12AU7 - what do you use them for?

[Baguette]

Hello there,

I see few "classic" amps using the 12AU7. Weird, it looks like a useful tube.

1/ Like, why aren't they used more often as a LTP to really bring some juice into a pair of 6L6s?

2/ Or, why did Fender use the 12AT7 as a reverb driver tube? Seems like a AU7 would have been a better choice since the tank requires big current, not voltage.

3/ Anyone's experimented with 12AU7 in LTP and as reverb driver tube?

4/ Would a 12AU7 reverb driver tube be fine with the usual 22K primary OT Fender's been using since 1963?

Thanks for the answers!

[sluckey]

Sunn used a 12AU7 for reverb driver in all their old reverb tube amps.
http://home.comcast.net/~seluckey/amps/sunn/sunn_sceptre_1971.pdf   (see page 2)

Magnatone used them for reverb drivers and phase inverters.
http://home.comcast.net/~seluckey/amps/magnatone/Magnatone_M10A.pdf

Hammond used bucketfuls in their organs. Lot's of them in the tone generators. Even the vibrato module used them. Here's a project I did that uses 3 of 'em...
http://home.comcast.net/~seluckey/amps/warbler/warbler.htm

[kagliostro]

Some opinions about

http://archive.ampage.org/threads/4/gagd/267061/12au7_for_reverb_Why_not-1.html

http://www.tdpri.com/forum/amp-central-station/101973-12au7-reverb-driver.html

http://music-electronics-forum.com/t24092/

Kagliostro

[tubenit]

They are useful for active effects loops.

with respect, Tubenit

[blues man]

I set up a 5f8a for bass and used a 12ay7 first stage, 12at7 second stage, and a 12au7 for the pi. It sounds great for bass and my late 50's national archtop for clean playing.

[RicharD]

12AU7's make fantastic PI's and cathode followers.  Capable of driving a very low impedance.  They make a sweet clean gain stage when looking for only a gain of 10.  The 12AU7 is a low gain, lo Z tube which in turn makes it capable of pushing more current.  The 12AX7 is a hi gain hi Z tube which is happy as a clam at 1mA.  You don't see 12AU7's much in geetar amps cuz geetar amps are mostly about gain.  12AX7 is the right bottle for gain.  12AX7's were used as PI's too not so much because they're the best bottle for the job but so as to limit the variety of tubes needed.  12AU7's are much more prevalent in hifi and musical instruments.  It's a very useful tube but not so much for geetar amps.

[Geezer]

I've got ~~50 12AU7's from an old Conn organ........I've often toyed with the idea of cascading a bunch of them to get the the gain needed for guitar.

They are sometimes useful in Overdrive circuits to get the OD under control (without futzing with the bias, gridstoppers & such).

G

[Baguette]

Is there anything I should do tune a LTP or a reverb driver for 12AU7 instead of AT7?

(I'm thinking classic AB763 here).

Thanks!

[Tone Junkie]

Cant you take one side and run it into the other for about the gain of 1 side of a 12ax7  10 times 10 equaling 100.
Thanks Bill

[firemedic]

I like the 12AU7, the clean tone is real nice. I have a 5814 (same thing) as a PI in my TOS.

[RicharD]

>Is there anything I should do tune a LTP or a reverb driver for 12AU7 instead of AT7?

It should work just plug-n-pray.  An LTPI has gain = to about 20 in the AB763 circuit.  Obviously, that's gonna go down.  I suspect about AV=7ish.  I'd tweak the cathode resistor for a bit more current and better symmetry.  You're gonna lose a lot of gain using it as a reverb driver.  Probably so much the reverb will be barely audible.

>Cant you take one side and run it into the other for about the gain of 1 side of a 12ax7  10 times 10 equaling 100.

That's how it works and actually it would be hotter than a single 1/2 of a 12AX7.  The typical Leo 100k/1k5 gives you an AV of almost 60.  The Marshall 220k/2k2 delivers an AV of 75. 

[firemedic]

I tried one in the reverb send socket of my AB763 when it still had both stages paralleled (it's a single stage now) and it worked fine. Of course there was less reverb but more than enough for me....

