You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

449 lines
19 KiB

<!doctype html public "-//w3c//dtd html 4.0 transitional//en">
<html lang="en">
<head>
<meta http-equiv=Content-Type content="text/html; charset=windows-1252">
<title>The Class-A Amplifier Site - A JLH Amp for the Quad ESL57</title>
<link rel="StyleSheet" href="class-a.css" type="text/css" media="screen, print">
<link rel="shortcut icon" type="image/ico" href="favicon.ico">
</head>
<body background="slate.jpg" bgcolor="#E8E8E8" lang=EN-GB link=blue vlink=blue>
<div class=Section1>
<p align=right style='text-align:right'><b>The Class-A Amplifier Site</b></p>
<p align=right style='text-align:right'><span style='font-size:10.0pt'>This page was last updated on 4 February 2002</span></p>
<p><b><span style='color:blue'><a href="index-1.htm" title=index-1.htm>[ Back to Index ]</a></span></b></p>
<p>&nbsp;</p>
<p align=center style='text-align:center'><b><span style='font-size:20.0pt'>A
JLH Class-A for the Quad ESL57</span></b></p>
<p align=center style='text-align:center'><b><span style='font-size:20.0pt'>&nbsp;</span></b></p>
<p align=left style='text-align:left'>Credits: Original design - John Linsley Hood</p>
<p align=left style='margin-left:50.2pt;text-align:left;text-indent:21.8pt'>Circuit modifications - Geoff Moss</p>
<p align=left style='margin-left:50.2pt;text-align:left;text-indent:21.8pt'>Layout, pcbs and construction - Nick Gibbs</p>
<p>&nbsp;</p>
<p>This version of the JLH Class-A amplifier is the result of a series of
emails and design discussions which culminated in the subsequent construction
of two high current amplifiers specifically optimised to drive Quad ESL57
electrostatic speakers.</p>
<p>&nbsp;</p>
<p>Some months ago, I received an email from Nick Gibbs regarding his 1969 JLH
and a possible upgrade to a 1996 version with a higher quiescent current.
Nick's 1969 JLH was about 16 years old and had been in almost daily use. It had
a 27V supply rail and a quiescent current of 1.2A and Nick was using it to
drive his Quad ESL57s, since neither his Quad 405 nor JLH MOSFET amps would do
so without tripping the protection circuits. The little 10W JLH worked well
with the ESL57s, albeit with some occasional clipping on louder passages, and
Nick felt that a higher current version would best meet his needs.</p>
<p>&nbsp;</p>
<p>The ESL57 is a difficult load to drive in that it is capacitive and its
impedance drops to a low of around 2ohm at 15kHz. A high current delivery is
therefore required but, to offset this, the maximum voltage that should be
applied to the ESL57 is only 33Vp-p. It seemed that a JLH Class-A with a
reduced supply rail voltage and a higher quiescent current would be ideal.</p>
<p>&nbsp;</p>
<p>After we exchanged a number of (sometimes lengthy) emails, the final design
evolved. It was a cross between the 1969 and 1996 versions (hopefully with the
better parts of each <span style='font-family:Wingdings'>J</span>) operating
off +/-20V supply rails and with a quiescent current of between 3.5A and 4A.
