How do you decide that an amplifier is complete and satisfactory?

[Snfoilhat]

If anyone is interested in sharing their methods, I'd like to hear how people judge the completeness of an amplifier build. Almost every build thread includes some voltage test points, idle power tube plate dissipation, and every build includes some kind of player/listener ear test -- do you like what you hear?

Some builders investigate how the amp behaves with an oscilloscope, though I gather from reading here that that is not the majority approach. One thing I'm especially interested in hearing about is noise floor, since even in the best instrument amps it is > 0, so every builder must decide at some point to stop taking action to lower it. The other facet of this it would be cool to hear is how home builders compare to something off the floor of a music retailer, and I'd really rather not hang out at Guitar Center for an afternoon cracking all the new Fenders wide open and listening to self noise.

Do you listen to the amp w/ input grounded volume 0? Volume 10? All the tone controls at 7? At 10? How much noise do the Fender Custom $$Handwired$$ whatevers make with the Reverb control at 10? Do you know what I mean?

What benchmarks do you think are the most useful to you? Thanks in advance!

 

[Uncle Daddy]

I've built some Brownface designs lately, and they are the most quiet amps for floor noise I've ever built. The lack of reverb circuit makes them super quiet. I couldn't get a 5e3 that quiet; I think the chassis is too cramped. Using a low SPL speaker helps too- I put a Weber 12 Alnico Signature in this latest build; I ought to call it the Red Oktober it runs that quietly.

 

[schmee]

I'd say 90% of it is "are there any noises?" Hiss, crackle, hum.
Outside of that, it's volume 3-7 or so and listening for rattles, bad speakers, or good tone.
Then it's evalutaion of how it sounds overall. Bright/spikey? too much break up? too cold on bias?

 

[jjlemon]

As it happens, I've just (yesterday) finished building a 6v6 P_P amp. First measurements taken after realising there were no obvious problems (after bulb limiter test) was to go through and write up voltages at each power supply node then quiescent voltages at each stage, plates, cathodes and screen for 6v6's.
All seemed within a few volts of LTspice simulations, and dissipation calculations all good.
I then did a quick scan through with scope of each stage after injecting 200hz and 1khz signals, wiggling volumes and tone controls and checking each stage working correctly.
Happy with all that, it was time to see what it sounded like. All good, so I'm chuffed to bits that it's working, despite my failing ability to consistently make good solder joints. I should go back a re-do a few of those.

I would like to go back over the scope tests and take pics of the traces with controls at say 12 o'clock, just for baseline reference. But it's too tempting to just sit and play guitar.

 

[tubeswell]

I'd like to hear how people judge the completeness of an amplifier build.

Other musos come out of nowhere between breaks at gigs and compliment you on toan. Until then, its just a bunch of working parts.

 

[Phrygian77]

I've handed completed amps off to customers within a day of finalizing the chassis. I don't usually do complete voltage checks. I just bias and make sure the amp is working as expected on the bench. I'm always double checking my solder joints during construction, so I'm not usually too concerned about rechecking them. Although, I will give chassis a good wack while bench testing just to make sure nothing is loose before it goes into the cabinet.

 

[King Fan]

Interesting question, @Snfoilhat . Are we seeing the answers sort out according to who built what for why?

The unlucky first builder decides it's done after a variable and scary period of trouble-shooting ends. Sometimes this takes hours, sometimes weeks, sometimes forever.

The early-build, build-to-plan builder is done when it fires up and sounds good. If he's on TDPRI, he may also measure voltages. Done.

The professional builder-for-sale knows the amp will work, gives it a test listen to make sure, and heads it out the door. Done.

The experienced builder who's building to capture a certain *sound* in a known amp tests each phase as he builds, confirms fire-up and voltages, and then *may* start to tinker. This coupling cap, that bright cap, this switch, etc. May never be fully done; we see lots of folks here re-design an amp after a year or several.

Then there are guys like you who envision a marriage between this power section and that preamp, with these uncommon tubes, and that plot twist borrowed from a different opera. You'll have to tell us how you know when an amp like that is done, but we're impressed in any case.

