interesting.  I'll have to try that, but why would that affect hiss I wonder?

[Willabe]

So you set the amp on a steel plate??  Interesting.  I'll have to try that, but why would that affect hiss I wonder?

It's shielding. It helps greatly to reduce any air born junk from getting into the circuitry.

All the old Fender BF amps had a piece of steel window screen stapled onto the cabinet above the chassis opening. (I don't know if the brown face and tweed had this also.) Later Fenders had a solid piece of thin metal stapled in the same place. 

Same exact concept as shielding a guitars control cavity.

[markmalin]

I see.  I've done that on other builds to cut down on buzz, but I thought hiss was more thermal reactions in the components.  I'm going to get some sheet metal and see if I can make any difference in this.


Anyone messed around with testing the outside foil on coupling caps?  Probably grasping at straws, but I can't get it out of my head that something inside the amp is literally generating this noise.  I'm going to put a master volume in which will help keep the noise floor down, but it's just so bad I wanted to reduce it before adding the mv.

[Willabe]

Anyone messed around with testing the outside foil on coupling caps? 

Yes, look at this thread link started by Tubenit;

http://el34world.com/Forum/index.php?topic=11427.0

[markmalin]

Anyone messed around with testing the outside foil on coupling caps? 

Yes, look at this thread link started by Tubenit;

http://el34world.com/Forum/index.php?topic=11427.0

Good stuff.  I read the article on Aiken the other day and printed it out.  As my last ditch attempt (...now second to last ditch since I'm gonna get a big sheet of steel) I'm going to go through, desolder all the caps, test/mark outside foil, and re-install them.  I'll work on that tomorrow...I need beer...

[Willabe]

Have a look at this, if you haven't yet, it's on grounding schems;

http://el34world.com/Forum/index.php?topic=18828.msg193010#msg193010

And, I only see 1 shielded piece of wire, might be more?

Grid wires need to be as short as possible or they need to be shielded to keep out noise.

Any signal wire AFTER the plate coupling cap is now a grid wire. So wires after the plate coupling cap going to a volume/MV control then going to the next tubes grid might need to be shielded, also the TS wires after the TS caps as these caps take the place of the plate caps and after the PI coupling caps. These are all common places you'll find shielded wire runs.     

[markmalin]

Have a look at this, if you haven't yet, it's on grounding schems;

http://el34world.com/Forum/index.php?topic=18828.msg193010#msg193010

And, I only see 1 shielded piece of wire, might be more?

Grid wires need to be as short as possible or they need to be shielded to keep out noise.

Any signal wire AFTER the plate coupling cap is now a grid wire. So wires after the plate coupling cap going to a volume/MV control then going to the next tubes grid might need to be shielded, also the TS wires after the TS caps as these caps take the place of the plate caps and after the PI coupling caps. These are all common places you'll find shielded wire runs.     

Thanks, Willabe.  This is good stuff.  I've read GW's stuff about grid wires being short.  I think he said plate wires should be on the long side (?).  I added a shield on the volume to V2 wire.  I may add others.  I appreciate the explanation "any signal wire after the plate coupling cap is now a grid wire".  Good point.  I'm going to keep hacking ;)

[Willabe]

I think he said plate wires should be on the long side (?). 

I think your right on that, if he did, he gave an explanation as to why.

Think about it this way....  Grid wires act like an antenna and pick up air born hash/noise/radio signals, etc. Plate wires tend to reject them, it's an impedance thing.

So, we want to go from point A (plate) to point B (grid). And the tube sockets are on the back side of the chassis. But we have a volume control in front of the 2nd/next tubes grid and it's on the face of the chassis. So we have to go from the 1st triodes plate/back of chassis, to the volume control/front of chassis then back to the grid of the 2nd triode. Depending on the chassis and where the tube socket(s) and volume control are located AND where we put the coupling cap, now we have 6", 8", maybe 10" of grid wire. Could be a problem.

If we mount 1 end/far end/AC side of the coupling cap directly to the pot then we don't need shielded wire going to the pot, we only need it going from the pots wiper to the 2nd triodes grid.

So it's that were trying to put the coupling cap closer to point B and further away from point A in our layout. But you can't always do it like that if you have a pre made eyelet/turret board. So just use shielded wire if you need to.

Some guys hang/mount the TS (tone stack) caps on the tone pots for the same reason.

