Back in the UK, stereo pre

[loogie]

I'm in London for a bit and my colleague and I would like to do a simple stereo pre-amp around the 12AU7.  Anyone out there have a favourite place for parts?  Especially 9-pin sockets and scraps of aluminium.

We're going to try the old 230v/12v, back to back transformer setup with DC heaters.  I was thinking about wedging a James tone stack between the two halves of the 12AU7.  We're still working on the plan.  Anyway, Maplin's will have just about everything we need except the sockets.

[Merlin]

Cricklewood Electronics is pretty popular, although being London you can expect to pay a bit more than usual.

http://www.cricklewoodelectronics.com/Cricklewood/search.php?mode=search&page=1

[loogie]

Just got back from Cricklewood Electronics.  I was impressed with the knowledgeable staff.  We're both feeling quite upbeat about our prospects regarding this latest project.  Thanks for the tip!  A little crowded on the tube from Baker Street to Charing Cross, but a pint at The Harp eased our anxiety.

[loogie]

To our complete amazement the thing works.  We ended up using a 12AX7 for each channel with a James tone stack in between.  The canonical configuration: 100k, 1.5k, etc.  I'm sure its not audiophile quality and no doubt the circuit is ridden with flaws, but it sounds damn good.  We were rolling on the floor like crazy monkeys.  No hum and you have to turn it all the way up and put your ear next to the speaker before you hear much in the way of hiss or buzz.

We're feeding it with a cheap cd player and a Mac.  Event YouTube sounds good.  We're sending it to a Samson power amp.

We tried everything from Miles to Mahler and just couldn't believe the clarity and definition.  Jeez, I wonder what a good tube pre sounds like? 

[RicharD]

I don't know about parts in the UK, but I'd be interested in seeing your schematic.  I'm working on a similar project with/for Kagliostro.

http://www.el34world.com/Forum/index.php?topic=8860.0

At one point, I had a passive James tone stack in the circuit, but it totally killed frequency response.  This weekend I am going to experiment some more.  My plan is to make the James stack active by putting in the feedback loop of a plate follower. 

[loogie]

This is just a line driver, I guess, with eq.  We'll add a mic input on the next trip.  I thought about using some negative feedback to clean up the noise I anticipated we'd have and to lower output impedance, but it sounded fine so I bagged it. 

The schematic is just on paper now, but I'll draw it up.  In the meantime just imagine a 12AX7, 100k, 1.5k, 22uf feeding a James which is loaded by a 1M pot (volume) which feeds a 220k pot (balance) and then to another triode stage setup the same way as the first one.  Its loaded by a 270k resistor.  All dual pots because its stereo.  Crude, but nice.  We did the two cheap doorbell transformer routine with DC for the heaters.

[kagliostro]

 

Butterylicious  is a very much likeable person

he always wants to joke

he is working on a plan to give help to me

not with me

of sure I'm not to its height 

Kagliostro

[RicharD]

>I thought about using some negative feedback to clean up the noise I anticipated we'd have and to lower output impedance, but it sounded fine so I bagged it.

A plate follower does does decrease distortion, dramatically improves power supply rejection ratio, and it lowers the output impedance.  It also increases input headroom.  The sacrifice is gain but fortunately our friend the 12AX7 has plenty of gain, so much in fact that I typically do not bypass the cathode when experimenting with "clean" circuits.  Frequency response and input headroom drastically improves.  PSRR gets worse but since your amplifying less, it's pretty much a wash.  Have you done any sort of frequency response test?  Sorry to drag on but I've gotten very interested in this corner of thermionics.  I'm just a tad burned out on geetar amps.  I attached a TubeCad comparison between a grounded cathode and a plate follower using the popular 100k/1k5 configuration @300V (coupling cap set at 2.2uF for acceptable bass response).  Fairly amazing the difference a single 1M cap makes.

[loogie]

By all means drag on.  Next trip we'll try to do some structured testing.  All I can tell you now is that the bass can be made to boom and the highs can be driven to stridency.  I wish I could tell you more, but it does a real nice job with everything we throw at it.  We tried the traditional settings for reggae, max bass and treble, and it was dead on.  It covered every inch of 'Kind of Blue' and Ives 'From the Steeples and the Mountains' with all its bells and cymbal crashes was brilliant.

