First step: Define requirements

These days, a preamp is almost redundant. Signal sources in my living room include CD, DVD, satellite/cable, a
dedicated phono stage, and MP3, not an untypical mix. All have roughly 2V output at a low source impedance.
My phono will get its own dedicated step-up/equalizer designed to output a similar voltage. So, we need at least
5 inputs, all high level. My power amps are of normal sensitivity (as probably are yours), so we really only need
unity gain. A10K input impedance is fine- there will be no weedy sources allowed in MY system!
There’s another reason for the choice of 10K as the input impedance- I’m sold on the use of input transformers.
Wait, wait, I see you starting to click to some other site, but hear me out: given the multiplicity and
uncontrollability of grounds in a multi-source system like mine, there is a decided advantage to galvanic
isolation. There’s also a decided advantage to high common mode rejection in a noisy (electromagnetically
speaking) environment. And at these signal levels and impedances, you can get good bandwidth and very low
distortion. As a bonus, the transformer will do a nice job of limiting bandwidth; though some people have a
religious belief that more bandwidth is better, there’s precious little (ok, NO) evidence that extending bandwidth
beyond 20-25kHz is audible. That’s no surprise- we don’t have sources that provide signal energy more than an
octave above that, at best (and rarely).
There is plenty of evidence that the consequences of transferring high frequency junk through the power amp
ARE often audible, even with the limitations of program material. The transformer specified will take us to 90kHz
or a little higher while not passing AM radio or our local CB operators. The canonical taxi drivers will not be
considered (when was the last time you saw a taxi driver use anything but a cell phone?). That bandwidth should
give us margin to accommodate any currently available signal source and any contemplated sources for the next
decade or two. See the white papers at http://www.jensentransformers.com for detailed and lucid discussion of
transformers and grounding.
As far as output drive abilities, we can calculate what we might need. The maximum length of interconnect I am
likely to use is 3 meters (say, 10 feet). That will also accommodate 99.9% of the rest of the world. With an
interconnect capacitance of 150pF/m (the worst I could find), we need to drive 450pF. Call it 500pF. The power
amp will have an input impedance of 10K at minimum, and might add another 500pF of input capacitance, if it’s
a particularly nasty design. So we have our worst case load: 10Kohm paralleled with 1000pF. The unit should
also be perfectly stable into this or any other likely load- this is a restricted club and we do not take kindly to
stray oscillations sneaking in the back door.
We want the unit to be quiet. Hum and hiss should be inaudible. Let the analog tape or mike preamps annoy
you, not your preamp. Noise below -80 dB from full output is acceptable in real rooms, so let’s go a factor of ten
better and insist on -100dB.
OK, we know what we want to accomplish. How do we do it?

Second step: Basic design outline

Starting at the input, my choice of transformer is the Jensen JT11-P1. It’s a 1:1 input transformer, optimized for a
10K load, with great balance and common-mode rejection, low distortion, and not terribly pricy in the US. In the
UK, Sowter makes a similar-looking unit, the 3575; the specs aren’t quite as nice-looking, but it appears pretty
satisfactory. The CMLI-15/15C from Cinemaq is supposed to be equivalent to the Jensen, but at a lower price
and with a slightly higher optimal load. If you really want to go on the cheap, I’ve heard mostly good reports
about the Edcor WSM 10K/10K, and they’re under $10 a pop. There are fine-looking units from Sowter and
Lundahl that would also work, but the terrible condition of the dollar may encourage American builders to think
domesticFinally, if you insist on gain (perhaps your phono stage’s gain is marginal), there are several options in
transformers that are (or can be strapped for) 1:2 stepup that retain most of the performance of the 1:1 Jensen
and Cinemags.
It seems trivially obvious to say this, but… the transformer is critical of its load and will only give its best if that
load is optimized. If we don’t do things right at the start, it will all be wrong at the end. By “wrong,” I mean ringing,
overshoot, or HF limitations. Not to mention distortion. For the Jensen, the optimal load is a 10k resistance with
as little shunt capacitance as possible. If you use a different transformer, you have two choices: figure out the
optimum load using a square wave generator and an oscilloscope (best) or following the recommendation in the
manufacturer’s data sheet (usually fine).
We don’t want to load the Jensen transformer with capacitance, we need a low output impedance, we need unity
gain- are you thinking what I’m thinking? Sure you are- we want a cathode follower. Ultra-high input impedance,
ultra-low input capacitance, high power supply rejection, easy to stabilize, ultra-low distortion, and low parts
count. Yes, I see you wriggling in your chair a bit- you read somewhere that cathode followers sound awful and
have all kinds of performance problems. Or a “knowledgeable” buddy of yours told you that. Whatever. It’s just a
crock, and it’s a crock that was fired in the kiln of incompetent design and filled with the ejecta of tragic
ignorance. Ask your buddy if he’d turn down an immaculate pair of Marantz 9s.
We will not do an incompetent design.
Let’s see what a proper design will entail. Worst case, we want to drive a 500pF load while retaining the
bandwidth that the input transformer allowed us. The source impedance and the cable’s shunt capacitance form
a first-order low-pass filter with a 3dB down frequency of f3 = 1/(6.28*Zout*C). Rearranging terms to solve for Zout,
we see that for a 90kHz bandwidth, we need a source Z of about 1.8Kohm. That’s pretty doable. But we also
have to consider how much current will be needed to drive the load to the maximum voltage. If we assume a 2
volt input sensitivity (that’s RMS; peak will be about 2.8V) for our power amp, the signal current needed to drive
the 1000pF load capacitance to the full voltage while maintaining the 90kHz bandwidth is about i = 1.6mA. Using
a rule of thumb, we arbitrarily dictate that the standing cathode follower current ought to be at least several times
higher, just to have margin and to minimize distortion. So let’s say we need 10-15mA running through the
cathode follower at minimum.