The circuit is shown in Fig.1, but it should be noted that Nick used MJ802
output transistors in place of the 2N3055s since he already had these devices
available.</p>
<p>&nbsp;</p>
<p align=center style='margin:0cm;margin-bottom:.0001pt;text-align:center'><img
border=0 width=738 height=433 src="jlheslfig1.gif"></p>
<p align=center style='text-align:center'>Fig. 1 - The Final Circuit.</p>
<p align=center style='text-align:center'>&nbsp;</p>
<p>As can be seen, the circuit is a mixture of the two original JLH versions,
with modifications to enable an increase in quiescent current. Parallel pairs
of output transistors have been used to keep the dissipation in each device at
an acceptable level. The 0R1 emitter resistors are included to ensure equal
cu+rrent sharing between each device. The quiescent current control is the
standard 1969 bootstrap method whereby C4 maintains a constant voltage across
RV2 and thus a constant dc current into the bases of the output transistors.</p>
<p>&nbsp;</p>
<p>The input stage of the 1996 version has been utilised, but for dc offset
control the 7815 has been replaced with a constant current source to avoid the
instability problems that have been encountered when the 7815 is operated at a
low current. Several capacitor values have been increased to modify the low
frequency &#8209;3dB point and to reduce low frequency distortion. High quality
components have been used throughout.</p>
<p>&nbsp;</p>
<div class=MsoNormal style='margin:0cm;margin-bottom:.0001pt'>
<hr size=3 width="100%" align=left>
</div>
<p>&nbsp;</p>
<p>Addendum - 4 February 2002 </p>
<p>&nbsp;</p>
<p>Note, care must be taken to ensure that R5 and RV2 are adequately rated. The
current through these components is slightly greater than the sum of the output
transistor base currents. The output transistor base current is the output
transistor quiescent collector current (Ic) divided by the current gain (Hfe)
of the device. The current through R5 and RV2 is therefore approximately equal
to 4 x Ic / Hfe and this should be calculated for the chosen output transistor
quiescent current and output transistor type. It is recommended that output
transistors with a gain of 100 or more at the working collector current are
used in this design to reduce the power rating requirements for R5 and RV2.</p>
<p>&nbsp;</p>
<p>Whilst it should not be difficult to obtain fixed resistors with the
required power rating, the preset potentiometer could be more of a problem
since the more common ones are only rated at 0.5W or 1W, though higher rated
devices are available. It must be remembered that the power rating of a preset,
when connected as a rheostat, is proportional to the length of track in use.
The required power rating must therefore be calculated from the current flowing
through the preset and the full preset resistance value. </p>
<p>&nbsp;</p>
<p>With high gain (&gt;100) output transistors and a quiescent current of 3A, a
1W device should be adequate for R5 and 2W for RV2, provided RV2 is no greater
than 500ohm. If a larger value of RV2 is found to be necessary, it will be best
to use a 2W fixed resistor in series with RV2 to avoid the need for a higher
power rated preset. (Note, the original value shown in the Fig. 1 for RV2 was
2kohm. This value has been changed due to the power rating considerations).</p>
<p>&nbsp;</p>
<p>If RV2 needs to be set to below about 300ohm due a particular combination of
quiescent current and transistor gain, I suggest that R5 be reduced to between
50 and 100ohm to avoid the need for increasing the size of the bootstrap
capacitor C4.</p>
<p>&nbsp;</p>
<div class=MsoNormal style='margin:0cm;margin-bottom:.0001pt'>
<hr size=3 width="100%" align=left>
</div>
<p>&nbsp;</p>
<p>The power supply (one for each channel) is shown in Fig. 2. This is
basically the standard LM338K circuit, included elsewhere on this site, with
some capacitor variations/additions.</p>
<p>&nbsp;</p>
<p align=center style='margin:0cm;margin-bottom:.0001pt;text-align:center'><img
border=0 width=766 height=542 src="jlheslfig2.gif"></p>
