Small Amp Design for Cleans

[N99]

Lately, I've found myself dissatisfied with the clean tone/headroom and bass response of my current amp (a Silvertone 1481), and have been working on a small amp design with a focus on improving both.

I've focused on making the circuit meet these criteria:

- reasonably simple and without unusual components/component values

- output of ~10W

- excellent low frequency response (I routinely have my guitar tuned down to C)

- clean at all but the highest volume settings (if I need big time distortion I'll use fuzz!)

I'm posting this primarily to get a sanity check; this is the first tube circuit I've designed and I'd like to make sure I've followed appropriate practices, particularly with regards to safety and robustness.

In particular, I'd like to know if the HT fuse values are adequate and if there is any further protection the circuit might need. I'm also a little unsure of the headroom on the 12AT7; am I likely to overdrive it more than the power amp?

[Steve Conner]

The design looks fine. My only worry is that it might not have enough gain with only half of a 12AT7. Even when these small amps are played "clean" they're still overdriving and clipping the peaks somewhat. If you try to avoid that, it'll sound quiet and wimpy.

I'm also unsure on a deeper level, whether a 10W single ended amp is a good fit to your goals. You say you want more headroom and bass, maybe that is subjective speak for "I want a big high-powered amp".

[Chuck H]

Also (and I know I'll get some flak for this), 12at7's don't sound all that great. They have impressive specs as far as dissapation AND amplification AND Vp max, but no tube can do it all and what the at7 has in versitility, it lacks in tone quality IMHO. I usually reserve their use to "function" stages like reverb drivers and trem oscillators. And they're fine in later stages where small signal amplification isn't the goal, like the PI, where your likely to have higher Vp and most of the amplifying is done. It's been argued that tubes don't have a tone. But there are prefered audio tubes just the same.

As Steve mentioned, with only one triode you won't get much amplification. Maybe not enough. Looking a your schem I'll predict about six to ten VAC of signal at the power tube grid. I don't think you'll ever get your ten watts like that. So maybe use both triodes of a more audio accepted tube like a 12au7 which sould get you a little over twenty VAC at the power tube grid (still a very conservative amount). If you use something like a 12ay7 (not commonly available but still easy to find) you'll have enough gain to implement a tone control. That would complicate the design a little but the tonal perception of bigger bass can be greatly affected by the balance of frequencies.

I'm going to guess that this is to be a jazz amp. The idea springing from the little single ended big bottle amps that have become popular in that genre. The idea being that the tone should be pure and the guitar can shine through. Fine. But as Steve mentioned, you do need to get a little of the tube mojo happening by distorting the peaks. This isn't a notable amount of clipping but more like a mild reshaping of the pure sine wave and it warms up the tone and keeps it from seeming whimpy and small.

If this is the type of amp your hoping to "improve" on, and "I" were designing it, I would probably use a 12ay7 preamp, implement a simple low pass tone control and use a 6550 power tube with a VP around 420V. I've always liked 12ay7's for round clean tones. They just seem to "sound" tubey and cool (I'll expect the opinion police to tell me they don't have an inherently different sound than other tubes).

JM2C

[Steve Conner]

Chuck, the "tubes don't have tone" thing is a subtle (EE weenie) point. They don't have anything inside that can shape the frequency response. A tube by itself will amplify all frequencies the same from DC to ultrasonic.

But, different tubes distort the signal differently when overdriven. That means that they generate different harmonics, and in that sense they do have a tone. One that clips harder and generates more high-order harmonics will sound "brighter", even though you wouldn't observe a treble boost if you put on your EE hat and tested it with a signal generator and scope.

To restate in a different way: tubes don't make their tone by shaping frequencies in the signal, as a tone stack does, but by generating new frequencies that weren't in it before.

On the original subject, I don't have anything to add to what Chuck said.

[loudthud]

I agree with the opinions above that you will need more gain. The OT you have chosen should be big enough, but may have too high of a primary impedance for the lowish B+. You'll get more headroom from more B+.

[Wombaticus]

Chuck, the "tubes don't have tone" thing is a subtle (EE weenie) point. They don't have anything inside that can shape the frequency response. A tube by itself will amplify all frequencies the same from DC to ultrasonic.

But, different tubes distort the signal differently when overdriven. That means that they generate different harmonics, and in that sense they do have a tone. One that clips harder and generates more high-order harmonics will sound "brighter", even though you wouldn't observe a treble boost if you put on your EE hat and tested it with a signal generator and scope.

To restate in a different way: tubes don't make their tone by shaping frequencies in the signal, as a tone stack does, but by generating new frequencies that weren't in it before.

