Im almost there... but maybe not... Valve circuits? [noobie]

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

Sammas

Well-known member
Joined
Sep 30, 2004
Messages
547
Location
Sydney, Australia.
today i started investigating some guitar amp schematics to try and get a grasp of whats actually happening. Big happening for someone quite unfamiliar with electronics compared to many here.

In the following schematic i can kinda see what is going on now... compared to this morning. Two preamp common cathode (or fully bypassed cathode) gain stages before the output stage. I seem to understand the purpose of the coupling caps, grid resistors, cathode resistors and bypass cap, plate load resistors, load resistor (or grid resistor of the next gain stage)... I've even looked into some of the maths about biasing and the gain of each stage...

Now for the part which confuses me? probably due to a lack of knowledge and an overwhelming large amount of infomation for one day. In the schematic there is 240v supplied to the plate load resistor - The plate voltage supply. I've noticed on several different schematics that they not only vary between designs but between stages as well. For example in one design there are 3 different voltages supplying 4 different gain stages... One (the last one before the output stage) had a plate voltage supply of more than 300v (all were 12AX7's). I don't understand why, and have been searching for sometime trying to figure out what determines the voltage supplied to each stage... i've looked for equations or explainations but can't seem to find an answer. I'd be at the library right now looking for some answers except its almost 10pm.

the coupling caps obviously isolate the grid in the next stage from the high voltage DC... can anyone point me in a direction towards an answer? or more infomation? or anything?

It would also "kill two birds with one stone" if someone could point me to a source of infomation regarding the actual voltage supply in the bottom schem. I can seem to find any answers on the maths (maybe?) behind creating stepped voltages like done in the schematic. For example if i wanted to add 210v supply after the initial 285v & 240v what cap/resistor configuration would get me there...

Im probably in over my head, but its been a productive day and it would be cool if you guys could give me maybe even just a phrase to start the ball rolling again. Tomorrow im off to the library to get me some books on the matter also... but times-a-wasting :wink:


Firefly_sch_R1.gif
 
The prime goal is not different voltages, but different levels of PSu filtering. The preamp and gain stages that need clean supply is fed from supply with additional R/C decoupling for lower noise, the unchritical driver and output stages gets dirtier but higher supply...

Jakob E.
 
arrrgh! i see. That probably explains why in almost every guitar amp schematic i've looked at there has been that similar "slope" of plate supply voltage. The earlier stages are supplied by the lower more filtered voltage then it steps its way up towards higher, less filtered voltages that are used in the output stage and transformer...

I guess the next thing to do is study up on power supplies and the output stage.

Thank you very much :thumb: :guinness:
 
[quote author="gyraf"]The prime goal is not different voltages, but different levels of PSu filtering. The preamp and gain stages that need clean supply is fed from supply with additional R/C decoupling for lower noise, the unchritical driver and output stages gets dirtier but higher supply...[/quote]

Which they need, because they're expected to handle larger signal voltages.

So in effect you have input stages that need extra filtering but can get along with lower voltages (especially if you plan to have them overload a little for effect), and driver/output stages that need higher voltages but can get along with less filtering. Everyone winds up happy.

Peace,
Paul
 
> The plate voltage supply. I've noticed on several different schematics that they not only vary between designs but between stages as well.

More voltage means more output, or less distortion for the same output.

Less voltage is cheaper. "Cheap" dominates commercial design.

Output stages need more voltage. Input stages work with smaller signals, so they can work at lower voltages.

The large levels in an output stage will not be harmed by a little power buzz. The small signals in input stages would be harmed by even a little-little power buzz.

Output stages need lots of power; input stages only a little.

So you start with your biggest problem: supplying enough voltage and current to meet your output goal.

Then you figure how clean that big-power has to be, and design the filter. For many push-pull stages, which tend to cancel power buzz, you can get away with just one capacitor. For many single-ended stages, which don't cancel power buzz, such a design leads to an impossibly large (or un-cheap) capacitor, so a CRC or CLC filter is picked.

Then you figure how much buzz the driver stage can stand, compare to the buzz in the big-power supply, and (almost always) design a power filter to reduce the buzz. It is much easier to clean-up the few-mA of the driver (and maybe earlier stages) than to clean the big-power filter to the cleanliness needed for driver and smaller stages.