We’re now in a position to select a tube. There are a lot of candidates which are happy at the desired current,
two very common ones being the 6SN7 and the 6DJ8/ECC88. These also have the considerable virtue of being
cheap and easy to find. Recalling that the output impedance of a cathode follower is about 1/gm, we can check
the suitability. The 6SN7 has a transconductance of 2.5-3 mA/V, which translates to an output impedance of
300-400 ohm. Add a cathode stopper (more about that later) and we’re up to roughly 1K, well within what we
need. Similarly, a 6DJ8/ECC88 at 10mA will have a transconductance about three times higher, resulting in a
base source impedance of roughly 100-150 ohm. So from this standpoint, either will work. The 6SN7 is a lower
distortion tube, but with the inherent feedback of a cathode follower, both tubes are likely to have exceptionally
low distortion at these signal levels.
The ECC88 is a winner in this application because of its plate voltage requirements: less than half what is
needed for the 6SN7. So low, in fact, that the B+ supply can be made with a standard isolation transformer, and
still have enough headroom for active regulation. As a bonus, the higher transconductance translates into a
lower noise floor.
A constant current source as a cathode load will help make biasing easier. It will also ease some of the voltage
requirements as we will see in the detailed design description. I have been leery of their use in the past, based
mostly on unsatisfactory experiences with FET current sources that were fashionable back in the days of Jimmy
Carter. Several writers convinced me to use discrete bipolar CCS and I’ve been delighted with the results.
They’re cheap, perform well, and are quite reliable, in all ways a vast improvement on those awful two-legged
devils of my youth. A pentode also can make a fine current source, but it needs lots of voltage, more chassis
room, more heater supply current, higher cost, higher parts count, lower reliability, separate heater supplies…
but it’s a perfectly valid choice. I just didn’t choose it. Either way, it will be made adjustable: adjustability gives
the tweaker loads and loads of fun opportunity and flexibility to play with operating points and alternative tube
choices.
Conceptual Diagram of The Heretical Preamplifier

Next -> Heretical, p2

Third step: Design details

Looking at the plate curves for the ECC88, we see that with 90V on the plate, a cathode-to-grid voltage of 1.7V
will give us the desired 10mA minimum current. The significance of this choice will be clear later. Nonetheless,
this is a common spot to run these tubes and for good reason: it’s a low distortion point which doesn’t drive the
plate dissipation too high. The tradeoff in choosing a lowish plate voltage is the limitation on swing, but that’s not
an issue for our requirements here.
The volume control is arbitrarily set to 100Kohm because that’s an easy value to source and I happen to have
some Alps Black Beauties in that value. I will not sneer if you use 50K or 250K instead. A stepped attenuator
would be lovely here. Whatever you choose, you will need to parallel it with whatever resistance is necessary to
load the transformer secondary with 10K. If you’ve got a high quality 10k attenuator in hand (15k if you’re using
the Cinemag), you’re set!
Note that the grid is at DC ground.
The bipolar current source may be done several different ways: I’ve used single bipolars, ring-of-two, and
cascodes. Single bipolars have mediocre performance. Ring-of-two is a good choice, but I’ve had one oscillate
on me. So in a fit of pique, they have been sent down to the minors. And I sent them down because the cascode
is a perfectly good alternative and has behaved well for me. This particular cascode will be powered by a +/-12
volt supply, a complication which benefits us slightly here and a great deal in the next couple of implementations.
The -12V supply allows the tube to easily swing the required 6-ish volts peak-to-peak while allowing for sufficient
voltage to keep the current source in compliance.
The easy bits are the choice of grid stopper and cathode stopper resistors. They tend to be non-critical
(sometimes even omitted entirely, as they are in the simplified schematic above) and usually picked by resorting
to tradition and experience. I will not omit them because I want this preamp to be rock-stable. The grid stopper
can typically be between 1K and 10K. Since I have a box full of tiny 1K resistors and the input capacitance is
low, that’s the value I use. The output stopper is slightly more problematic since it adds to the preamp’s source
impedance, but I’ve never had a problem with 200 ohms in that spot. Half that would no doubt be fine if you’ve
got some real drive issues to worry about. The main thing about stoppers is getting them as physically close as
possible to the tube pins.
The output coupling cap is chosen to give a low frequency cutoff at least a decade below the lowest frequency of
interest with the lowest anticipated load. Using the 10Kohm figure previously bruited about, for a 2 Hz cutoff the
output capacitor needs to be 8uF. I have a coffee can full of 4.7uF polycarbonate caps, so a pair of those in
parallel will more than suffice.
Let’s look at a few of the remaining values. The cascode current source uses a red LED for a 1.7V reference;
these parts have relatively low noise, low impedance, and provide a nice visual indication that all is well in their
portion of the circuit. The resistors feeding it are chosen to provide about 5mA of current. The transistors can be
any general purpose high beta bipolars. Double checking, the current source will be set to 10mA, so with betas
on the order of 100, we will only lose 0.2mA to the transistor base. This will not appreciably disturb the LED
string. With a reference voltage of 1.7V, a quick Ohm’s Law check indicates that the emitter resistor will need to
be about 100 ohms. A trim-pot of 250-500 ohms will allow plenty of range. If you’re the fussy type, use a 100
ohm trim-pot in series with a 47 ohm fixed resistor. If you’re not and you just want a box you can put in your
system and use for the next 20 years, skip the trimmer completely and just use a fixed resistor of 100 ohms and
an ECC88 for the tube.