I have seen on a few forums that a 12au7 is not appropriate for use in a guitar amp but as with many aspects of this hobby, if you try it you may like it, or not. It was my favorite input tube on my 1st build, it just sounded the best to me. In fact the 5751 is about the highest gain preamp tube I have, in any amp, handling the signal. Usually a 12ay7 is quite enough.
Sometimes your amp may be a little too gainy and it's a great way to easily tweak & tame it.

[alerich]

I bought a Blackheart Handsome Devil head and two cabinets this weekend on Craigslist. The HD is a two 12A_7 preamp tubes and two EL84 power tubes amp. The original owner wanted as much headroom as possible and wanted to be able to turn the drive up higher without it breaking up and he put a 12AU7 in V1 for the two gain stages and a 12AT7 in the PI. It sounded pretty good and broke up far later in the curve but I thought the tone suffered as a result. Too weak. I exchanged the 12AU7 for a 12AT7 and it made the clean tone much fuller. The clean tone on these amps is outstanding. The OD sound is quite nice, as well. More old school than over the top.

That Firefly one watt amp that is floating around the net uses a self split 12AU7 for the power tube.

[Ritchie200]

Yup, Firefly uses them.  Also used as a driver tube in my Major to swing the big current to the KT88s.  I've also used 12A_7 tubes to calm a nasty higain preamp like a 5150.  However, the U7 is just to wimpy to use IMHO.  Use a GOOD U7 in the Firefly and it sounds great.

Jim

[SoundmasterG]

They used a 12AU7 on the input in a Vox AC100, and they also used one as the PI using a floating paraphase inverter. These are clean amps that just grit up a bit at max volume, but the sound is fantastic for what it does.

Greg

[jeff]

The original owner wanted as much headroom as possible and wanted to be able to turn the drive up higher without it breaking up and he put a 12AU7 in V1 for the two gain stages and a 12AT7 in the PI.

 Is there a difference between using a lower gain tube vs. just keeping the gain knob at a lower setting? In other words, is using a 12AU7 on ten the same as using a 12AX7 on 3 or 4? Is it just psychological because it's cleaner at 10 but in reality it's the same amount of gain you'd get with a 12AX7 on 3 or 4?

[tubenit]

Is there a difference between using a lower gain tube vs. just keeping the gain knob at a lower setting? In other words, is using a 12AU7 on ten the same as using a 12AX7 on 3 or 4? Is it just psychological because it's cleaner at 10 but in reality it's the same amount of gain you'd get with a 12AX7 on 3 or 4?

Great question!!  My opinion is the tone is different.  I have used 12AX7, 5751, 12AT7, 12AY7, 12AV7 and 12AU7 tubes and tried to regulate the amount of gain or volume or overdrive and still found a total difference even when the volume or OD sounded about the "same".  I acknowledge my answer is based on some subjective experience.

With respect, Tubenit

[jeff]

I guess on ten you don't have the series resistance of the pot? Maybe? Don't know. You lose high end as a pot is turned down, don't you? I guess if three is all you need the lower gain tube would be better. But I like my amps to go to 11. Cuz' if you need that extra little push over the edge and you're on 10 where can you go? The answer is none.

[simonallaway]

Cuz' if you need that extra little push over the edge and you're on 10 where can you go? The answer is none.

I never get tired of that quote

[firemedic]

A true Zen saying, as frank would say.
I like all the posts on this thread. Fact is I still dig my 12AU7s- I did some tube rolling today in the TOS & the U7 still came out on top in the PI.

[RicharD]

>Is there a difference between using a lower gain tube vs. just keeping the gain knob at a lower setting? In other words, is using a 12AU7 on ten the same as using a 12AX7 on 3 or 4? Is it just psychological because it's cleaner at 10 but in reality it's the same amount of gain you'd get with a 12AX7 on 3 or 4?