 

[2L man]

I test amp noise and hum all pots dimed without input plugged because jack(s) have switch(es). I have made a shorting female jack and can test instrument cable effect to hum. Most hum and noise comes from dimed single coil quitar where for example fluorest light tube light On in home has huge effect.

There are few things I always do whem I build amps:

1. Proper grounding which to me is make all return current flow thru single ground bus ot few Star wires and nothing thru chassis.

2. Shielded cable(s) from input jack to first tube control grid and grid leak resistor to tube socket and solder shield only to jack end.

3. First I quit using carbon comp resistors around first tube because of hissing they cause, but now I only sometimes install CC grid block resistors for power tube(s) and PI anodes where their noise might have some positive effect to distortion. CC resistors lower capacitance should also have some positives?

4. When I use plastic coated wires I first put populated board to amp and cut all the wires as short as possible, only about 1/4 longer than lug hole where it go, take board out and strip wire ends, mount the board, bend wires to lugs and solder them. When I use cloth wires I first mount the board and make wires as short as possible when installing them. I use pushing sharp blade to cut the cloth rolling wire against the table. Perhaps 90% or more of the wires come straight and few must have small curves. I think this might have positive effect to lessen the noise which come because of interference between stages. I also like how the amp insides look.

5. When amp has NFB I use shielded cable and install resistor to jack end and there I use both ends of shield as a reference ground.

6. I have becan installing 3, 4 or even 5 RCR filter stage power supplys where are smoothing resistors also on return current. This seem to lessen SE amp hum more than a single Choke power supply.

There is also a possibility that all this make amp sound more dull because some interaction and noise might be the thing?

How amp sound is where I confess I have got bit lost but it did fill the times I had no work because of Covid19. I have test lots of different manufacturers tubes and different tube types even different PP power tube types same time when biases are really different and there is potential but I have got bit tired to it. When I began I thought I remember what I do and did not make good notes but then it spread so much that I have forgot most of it. Also I should have record different settings to really get good comparison but can't see doing that in near future.

I sometimes tune EQ circuit and pre amp biases but them also often are so subtle that changes would require recording. Now I have an oscilloscope at home but use use it around power supply and power stage where happen more interesting things and which I don't yet fully understand. Tuning bias and operating point is what I like to do, especially in SE amps.

 

[tele_savales]

I check these noise floor by slowly diming the amp with nothing plugged in, and then w a guitar plugged in. I also look at the signal on a scope. I have to be careful not to "accidentally" hammer an Em chord when doing this as I live in an apartment building.

I came to the realization w my last build that I am sick and tired of single coils and was going to wholesale replace every pickup in the house. even thinking of replacing my beloved P90's. Most of my more recent builds sound pretty quiet w humbuckers or in humbucking position.

I cannot do it at home, but I have been making trips to the local rehearsal studio to blast my builds. If I couldn't feel safe taking them on a gig, letting anyone use them, or selling anything I've ever built, then they're not ready, and the only way to find that out is to crank them for several hours w a couple different guitars.

 

[Phrygian77]

I check these noise floor by slowly diming the amp with nothing plugged in, and then w a guitar plugged in. I also look at the signal on a scope. I have to be careful not to "accidentally" hammer an Em chord when doing this as I live in an apartment building.

I came to the realization w my last build that I am sick and tired of single coils and was going to wholesale replace every pickup in the house. even thinking of replacing my beloved P90's. Most of my more recent builds sound pretty quiet w humbuckers or in humbucking position.

I cannot do it at home, but I have been making trips to the local rehearsal studio to blast my builds. If I couldn't feel safe taking them on a gig, letting anyone use them, or selling anything I've ever built, then they're not ready, and the only way to find that out is to crank them for several hours w a couple different guitars.

Build a reactive load (or buy one) and use an IR loader (TC Impulse, Mooer Radar, NUX Studio, etc).

