Hi does a sag in the b+ add noise to the amplifier ?
Adding a sag resistor will add more noise then ?
Adding a sag resistor will add more noise then ?
No except the Johnson noise of R but no additional noise because of a technology like switching power supply because the PSU in a tube amplifier is a linear PSU.
You get a little less hum, whereas you may have 8V ripple without a sag resistor you might get 6V ripple with one. Not hard numbers just as an example, your mileage may vary.
I think it depends on what exactly is meant with "sag".
If sagging here just means short drops of the value of B+, caused by the resistance in the B+ line and the higher current consumption of the power stage at peaks (especially in class AB and class B push-pull amplifiers) than I would think that the ripple on B+ gets a bit higher during these moments of sag. During those moments part of the current that normaly charges the reservoir/smoothing capacitor(s) during every AC cycle now flows to the power stage sot the reservoir/smoothing action is a bit less.
But this is not likely to be a problem since you probably won't be able to hear the somewhat larger ripple on the B+ line during peaks (and push-pull output stages tend to cancel this type of ripple out anyway).
If sagging here just means short drops of the value of B+, caused by the resistance in the B+ line and the higher current consumption of the power stage at peaks (especially in class AB and class B push-pull amplifiers) than I would think that the ripple on B+ gets a bit higher during these moments of sag. During those moments part of the current that normaly charges the reservoir/smoothing capacitor(s) during every AC cycle now flows to the power stage sot the reservoir/smoothing action is a bit less.
But this is not likely to be a problem since you probably won't be able to hear the somewhat larger ripple on the B+ line during peaks (and push-pull output stages tend to cancel this type of ripple out anyway).
Sag usually occurs when the amp is overdriven and draws a lot of current. Because current draw is high filtering caps drain more at each cycle and ripple is alleviated. Because the clipping threshold is the plate supply voltage with significant ripple the ripple ends up modulating the clipping and introduces a "ghost note".
Sag occurs all the time if the power supply is not regulated. All unregulated supplies drop voltage as a function of current. Some more than others, but it always occurs. It depends on the impedance of the transformer, filter chokes (if any), rectifier, and dropping resistors (if any) - working against the load current. If the load is variable and the power supply impedance is high relative to the maximum load, you get a lot of sag. Resistive and reactive impedances behave differently in terms of the profile of the sag vs. load current. You may have one supply that drops rapidly the instant you start drawing anything significant, and then levels out dropping al a lower rate. Or one that drops at a more constant rate as load is increased. The same amplifier on the two different supplies will behave differently once you drive it from small signal into clipping, and from light clipping to heavy clipping.
The sag and ripple can be somewhat independent of one another. You do get more ripple on any supply with more load, but it doesn’t necessarily follow that a supply with 30% more sag than another will have 30% more ripple. You can have a transformer with a high resistance and reactance, with a LOT of capacitance. It will still sag, but will take time to do so and have low ripple.
The sag and ripple can be somewhat independent of one another. You do get more ripple on any supply with more load, but it doesn’t necessarily follow that a supply with 30% more sag than another will have 30% more ripple. You can have a transformer with a high resistance and reactance, with a LOT of capacitance. It will still sag, but will take time to do so and have low ripple.
Ripple leads to intermodulation distortion, which is an audible part of amplifier sound.
See Class AB Ripple (Ampbooks) which references and mostly translates a 2007 study called "Wildcat" (Thomas Schmidt & Henry Westphal original in German here)
They quote (in English)
Original German (Pg 10)
Lots of pictures of what's going on.
It's the usual story: non harmonically related artifacts are very audible, even if a long way down.
See Class AB Ripple (Ampbooks) which references and mostly translates a 2007 study called "Wildcat" (Thomas Schmidt & Henry Westphal original in German here)
They quote (in English)
Original German (Pg 10)
Lots of pictures of what's going on.
It's the usual story: non harmonically related artifacts are very audible, even if a long way down.
Agree.
The problem is exacerbated in vintage guitar amplifiers, because the filter capacitance was already very small by modern standards, so there was a lot of ripple at the best of times.
With a modern low-resistance silicon rectifier, peak-to-peak ripple voltage is very nearly directly proportional to (DC) current drawn from the filter cap. Draw twice as much current, get twice as much ripple. That means lots of ripple at the same moments as lots of "sag", because both phenomena are caused by larger current draws from the output tubes.
The relationship (between ripple voltage and current draw) is less simple when the diode (tube rectifier) itself has lots of its own resistance. But in general, more current draw causes more ripple voltage. This is why we need bigger filter caps in higher-current power supplies. And yes, it means more ripple when you draw more current, ergo, more ripple at the same time as more "sag".