And if you are building this version as a final unit and will not be pursuing the later enhancements, you can get
rid of the +12V supply and run the resistor feeding the constant current source to the cathode follower’s anode
supply. In this case, to get 5mA flowing, the resistor’s value must be increased to 20K:
The 1M resistor from the volume control wiper to ground is intended to keep things stable in the event the
control’s wiper goes bad, a common occurrence. The 15K resistor shunting the control serves to load the
transformer with its optimum 10K load.
The circuitry involving the MOSFET and relay is a dethumper- when the preamp is powered up, the output can
have some dramatic bangs as voltages come up and the tube heats. This portion shunts all of that to ground
until things have had a chance to settle themselves. The parts choices were mostly dictated by stock-on-hand;
the 33uF cap is a 35V unit (a 47u will give a bit longer delay), the transistor can be any small MOSFET that will
handle the relay current. In my first go, the relay was a 12V unit, so I just used the 12V rail and ground as the
endpoints- that’s what’s indicated in this schematic. In another version, I used an IRF510, just because it was
under $2 at Radio Shack. In this version, the circuit was connected between the +12 and -12 rails and a 24V
relay was used. Either way is equivalent, your choice will depend on what relay you have in your own personal
coffee can.
Alternatively, one could use a 555 timer or even a CMOS inverter to fire the relay. This circuitry is optional if you
can guarantee that the preamp will not be powered up into a live power amp. This entire circuit is quite crude, so
if you want to use something a little more elegant and robust to things like power line bounces, I will not be
offended if you jettison my handiwork.
Power supply requirements are fairly nominal- the cathode follower runs a constant current, a lowish one at that,
and sports fine power supply rejection. DC ought to be used on the heater, and the +/-12V source can take care
of that duty. The series resistors are calculated on the assumption of a 300mA heater; if the tubes you use are
360mA, they will need to be changed to a pair of 100 ohm units.
I’m a crude, simple guy so I used a crude, simple cabinet- a $6.50 black plastic box with an aluminum lid,
sourced from Radio Shack. You may wish to do something nicer.
A note on inputs and grounding: the input transformer allows you to drive the preamp via balanced or
unbalanced sources. For balanced sources, the two ends of the primary are used as the inverting and
noninverting input. For unbalanced sources, the terminal marked “ground” is connected to sources’ ground- the
source ground is isolated from the rest of the preamp grounds. That’s the beauty of input transformers. If switch
contacts don’t bother you, you can easily use a DPDT switch on the primary side to do polarity inversion. That
will give you even more fun things to play with while listening to music.
To review, in Part 2, we considered a very basic transformer-coupled input, capacitor-coupled output unity-gain
preamp. That circuit is, as it stands, an excellent performer. We do need to power it and to build it, so those
circuits and options will be considered before we move on to some fancier variants. But this is a fine stopping-off
place if you just want a good sounding, reliable, easy-to-build preamp.
In this part, I’ll cover the raw supply, the regulators, some construction notes, and a few words on parts choices.
It’s a bit unfair to have you make decisions in the dark, so I’ll preview the next variation: I hate large capacitors,
so I trade the large output cap for a small input cap, direct-coupled output, and servo to clean up the offset. And
the last iteration will substitute a rather novel servo for the usual method, decoupling the servo even further from
the signal. This will come, as usual, with a degree of trade-off… If the idea of wrapping op-amp circuits around
your preamp gives you the willies, this first version ought to be your final destination.
Discipline
We need to power our circuit. To do so, we would ideally have a supply that gave us exactly the DC voltages we
need with the effortless silence of a skilled servant. We have made things somewhat easy for Jeeves by
perspicacious design choices: a cathode follower has excellent immunity to power line disturbances, the high
impedance current source isolates the cathode from most of the noise or wobbliness remaining on the -12V rail
after regulation, and the heaters can trivially be made blameless by wiring and grounding correctly. But the
supply needs to be good enough to not be the limiting factor in distortion and to achieve low enough noise such
that the preamp’s total signal to noise reaches or betters my -100dB criterion.
Cum on feel the noiz!
Let’s try to quantify our noise budget and see how that impacts power supply design. There is a school of
thought that power supplies can never be too good. In a sense, that’s correct, but in my world, once the power
supply is no longer the limiting factor, it’s Good Enough. I’ll try to lay out alternatives and options so that if my
approach is horrifying to you (“Ohmigawd, he’s using Radio Shack parts!”), you can at least understand the
rationale and make whatever changes strike your fancy.
What’s the overall target? For a 2V preamp output, -100dB equals 20 microvolts. So the RMS sum of all noise
sources should be at that level or lower. We can quibble about hearing thresholds, masking, Fletchers, and
Munsens, but at the end of the day, it’s unlikely that anything at the -100dB level will be audible in any
reasonably normal setup. If the total contributions of the power supply and signal circuits sum up below that
20uV target, the preamp will be blameless in that sense.
As a quick aside, let’s look at noise in the signal circuitry. The principle sources will be the tube’s noise, noise
from the cathode load current source, and thermal noise from the input circuit. We will start at the beginning. At
full bore, source noise will depend on (naturally) the source. But if we assume a source impedance of 1K from
whatever CD player we have hooked up, we can rough out a Thevenin equivalent. The 1K source is in series
with the (roughly) 1K resistance of the transformer. This is shunted by the 10K worth of resistance that the
volume control circuit represents. We’re at 1.67K. We have the grid stopper in for another K. 2.67K. The
Johnson noise is then seen to be well under a microvolt. With zip for DC across any of these resistances, excess
noise is buried. Whew!