Say we needed 20dB buzz reduction between big-power and driver supply point. And we don't want a lot of voltage drop: power output tubes need big grid swings, so drivers make pretty large plate-swings, which demand pretty big supply voltage.

Let us (idiotically) select an RC filter with R=1Ω, which will give a lovely low drop (0.010V at 10mA). To get 20dB or 10:1 ripple reduction, we need a capacitor that is 0.1Ω at 100/120Hz. This comes to 15,000-20,000uFd. Since the driver usually works at 200V-400V, this has to be a 300V-450V rated cap. 22,000uFd at 400V is bigger than a coffee-can and about $300: ouch!

So we have to take some voltage drop. If we want a minimum cost design we might even allow 40% voltage drop to a low-level stage. For a driver we want a bit more, but 10%-20% is a good starting place. 400V*12% is 50V drop; at 5mA, that is R=10K. Still assuming 20dB filtering, we need a capacitor that is 1,000Ω at 100/120Hz. This comes to 15-20uFd. Such a cap in 400V rating is affordable and compact. Indeed 10K+20uFd is a popular filter.

The driver has gain, so the stage before the driver has to get even cleaner power. Assuming the basic circuits have small supply rejection (pentode ~ 0dB, triode ~5dB), each stage needs to be cleaner by about the amount of its voltage gain. A 12AX7 stage, gain ~30-70, needs 30dB-37dB power filtering from what is acceptable for the stage it feeds. Since the current is more like 1mA, we can use a resistor 5 times larger. Since the signal level is much smaller than the driver, we can allow even lower supply voltage on this stage, and use an even larger resistor. 100K and another 20uFd cap will give more than 40dB filtering.

All of this leads to a "slope" or series of steps, going down from the big stage to the little input stage.

I'm assuming that signal frequencies are not a lot lower than power frequency. If you are going for bass lower than 20Hz you have another problem. The several amplifier stages increase the signal, but the power system leaks output signal back to the input. Feedback! With just 2 stages, it is negative feedback: your bass response is a little worse than you expect. Three stages under one power filter give positive feedback, and your bass response rises, possibly to oscillation ("motorboating": the "putt-putt" of subsonic oscillation sounds like a 1920 motorboat). Don't put three cascaded stages on one power point without doing analysis. Even with two stages, your power caps should probably be greater than coupling-cap times stage gain. 0.1uFd coupling cap, second stage gain=50, use at least 5uFd, and in this day of affordable Japanese caps, probably much more than that.

> In the schematic there is 240v supplied ...One ...had a plate voltage supply of more than 300v

Input stages work with very small signals. And while higher voltage gives lower distortion for the same signal level, it takes heroic voltages to get a major reduction in THD. So the HeathKit tube HiFi preamp only fed 90V to its input stage: at 0.1V signal that was plenty, and it gave ~200V of drop to use in filtering. With beam-power output tubes, it is sometimes possible to run the drivers with half the B+ of the output, but we usually feed them 80% so it won't be too fussy about component tolerance and drift. Some designers, especially when they have low-cost power caps, just like to run low-level stages at high voltage.

Guitar amps get into special flavors. Some run the input tube at barely-enough voltage so it will add color at moderate signal level (or just to save a buck in the filtering). Some run very high voltage so the input stage stays clean: then you can fiddle some knobs and force the second stage clean or dirty. The 12AX7 is a funny tube that actually works different at high voltages (OTOH, it will pass current at zero voltage, the grid is inside the cathode cloud).
 
None of this is specific to valves/tubes. Look at early transistor amps. Output stage ran on raw rectified/filtered 50V. Tone-control stage got fed ~30V with an RC filter. Phono preamp got 20V through another RC-filter. Only with the rise of cheap Zeners and LM7800 chips did we start seeing solid +/-15V rails throughout the non-power stages. Even then, the phono preamp might be fed +/-14V dropped through a couple RC filters to clen-up the Zener noise.
 
Wow! Thanks pstamler & PRR for talking to the time to share some knowledge. Things are certianly getting clearer :thumb: :guinness:

I've just ordered the parts to build that little amp... I figure it will be a good place to start and once its build and working i might tinker with it, maybe add a simple tonestack or an extra gain stage...

Thanks again! You help is very much appreciated!
 
I wish i knew about this site when i was learning...
you guys rock
 
Back
Top