I first read the reference Thoglette cites quite a few years ago. I find myself rather skeptical. The statements were so over-the-top that it is hard to take them seriously. Here are some of the statements made:
Clean tones HAVE to sound the same as before. The amplifiers frequency response hasn't changed, nor the THD below full overdrive. How can we take seriously the claim that "tonal characteristics are completely different"?Suspicious Reference said:
This is another bit of hyperbole that's impossible to believe. Every tube stage in the Bassman adds some compression to the guitar signal, most particularly the DC-coupled cathode follower stage. This compression causes "touch sensitivity".Suspicious Reference said:
Power supply sag adds to the compression / touch sensitivity, but the new PSU in the experiment was designed to produce the same amount of sag as before. So how could the touch sensitivity suddenly disappear?
Every tube stage produces overtones, and those cannot go away because of a power supply change.Suspicious Reference said:
It's certainly possible that IMD dropped dramatically, but that is only one part of the "overtones"; saying "overtones - zero" is just more hyperbole.
Pure hyperbole. The amp is exactly the same as before for clean tones - and clean tones is what Leo Fender designed it for. It is quite believable that the quality of heavy overdrive changed, but that is only a small part of guitar tone. "...more like a kitty cat" is such a huge amount of exaggeration that I find it impossible to take seriously.Suspicious Reference said:
More hyperbole...seriously, "heart and soul"?Suspicious Reference said:
Seriously? An all-tube amplifier famous for clean and slightly overdriven tones now sounded like a transistor amp because of a stiffer power supply? What about all the tube-generated distortion? Every tube stage adds harmonic distortion, could these guys simply not hear any of it? Or is this just B.S.?Suspicious Reference said:
I don't know if the guys who conducted these tests had tin ears, or maybe they were youngsters only interested in cranking the amplifier to maximum overdrive the entire time, or maybe they just enjoyed exaggerating. Whatever the cause, these stated claims are so over-the-top that it becomes impossible to take them seriously.
I believe the data - the oscilloscope measurements - but IMO the subjective conclusions have to await listening by musicians who use the entire dynamic range of the amplifier (not just full-on overdrive), and who are able to write down their observations without launching off into exaggerated hyperbole.
-Gnobuddy
By definition this is at a level of drive that the amp is starting to "rawk" So the comments on "entire dynamic range" are probably moot.
The CRO plots are straight forward: the difference is a pile of IMD from a non-musically related source.
Yet you're telling everyone that that can't be possibly be audible.
We know that amps-that-sag don't sound like amps with "stiff" power supplies. We know non-harmonically related stuff is audible 30db (and more) down. We know the "hi fi" guys stuff stupid amounts of capacitance on their power supplies and declare it audible.
For me, this is a logical explanation for the "Class A/B" sound (whether in the output stage or the pre-amp)
Certainly, there's the a risk of non-blind A-B testing impacting the team. However, the result they think they heard negates their initial thesis. Which is unexpected: they expected both amps to sound identical.
But, be my guest. Build it and tell us what you hear.
Well, no, that's not what I'm saying.
It's quite plausible that lots of IMD when the amp is driven all the way into hard clipping is audible.
What is not plausible, is that the amplifier sounds entirely different at all power levels - tonally different, lifeless, no touch sensitivity, et cetera - simply because PSU ripple is reduced. The claim is that a great amp is reduced to useless garbage simply by knocking down PSU ripple - and frankly, I find that claim quite ridiculous.
I've never built a Bassman, but have built a few small tube amps. I always use silicon diodes for rectification, and filter the heck out of the power supply. To my ears, they overdrive just fine, with stiff supplies and negligible PSU ripple.
As for "sag", I find that bypassing a pentode screen grid dropper resistor with a capacitor, produces as much sag as I would ever want. The power supply itself can be as stiff as anything.
I won't argue with anyone's subjective preferences - if someone likes the sound of IMD when the amp is wide open, fair enough.
But I do not believe that the entire sonic signature and behaviour of the amplifier, at all output levels, changed utterly, for the worse, simply because PSU ripple was reduced. Remember, these guys even went to great lengths to preserve sag - and they still claim touch sensitivity utterly vanished, in spite of PSU sag, if they reduced PSU ripple. There is a great stench of fish around that claim, IMO.
It all feels like one of those late-night TV car wax commercials. During sleepless nights in grad school, I watched as rusty junkyard cars were restored to showroom finish and lustre, simply by applying some car wax to the rusted sheetmetal. The spokesperson went on to shine a laser pointer off the freshly waxed surface, pointing out that it was now flawless, shiny, and perfectly restored.
Here we have the opposite: a famously revered tube guitar amp that suddenly turns into a rusty junkyard wreck if you wipe off the wax, err, supply ripple. 🙄
You don't think that's a little hard to swallow?