The current source might appear to be a worry, but it’s not, really. The looking-in impedance at the cathode is
about 150 ohms. Ignoring the favorable shunting action of whatever load the cathode follower is driving, the
noise from the current source is divided down by its source impedance (on the order of 10 megohm) and the
looking-in impedance. So to contribute a microvolt of noise at the output, the current source will need to produce
70,000uV (70mV) of noise. The prevalent source of noise is from the modulation of the LED reference voltage
via noise on the 12-volt rails. The LED has an impedance of roughly 10 ohms at the current we’re running it. It’s
fed by 2.4K (ignoring that pesky bipolar’s base), so the divider ratio is about 240. That means we can tolerate
nearly 17mV of ripple or noise on those 12V rails before exceeding a microvolt transferred to our output. This will
be relatively easy- in these days of modern times, we’re blessed with cheap, reliable regulators that sport noise
decades lower than that. If we want to adopt a suspenders and belt approach, we could, in principle, replace the
upper resistor with another constant current source, but the Escherian possibilities can quickly get out of hand.
“It’s turtles all the way down.” A resistor will work just fine there.
There’s the tube, too. We’re running it in a region where the transconductance is moderately high, about 8mA/V.
Using the old rule of thumb equivalent noise resistance = 2.5/gm, we can estimate the tube’s Johnson noise
resistance as about 300 ohms. Considering the input resistance, we can see the noise contribution as negligible.
So, we’ve got an intrinsically quiet circuit- if we don’t muddy it up with bad power supply design, we’ll hit our
design goals with room to spare. OK, let’s figure out a way to power up this bad boy
Raw, raw, raw, that’s the spirits we have here!
The raw supply will be in a separate box so that its emanations don’t piddle all over our nice, clean circuit.
Regulation will be closer at hand to where it’s needed.

We begin with consideration of B+ needs. Isolation transformers are a dime a dozen, we don’t need a lot of
voltage, so that is what we’ll use. There are many mansions in my father’s house; my favorite mansion is the old
fashioned E-I core. One can buy these new for about $16 (the Triad World Series will give us ten times more
current than we need, #VPS230110). Or, if you’re like me, you root around in the scrap box to find something.
When you do, you’ll have the heart of a 160V (or so) supply, which gives the filtering and regulation plenty of
room to maneuver while fighting the noise.
There are many religions regarding how the line voltage should be filtered. Much depends on your actual power
source- mine seems to be relatively clean. Do please use a fuse, and make sure that the safety (earth) ground
from the power line is securely tied to the chassis. You can kill three birds with one stone by using one of the
Schurter power entry modules, which combine an IEC power connector, a fuse, and an RFI filter. A suitable unit
is the 5200.0123.1, stocked by Mouser at about $21. If you don’t mind doing some extra metalwork, you can use
the 5110.0133.1, which has the IEC receptacle and RFI filter, and then add a separate fuse-holder. An SPST on-
off switch is optional. Terminal worrywarts will use a DPST and switch both sides of the line.
Following the transformer, we need rectification. Much bird is whipped debating the merits and demerits of
various schemes. My take: use high speed diodes- they’re reasonably inexpensive and while I think their merits
compared to old-fashioned 1N4007 types are minimal, their better performance certainly won’t hurt. The Vishay
UF5408/1 is ridiculously over-rated at 3A/1000V, but costs about $0.32. Buy a pile and use them for your next
twenty tube projects. Bypass the diodes, if that makes you happy. I didn’t bother- I’m paying for the following RC
filter and by god I’m going to make it earn its keep. But a 0.1uF caps paralleled across each diode is quite
customary.
What about tube rectifiers? I can grant their use for nostalgia reasons, but for me the trade-offs are too severe.
High resistance, inefficiency, heat production, and unreliability are the demerits. The one strong point is the
gradual warm-up, but it’s not much of an advantage here- the plate voltage is too low to induce cathode stripping
and we’ve got the output solidly dethumped. Use a tube if you like, but there’s just no question that if the goal is
solid and silent DC, silicon rules.
On to the filtration. We start off with a relatively honking 470u/200V cap, a common switching supply value. This
isn’t an application where we need the ultimate in ESR or DA or whatever. We do, however, want that cap to be
of good quality for the long haul, so my choice is a 105° C rated cap. If you’re like me, you’ll use a UCC
KMH200VN471M25X30T2. That’s a mouthful of a part number, but Mouser sells them for about $2.50. If your
wrist gets tired writing out that number, an equally good choice is the CDE 380LQ version, part # 471M200H022;
it has a lower temperature rating (which is probably OK since there’s no real heat sources in the raw supply
box), but has a very high ripple current rating. It’s up to you. Whichever we choose, before going any further,
we’ve knocked our ripple down to less than a volt. A couple of RC filters and a regulator should have no problem
scaling that back by another 80dB or so. We put in a 470K bleeder resistor for safety.
At this point, the raw B+ (about 160V) exits the power supply box in sufficient quantity to power two channels;
mission accomplished. We’ll come back to the B+ line when we consider regulation and filtering options later.
Now, we have a heater, dethumper, and current source to power up. I used +/-12V supplies because I knew
there would be a servo in my future. And even if there weren’t, the extra supply doesn’t represent any significant
further investment in money or chassis space. Our current requirements are relatively modest- budget 365mA
maximum for the tube heater (it will probably be 300mA, but worst case…), and another 5mA per channel for the
current sources. Our servo, when it takes a bow, will only need another 5mA or so. So if we plan for a 500mA
supply, we will have a comfortable margin.
