Silicon ChipBuilding The Ultra-LD 100W Stereo Amplifier; Pt.2 - May 2000 SILICON CHIP
  1. Outer Front Cover
  2. Contents
  3. Publisher's Letter: Do-it-yourself amplifiers: a new approach / The Dolby Heaphone story
  4. Feature: What's Inside A Furby? by Julian Edgar
  5. Project: Building The Ultra-LD 100W Stereo Amplifier; Pt.2 by Leo Simpson
  6. Order Form
  7. Feature: Dolby Headphone: Five Channels Of Surround Sound by Leo Simpson
  8. Back Issues
  9. Product Showcase
  10. Project: Build A LED Dice by Doug Jackson
  11. Vintage Radio: Making the obsolete useful again by Rodney Champness
  12. Project: Low-Cost AT Keyboard Translator by Steve Carroll & Bob Nicol
  13. Project: 50A Motor Speed Controller For Models by Ross Tester & Branco Justic
  14. Book Store
  15. Market Centre
  16. Advertising Index
  17. Outer Back Cover

This is only a preview of the May 2000 issue of Silicon Chip.

You can view 32 of the 96 pages in the full issue, including the advertisments.

For full access, purchase the issue for $10.00 or subscribe for access to the latest issues.

Items relevant to "Building The Ultra-LD 100W Stereo Amplifier; Pt.2":
  • Ultra-LD 100W RMS Stereo Amplifier PCB patterns (PDF download) [01112011-5] (Free)
  • Ultra-LD 100W Stereo Amplifier PCB patterns (PDF download) [01105001-2] (Free)
  • Panel artwork for the Ultra-LD 100W RMS Stereo Amplifier (PDF download) (Free)
Articles in this series:
  • Ultra-LD 100W Stereo Amplifier; Pt.1 (March 2000)
  • Ultra-LD 100W Stereo Amplifier; Pt.1 (March 2000)
  • Building The Ultra-LD 100W Stereo Amplifier; Pt.2 (May 2000)
  • Building The Ultra-LD 100W Stereo Amplifier; Pt.2 (May 2000)
  • 100W RMS/Channel Stereo Amplifier; Pt.1 (November 2001)
  • 100W RMS/Channel Stereo Amplifier; Pt.1 (November 2001)
  • 100W RMS/Channel Stereo Amplifier; Pt.2 (December 2001)
  • 100W RMS/Channel Stereo Amplifier; Pt.2 (December 2001)
  • 100W RMS/Channel Stereo Amplifier; Pt.3 (January 2002)
  • 100W RMS/Channel Stereo Amplifier; Pt.3 (January 2002)
  • Remote Volume Control For Stereo Amplifiers (June 2002)
  • Remote Volume Control For Stereo Amplifiers (June 2002)
  • Remote Volume Control For The Ultra-LD Amplifier (July 2002)
  • Remote Volume Control For The Ultra-LD Amplifier (July 2002)
Items relevant to "Build A LED Dice":
  • PIC16F84(A)-04/P programmed for the LED Dice [Dice.HEX] (Programmed Microcontroller, AUD $10.00)
  • PIC16F84 firmware and source code for the LED Dice [Dice.HEX] (Software, Free)
  • LED Dice PCB pattern (PDF download) [08105001] (Free)
  • LED Dice panel artwork (PDF download) (Free)
Items relevant to "Low-Cost AT Keyboard Translator":
  • AT Keyboard Translator PCB pattern (PDF download) (Free)

Purchase a printed copy of this issue for $10.00.

