Sequel to Low Cost SET Amp for Newbies

[John L Stewart]

At 130 views & counting it seems like we need a follow on. So here it is
with more power. This was originally published in the Glass Audio
Projects Book, April 2002, along with 16 other vacuum tube designs. The
space on AudioBanter is tight, so some of the information will appear in
a following post.

MORE POWER FOR THE AFFORDABLE SE AMPLIFIER

While working on the earlier Affordable SE Amp (GA 4/00) it occurred to
me there might be an easy way to get even more power with only a slight
increase in cost.

All single ended amplifiers need to run in Class A mode if they are to
deliver good quality sound. In order to do that they need to be capable
of dissipating considerable heat. Tubes in the list for the original low
cost SE Amp barely fit that requirement. However, several compactrons
with a beam tetrode & high mu triode in a single envelope are available
on the NOS market at reasonable cost. See Table One. They have power
handling ability at about 50% better than those on the previous list. To
further this project, I selected the 6LU8.

CIRCUIT FEATURES

1) Operation in Triode or Ultralinear mode.
2) Operation with or without feedback.
3) Power frequency hum cancellation.
4) Start/Standby & Run Modes.
5) Excellent performance.
6) Low cost has been maintained.

TRIODE OR ULTRALINEAR?

The original amplifier used a low cost Hammond 125E for it’s output
transformer. That transformer has a center tap which could be connected
to the screen of the beam tetrode, thus providing a degree of
ultralinear operation. I arranged layout of the various parts so that a
comparison of performance could be made between triode & ultralinear
topologies. Referring to Figure one see that the screen grid of the beam
tetrode is shown unhooked. You can make a connection either direct to
the plate (triode operation) or to the 125E center tap (ultralinear).

CAUTION- Don’t use a switch to make the circuit change. Energy stored in
the magnetic field of the 125E tries to keep current flowing. You could
cause a failure of the insulation in the transformer because of the
large induced voltage caused by switching. When you make the change your
amplifier should be completely disconnected from the power source.

At midband the resulting amplifier delivered 5 watts in triode mode & 6
watts when connected for ultralinear operation. Refer to Table Two for a
summary of performance of the various configurations.

THE MODIFIED AMPLIFIER

As shown in the schematic the amplifier is connected so that all of the
gain is used by feedback. If you would like to run without feedback then
unhook R4 from terminal one of the 125E & connect to COMMON. You will
also need to remove C3 & R5 from the circuit & connect C2 to the
junction of R6 & R7.

The coupling cap between the voltage amp & the power amp becomes more
critical as more feedback is used. You are normally told to use as large
a coupling cap as practical at this position. However as feedback is
increased
a hump will appear in the low frequency response. I measured up to 4 db
rise at 30 to 40 Hz. C3 & R5 provide a step in the low frequency
response. This allows more feedback to be used without instability.

An important addition to this amp are three grid stopper resistors. This
tube has very high transconductance & will likely oscillate on it’s own
if given a chance. I knew I had problems when my FM receiver was
interfered with
while the amplifier was running. I observed spectrum up to 150 Mhz. Grid
stopper resistors lower the Q (Quality Factor) of the circuit according
to the formula

Q = (1/R)  SQR ( L/C )

As little as 100 ohms in a conductor interferes with the RF to the
extent that it is eliminated. The audio is left intact.

The remaining LF time constant in this amplifier is the filter section
of the power supply. Many experimenters either ignore this one or simply
follow the lead of others. However, it is in the audio path & does have
considerable effect on the final result.

For triode operation connect an 8 ohm loudspeaker to terminals 2 & 6 of
the 125E. If you would like to use the ultralinear mode, then you can
connect an 8 ohm loudspeaker to terminals 3& 6.

THE POWER SUPPLY

You will need more power for this version of the SE Amp. Here it is
supplied by a Hammond 271X transformer. Filtering is provided by a PI
section consisting of a Hammond 156L, 5 Henry choke & a pair of 100
microF caps.
This time the caps are rated at 450 volts working. Refer to Figure 2.