<p align=center style='text-align:center'>Fig. 2 - The Power Supply</p>
<p align=center style='text-align:center'>&nbsp;</p>
<p>Nick initially adjusted the quiescent current to 3.5A but found after
listening tests that increasing this to 4A gave a noticeable improvement. Even
at an Iq of 4A the amps run cool due to the substantial heatsinking (0.5degC/W
for each output device and each LM338K). Variations in quiescent current and dc
offset with temperature are minimal, with an Iq of 3.8A and a dc offset of less
than 35mV at switch-on. </p>
<p>&nbsp;</p>
<p>As for the sound quality, Nick's initial comments are summarised below:</p>
<p>&nbsp;</p>
<p align=left style='margin-top:0cm;margin-right:1.0cm;margin-bottom:0cm;
margin-left:42.55pt;margin-bottom:.0001pt;text-align:left'><span
style='font-size:11.0pt'>I have just spent two hours listening ... I cannot
believe the improvement over my old JLH. I have ended up with 20V rails and an
Iq of 4A. You may well understand that I am feeling a little emotional at the
moment so I will attempt to quantify the sound in a point form:-</span></p>
<p align=left style='margin-top:0cm;margin-right:1.0cm;margin-bottom:0cm;
margin-left:42.55pt;margin-bottom:.0001pt;text-align:left'><span
style='font-size:11.0pt'>&nbsp;&nbsp;</span></p>
<p align=left style='margin-top:0cm;margin-right:1.0cm;margin-bottom:0cm;
margin-left:42.55pt;margin-bottom:.0001pt;text-align:left'><span
style='font-size:11.0pt'>Female vocal - incredible, makes the hairs on the back
of my neck stand up.</span></p>
<p align=left style='margin-top:0cm;margin-right:1.0cm;margin-bottom:0cm;
margin-left:42.55pt;margin-bottom:.0001pt;text-align:left'><span
style='font-size:11.0pt'>&nbsp;</span></p>
<p align=left style='margin-top:0cm;margin-right:1.0cm;margin-bottom:0cm;
margin-left:42.55pt;margin-bottom:.0001pt;text-align:left'><span
style='font-size:11.0pt'>Instruments and singers now appear as solid 3D
objects, they have constant depth, if that makes sense?</span></p>
<p align=left style='margin-top:0cm;margin-right:1.0cm;margin-bottom:0cm;
margin-left:42.55pt;margin-bottom:.0001pt;text-align:left'><span
style='font-size:11.0pt'>&nbsp;</span></p>
<p align=left style='margin-top:0cm;margin-right:1.0cm;margin-bottom:0cm;
margin-left:42.55pt;margin-bottom:.0001pt;text-align:left'><span
style='font-size:11.0pt'>No blurring of image or loss of depth on loud moments.</span></p>
<p align=left style='margin-top:0cm;margin-right:1.0cm;margin-bottom:0cm;
margin-left:42.55pt;margin-bottom:.0001pt;text-align:left'><span
style='font-size:11.0pt'>&nbsp;</span></p>
<p align=left style='margin-top:0cm;margin-right:1.0cm;margin-bottom:0cm;
margin-left:42.55pt;margin-bottom:.0001pt;text-align:left'><span
style='font-size:11.0pt'>Acoustic guitar - real!</span></p>
<p align=left style='margin-top:0cm;margin-right:1.0cm;margin-bottom:0cm;
margin-left:42.55pt;margin-bottom:.0001pt;text-align:left'><span
style='font-size:11.0pt'>&nbsp;</span></p>
<p align=left style='margin-top:0cm;margin-right:1.0cm;margin-bottom:0cm;
margin-left:42.55pt;margin-bottom:.0001pt;text-align:left'><span
style='font-size:11.0pt'>Bass - although the ESLs are 3dB down at 55Hz
everything is so well defined, I would say at this point that my old JLH was
brilliant here too.</span></p>
<p align=left style='margin-top:0cm;margin-right:1.0cm;margin-bottom:0cm;
margin-left:42.55pt;margin-bottom:.0001pt;text-align:left'><span
style='font-size:11.0pt'>&nbsp;</span></p>
<p align=left style='margin-top:0cm;margin-right:1.0cm;margin-bottom:0cm;
margin-left:42.55pt;margin-bottom:.0001pt;text-align:left'><span
style='font-size:11.0pt'>I can hear more hiss from the source material,
although I cannot as yet fault the top end reproduction, the ESLs are very
revealing.</span></p>
<p align=left style='margin-top:0cm;margin-right:1.0cm;margin-bottom:0cm;
margin-left:42.55pt;margin-bottom:.0001pt;text-align:left'><span
style='font-size:11.0pt'>&nbsp;</span></p>
<p align=left style='margin-top:0cm;margin-right:1.0cm;margin-bottom:0cm;
margin-left:42.55pt;margin-bottom:.0001pt;text-align:left'><span
