[AtomBoy]

There is a quiet, but distinct (60hz?) buzz. I thought it might have been because I skipped the film caps just to see what would happen so I added them and it made no difference. Then I thought that the foot long bias test point leads might be an antena, so I made them much shorter.

The buzz is audible on both left and right using some cheap earbuds. Using 68ohm cans, it's louder on the right. I've swapped the tubes and double checked the bias (13v).

It doesn't change when I touch the case or move it, and I don't think its anything to do with the input because I can unplug the cable and its still there. I can leave it going for hours and the buzz doesn't change.

I can't hear the buzz in anything but the quietest parts of music and it otherwise sounds pretty good to me. I am no golden ear, but it sounds a little bit less detailed than my CMoy. It seems very dependant on the type of music, because some electric guitar stuff and vocals sound amazing.

Any troubleshooting advice would be greatly apreciated. I'm happy to take and post other angles if it would help.

 

[drewd]

Here are a couple of things to try: make sure that any power wires are routed as far away from the signal wires as possible (it looks like you've made a pretty heroic effort to do that, though).

Check your grounding technique. If at all possible, use a star ground - ground everything at one point. It's possible to end up with a ground loop, particularly with the tubes and the buffers.

It's also possible that the hum is coming from your power supply. The Millett amp has very little power supply filtering of its own and no PSRR to speak of, so if there is any 60Hz (or, more likely, 120Hz) leaking through, there's a greater chance of hearing it in this amp than with an opamp-based amplifier. Like any single voltage amplifier, it really cries out for a good power supply.

Let us know if either of those work (or not). Good luck!

-Drew

 

[AtomBoy]

My electrician buddy came over and we got my Ebay super nice Condor power supply wired in and that made the buzz go away. Well, duh.

I'm still thinking it shouldn't buzz so badly.

I do have two grounds- one on each terminal strip. They are each bolted to the base plate. Would wiring them together help? Do I need to insulate one from the plate? Or both?

I've also got the proto board with the buffers grounded twice. Each signal wire has a ground twisted around it (bits of cat-5). Should I snip one?

I did know better than to let power run parallel to signal for too far and I don't see much I can do to make that much better without a major rework.

Although, I did have fun making this and I see a couple of things I would change. If I did do the whole thing again- any sugestions?

Another clue maybe- this thing is highly 'microphonic'. Tap the bench next to it and you can realy hear the tubes ring in the phones.

 

[rickcr42]

Quote:

I do have two grounds- one on each terminal strip. They are each bolted to the base plate. Would wiring them together help? Do I need to insulate one from the plate? Or both?

Quote:

I've also got the proto board with the buffers grounded twice. Each signal wire has a ground twisted around it (bits of cat-5). Should I snip one?

grounding is the hardest thing to get right in a hard wired design but from your descrition it sounds like you are looking for trouble and ground loops.

Think "sectional" grounds and then "overall" ground.

Consider the two channels not as two parts of the same thing but think of them as individual devices,each requiring its own ground.So build the channel and run ALL the grounds to a single point locally (the circuit group) then using just ONE wire run it to your star grounding point.Do this for each channel.

You want total chassis isolation so it is YOUR CHOICE where the ground gets connected.That means insultaed input and output jacks or at worst,insulted output jacks but chassis grounded input jacks.

option #1 total isolation : right next to where the input jacks enter the chassis bolt a three rung terminal strip,the type that has the center post making contact with the chassis.The two "isolated' terminals will become your left channel and right channel ground points (one for each channel)
to one attach all the grounds for that channel-input jack ground,ouput jack ground,circuit ground (the single lead mentioned before)
Do the same with the other floating lug for channel 2.

Now add a 10 ohm 5 watt resistor in parallel with an 0.1 uf ceramic cap from this channel ground point (one R/C set for each channel) to the center lug which is grounded already to the chassis.Finally add the power supply ground lead right there and THERE ONLY ! No other connections to the citcuit grounds !

Option #2 non isolated input jacks : the same basics as above but instead of a three lug terminal strip use a four lug strip with all isolated terminals.
Your "New" star grounding point will now be individual left and right grounds with the only actual chassis connection being made by the non isolated input jacks.
As before add the R/C but instead of to one lug each channel now has its own lug grounding point.The power supply ground should still be a single wire and best if 20 guage solid.

you are looking for one actual connection to the chassis or at worst (best ?) one for each channel and the best way to get there is to isolate eveything and choose your grounding point.The R/C is a mini-ground-braker and also reduces DC ground currents at RF freqs from flowing back to the circuit.The $1.50 or so it adds is well worth it.

good luck-and tidy up those wire dammit !

 

[AtomBoy]

Sir, yes, sir! I'll get those wires tidied up real purdy as soon as it works right!

Do I really need 20ga wire and 5 watt resistor for a 24v ~400ma application?

I've torn the whole thing to bits and will be taking a shot at option #2. It might be a while as I have to go back to the day job tomorrow.