The filter section is resonant at 10Hz. That has an effect on the low
frequency performance of the amp. A simple regulated power supply would
eliminate this problem. Perhaps next time.

A current sampling resistor of 10 ohms is in the center tap lead of the
HV winding. Using a X10 probe I measured peak current of 250 mA on the
scope. Average current was 70 mA.

The pilot light has been moved to the 5 volt winding to somewhat relieve
load on the 6.3 volt winding. However, total load is well within the
capabilities of the 271X. I measured input to the power transformer to
be
46 VA* in the run mode. In standby it was 20 VA. The 271X is rated for
63 VA input.

A Start/Standby mode is provided by the 3PDT, center off switch S1
sections A, B & C. Diode ring D3 thru D6 sets the heaters to about 4.9
volts when S1C is open. C105 & C106 are Mallory part number UN103M.
These caps are Underwriters Laboratory approved for across the line
applications.

POWER FREQUENCY RIPPLE CANCELING

In a previous article (GA 3/98) I had described a simple addition to the
power supply circuit which would almost eliminate the power frequency
hum component. That would be 60 Hz in North America & 50 Hz in Europe &
elsewhere. This component of hum results often when a centertapped
transformer is used in the power supply.

Two halves of the HV winding may not be balanced for resistance. For
example, in the Hammond 271X the two sides of the HV winding measured
167 & 142 ohms. The power frequency ripple happens to be in the same
range as the resonance(s) of most loudspeaker systems. This sometimes
causes problems in systems using SE triode amps. Here is a practical
application of that circuit.

You will find the fix in this amplifier easy to make. The 5 volt winding
is connected in series with the center tap of the HV winding. The extra
5 volts will be series aiding for one side of the HV winding & series
opposing for the other side. There are no polarity markings on most
power transformer leads. On the first try there is an even chance the
hum will get worse.

Your best check for the proper connection is by using an oscilloscope.
Connect the scope probe to the ungrounded end of the 10 ohm resistor (TP
A) in the center tap return lead. Proper connection of the 5 volt
winding results in the current pulses all having about equal amplitude.
If you have connected the 5 volt leads the wrong way, alternate current
pulses will have somewhat different heights. Refer to Figure 4.

When connected properly I measured a 12 dB reduction of the power
frequency ripple as compared to the normal hookup. It cost us nothing.

If you don’t have an oscilloscope to make this adjustment, two other
kinds of measurement are possible. A peak responding AC voltmeter such
as the HP 410A, 410B or 410C could be used to measure the voltage peaks
produced across the current sampling resistor. Correct connection of the
5 volt winding leads is that which produces the smallest voltage. This
will vary depending on your circuit but will be in the range of 2.5
volts.

Finally, you can build your own simple peak responding voltmeter. See
Figure 4. Most any signal or power diode will work. From my junk box I
tried a 1N4007 as well as a very old 1N56A. The 1N56A is a Germanium
device so has a smaller forward drop. It worked best. The external VOM
of 20000 ohms per volt was set to the 3 volt range. Connect the 5 volt
winding leads to give the lowest reading when connected to the test
point on the 10 ohm sampling resistor. Indication will be about 2.5
volts.

LINE STAGE

When you use the feedback version of the amplifier you will find the
gain to be somewhat low. In order to get enough gain most experimenters
will try three stages of gain. Instability often results when feedback
is connected. Even the famous Williamson amplifier exhibits some low
frequency problems since there is too much gain & phase shift inside the
feedback circuit.

A simple line stage can easily fix this problem. Refer to Figure 3. This
is an application were a pair of feedback amps are connected to provide
enough gain without instability. As shown the line stage has a gain of
8.

CONCLUSION

Was the project worth the time & effort required? I would say yes. For a
very reasonable cost the amateur can get some experience of what vacuum
tube audio is really about.