style='font-size:11.0pt'>I think more than anything else it's the fact that the
amps&nbsp;NEVER appear to get confused (?) (increasing the Iq from 3.5A to 4A
prevented slight confusion/clipping (recovery)&nbsp;with heavy mid to upper
band periods, at the levels I listen at). Constant image solidity,
depth&nbsp;and remarkable detail at all times are what these amps are about,
bloody brilliant!</span></p>
<p align=left style='margin-top:0cm;margin-right:1.0cm;margin-bottom:0cm;
margin-left:42.55pt;margin-bottom:.0001pt;text-align:left'><span
style='font-size:11.0pt'>&nbsp;</span></p>
<p align=left style='margin-top:0cm;margin-right:1.0cm;margin-bottom:0cm;
margin-left:42.55pt;margin-bottom:.0001pt;text-align:left'><span
style='font-size:11.0pt'>I have broken into my Cambletown Whisky as celebration
</span></p>
<p style='margin-top:0cm;margin-right:42.55pt;margin-bottom:0cm;margin-left:
42.55pt;margin-bottom:.0001pt'><span style='font-size:11.0pt'>&nbsp;</span></p>
<p>And a few days later:</p>
<p>&nbsp;</p>
<p align=left style='margin-top:0cm;margin-right:1.0cm;margin-bottom:0cm;
margin-left:42.55pt;margin-bottom:.0001pt;text-align:left'><span
style='font-size:11.0pt'>The amplifiers just get better the more you listen,
real instruments and human voice are truly superb and very involving, plus of
course the sensational 3D solid imaging. I wish you could hear them. </span></p>
<p align=left style='margin-top:0cm;margin-right:1.0cm;margin-bottom:0cm;
margin-left:42.55pt;margin-bottom:.0001pt;text-align:left'><span
style='font-size:11.0pt'>&nbsp;</span></p>
<p align=left style='margin-top:0cm;margin-right:1.0cm;margin-bottom:0cm;
margin-left:42.55pt;margin-bottom:.0001pt;text-align:left'><span
style='font-size:11.0pt'>I am slowly working my way through my CD collection
with the new amps, and it just gets better!</span></p>
<p align=left style='margin-top:0cm;margin-right:1.0cm;margin-bottom:0cm;
margin-left:42.55pt;margin-bottom:.0001pt;text-align:left'><span
style='font-size:11.0pt'>&nbsp;</span></p>
<p>And Nick's most recent comments:</p>
<p style='margin-top:0cm;margin-right:1.0cm;margin-bottom:0cm;margin-left:42.55pt;
margin-bottom:.0001pt'>&nbsp;</p>
<p align=left style='margin-top:0cm;margin-right:1.0cm;margin-bottom:0cm;
margin-left:42.55pt;margin-bottom:.0001pt;text-align:left'><span
style='font-size:11.0pt'>The combination of Marantz CD17 MkII + new amps +
ESL57 is the first system that I have EVER heard that can do justice to the
sound of a piano. I have been listening to a Deutsche Grammophon recording of
Liszt's Hungarian Rhapsody, a little hissy, but for the first time, the attack
(?, I don't know how to describe this), the first instances of a piano note and
all that goes with it to convince you that you are listening to a piano, is
there. I have friends who play the piano and so I often listen to the real
thing. I consider this ability of the new amps very important. I thought the
ESLs would give me this with pretty much any amp, but it has taken the new JLH
amps to actually do it. Additionally, the insight into Bizet's Carmen, again on
Deutsche Grammophon, is exceptional. </span></p>
<p align=left style='margin-top:0cm;margin-right:1.0cm;margin-bottom:0cm;
margin-left:42.55pt;margin-bottom:.0001pt;text-align:left'><span
style='font-size:11.0pt'>&nbsp;</span></p>
<p align=left style='margin-top:0cm;margin-right:1.0cm;margin-bottom:0cm;
margin-left:42.55pt;margin-bottom:.0001pt;text-align:left'><span
style='font-size:11.0pt'>Increasing the size of the electrolytics from 220uF as
in my original JLH to 470uF has very noticeably extended the bass response.</span></p>
<p align=left style='margin-top:0cm;margin-right:1.0cm;margin-bottom:0cm;
margin-left:42.55pt;margin-bottom:.0001pt;text-align:left'><span
style='font-size:11.0pt'>&nbsp;</span></p>
<p align=left style='margin-top:0cm;margin-right:1.0cm;margin-bottom:0cm;
margin-left:42.55pt;margin-bottom:.0001pt;text-align:left'><span
style='font-size:11.0pt'>The original JLH is a magnificent amplifier, but with
the modifications it has become outstanding.&quot;</span></p>