I have separated out the filament and plate transformers. You can, in theory, use a transformer with both
windings on the secondary, but there’s a penalty to pay- unless there’s an electrostatic shield, the interwinding
capacitance will do a smashing job of coupling common mode noise onto the heater circuits. You can get a
transformer with an interwinding shield and add common mode filtration; a suitable filter is shown in Morgan
Jones’s “Valve Amplifiers,” figure 5.48, or you can improvise with an RF 1:1 transformer and a couple of 0.01uF
caps. Hams with coffee cans full of 2.5mH chokes can use a pair of those to good effect. I didn’t use a common-
mode filter in either of my units and saw no ill effect. But if you’re going to use a single transformer for high and
low voltage or experience any noise coupling via the heater, spend the extra few dollars and put in a common
mode filter. The transformer should be, at minimum, 30V CT (15-0-15) to allow the regulators a bit of room to
breathe. 500mA current rating is more than enough. At the expense of a bit more regulator dissipation, a 36VCT
unit was pressed into service; I used a scrapbox transformer, but suitable new units would include the Triad
World Series VPP36560 (PC mount) or VPS36700 (chassis mount).
As with the high voltage supply, high speed diodes are a Good Trick, and a cheap one, too. If you’re a bug about
power supply efficiency, Schottky rectifiers have a somewhat lower voltage drop than normal junction diodes.
But, really, this is only half an amp we’re talking about. Since I was buying the UF5408/1s for the high voltage
rail anyway, I just used the same thing for the heater supply. With a 3A continuous and150A surge rating, they’ll
do fine.
This is a good time to consider reasonably big caps for the supply. A pair of low ESR 3300uF electrolytics will
knock the ripple down to about 600mV. And they’ll be cheap. An exemplar is the Nichicon UPW1V332MMH,
featuring 15 milliohms of ESR for under $3.00.
And that pretty much wraps up the raw supply.
Plate Job
With solid raw supply in hand, we move on to the next two inter-related problems- attaining the proper DC level
and getting rid of the ripple and noise. This will require some regulators, or at the very least, more filtration.
Let’s start with the anode supply. The cathode follower will have a power supply rejection of roughly mu +1, so
we can tolerate at most 600-700 microvolts of noise before running afoul. Ideally, we’d like it even lower to leave
room for other noise sources. There are two basic approaches: active regulation and passive filtering. I’ll present
both options and let you decide.
The passive supply is quite a good option for circuits (like this one) that draw a constant current. This approach
has the value of simplicity but, as we shall see, the disadvantage of size and cost. Basically, it consists of a
series of RC low-pass filters. One can opt for chokes instead of the resistors for a further cost and size penalty,
along with radiation issues, for a circuit which does not work better in this application. We will use resistors.
The preamp with passive RC filtering looks like this:

In this implementation, the raw supply’s ripple is knocked down to about 150uV, well within the noise budget.
Separate filters should be used for each channel to prevent crosstalk. Separate raw supplies are luxurious and
totally unnecessary. The last cap is bypassed with a film capacitor to ensure good decoupling at high
frequencies. The filters should be placed as close to the load (the tube) as possible.
For those with a sense of adventure, active supplies are wonderful. They don’t cost much, don’t take up much
room, and perform spectacularly well when properly designed. There are a lot of good options to choose from;
the Maida regulator, a floating LM317 is a particularly popular choice. Fancier regulators using floating op-amp
circuits can also perform superbly, albeit with greater complication. Here, I used instead a very simple two-
transistor regulator, chosen for ease of construction and with an eye toward some adventures down the road.
This circuit is set up as a differential amplifier of sorts, the lower transistor acting as an error amplifier comparing
a reference voltage from the zener to a divided down output voltage appearing at its base. Its collector sets the
base voltage of the pass transistor and hence the output voltage of the regulator.
The divider string features a capacitor bypassing the top resistor in order to increase AC gain of the error
amplifier. This cap does not need to be very large to be effective. The Thevenin equivalent of the divider string
(with a little Kentucky windage to account for base current) is about 6.5K. A capacitor of 2u will then get within
3dB of ideal bypassing. I didn’t have any appropriately rated capacitors of a size that would fit my little perfboard,
so I stepped down to 1u. With that value, ripple voltage on the regulator output was under 100uV, so that wasn’t
too bad a let-down from ideal.
The zener is fed by the current flowing through the error amplifier’s emitter- it may be augmented by another
5mA via a 15K resistor tied to the rail if your zener is balky when asked to wake up or it it needs to be worked
harder to quiet it down a bit. Build the circuit without it, and if you experience any start-up glitches, put it in. A
real stud will eschew a zener and, instead, stack a pair of LM329 references. Either way, the 220u bypass cap is
your friend. The noise from the regulator I built was quite comparable to the RC passive filtered version.
The MPSU10 I used for the error amplifier is no longer available. But you can substitute any bipolar transistor
with a 1W or better power rating, 200V or higher breakdown voltage, and as high an ft as you can manage.
Another MJE340 can be used in a pinch.
The low voltage regulators are quite generic. Please note that there is a good reason not to separate the heater
circuit from the other 12V parts of the circuit- the LM317/337 perform better regarding source impedance and
noise as the current draw increases (see Errol Dietz’s paper in Electronic Design, December 14, 1989, and
reprinted in “Troubleshooting Analog Circuits” by Bob Pease). Since the current sources and the (soon to be
added) servo op-amps draw a piddly few 10s of milliamps, the 300-360mA heater draw really helps to make that
regulator perform. The input cap is necessary when we put the raw supply in its own box; the 220uF bypass
caps on the adjust pin and the output get the noise down to the sub-millivolt level, well under any reasonable
need in this application. The most appropriate versions of the 317/337 are the “T” TO-220 packs. They’ll
dissipate 3 or 4 watts, so a small heatsink is appropriate. I used 1N4007 diodes for shutdown protection, but
nearly any good medium power rectifier will work. If you want to be extra-geeky, use the same high-speed
diodes that are used in the rectifier portion of the circuit.