In my various designs I try for synergy. That is to say the various
parts should somehow compliment each other so that the final result will
easy to implement, safe & reasonable cost. All designs will have some
tradeoff’s.
The designers job is to separate those which are important from the
rest.

The primary objective I had in mind with the Affordable SE Amp was to
illustrate what is possible on a tight budget. I wanted to avoid exotic
& otherwise expensive components which seem to be running rampant in the
present vacuum tube audio revival.

For those who would care to spend a little more money the most effective
improvement they can make to this design is a better output transformer.
The Hammond 125E was not designed with high fidelity in mind. I would
recommend the Hammond 1628SE. You will get a large improvement in the
low frequency response. Distortion at all levels would be reduced since
the 1628SE is specially designed for single ended applications. Refer to
Table 3 for some comparative distortion measurements made at 100 Hz.

*VA- Volt-Amperes


+-------------------------------------------------------------------+
|Filename: Power Amp 5H 75R.jpg |
|Download: http://www.audiobanter.com/attachment.php?attachmentid=379|
|Filename: Power Supply 5H 75R.jpg |
|Download: http://www.audiobanter.com/attachment.php?attachmentid=380|
|Filename: Table One Summary of Power Tetrodes.doc |
|Download: http://www.audiobanter.com/attachment.php?attachmentid=381|
+-------------------------------------------------------------------+



--
John L Stewart

[Stephie Bench]

Hi John,

Very nice. A couple of comments. On the screen stopper, sometimes you can get a "free" improvement in performance by tailoring the value of the stopper resistor. I think, but I am not familiar with that tube, that a slightly larger resistor may improve the distortion, particularly at moderate power levels. The second comment is the cathode bypass cap. Depending on its overall quality, bypasing that cap with another non electrolytic, say 1 or 10uF polyprop will both improve HF performance, and often lower distortion, since it removes the nonlinear effects of the cap.

Thanks for publishing those.

One further note in general note on low cost SE amps. By using a tetrode or pentode connected in triode (or UL) mode, you often get the benefit of a subtly better (more precisely, lower order) plate resistance linearity. This often doesn't particularly show up in static distortion measurements, but under the condition of a reactive load, it produces lower distortion. Nice benefit from using a "cheap" tube. You can ve4ify this by making your static load reactive somewhat, and measuring the distortion in that condition.

Stephie

[John L Stewart]

And still more.

Cheers to all, John L Stewart


+-------------------------------------------------------------------+
|Filename: Ripple Checking Circuit 75R.jpg |
|Download: http://www.audiobanter.com/attachment.php?attachmentid=385|
|Filename: GA28PARTS Z.txt |
|Download: http://www.audiobanter.com/attachment.php?attachmentid=386|

[Patrick Turner]

John Stewart proposes a low cost SE amp using 6LU8.
http://www.audiobanter.com/attachment.php?attachmentid=379

It is many years since that tube has been made. Maybe not everyone could obtain one. Pentode Pda rating = 14Watts, so say 12W max in practice for SE class A1. This means max Po = 5 Watts with low OPT losses, at 42% anode efficiency.

I suggest it would be easier to use say 6BQ5 driven by 12AX7 with its 2 triodes in parallel. Both are made now, and old stock is available.
The above schematic has the V1 triode cathode with 1k8 connected to whole OPT sec winding and for GNFB network. I humbly suggest there's a resistor missing between cathode and 0V; the whole GNFB network needs to be changed.
Of course if one does want a nice cheap 5 Watts, it is even better to use a single EL34 in triode, idling with Pda = 16W, and same paralleled 12AX7 as input / driver, and with say 12dB GFB. Then Ea might be higher, and RLa higher, but then any cheap amp with low Po should be extremely well thought about to compensate for its basic shortcomings of having not much Po. UL with EL34 is good - and for reasons Stephie mentioned; although they'd go right over most ppl's heads. UL Po could be 7Watts with Pda at same 16W at idle, same Ea and Ia. But RLa changes, so thought must be used if any benefit is to be heard.
If EL34 is considered, then minor bias change will allow 6L6GC, and the world is full of 6L6, EL34, 12AX7. I'm not sure how much goodness resides in the probably small cheap Hammond OPT.
Patrick Turner.