<p style='margin-top:0cm;margin-right:1.0cm;margin-bottom:0cm;margin-left:42.55pt;
margin-bottom:.0001pt'>&nbsp;</p>
<p><b>Circuit Boards</b></p>
<p>&nbsp;</p>
<p>Nick has kindly supplied me with a copy of his pcb layout for both the
amplifier board and the regulator board in case they are of interest to other
constructors. These are reproduced below at full size. It should be noted that
the amplifier board is laid out for Caddock MP930 series power resistors, on
heatsinks, for R5, R9, R10, R11 and R12 and also that Nick has used two
resistors in series (100ohm and 50ohm) for R5 as these were more readily
available (and cheaper). Component overlay diagrams have also been included
after the pcb diagrams.</p>
<p>&nbsp;</p>
<p align=left style='text-align:left'>The actual board sizes are: Amplifier board 8.55 x 5.25 (217mm x 133mm)</p>
<p align=left style='text-align:left'>Regulator board 3.3 x 2.9 (84mm x 74mm)</p>
<p align=left style='text-align:left'>&nbsp;</p>
<p>&nbsp;</p>
<p align=center style='text-align:center'><img border=0 width=501 height=814
src="jlheslampboard.gif"></p>
<p align=center style='text-align:center'>Fig. 3 Amplifier pcb (viewed from copper side)</p>
<p align=center style='text-align:center'>&nbsp;</p>
<p align=center style='text-align:center'>&nbsp;</p>
<p align=center style='text-align:center'><img border=0 width=314 height=276
src="jlheslregboard.gif"></p>
<p align=center style='text-align:center'>Fig. 4 Regulator pcb (viewed from copper side)</p>
<p align=center style='text-align:center'>&nbsp;</p>
<p align=center style='text-align:center'>&nbsp;</p>
<p align=center style='text-align:center'><img border=0 width=501 height=814
src="jlheslampoverlay.gif"></p>
<p align=center style='text-align:center'>Fig. 5 Amplifier pcb overlay (viewed from component side)</p>
<p align=center style='text-align:center'>&nbsp;</p>
<p align=center style='text-align:center'>&nbsp;</p>
<p align=center style='text-align:center'><img border=0 width=314 height=276
src="jlheslregoverlay.gif"></p>
<p align=center style='text-align:center'>Fig. 6 Regulator pcb overlay (viewed from component side)</p>
<p align=center style='text-align:center'>&nbsp;</p>
<p align=center style='text-align:center'>&nbsp;</p>
<p><b>Finally</b>, for those of you interested in seeing the results of Nick's labours:</p>
<p>&nbsp;</p>
<p align=center style='text-align:center'><img border=0 width=642 height=422
src="jlheslphoto1.jpg"></p>
<p align=center style='text-align:center'>Photo. 1 The pcbs</p>
<p>&nbsp;</p>
<p>&nbsp;</p>
<p align=center style='margin:0cm;margin-bottom:.0001pt;text-align:center'><img
border=0 width=642 height=610 src="jlheslphoto2.jpg"></p>
<p align=center style='margin:0cm;margin-bottom:.0001pt;text-align:center'>Photo. 2 Nearing completion, a plan view</p>
<p align=center style='margin:0cm;margin-bottom:.0001pt;text-align:center'>&nbsp;</p>
<p align=center style='margin:0cm;margin-bottom:.0001pt;text-align:center'>&nbsp;</p>
<p align=center style='margin:0cm;margin-bottom:.0001pt;text-align:center'><img
border=0 width=642 height=478 src="jlheslphoto3.jpg"></p>
<p align=center style='text-align:center'>Photo. 3 The finished amplifier with top cover removed</p>
<p>&nbsp;</p>
<p align=left style='text-align:left'>&nbsp;</p>
<p><b><span style='color:blue'><a href="index-1.htm" title=index-1.htm>[ Back to Index ]</a></span></b></p>
<p>&nbsp;</p>
<p>&nbsp;</p>
<p><b><span style='font-size:8.0pt'>HISTORY:</span></b><span style='font-size:
8.0pt'>Page created 04/11/2001 </span></p>
<p><span style='font-size:8.0pt'>18/11/2001 Credits, pcb details and photos
added</span></p>
<p><span style='font-size:8.0pt'>19/11/2001 pcb details corrected</span></p>
<p><span style='font-size:8.0pt'>21/11/2001 pcb overlays added</span></p>
<p><span style='font-size:8.0pt'>31/01/2002 R5/RV2 power rating notes added</span></p>
<p><span style='font-size:8.0pt'>04/02/2002 R5/RV2 notes revised and separated
as an Addendum. RV2 value changed to 500R in Fig. 1 (was 2k)</span></p>
<p><span style='font-size:8.0pt'>&nbsp;</span></p>
</div>
</body>
</html>