 

[chas.wahl]

Other musos come out of nowhere between breaks at gigs and compliment you on toan. Until then, its just a bunch of working parts.

^^^This. When I was in architecture school, my first design studio professor, a very thoughtful guy and a refugee from Eero Saarinen's office in Detroit, said to us "it takes two people to design: one to do the design, and the other to hit 'em over the head when it's done." That has stuck with me for 4 decades.

 

[drew1d]

The only things that should matter to me would be 1. Does it sound good at the volume I play at? 2. Is it constructed to last and not stressing any parts.

So purpose and reality. It's purpose is a little breakup machine at a reasonable volume. The reality is that it's a no frills, no flexibility, no BS amp that sounds good when I play at home.

So the amp is done when the purpose and reality are close enough.

What happens though, I like to tinker, so I'll get in my head that if I change this "one little thing" the amp "will be more" of something. In my champ build, I keep thinking, "Well...if I used a better transformer it could've..." then I play it and I'm like, "I'd be nuts to change this." Then I start thinking "Well...I could sacrifice an input and put a tone knob..." then I play it, and I'm like, "I got too many knobs on all these pedals, why would I want more on the amp? it sounds good as it is."

 

[Snfoilhat]

I really appreciate all these replies!

In the few years that I've been building, I've worked exclusively in what Uncle Doug calls "fantasy amps"

What if high performance, reasonably priced PA and effects pedals had both been available in the mid 1960s -- what would makers like Fender and Marshall and Vox had made? I'm one of many working with this alternate 1960s pallet, including Fender themselves! The new faux-1968 'Pro Reverb' with only one channel and one 12" looks to me like something they took right off Shock Bros or EL34 World.

So I haven't learned any of the innovations of say Mesa/Boogie or Dr. Z -- no ferrite beads, no DC heaters, no aerospace lead dress, just a couple extra caps or resistors for stability and noise reduction taken from the silverface era if anything.

Because I've never owned original vintage gear and neither have most of the people I've played with or even most of the bands I've seen, I don't have a sense of just how good to make my stuff before calling it a day. I don't want to pick up local radio on my reverb recovery and I fix that when it happens, but if I can't get the amp to be silent with both the volume and the treble dimed, it matters to me if Leo & Co. could (in which case I need to work harder) or couldn't either (in which case I can take a break).

The problem I've had trying to suss this out through forums is that so many posts about gear are in the style of ad copy. I'm stoked for every new 5F1 and Princeton Reverb a newbie builder fires up, but it's not helpful just to read that it's quiet as the tomb or whatever. It's fine to say, it's awesome to see people's success, it just doesn't also provide a good data point. Outside the hobby builder threads things get even trickier, because then you have boutique builders protecting their brand and occasionally making any kind of wild claim (and sometimes it's even true), or the buyers of boutique builders trying to protect their investment. Which again is fine and fun, just not useful.

So it's great to hear all these different approaches to personal targets or quality control or what to prioritize. Thanks!

 

[loopfinding]

Re: noise

I use no less than 2W metal film resistors. 3W on the plates and screens. Film caps everywhere in the signal path. I shield all signal runs. I opt for an RC filter instead of just a stopper on the input, and I place it right at the tube socket. Other runs as short as possible and against the chassis. If I have a spare triode half, I parallel the input stage.

I usually bump the first filter cap up to the max the rectifier will tolerate. Bump up the others slightly. Sometimes I add an additional RC filter for the input stage. I put my caps next to the stage and not just lump them all at the end of the board. I use a bus grounding scheme, with the components of each stage grounded directly to the filter cap ground of the stage and then connect to the next stage. Sometimes terminating the bus at the input jack end or the transformer lug end makes no difference. CT of the power transformer is directly soldered to the ground of the first filter cap. I keep the mains and power switch as far away as possible from the board or panel controls, in a far corner.

Two different color heater wires twisted with a drill and tucked right against the chassis corner or lip. I always add a humdinger. I always float the heaters in a cathode biased amp.