[maxhifi]

Decided to check out RAT after years absence.. not much left, but at least the arguments appear to be gone. I'll add my two cents to this topic, since it looks interesting (and maybe to enjoy contributing to usenet while it still exists!).

I think the amplifier looks like a great idea for a project... Sure someone may have a little trouble getting the 6LU8, but on the other hand old TV tubes are a great way to get NOS tubes on the super cheap. Supply is fairly minimal, but demand is even smaller still, so unless it becomes a national pastime to assemble cheap as dirt tube amplifiers, someone will be able to secure a pair of RCA or GE or Sylvania 6LU8 for well under the cost of a single quality 12AX7 (Perhaps living in North America is an advantage here). eBay is flooded with the 6LU8, and other similar tubes, and it avoids being forced to go
with either inferior modern production tubes, or high dollar desirable types of NOS which have likely been picked over. I understand it may be somewhat controversial to call modern tubes inferior to NOS, but a quality NOS tube is quite a safe bet from a reliability standpoint, considering they were made in the days tube tech was to be relied upon for communications between humans, rather than a pastime enjoyed mainly by eccentrics, people overcome by nostalgia, "connoisseurs" and obsessive musicians.

Also, it's kind of elegant to have a two tube stereo amp.

A salvaged output transformer from an old TV chassis would make this even cheaper, and it would be possible to figure out how much NFB it can take before instability sets in. Those Hammonds are the right ticket for this sort of cheap and cheerful effort, but I think they're priced about double what their real value is.

This amplifier would be right at home driving a paper cone speaker, for example, as a small scale computer hi-fi setup. Naturally the limited power output and small output transformer core would severely limit applications in a serious hi-fi context, but nobody would expect a <$100 amplifier built of junk-bin parts to perform in the league of serious dollar equipment.

I would submit that the problem with a low cost SE 6L6 or EL34 amplifier, is to do it properly you would exceed the cost of a push-pull amplifier and at the same time suffer worse performance. I don't really see any value in it. A push pull transformer won't be gapped, so you avoid this cost plus get free cancellation of second harmonic distortion, and increased hum rejection all in one stroke by moving to push-pull. The special single ended sound (if it indeed exists) that some people rave about won't be captured at this price either, so all you really end up with, when you make a cheap SE 6L6 amplifier is an
amplifier that costs more than a push pull amp and performs worse.

[mick]

On Tue, 15 Jul 2014 18:58:23 -0600, maxhifi wrote:

> Patrick Turner wrote:
>
>> John Stewart proposes a low cost SE amp using 6LU8.
>> http://www.audiobanter.com/attachment.php?attachmentid=379
>>
>> It is many years since that tube has been made. Maybe not everyone
>> could obtain one. Pentode Pda rating = 14Watts, so say 12W max in
>> practice for SE class A1. This means max Po = 5 Watts with low OPT
>> losses, at 42% anode efficiency.
>>

<snip>


The 6LU8 is all but unobtainable here in the UK. It often costs far more
for the postage than it does for the valve - and we always get hammered
on postage from the US. (The postage is far cheaper from China). 1/2
ECC88 / 6L6 seems to work quite nicely but has a bit less gain than
ECC83/12AX7 for the first stage. The Hammond 125SE OPT also seems to be
quite reasonable for the money - especially when you look at how much it
costs for "proper" iron over here. They are probably on a par with
similarly specified Edcor.

[maxhifi]

For you guys in the UK how about a PCL86 based version of the amplifier?
Should be able to find some cheap NOS over there, and it even uses a more
conventioal tube socket.

[maxhifi]

Also, Radio Spares has an oputput transformer I would use for this design if I
was a UK citizen... http://uk.rs-online.com/web/p/audio-transformers/2106475/
That with an adapted circuit for a PCL86 would be a good start.