So, with a bit of thought and some pretty conventional circuitry, we have made a power supply that is far better
than our requirements. Let’s go have a drink to celebrate.
Men With Toolbelts
Given my total lack of skill in the mechanical aspects of fabrication, I feel a bit like a rabbi recommending brands
of bacon. But at least I can explain my worst-case methods. As mentioned previously, I used two Radio Shack
plastic boxes to house the preamp and raw power supply, with the aluminum covers providing a crude ground
plane. This will probably be upgraded at some point- I think I’ll be living with this unit for quite a while. But for the
moment, it works.
The circuitry was built on six small perf boards, each containing a functional block for one channel (current
source, B+ regulator/servo, and signal circuitry). This was done so that I could switch various circuit chunks in
and out during development. If you want to use this as an experimental base, that’s fine, but a circuit board or
proper point-to-point would be a major step up. I’ve got some of those nice Tektronix ceramic/silver terminal
strips that will be pressed into service when I do a rebuild. A good circuit deserves more than I’ve given it.
Input and output jacks should, logically enough, be isolated from the chassis- and from one another. Use a high-
quality connector here where it counts.
The umbilical from the power supply is terminated with a 5-pin AMP female plug. The corresponding male AMP
connector is mounted on the preamp chassis. There’s no technical reason why audio-type connectors couldn’t
be used, but I highly recommend you don’t do that- it’s not a matter of if, it’s a matter of when someone will plug
the wrong thing in and let out the smoke.
I won’t specify wiring. I don’t think it’s terribly important, but if you do, use whatever flavor you like. To a great
extent, I used Morgan Jones’s suggestion (in “Building Valve Amplifiers”) of solid silver wire with Teflon sleeving.
I must admit that, with silver-bearing solder, it makes a lovely joint.
Grounding can be done in several ways- some prefer to use a single bus, connected to the chassis at one point.
Some prefer star points. Others like to combine the two. Whichever you choose, the isolated nature of the inputs
means that the usual ground point near the input jack is unsuitable. The best place to set the ground for this
design is at the ground point for the input transformer secondary.
I haven’t diagrammed input switching since that will vary a bit from installation to installation. Use a good quality
switch; I used a surplus Cinema Engineering 5 position two deck rotary switch with shorting between positions
for optimal isolation. The input grounds are all tied together and kept completely separate from the preamp
ground.
As I mentioned before, I used an Alps Black Beauty volume control. That’s a very, very good second choice. My
first choice would be a stepped attenuator. If you’ve got the time and resources to build one, I’d highly
recommend it.
I haven’t shown any pilot light circuitry, if you want one, I’d suggest putting a green LED in series with a 4.7K
resistor, then putting that string between the +12 and -12 rails. You can also put in an indicator showing when
the dethumper fires; I’d probably use a different color LED, put it in series with a resistor as before, and connect
that string across the relay coil. When the dethumper unmutes the output, the light will go out.
When building this circuit, it’s best to start it up without the tube in place, loading the current sources with a 1K
resistor tacked between the collector and the +12 rail. Adjust the trim-pot to get 10V across the resistor. Then
disconnect the resistor, warm in the knowledge that your current source is properly adjusted and may safely be
used with your tube.
OK, enough for now. You have all that you need to build a good-sounding unity-gain tube preamp that will be
quiet and unobtrusive. If you want to hang on for the ride, things will get a bit more novel in the next part.
In earlier parts, we considered a unity gain preamp with an ECC88 at its heart. The design was mostly
conventional, but withal, a good performer, a unit that will not be the limitation in any reasonable system. Now
it’s time to get unreasonable. And, unreasonable guy that I am, my target is that big, fat output coupling
capacitor.
Flow and Eddy- A Philosophical Diversion
It’s time to talk about capacitors. Next to wire, caps are the bits which are most likely to induce audiophile
neurosis. Living in dread fear of dielectric absorption, ESR, dissipation factor, and insufficient mysticism, the
audiophile is a perfect target for those who would encourage the neurosis in the hope of selling a few plastic
jellyrolls at platinum prices. On the rare occasions when the jellyroll peddlers (or merchants of similar energy-
storage desserts) provide actual data to show the superiority of their particular polymer pastry, one thing seems
to be universal- the data are shown for high signal currents, usually with an annotation that this is done to show
the results more clearly. Well, THAT is interesting and points to a bit of electronic jiujitsu: if increasing signal
current makes errors in cap vendors’ tests greater, then it makes sense that reducing the signal current should
reduce such errors. No cap vendor seems to make their case at microamp levels…
Many lifetimes ago, I served a rigorous apprenticeship with the great designer Murray Zeligman. Murray is a very
intuitive guy and one of his rules of thumb (and the man must have had about a million thumbs) was to use the
smallest coupling cap possible in any circuit. In his view, the large capacitors in fashion at that time tended to
“thicken” the sound in an unnatural way. As the Jesuits say, “Give me the boy and I’ll give you the man,” so
Murray’s prejudice has never left me. And here we have a clue about why he might have been right.
Consider a basic RC coupling circuit. Very typical values in a tube circuit would be 0.1uF for the cap and 1M for
the resistor. This results in a high-pass filter with a -3dB frequency of 1.6Hz, a decade below the (usual) lowest
frequencies of interest. With that much margin, phase shift within the passband remains quite low. Following the
dictum of “smaller is better,” we reduce the size of the capacitor tenfold to 0.01uF. We see that our -3dB
frequency has likewise risen to 16Hz and the phase shift in the passband is now quite significant. Audible?