I use closely matched power tubes, paying the supplier to match them (factory matched are always loosey goosey). I buy JJ preamp tubes and pay for the low noise/microphonics testing.

Then the rest basically I monitor it on idle dimed on the scope and futz with the grounding runs or doing other stuff like rotating the OT slightly, moving wires with chopsticks, separating cathodes, or whatever as necessary.

I also always shield the cabinet or backplate of the amp with shielding tape, and to prevent some noise on the bench, I usually lay a piece of wood with shielding tape on the lip of the chassis.

 

[tele_savales]

I use no less than 2W metal film resistors

Ditto. Building Rob's Jcm 800 made me a believer. Incredibly low noise floor. I also put the 68k grid stop resistors right on the tube socket.

 

[tele_savales]

Build a reactive load (or buy one) and use an IR loader (TC Impulse, Mooer Radar, NUX Studio, etc).

I built both a 100 watt dummy load box and a 200 watter, but they're simple, not reactive. That's probably my next project.

 

[loopfinding]

I cannot do it at home, but I have been making trips to the local rehearsal studio to blast my builds. If I couldn't feel safe taking them on a gig, letting anyone use them, or selling anything I've ever built, then they're not ready, and the only way to find that out is to crank them for several hours w a couple different guitars.

Yeah this is also important. I wouldn’t take an amp out that I hadn’t confirmed is not crapping out for a few dozen play hours.

I remember adding a fixed bias scheme to a cathode bias amp once to get a little more headroom for a gig. It was fine for a few days of playing on it briefly. Spent the day or two before a gig hammering stuff out on it and something was faulty with the wiring, so right back to cathode bias for the gig it went, “last known working configuration.”

 

[Phrygian77]

Re: noise

I use no less than 2W metal film resistors. 3W on the plates and screens. Film caps everywhere in the signal path. I shield all signal runs. I opt for an RC filter instead of just a stopper on the input, and I place it right at the tube socket. Other runs as short as possible and against the chassis. If I have a spare triode half, I parallel the input stage.

I usually bump the first filter cap up to the max the rectifier will tolerate. Bump up the others slightly. Sometimes I add an additional RC filter for the input stage. I put my caps next to the stage and not just lump them all at the end of the board. I use a bus grounding scheme, with the components of each stage grounded directly to the filter cap ground of the stage and then connect to the next stage. Sometimes terminating the bus at the input jack end or the transformer lug end makes no difference. CT of the power transformer is directly soldered to the ground of the first filter cap. I keep the mains and power switch as far away as possible from the board or panel controls, in a far corner.

Two different color heater wires twisted with a drill and tucked right against the chassis corner or lip. I always add a humdinger. I always float the heaters in a cathode biased amp.

I used closely matched power tubes, paying the supplier to match them (factory matched are always loosey goosey). I buy JJ preamp tubes and pay for the low noise/microphonics testing.

Then the rest basically I monitor it on idle dimed on the scope and futz with the grounding runs or doing other stuff like rotating the OT slightly, moving wires with chopsticks, separating cathodes, or whatever as necessary.

I also always shield the cabinet or backplate of the amp with shielding tape, and to prevent some noise on the bench, I usually lay a piece of wood with shielding tape on the lip of the chassis.

Sorry, but what does any of this have to do with the completeness of your build? Making sure it is as low noise as you expect, I get that, but you basically tacked your whole build philosophy onto your post.

 

[loopfinding]

Sorry, but what does any of this have to do with the completeness of your build? Making sure it is as low noise as you expect, I get that, but you basically tacked your whole build philosophy onto your post.

nothing, it's mostly in response to:

One thing I'm especially interested in hearing about is noise floor, since even in the best instrument amps it is > 0, so every builder must decide at some point to stop taking action to lower it.

hence "re: noise." make your life easier from the get go. eliminate the need for hunting gremlins as a "finishing" task.

 

[Phrygian77]

@loopfinding you're right. I wasn't thinking about that part of his post.

Tubes themselves are inherently noisy. That part is unavoidable, and there's definitely a point of diminishing returns with everything else.