I'm sure mathematically you could dial in equal gains.  Would it sound "exactly" the same, maybe but probably not.  Does a 12AX7 with Rp=220k and a Rk=2k2 sound the same as one with a 100k Rp and 1k5 Rk?  In most cases you can simply plug a 12AU7 into a circuit designed for a 12AX7 and it will work, but it's usually less than ideal.  Typically the first thing to noticeably suffer is high frequency response.  Perhaps this is why 12AU7 have a reputation for sounding dark?  In a properly biased circuit, the 12AU7 is not at all dark.

A picture is worth a thousand words. 
Fig1 = 12AX7 in typical Fender configuration
Fig2 = 12AU7 in the same configuration
Fig3 = 12AU7 in a "more correct" configuration

[darryl]

Is there a difference between using a lower gain tube vs. just keeping the gain knob at a lower setting? In other words, is using a 12AU7 on ten the same as using a 12AX7 on 3 or 4? Is it just psychological because it's cleaner at 10 but in reality it's the same amount of gain you'd get with a 12AX7 on 3 or 4?

Using a lower-gain valve to raise the clean headroom may reduce the overall treble response if there is a bright cap across the volume control. Once a volume control is turned up past the 12 o'clock position, the bright cap becomes steadily less effective.

[alerich]

Is there a difference between using a lower gain tube vs. just keeping the gain knob at a lower setting? In other words, is using a 12AU7 on ten the same as using a 12AX7 on 3 or 4? Is it just psychological because it's cleaner at 10 but in reality it's the same amount of gain you'd get with a 12AX7 on 3 or 4?

Great question!!  My opinion is the tone is different.  I have used 12AX7, 5751, 12AT7, 12AY7, 12AV7 and 12AU7 tubes and tried to regulate the amount of gain or volume or overdrive and still found a total difference even when the volume or OD sounded about the "same".  I acknowledge my answer is based on some subjective experience.

With respect, Tubenit

I'm with Tubenit on this one. It's not just about the gain factor of all these different tube types (let alone mixing different manufacturers in, as well). When you start looking at factors like current capability and transconductance it almost starts to be an "apples and oranges" things between the different tube types. They all essentially do the same thing but how they do it and what you end up with is the secret. And this is just within the confines of one amp/design. Move to another amp and it all changes again. That's why tubes are such wonderful magical things.

[Tone Junkie]

Guys Im kind of a dork i have a hard time reading the differant values set into that load line curve program you use.
 Are there any schematics that use that particuler set of values I could study, it helps me to see whats what and were it goes, what I really need to do is get more familier with the program so I can see it in my minds eye when I look at it.  minds eye is that kind of like a belly button window from the old hendrixs song.
Thanks Bill
PS I think I need more coffee

[jeff]

 I'm gonna eat up all your chocolates.
I assume Ra is plate resistor Rk is cathode resistor bypass cap is on the right in the middle above the graph. What does Cap refer to coupling cap?


What program is that it looks cool.

[RicharD]

You don't really need software, but TubeCad is economical and quick once you get the hang of it.  The first example is a typical Fender V1 stage.  I'll run through the parameters and they are first.  Obviously you want to first select the circuit type.  In this case it's a grounded cathode amp.  Here is the basic schematic so you can see where all the parameter components are:

Going down the list, first you select the tube (12AX7 yeah).  Next is the coupling cap.  This sets the -3dB lo frequency.  Worry about that later.  The next 2 are desired current and voltage available.  Guess at the current because you'll have to go back and fine tune it at the end if you want precise estimates.  Now it get's fun.  Ra is the plate resistor.  Rule of thumb for symmetrical output is Ra=2xRp where Rp is the plate impedance of the tube from the spec sheet.  The 12AX7 has a Rp of 62500.  Leo used a 100k, not quite 2x but close enough.  

Now we are ready to plot a load line.  You don't have to have software to do this.  Print out a curve sheet and use a pencil and ruler.  The point on our X axis (voltage) is VB+.  The point on our Y axis (current) is VB+/Ra.  Now you plot a line connecting the 2 coordinates.  This is the load line or output line.  Obviously playing with Ra & VB+ changes the slope of the line.  For most "symmetrical" output, you want the load line to be perpendicular to the curve lines.  For more gain, you slope more horizontal.  