Maybe, maybe not, but why have that phase shift there in the first place if it can be avoided? The curious reader
might want to know why this bothers me, but the corresponding phase shift from the transformer doesn’t.
Fortunately, the rolloff from the specified transformer is nearly two orders of magnitude lower, an astonishing
0.25 Hz. That coupling capacitor is indeed the choke point (bad pun). So the resistor must be correspondingly
increased to 10M, a rather large value for most tube circuits, so large that grid leakage currents will often cause
the biasing to go a bit haywire. In some cases, the resistance can be bootstrapped, and that can be useful, but
it’s not a universally available option.
Since we entered with a notion about signal currents, let us consider them for these RC combinations. At a
2VRMS signal level, the current through the 0.1uF/1M combination is 2 microamps. Likewise, the current
through the 0.01uF/10M high-pass is 0.2 microamps, a tenfold improvement. But in an absolute sense, both of
these values are quite low.
Looking at the uglier part of the circuit, we have specified a load of 10kohm and have sized the output cap
correspondingly: the -3dB frequency for a 10uF output capacitor and a 10k load is the same as the 0.01/10M just
considered. But, look what has happened to the signal current- Ohm’s Law tells us that it has gone up by
3 decades, from 0.2 microamps to 0.2 milliamps. That is clearly not the direction we would want to go, and a
reasonable assumption is that any defects caused by the capacitive coupling have likewise increased a
thousandfold.
Making things worse is bulk. A 10uF capacitor is BIG. There’s a whole steaming PILE of dielectric that must
have its polarization sloshed around several thousand times a second. Try this: take two capacitors with the
same voltage rating but separated by a decade or more of capacitance. Unroll them and measure the relative
area. Which one do you think will have less effect on the signal?
We’re unfortunately stuck with a 10uF cap if we want to drive a 10k load- or are we?
Pinning the Tail on Don Quixote
Name something better than a capacitor for getting a signal from point A to point B. Take your time, I’ll wait for
you to- what’s that, you’ve got an answer? A piece of wire? That’s right! Even in the most fevered fantasy of the
most deluded audiophile, the defects of a piece of wire are in a whole different league (a better one) than those
of even the finest capacitor. The problem is DC. If the grid of the cathode follower is held at ground, the cathode
must be offset by a couple of volts of DC above ground, not a healthy thing to thrust at an unsuspecting power
amp. Likewise, if we hold the cathode at ground, the grid must be a couple of volts negative. There is no third
choice.
Choice One (the grid at DC ground) is where we start from. Blocking DC from the output by using a cap is what
put us in this quandary. So we must level shift the output. There are several ways to do this, but all have the
disadvantage of running the signal through a rather non-ideal set of components (like a bypassed resistor or a
battery or an LED).
I think we can set aside Choice One for the moment. Or longer.
Choice Two (the grid biased negatively with the cathode at DC ground) can be implemented a few different
ways. One is the bootstrap, as shown here:
In this circuit, the input grid leak resistor is returned to a point along the cathode load chosen so that the cathode
side of the tap corresponds to the resistance necessary to get the desired DC grid-cathode bias voltage with the
constraint of the cathode being at 0VDC. As mentioned above, this topology has the salutary effect of increasing
the effective load that the grid-leak resistor presents to the coupling cap. But it will take some trimming to get any
offset nulled; worse, from tube to tube and from degree Celsius to degree Celsius, the DC voltage on the
cathode WILL wander about, no matter how carefully the circuit is trimmed.
We can also put a negative voltage on the grid via its resistor. Symbolically, it would look like this:

In this sketch, the bias comes from a battery, but there are a large number of ways to derive that voltage; the
large grid resistor makes the impedance of the negative voltage source rather nugatory for any reasonable
values, so low noise, rather than dynamic impedance, is the primary design desiradatum. There are two
problems to overcome: first, the trim/drift issues we just discussed. And now, a second, the lack of bootstrapping
of the grid leak resistor, which loses us the advantage of using a smaller input capacitor. So, this would
seemingly be a nonstarter, having doubled the number of design problems.
A hint that there might be a single solution available to solve both problems may be gleaned from the
observation that the element to be trimmed in this latter implementation is a voltage rather than a resistance tap.
In other words, if we can control the applied grid voltage, we can keep the cathode voltage at zero. And that
inevitably leads us to the word “servo.”
The first implementation of DC servo-controlled cathode followers I ever saw was from David Berning, in some
PAS-3 mods that eventually evolved into the Berning TF-10. He used an inverting IC integrator to sense the
output of his buffer and feed back a signal that had a very high DC gain, but very low AC gain at audio
frequencies. That signal was applied to the grid so that if the output tried to wander positive, the integrator would
bite back fiercely with a negative response to the buffer’s input. And thirty years later, I still think that’s the best
way to do it. IC op-amps are terrific at DC tasks and can clamp down that offset in a heroic fashion.

We Can Rebuild It, We Have the Technology

The application of a servo to our Heretical line stage is shown here (click for larger image):
We have jettisoned that awful output cap and replaced it with a bit of wire. We have moved the RC coupling to
the input side of the tube, where a satisfyingly small capacitor will give us the desired low frequency response.