Remember back in the 50's/60's, when most of the classic amps were designed, they didn't have anywhere near the air born 'hash' we have now. There we no light switch dimmers, florescent lights were much less prevalent, if at all yet, PC's/lap tops, cell/smart phones, things like that. (Also AC power grid line noise was very low compared to now. They just didn't have the amount of traffic and devices plugged into the AC grid.)

Plus they weren't building hi gain amps either, so they didn't have to deal with shielded wire for the most part. A good steel chassis with some type of shielding over the back/open side of the chassis was enough.   

[Ritchie200]

I don't know about that!  All the neon beer signs in the bar would make quite a racket when the back of the stage was next to the front window!

Jim

[Willabe]

     That's true, I forgot about those.       

But at least we'd have that beautiful neon signage to look at!

[markmalin]

great write up, Willabe!  Thanks :)

[John]

Yeah Willabe, thanks! I read your post and changed the layout for my re-wire job I'm doing a little bit to get the coupling caps close to the pots.

[markmalin]

I'm checking the outside foil on all the caps now.  Wondering, in the tone stack how do you orient the outside foil?  Most descriptions talk about coupling caps (outside foil toward the plate) and caps to ground (outside foil toward ground), but what about in the tone stack??

[markmalin]

Ok,  I've done all the following and it still hisses like a mad man:


- I went through all the caps, desoldered them, checked for outside foil and re-installed them.  (only the orange drops, that is).
- I tried replaced the input jacks with isolated jacks (and rerouted the grounds to accommodate)
- I actually went through and did more of a star ground while I was at it.
- I changed out the 47pf cap for a larger 120pf cap to see if it toned down the hiss.  nope
- for grins I attached that 120pf treble cap at the pot.  nope
- I took the 500pf cap right after the plate of V1 and soldered on another 500pf to double it just to experiment.  same thing


Nothing seems to touch this hiss. 


I was reading something Tubenit wrote about hiss years ago - he mentioned increasing the treble cap, and some other ideas like shunting high freq to ground.  I need to go back and re-read that to try some things 'cause noffin is working yet.

[shooter]

other ideas like shunting high freq to ground

That's what you did by running caps from plate to ground or cathode.  They can also be paralleled with the plate R.

Did you try hanging a BIG cap - 100-250uF - cathode to ground on both pin 3 and 8?

[sluckey]

Does this amp follow Hoffman's schematic? IE, it has a normal and a top boost channel? Is the normal channel hissy?

[markmalin]

Does this amp follow Hoffman's schematic? IE, it has a normal and a top boost channel? Is the normal channel hissy?

Sluckey - Yes, this is his schematic except for the half power switch. 

The top boost channel is hissy.  The normal channel is real quiet!

That being said, I'm wondering what happens if I play around with the first coupling caps.  The normal channel has like a .047uf I think, while the top boost has a 500p followed by what is effectively a "bright switch" cap of 100p across the volume.  I mean this amp is SILLY bright, I wonder about modifying these caps.

  This is off topic - but what's with the 0.1uf cathode bypass cap in the top boost?

1) Normal cathode bypass caps in the first gain stage are like 25uf, 50uf, 10uf.   If I understand correctly, the smaller the cap value the more high frequency will be allowed to pass through the signal path.  Maybe 0.1uf is just too small?  What if I put something more fenderish in there, like 25uf or 5uf cap?  Maybe it would tame it down??
2) Why is this cap non polarized, is it just that it's such a small value?  I've got a big honkin' 0.1 600v polyester in there!

[markmalin]

other ideas like shunting high freq to ground

That's what you did by running caps from plate to ground or cathode.  They can also be paralleled with the plate R.

Did you try hanging a BIG cap - 100-250uF - cathode to ground on both pin 3 and 8?

Shooter - I didn't try a big cap on both 3 and 8 to ground.  I'll give that a try.  I'm not sure if i have anything that big, I juse used like 220 pf.  I may have some 100uf's left over from bias circuits.  Did you mean pins 3 and 8 of V1 or V2??  And from 3 and 8 to ground? So I'm basically increasing the size of the already present cathode bypass cap as a test, yes?

Thanks!
Mark.

[sluckey]

Remove the .1 cap from V1B pin 8. Also remove the 22µF from V2A pin 3. Replace the 220K resistor on V1B pin 6 with a 100K.