Looking at the Leo line at we go into cutoff at 2mA & 90V.  Sticking with symmetrical performance, you want to set your operating point or Vtube 1/2 way between Vcuttoff and VB+.  In this case  about 185V  Plot a line straight up.  Now we know the current and the bias voltage of the tube so we can calculate the cathode resistor.  1.5V/1mA=1500 ohms.   We can also calculate gain.  Our operating point is -1.5V.  Plot a line straight down from the point where -2 crosses our load line.  Call it 215V.  Now do the same for the -1V point.  Call it 155V.  215-155=60 so for 1 volt of input swing you get 60V out.  AV=60

Back to TubeCad, We got Rk from the load line so Rin is next.  This effects -3dB high frequency.  Let's worry about this later.  Rl is the volume pot.  Rule of thumb is for Rl to be 5x Ra.  Rl and Ra are in parallel.  If Rl get's too close to Ra, then their parallel resistance must be taken into consideration when plotting the load line.  If Rl is large enough, it's affect on gain is negligible.

[RicharD]

Here is a step by step load line procedure.  I used a 12AX7 again but this time I am shooting for more gain by increasing Ra.

OK, so let's take this 1 step further now and see what the input threshold of this circuit is.
Vcutoff is 15V, Vtube is 155V, this is a difference of -130V.  VB+ is 300V which is  +145V above our operating point.  This means the negative cycle will start to clip a little before the positive cycle.  Vout max is 2x the smaller of the 2 numbers, in this case 130x2=260V
Vout max/AV=Vin max or Vthreshold.  260/83=3.1V    That is slightly greater than 2x Vbias.  The most you can hit a tube with w/o distortion is 2xVbias so Vthreshold is 3V.  It'd be pretty hard to clip this tube with just a geetar as an input but if it were duplicated and installed as V2, you could knock it silly.  Let's say your guitar is putting out 1 Volt and it's feeding into V1 so you're getting 83V out.  V2 is a duplicate and has a 3V threshold before it starts to clip but we're hitting it with 83V, that's over drive central, or you dial the wiper of Rl of V1 back until you get to 3V for clean undistorted signal.

Now let's look at the curve sheet for the 12AU7.

Right off the bat, notice that plate current tops out 10x higher than a 12AX7. It has curves from 0 to -30V vs. 0 to -5V of the 12AX7.  Right off the bat you can plan on drawing more current but being able to hit the tube harder w/o it distorting.  Cool so let's design a V2 circuit.  Let's get a little more B+, 350V sounds good.  Now I drew the load line approximately perpendicular.  350V/23mA is 15k ohms.  V cutoff = 120V so  Vtube=235V, Vbias=-9V, Itube=7.5mA, Rk=1200  For simplicity sake I plotted Vin=2V and Vout=22V therefore AV=11.  vtube is right dead center of VB+ and Vcutoff so Vout max = 230V.  230/11=21, again greater than 2x Vbias so call Vthreshold= 18V  Your mileage may vary.  I've heard the rule of thumb for Vthreshold is to be equal to Vbias.  None the less, you can smack a 12AU7 a lot harder than a 12AX7 before distortion so it is a viable candidate for V2 in a clean amp.  Another benefit is the lower output impedance.  It's not as low as a cathode follower but low enough to drive a tone stack efficiently.  Lok at many of the popular guitar amp tone stacks.  The average loss is -10dB.  This little circuit is making 21dB.  Sandwich a -10dB tone stack between 2 of these and you walk away with +32dB.  V1 described is about 38dB.  This totals +70dB or AV=3100.  You got plenty of output swing to drive any PI or single ended bottle.  It'd probably be a nice clean amp.

[Tiny_Daddy]

Just got a Traynor Guitar Mate that uses 12AU7 as the first two gain stages. Hoffman site has the schematic.

[thermion]

Well explained, Buttery.
Send this one to the archive?

I really dig a good 12au7 as V1 in a bf bassman.