-Gnobuddy
Gnobuddy, the Wildcat project made imho a pretty good effort at configuring a power supply with the same sag characteristics as original, but suppressing the mains ripple modulation added to notes. Yes their comments were subjective and likely biased to accentuate any differences as it was their project, but afaik they were mostly guitar enthusiasts. Reconciling why they noticed a difference may require a little more than blanket rejection of their observations, and a small number of personal observations on amps you've built. If you had set up a similar comparison and observed no noticeable difference then that would be another sample point.
There may be great subjectivity of what people are tuned to hear, or more specifically at what magnitude. For example...
Low damping factor. No doubt its effects are audible, but how audible? Some call it a merely subtle effect, audible in A/B test but not that significant, some claim the very same effect is substantial and an amp without it sounds lifeless and lacks dynamics.
Dynamic crossover distortion. We all likely hear its "swirl" and garbling blocking distortion. Some label it as essential characteristic of class-AB tone and dynamics, some totally hate the very same characteristics and state that amps will sound just fine without them.
Ripple IM. There are camps that label it essential and swear to "fatness" it adds to tone (and there are even patented schemes that add it to tube emulations). Some people go as far as argue that the tone changes at 50 Hz vs. 60 Hz mains frequency. Then, some others hate the out-of-tune ghosting and find it very disturbing while some can't distinct a difference between modulated and unmodulated clipping at all. (I'm in this camp, yes, have experimented with it).
So, there is a chance that no one is "right" here but we as individuals are simply tuned to hear all these effects at different magnitudes and give them different weighs.
Low damping factor. No doubt its effects are audible, but how audible? Some call it a merely subtle effect, audible in A/B test but not that significant, some claim the very same effect is substantial and an amp without it sounds lifeless and lacks dynamics.
Dynamic crossover distortion. We all likely hear its "swirl" and garbling blocking distortion. Some label it as essential characteristic of class-AB tone and dynamics, some totally hate the very same characteristics and state that amps will sound just fine without them.
Ripple IM. There are camps that label it essential and swear to "fatness" it adds to tone (and there are even patented schemes that add it to tube emulations). Some people go as far as argue that the tone changes at 50 Hz vs. 60 Hz mains frequency. Then, some others hate the out-of-tune ghosting and find it very disturbing while some can't distinct a difference between modulated and unmodulated clipping at all. (I'm in this camp, yes, have experimented with it).
So, there is a chance that no one is "right" here but we as individuals are simply tuned to hear all these effects at different magnitudes and give them different weighs.
In my youth I moved from a country with 50 Hz AC to a country with 60 Hz AC, and it took me over a week to get used to the disturbingly different sound of the buzz from the fluorescent light over the sink in my tiny college dorm room. It was too high-pitched, and just sounded really wrong. 🙂
The new 60 Hz hum also put the kibosh on my little trick for tuning my guitar without a tuner: I used to tune the 3rd-fret G on the 6th string to the sound of 50 Hz fluorescent-tube buzz (which is at 100 Hz). That G is supposed to be tuned to 98 Hz (in A440 tuning). Tuning the low string G to 100 Hz got me within 2% of the proper frequency. 🙂
Fortunately, tuning the 5th string, 2nd fret "B" to the sound of 120Hz hum also works pretty well.
Electronic guitar tuners were expensive back then, and I had no money.
-Gnobuddy
Years ago I made a 3W 5E3 and adjusted the bias until I got it to sound right. Found that I just got to the point where there was no crossover notch under heavy drive. How many notes you have ringing out might also cause you to love or hate some power line frequency modulation, along with how overdriven. Single notes or uncomplicated chords might have the power line thicken the sound up for R&R, might just sound garbled if too much is going on. That includes other instruments playing in the same space.
Exactly. What you're saying makes complete sense - IMD caused by power supply ripple can affect overdriven sounds, and may or may not sound musical depending on context and personal preference.
On the other hand, wildly exaggerated claims from the Wildcat project like "...the amplifier is no longer recognizable...limp...cheap...like a ...transistor amp from an...discount store..." are far too overblown to take seriously.
Since the crossover notch under heavy drive appears because the bias point shifts colder once grid current starts to flow in the output tubes, I wonder what would happen to the sound quality, if other measures were taken to reduce the amount of bias shift (bigger grid stopper resistors, smaller grid bias resistors). Maybe one could keep the famous 5E3 sound, without having to bias the output tubes quite so hot.
Tubes are looking to be a critically endangered species right now, so anything that can be done to lengthen tube life is a step in the right direction.
Interestingly, at least some "18-watt" fans seemed to like the sound of heavy crossover distortion during overdrive, so there's that musical preference thing again. Perhaps these guys are Pantera and Slipknot fans.
-Gnobuddy