Thanks for the TubeCad capture.  Someday I'm going to get that app.  I've used plate followers before in several amps and while I'm not yet burned out on geetar amps I'm definitely keen to branch out especially after my experience with an Audio Research amp.  I want to do a real hi-fi pre and a power amp -- and an RIAA section.  In fact on the list of TODOs next time I'm in London is to build a power amp for our pre.  I also want to do some synth modules.

I'm sure the pre we built looks awful on paper, but neither of us is an audiophile; just one old rocker and a numbnutz from the mid-west

[cigarboxblues]

like the topic ,makes me feel i can acheive a guitar amp one day thanks for the read guys 

[HotBluePlates]

I thought about using some negative feedback to clean up the noise I anticipated we'd have and to lower output impedance, but it sounded fine so I bagged it.

Negative feedback could be used to alter the apparent impedance of the amp, but I don't think it would do anything for noise except make it worse.

I had just been reading High Fidelity Circuit Design, by Norman Crowhurst, and he notes that since noise is a random event, feedback would have to be applied instantaneously to reduce it. There's always a very tiny bit of delay in feedback reaching from the output back to an earlier stage, so "instantaneously" is ruled out.

In any event, the amp has an overall gain before feedback, and we all know that as gain goes up, noise almost always goes up as well. With feedback, you still have the same amount of gain within the amp, but the apparent gain is reduced by the amount used for feedback, to alter impedance and/or reduce distortion.

So after feedback is applied, you'd wind up turning your amp up higher than you would without feedback, because the apparent gain is less. And noise comes up as you turn up the volume.

So an amp with a good bit of feedback will seem noisier than the same amp with no feedback.

Norm notes that the solution is usually to use feedback around fewer stages, and to design the amp to be as low-noise as possible before applying feedback. So no free lunch today...

[RicharD]

>So after feedback is applied, you'd wind up turning your amp up higher than you would without feedback, because the apparent gain is less. And noise comes up as you turn up the volume.

The TubeCad report I posted above totally supports this..... or does it?  The plate follower has 3dB less overall gain but 3dB better PSRR (power supply rejection ratio).  Now what I am unsure of is if you turn up the volume 3dB to compensate, does the power supply noise increase 3dB or are we dealing with a ratio?  Makes me wish I had better test equipment.

What I do know is that a plate follower increases input headroom.  That's so beneficial in the studio because >50% of musicians don't get the concept of turning down in the studio.  What you hear and what the mic hears are 2 entirely different thangs.  The lower output impedance thang is nice too, but a cathode follower will really lower the output impedance.

[HotBluePlates]

The TubeCad report I posted above totally supports this..... or does it?  The plate follower has 3dB less overall gain but 3dB better PSRR (power supply rejection ratio).  Now what I am unsure of is if you turn up the volume 3dB to compensate, does the power supply noise increase 3dB or are we dealing with a ratio?

We are actually talking about 2 different things.

PSRR is all about how much of the ripple present at the power supply node will get into the output signal of the stage. The thing to consider is how much is too much? If 5mV of ripple is passed along with a signal of 45v, it is so small as to be "zero". But 30mV of ripple added to a signal of 100mV is noticeable. So PSRR is really telling you "how sloppy" your power supply node can be for that stage. Poor PSRR means a cleaner supply is needed.

As a related side-note, many service-grade VTVMs have a power supply consisting of a half-wave rectifier and a single filter cap. Their circuit doesn't amplify much (if at all) and the one stage that might be amplifying is a 2-triode balanced stage which rejects power supply noise common to both triodes. However, lab-grade VTVMs that measure ac usually have a good deal of amplification to measure very small voltages that the service-grade ones can't, and they use very well-regulated power supplies (for more reasons than just noise).

But the noise I was talking about is the hiss coming from thermal noise of resistors, johnson noise in tubes, etc. Meaning "high gain amp hiss". Feedback will not help that.

Also, 3dB of feedback is not very much. Depending on how you measure (meaning what your reference is for the dB ratio), an unbypassed cathode resistor will give 3 or 6 dB of feedback. But I was really talking about feedback over 2 or more stages, rather than local feedback at a single stage.

Oh... one more point. What the book was talking about with regard to noise was not just the delay of arrival of feedback, but that a signal that is cyclic in nature (audio tone, hum, sine wave component of a complex signal) can be reduced by feedback, but that hiss is not cyclic... it is truly random occurances of a broad range of frequencies.