[jeff]

We're used to seeing 1/2W plate resistors.
Before we all start ripping out our 100Ks and putting in 15Ks and 12AU7s:

Would you use a 2W 15K resistor?

We're dropping 115V(350-235) at 7.5mA
115V X .0075mA=.8625W

Doubled to be safe 1.725W

[DummyLoad]

We're used to seeing 1/2W plate resistors.
Before we all start ripping out our 100Ks and putting in 15Ks and 12AU7s:

Would you use a 2W 15K resistor?

We're dropping 115V(350-235) at 7.5mA
115V X .0075mA=.8625W

Doubled to be safe 1.725W

why are you doing that?

take a look at resistance coupled amp charts here...

scroll down to page 7 - what you'll notice is the load resistance values (Rg in charts) are lower - in most cases this the volume pot or tone stack - since Zout of the 12AU7 is much lower than that of the 12AX7. this is for a reason; HF roll-off - as you decrease Zout, then small values of load capacitance have a greater affect on the HF corner frequency roll-off with larger values of Rg that your used to seeing with hi-mu tubes.

in summary: it's not necessarily the plate or cathode resistors you'll need to replace, but the load components e.g. volume pots, coupling caps, and bypass caps. scroll back up to the top of the document - there's a blurb on the corner frequency for the values given in the tables for Cc and Ck - IIRC it's 100Hz.

notice in the charts for the 12AU7 that there's little deviation in the value of Av, so just find the sweet spots with trial-and-error and use the tables to get you on the dart board.

on tube-cad sim with the following parms:
12AU7
300V B+
100K Ra
1K5 Rk
500K Rload - the pot
Ib =~~2.1mA
Vgk =~~ -3.1V
Cc - .005uF - for roll-off at around 65Hz; .0068 get down to ~~45Hz.
Ck is 22uF

with the static conditions above, Vra is ~~ 200V. the dissipation is then roughly 200V*2.1mA = 420mW - so yes, i'd probably use 1W resistors - or run with a warm 100K half-watt.

--DL

[jeff]

I'll have to read this one again. The 3RD graph lists Ra as 15K and the schematic lists Ra as the plate resistor. I thought he was suggesting using a 15K plate resistor. I guess jumped the gun.

[Tone Junkie]

Thanks Buttery now that i can sink my teeth into.
I will print those out and put them on my wall to study you guys give me the best stuff.
Bill  

[jeff]

I'll have to read this one again. The 3RD graph lists Ra as 15K and the schematic lists Ra as the plate resistor. I thought he was suggesting using a 15K plate resistor.

in summary: it's not necessarily the plate or cathode resistors you'll need to replace, but the load components e.g. volume pots, coupling caps, and bypass caps.

Can you show me what I'd need to change to modify this for better use with a 12AU7 in V1.
http://www.el34world.com/projects/images/stereopreampschematic.gif
I've got a couple 12AU7s and I'd like to try them out.

  Are you saying just use a different pot instead of a 1M and change the caps and 470K resistor?

[DummyLoad]

Are you saying just use a different pot instead of a 1M and change the caps and 470K resistor?

yes.   

[jeff]

 OK I'll experiment with caps but what's a starting point for the volume pot to make a 15K load?

Should I use a 25K pot as a volume?

[HotBluePlates]

More general rules of thumb:
If you have a given plate load resistor, you want the following resistance to ground to be at least double that, while 4-5 times or more is better. If the volume pot or grid reference resistor smaller than recommended, it will actually lower the gain of the stage prior.

Example: 100k plate resistor, use 220k (or preferably bigger) resistance after the coupling cap.

[jeff]

 Thanks that's good to know.

But if I'm understanding this correctally were trying to change the load line to 15K  without changing the plate resistor.

........why are you doing that?(in regaurds to using a 15K as a plate resistor instead of a 100K)
in summary: it's not necessarily the plate or cathode resistors you'll need to replace, but the load components e.g. volume pots, coupling caps, and bypass caps.......--DL

We're trying to load it down. What volume pot would load down the tube to get a 15K load line using a 100K plate resistor?