Just to see what would happen, I snipped some of the ground connections I had on the signal wires in order to give one path to ground. I worked it so the snipped wire still had a connection to ground. Buzz city on the channel that I snipped the wires on and the original buzz was still on the other channel. The new buzz was very dependant on where my hand was and would react to the vol pot whereas the old buzz didn't. Facinating

You learn more from your mistakes though - right?

Tomorrow is payday so I'm ordering The Art of Electronics.

 

[rickcr42]

Quote:

Sir, yes, sir! I'll get those wires tidied up real purdy as soon as it works right!

I like your style man

Quote:

Do I really need 20ga wire and 5 watt resistor for a 24v ~400ma application?

I meant 1W for the resistor but for the ground wires it is about low impedance connections and even when i use 30 guage for internal wiring i do the ground buss with 20

 

[AtomBoy]

OK, I did another one. This time with option #1 for grounding.

On the right where those two big resistors are is the three lugs for option #1. The circuit for the filaments grounds to the center lug. So do the diode and cap for the power. The center lug is the only place where the circuit meets the plate.

I wasn't sure about a few things like where the grounds for the jacks needed to go.

It has a new, much quieter buzz in both sides now.

I'm running it on the six dollar wall-wart. I'll try it on the nice PS later, but I'm learning more going with the worst case scenario here.

The plastic doesn't provide any shielding so I'm thinking of lining it with aluminum tape (no, not duct tape, but this really thick stuff I have that's made to patch pipes) and giving it a bottom.

 

[rickcr42]

much much tidier layout,nice job.....

Quote:

The plastic doesn't provide any shielding so I'm thinking of lining it with aluminum tape (no, not duct tape, but this really thick stuff I have that's made to patch pipes) and giving it a bottom.

Did not realise the enclosure was all plastic !

better than aluminum to line it would be,if you have access to it,is copper flashing like used for roofing and found in any hardware store for cheap.Easy to cut and bend,great shielding properties,not a bad look either plus you can solder to it if you need to unlike aluminum.
But ONLY connect this at the single ground point or you will again cause a ground loop.

Your bottom is open ? you may consider a wood base for this.would look pretty cool and add some weight to the amp at the same time.

Quote:

I'm running it on the six dollar wall-wart.

should not matter unless the current rating is too low for the stereo circuit.would matter to end sound quality but not if it hums or not.

any buzz still there must be layout related or a bad connection.Have you tried just moving the internal AC wires (from the wallwart input) around while it is "on" ?
a small movement can have results way beyong expectations at times.

BTW-being a low voltage amp not deadly like most tube amps but still it is a good idea to do any probing of a live circuit with a material that does not conduct electricity.A small dowel or even a popsickle stick with a notch in the tip works great.

 

[AtomBoy]

Quote:

Did not realise the enclosure was all plastic !

The top is 3/16in aluminum plate, the sides are 1/2in machinable plastic. It's stuff I had kicking around and I realy like the look of the plastic. It's so old, scratched up and yellowed that it looks like ivory or marble.

I'll go have a popscicle and poke around. My first target will be that blue wire that feeds the LED!

 

[rickcr42]

I think the look is incredibly unique and I love it.Same old is boring and anytime I see something that stands out by being different I am impressed.

Cosmetics are also part of the overall satisfaction level of a project.to some more important than the actual sound or functionality.

don't worry man,we will get that sucker hum free eventually.....

 

[rickcr42]

Must be time for some new glasses or a course in paying attention

looking at the pic again-

I can see a glaring problem with the wiring method.In the lower left hand corner you have a wire bundle being held together by a wire tie.This is a serious no no and what you have created is a classic loop antenna which is capable of picking up radiated noise from just about any source nearby.

Shorten the wires and get rid of the loops and the noise just may go away.

 

[AtomBoy]

That bundle is a pair of wires going to the bias test points. I pull it out to clip to the multimeter. I've been waiting for inspiration to strike for a better way to rig it and I think I have a way to do it that won't stink.

Rick, thanks for the tip! I realy apreciate all your input on this.

I listened to it for a while last night and just moving it from one spot to another made the buzz almost disapear. I'm guessing that means it's picking stuff up from outside and shielding the guts with the al tape or copper or an empty Doritos bag should help.

It will be interesting to get rid of my loop antena before doing any shielding and see what that does.

 

[rickcr42]

Quote:

Rick, thanks for the tip! I realy apreciate all your input on this

my pleasure,really

 

[AtomBoy]

Like... rilly, dude.

I found a somewhat understandable explanation of grounding here:

http://www.sonic.net/~ktstrain/Groundtutor.htm

I may have to change where some ground wires are. I've got the proto board with the buffers grounded at the same lug as the plate current stuff for the tubes and it looks like it should be separate.

Still waiting for the mailman to bring me The Art of Electronics...

 

[rickcr42]

actually you should be alright where you are at with maybe just minor wiring moves.

A little movement oin the scale of 1/2 inch can mean a lot when the entire circuit and radiation patterns is considered.
Kurt's primer is a good one though.

He actually had a web page up once with his designs and projects but that has been gone for about three years and all that is left is his Headwize headphone amp project and his SET 45 Amp project at the magnequest site.