We have wrapped an inverting integrator around the tube to keep the DC voltages nice and tidy and provide the
bias for the grid. The integrator is referenced to ground and it will do its damnedest to provide whatever negative
grid bias is necessary in order to keep the output at DC ground. The response of the integrator falls
monotonically at 6dB/octave, so that by the time we reach the audio band, it has stepped aside quite politely. To
ensure maximum politeness, an additional passive RC network can optionally be interposed between the
integrator and the bottom of the grid leak resistor.
This point cannot be overemphasized: the integrator does NOT contribute an in-band signal! Let us see why that
is so. First, there is the response of the integrator itself. For an integrator turnover frequency of 1Hz, the
response will be down by 60dB at 1kHz.
Second, there is a reduction of the integrator’s output at AC because of the voltage divider formed by the very
large grid leak resistor and the input impedance which shunts the tube’s grid to AC ground. We can estimate
how much this helps as follows: At 1kHz, a 0.01uF capacitor has a reactance of about 16k. The 10k input
resistor is paralleled by the source impedance of whatever is driving the preamp. Let’s do a worst-case and drive
the preamp from an infinite source impedance (open circuit). The lower leg of the divider is then the quadrature
sum of the capacitive reactance and the input resistor, which comes to about 18K. The voltage divider thus
attenuates any 1kHz at the integrator output by 18k/(10M +18k) or roughly 0.018. Translated to decibels, we
have a reduction of -55dB. So between the dropping response of the integrator and the voltage divider effect, the
1kHz worst-case contribution of the integrator is -115dB. With a typical low source impedance of a CD player or
similar device, we can knock the servo signal down another 6dB…
I believe that we can forego any worries about the servo fouling our pristine midrange and treble; even at bass
frequencies, it contributes a ridiculously small signal. The tradeoff for getting rid of the output capacitor is then
only one of complication- there is now a higher parts count, but on the plus side, that output capacitor was
probably not cheap. A handful of Radio Shack resistors, caps, and an op-amp cost a small fraction of a Teflon
capacitor.
To Servo Mankind- A cookbook
There is a bit of a delicate matter to determine- the relationship between the servo time constant and the time
constant of the input network. No doubt there is an analytical way to optimize their ratio, but in this case, I took
the lazy empiricist’s way out- I moved the servo time constant around while listening to music with well-recorded
bass. The effect is quite fascinating, with the sound ranging from bloated to anemic. The values shown in the
schematic produced a very neutral sound. There is a relatively broad “sweet spot” where the effect of change
was somewhere beyond evanescent. So don’t sweat it, especially if you do not have a monster woofer system.
Experimenting with that time constant will give the curious constructor hours and hours of amusement; for me, I
would be happy not to hear “This One’s For Blanton” again for a long, long time. Analytical listening is tedious.
The op-amp is not terribly critical, but it does need to be unity-gain stable. If you use the suggested FET input
device or something similar, you can connect the non-inverting input directly to ground without fear of offset
caused by bias currents. If you opt for a bipolar op-amp (and something like a 1458 works just fine here),
connect the non-inverting input to ground via a 100k resistor. Because this is basically a DC circuit, the parts are
not critical, but it is probably worthwhile to use a good quality (though non-fancy) capacitor for the integrator. I
used an inexpensive polypropylene unit. The input cap to the buffer can advantageously use that fancy cap, but
on the bright side, a 0.01uF Teflon capacitor is much, much cheaper than a comparable 10uF unit, and is on a
par with the price of a 10uF polypropylene of moderate pretension.
My construction technique is, as mentioned before, quite laughable. But it is functional. The circuit elements
were each constructed on separate perf boards, tested, then interconnected. Regardless of the chosen method,
it is a good idea to place the grid leak resistor close to the grid connection. Thus the lead length between the
servo and the buffer circuitry becomes much less important, a consideration for constructions that have the
servo circuit on an outrigger board.
The Proof of the Pudding

This is a plot of the spectrum from a 2VRMS 1kHz sine wave measured using an Audigy 2ZS soundcard and
RMAA software in a noisy environment. The green plot is a straight wire loop from input to output. The white plot
is the same but with the preamp interposed in the loop, and is transposed upward by 20dB for clarity. There are
two striking things about these plots. First, the preamp, by virtue of the input transformer, actually improves
matters. The straight wire loop has picked up all kinds of noise and distortion from the rather noisy test
environment- this is greatly reduced by including the preamp in the loop. Second, where’s the distortion? It’s
below the test residual (about 0.003%). And although the noise can be seen, at no point does it threaten to rise
above the -100dB design target set out in the beginning. The 120Hz ripple is nowhere to be seen, and there is
only the faintest trace of 60Hz hum pickup.
Although the op-amp’s offset voltage (often a bit worse for FET input compared with bipolar input) would seem to
be a limitation, I was unable to measure any output offset from the preamp. Not bad. IM at 10kHz/11kHz was
seen to be near residual, 0.005%. Swept IM was below test residual, 0.02%. THD checked at 20Hz and 20kHz
was, within the uncertainty of measurement, identical to the 1kHz numbers.
The preamp works as advertised. Heresy has its virtues.

Acknowledgements
As usual, I tapped the brains of a lot of smart people. Steve Eddy put the transformer bug up my, errr, nose, and
was kind enough to do the nice drawings and host them. Joel Tunnah’s schematic symbols made it easy for an
old fart like me to draw the dodgier-looking schematics. John Curl lectured me at length about servos. Various
individuals at diyAudio.com helped me fix my errors and improve the presentation; of special note was the input
from Brian Beck. And finally, Morgan Jones was kind enough to look this over and make some pertinent
suggestions; his repeated application of some excellent English ales to THIS author was an inspiration.
This is an expanded and edited version of an article that originally appeared in Bas Horneman’s excellent on-line
publication, diyMag.