The example is a 12AU7 with a 15K load line.

If I'm missing the point again please explain.

[Tone Junkie]

Jeff Im glad your asking all the hard questions Im sitting here following along, you just keep asking what Im thinking.

Thanks Bill

[jeff]

Tone Junkie, It's good to know I'm not the only one. Sometimes I feel like I'm being a big p.i.t.a.

 I'm curious to see what Butterylicious has to say. I think I follow what he's saying, great explination. I just don't know what Butterylicious had in mind to arive at a 15K load line if, as Dummyload suggests(if I'm reading this correctally), we're not changing the 100K plate resistor.

I have a bunch of 12AU7s. I'd like to get the best use out of them but I am a little unsure of what's being suggested here.

[RicharD]

Dummy Load and I are giving 2 different examples.  My example is designing from scratch using a curve sheet and load line. In my example I am trying to maximize for a low Z out and symmetrical performance.  Dummy Load's example is using the RC chart and working around existing Leo Fender values.   Both methods and both circuits will work.  Both circuits will deliver about the same results because the 12AU7 is a very linear tube.  Since the 12AU7 is a low gain tube, changing how the tube is loaded and biased isn't going to change the gain all that much.  There is more than 1 way to skin a cat and this is what is confusing you.

[jeff]

There is more than 1 way to skin a cat and this is what is confusing you.

Always does. Thanks
      Jeff.

[jojokeo]

Late to the party but after reading all of the terrific in-depth info there's something worth pointing out. Biasing to the left side of the load line or 0 volts is to bias towards grid-current limiting and to the right toward HT is termed cut-off.

This is important in understanding what is happening in regards to input sensitivity, headroom, how the type of tone of distortion varies by the introduction of even order harmonics vs odd, how the input and output signals are being clipped and the resultant tone characteristics associated with this.

One of my favorite quotes from Merlin's books, "Because the way each valve is biased affects the way each stage will distort, biasing has a large effect - perhaps the greatest effect - on the overall tone of the amp, whether that be clean and jazzy, warm and bluesy, dirty, gritty, or even hard and fuzzy."

[jeff]

"Because the way each valve is biased affects the way each stage will distort, biasing has a large effect - perhaps the greatest effect - on the overall tone of the amp, whether that be clean and jazzy, warm and bluesy, dirty, gritty, or even hard and fuzzy."

Amen to that.

[RicharD]

>Biasing to the left side of the load line or 0 volts is to bias towards grid-current limiting and to the right toward HT is termed cut-off. 

Ayup, I totally bombed on some terminology. 

[jojokeo]

Always very much appreciate yours and DL's stupendous explanatious terminolysis. 

[RicharD]

Always very much appreciate yours and DL's stupendous explanatious terminolysis. 

You wanna do some proof-reading?

http://www.sotxampco.com/Temp/Grounded%20Cathode%20Amplifiers.pdf

[jojokeo]

You wanna do some proof-reading?

http://www.sotxampco.com/Temp/Grounded%20Cathode%20Amplifiers.pdf

The word play was just that - fun and nothing personal - it actually was complementary, however - the link also discussed the same saturation (grid current limiting) & cut-off yeah? What's to read (that I haven't read before)? That was rather ambiguous & I don't get it unless...(I will take the high road).

Please check out page 7 and page 9 and gives us some feedback if you disagree on something?
http://www.freewebs.com/valvewizard1/Common_Gain_Stage.pdf

[RicharD]

I didn't take any offense.  I'm seriously looking for someone to proofread this to make sure my terminology is correct.  I am dyslexic and I flip flop stuff all the time.  I want to help people not confuse them.

[Willabe]

So you did write that. Please finish your work, I'd buy it!

I don't know about you guys, but at least for me, I find it very helpful to read the same type of thing, --but-- from different points of view. Thats not to say you don't have anything new to say, I think you do. I've read a lot of posts that you and Dummy Load have done with a lot of very interesting builds with different types of tubes.    

IMHO, I think you have a lot to offer in writting a book on tubes/tube amps. I think you'll do a great job, go for it.      


                 Brad