All this will reduce the gain of the top boost channel which will also reduce the hiss. Does this take you in the right direction?

[shooter]

Did you mean

I didn't realize one channel was good.  Follow Sluckey's lead, I was just looking for ANY change, so on the bad channel, v1 Big cap across the cathode R, and it can be in || with whatever other cap you have there.  It should be a quick, clip n test.

[Willabe]

1. If I understand correctly, the smaller the cap value the more high frequency will be allowed to pass through the signal path.

2. what's with the 0.1uf cathode bypass cap in the top boost?

3. Why is this cap non polarized, is it just that it's such a small value?

1. Close, the smaller a caps value the less mid range/bass it will let pass through, so it sounds brighter by subtracting mids/bass end. Caps and resistors are passive, they can only subtract, they can't add anything.

2. The K bypass cap value is used to 'tune' the amps signal. The smaller the K bypass cap the less bass/mids that gain stage will amplify.

3. Large value caps are usually made of electrolytic construction to keep their size and price down. Large value  film caps are made but their huge and way more $$. E-caps are polarized, film caps are not.

used like 220 pf. 

I think you mean you used 22uF?

220pF = 0.00022uF.   

[markmalin]

used like 220 pf. 

I think you mean you used 22uF?

220pF = 0.00022uF.

No, I only used 220pf.  I misunderstood shooter's original test suggestion and was thinking 100-200 picofarad from plate to grid.  I may have tried 500pf as well, but nothing that big.  My bad.

Thanks guys.  Sluckey, Shooter, Willabe, I'll try these things tonight.  I'd like to hear some difference with these big change tests so I understand how to head in the right direction.

[markmalin]

Hm... nothing! 
I took put a 25 uf cap in parallel with the cathode resistor of V1 (the Top Boost Channel) in place of the 0.1 uf cap.  Still hissed.
I soldered a 100uf cap in parallel with that, so 125uf and still the hiss was the same.

Sluckey, I removed the 0.1 uf cap, replaced the 220k cathode resistor with 100k, and pulled up one side of the 22uf on the cathode of the other half of V1 (the "normal" channel) to remove that from the circuit.  Still hisses!    It may be a touch less hiss, but it's still enough to where you could never record with this amp.  If I turn the treble down to 2, it's livable, but as soon as you turn the volume up to maybe 3 or 4 it's there again.  Or if the volume is down (livingroom volume, maybe 1 or 2) and the treble is up to 5 or more, it's just too much hiss.

My test case (listening) has been to run the amp with the volume on 3 and the treble on 8 (treble and volume knobs are pointing in opposite directions).  This has been my test each time I make a change.  If I turn the Cut control up, the last 15% of rotation of the cut silences the hiss.  This has always been the case.

[sluckey]

and pulled up one side of the 22uf on the cathode of the other half of V1 (the "normal" channel)

Wrong cap. I said "Also remove the 22µF from V2A pin 3". Pulling the correct cap should reduce the gain of the TB channel. Hopefully hiss will go down too.

You can also try rolling a bunch of 12AX7s thru V1 and V2. If you have some 12AY7s and/or 5751s try those too.

[markmalin]

and pulled up one side of the 22uf on the cathode of the other half of V1 (the "normal" channel)

Wrong cap. I said "Also remove the 22µF from V2A pin 3". Pulling the correct cap should reduce the gain of the TB channel. Hopefully hiss will go down too.

You can also try rolling a bunch of 12AX7s thru V1 and V2. If you have some 12AY7s and/or 5751s try those too.

Right, that makes more sense.  I did a quick check before calling it a day, lifted that cap and powered it up.  The hiss level is less with that cap out.  Still there but reduced due to the gain change. 
I've tried a few sets of pre-amp tubes, but this weekend I thought I'd run a bunch of them in V1 and V2.  I have Mullards, JJ's and EH's, as well as some old stock.

[markmalin]

I played the amp some last night and the hiss is under control.  The amp is less gainy, obviously, I'm tempted to add gain back in in small steps until just before the hiss starts getting annoying. 

[sluckey]

FWIW, most, if not all of the schematics with the top boost channel show a 25µF cathode cap on the first stage and no cathode cap on the second stage. The only place I've seen that 22µF cap on the second stage was on the original 'top boost' mod that was a separate module to be added to an amp. Looks like by the time the top boost channel was incorporated into the AC30, that second stage cap had been eliminated.