On the original subject, I don't have anything to add to what Chuck said.

The 12AT7 has a peculiarity relative to the other members of the 12A*7 family, namely that it begins to draw grid current very early and thirstily. This most assuredly results in unique (and many claim unpleasant) overdrive characteristics.

[Steve Conner]

I've tried overdriving 12AT7s, and I have to agree. I've never got them to sound that great. They can sound OK if only overdriven a little, but if you try to drive one hard and get lots of sustain, the feel isn't quite right.

I heard someone say that they make your guitar strings feel sticky, and while that sounds ridiculous, it struck a chord (if you'll pardon the pun ) with me.

[Gingertube]

Yup,
My experience of 12AT7 too. I use them for Schmitt Splitter in a PP amp and as PP reverb driver, else if I need that sort of (reduced) gain, then its back to octals and I use the 6SL7 or modern production 6N9P.
Actually I'm in the throws of a 6SL7 revival with 2 guys wanting me to build them a copy of my own all octal amp (All 6SL7 preamp section and concertina 6SL7 to 6V6 PP output).
Cheers,
Ian.

[N99]

A Revison

Thanks for the thorough responses!

I wasn't aware of the 12AT7's poor overdrive character. Oh well, at least I learned something new today!

I was careful to be conservative in the initial design, and it looks like I went overboard on that front. 10W probably isn't enough, that was wishful thinking on my part. I've been pondering this for a while and have come up with a new design which may work better. Here are the schematics:



These are the major changes:

-- Preamp is now both sides of a 12AX7; with that much B+, it should have plenty of voltage swing. Not quite sure if the 2nd stage needs a grid stopper, but it can't hurt, right?

-- Power tube is a 6550 (the earliest versions of this circuit used one but I decided to try lower power)

-- Power transformer is 330-0-330 (same as AB165 Bassman), bringing the B+ up to ~467V. The initial design kept the B+ low so I wouldn't have to put the filter caps in series.

-- B+ is now taken from 1st filter cap to get all the voltage available. The power tube screen is at end of the supply to keep it below its max voltage of 400 without bringing the preamp voltage down (why do 6550's need such a low screen voltage?).

Any feedback on this revision would be appreciated; hopefully I haven't made things any worse!

[mooreamps]

I'd set the grid bias to -24 volts, and start from there.. At -52 volts, you may not see much power coming out it ; if running Class A.

-g

[N99]

I'd set the grid bias to -24 volts, and start from there.. At -52 volts, you may not see much power coming out it ; if running Class A.

Whoops, accidentally left that on there. It was just a reminder to myself that the bias tap is 52V. That's not the actual bias I intend to use.

[Gingertube]

Gary is right,
See this
http://tdsl.duncanamps.com/dcigna/tu...sc/6550-2g.gif

Don't worry about 2K5 load vs 3K recommended. 20W at 13% THD is NOT Clean but should be better at lower power. From the data it looks like a voltage double power supply is the go with the screen supply take off the 1/2 voltage point.

Cheers,
Ian

[Chuck H]

WOW! That's a ton more gain. I think you may actually want to tame it down now, considering your goal of a clean amp. There may even be enough gain in that first stage to OD the second BEFORE power tube clipping. Four things I might consider would be to lower the plate loads to 220k or 100k (with a corresponding cathode resistor change to 2.7k or 1.5k respectively). Leave the second stage cathode unbypassed. If this still has too much gain or if the second stage is clipping before the power tube try taking the bypass cap off the first stage and adding it to the second stage. Change the preamp dropping resistor in the B+ rail to a higher value to lower the Vp there. And you'll definitely want to add a 0V reference resistor to the grid of the second stage. Just hang a 100k to 1M resistor right from the grid to ground.

A couple of things with the new design simplify the build. Like the fact that you no longer have plates drawing through the choke. So your choke can be MUCH smaller since it only supplies the screens and preamp. And I don't think you'll need to totem pole the caps beyond the main filter. The screen and preamp tube filters should be fine as a single cap.

And thank you Gary for the good catch on the bias voltage.

[J M Fahey]

What's wrong with building a tried and trusty classic amp which already does what you want?
Designing something really new usually (as in 99.99% of the time) means loooooooong time building, testing, not liking, rebuilding, modding, tearing everything down and starting with a clean slat, repeating the first 3 or 4 steps with *something* changed, etc.
Repeat ad nauseam , which in good English means "repeat until you throw up just *thinking* about it".
jm2c

[Jazz P Bass]

[QUOTE=J M Fahey;25623
Repeat ad nauseam , which in good English means "repeat until you throw up *.
[/QUOTE]
Wow.
Thanks for the Latin class