Building the Ul Stereo Amplifi In March 2000, we described the circuit of the Ultra-LD 100W amplifier module. In this article, we describe the construction of a complete stereo amplifier using two amplifier modules. It has input facilities for three program sources and a stereo volume control so no preamplifier is required. By LEO SIMPSON A S PRESENTED, this Ultra-LD 100W per channel stereo amplifier can be the basis of a very fine “no-frills” stereo sound system. You can plug in a CD player, tuner or tape deck and the selected program source is switched straight through to the volume control and then to the power amplifiers. This is equivalent to the “CD-Direct” mode on some stereo amplifiers. The Ultra-LD stereo amplifier takes the purist approach – no preamplifiers, no tone controls, no balance control or anything else to affect the signal quality before it goes to the power amplifiers. The result is the cleanest possible sound quality, rivalling that of the very best commercial amplifiers, regardless of price. In the “Publisher’s Letter” for the March 2000 issue we indicated our intention of building the new stereo 16  Silicon Chip amplifier into a computer case. In fact, we mentioned that we intended using a “clam-shell” desk-top computer case. Well, when we came to do the job we decided that the selected case looked a bit tatty and so we purchased a brand new ATX tower case for the princely sum of $66. This actually came with a power supply which is not needed for this project but which will be pressed into service elsewhere. Appearance aside, the clam-shell desktop case would have been adequate for the job but the tower case has substantially more room and has the bonus of slide-on top and side panels and effectively a sub-chassis inside with channels at the top and sides. These channels make it easier to tuck the mains wiring neatly away and the space between the sub-chassis and one of the side panels means that ltra-LD Ultra-LD fier Part 2 May 2000  17 they will be completely enclosed in the tower case and if you are playing music such as pipe organ at high levels, the dissipation can run as much as 50 watts per channel or more and this cannot be handled for more than a few minutes without forced air cooling. Very conveniently, just as we were going to press with the March 2000 issue, a neat tunnel heatsink extrusion to suit an 80mm fan became available from Jaycar Electronics and we elected to incorporate this into the design, as you can see. With the fan running at a low speed, the heatsink is very effective. Interestingly, 80mm heatsink tunnels now appear to be the standard cooling approach in 100W+ 5-channel home theatre receivers. All of the foregoing explains the assembly approach and also is the reason for the delay in presentation of this article which was originally intended for last month’s issue. Performance of stereo version The finished amplifier in the ATX computer case. If you really want to dress it up you could place a dress panel over the plastic in-fill panels and perhaps use much more upmarket knobs. Maybe you could spray the case champagne gold or . . . you can run wiring between them, for better shielding and a neater layout. Another bonus of buying a completely new case is that you get matching in-fill panels for the disk drive openings and this gives a neater finished appearance. In fact, we mounted the selector switch and volume control on one of the in-fill panels and the headphone socket on another. Using the tower case also gives more options in the way the power transformer is mounted although, 18  Silicon Chip as it turned out, that did not present a problem. As you can see from the photos, most of the power supply components are mounted on the base of the case. Fan cooling As part of our approach in using a computer case, it was always our intention to use a small fan to cool the heatsinks for the two amplifier modules. While normal running may not produce a lot of heat in the modules, We published a number of graphs which showed the performance of the prototype module in the March 2000 issue. However, for the stereo version we built two completely new modules and when they were finally installed in the tower case we ran the whole battery of tests again. It’s nice to confirm the results but in some instances the performance was even better with the new modules. Fig.1 shows the total harmonic distortion (THD) versus power at 1kHz when both channels are driven simultaneously into 8Ω loads. Power tops out (the onset of clipping) at just on 90W in both channels and you can see that one channel (right) was slightly lower in distortion at the higher powers. This measurement was taken with a bandwidth of 10Hz to 22kHz. Fig.2 shows the THD versus frequency for both channels driven into 8Ω loads at a power level of 90W. Here, one channel is slightly better at the midrange frequencies but it is higher in distortion at 5kHz and above. This is a result of the wiring layout. This is always a very careful compromise and here you will need to duplicate the power supply wiring details that we will discuss later in the article. The measurements of Fig.2 were taken with a bandwidth of 10Hz to 80kHz. Interestingly, Fig.2 stands up very well by comparison to our benchmark 15W Class-A amplifier when driven at 15 watts (see Fig.3, page 57, July 1998). We’re not going to claim the Ultra-LD stereo amplifier is better than the 15W Class-A design (that’s just not possible) but it indicates that the 100W amplifier is pretty good in this department. And of course, it has a great deal more power. Finally, Fig.3 shows the separation between channels across the frequency range from 20Hz to 20kHz, with both channels connected and alternately driven from the Audio Precision System One test set. This gives a result of better than -60dB over the whole audible spectrum for both channels. While this is a fair way short of the 90dB (typical) separation of a CD player, it is a good “real world” measurement, not the artificially enhanced result produced by the standard IHF-201 separation test. All other performance parameters of the Ultra-LD stereo amplifier are the same as published in the March 2000 issue. Now let us discuss the assembly of the amplifier module and then we will proceed to the power supply details and the rest of the amplifier assembly. Amplifier board assembly The component overlay diagram of the PC board is shown in Fig.4. Before starting the board assembly, it is wise to check the board carefully for open or shorted tracks or undrilled lead holes. Fix any defects before fitting the components. Start by inserting the PC pins and the resistors. When installing the 3.3V zener diode, make sure that it is inserted with the correct polarity. Also take care when installing the electrolytic capacitors to make sure that they are installed the right way around. Note that the 100pF compensation capacitor from the collector of Q8 to the base of Q7 should have a voltage rating of at least 100V while the 0.15µF capacitor in the output filter should have a rating of 400V. Another point to be noted is that if the amplifier is intended for continuous high power delivery at frequencies above 10kHz, then the 6.8Ω resistor in the output filter should be a wirewound type with a rating of at least 5W, otherwise it may burn out. Choke L1 is wound with 23.5 turns With the fan mounted at one end, this is what the two modules look like before they are mounted in the case. The inset shows how the two heatsink extrusions slide together to form a tunnel heatsink with the fins on the inside. May 2000  19 AUDIO PRECISION SCTHD-W THD+N(%) vs measured LEVEL(W) 10 28 MAR 100 16:42:57 AUDIO PRECISION SCTHD-HZ THD+N(%) vs FREQ(Hz) 5 1 1 0.1 0.1 0.010 0.010 0.001 0.001 .0005 .0005 0.5 1 10 100 200 Fig.1: total harmonic distortion (THD) versus power at 1kHz when both channels are driven simultaneously into 8Ω loads. The onset of clipping is just on 90W in both channels. This measurement was taken with a bandwidth of 10Hz to 22kHz. of 1mm enamelled copper wire on a 13mm plastic former. Alternatively, some kitset suppliers will provide this choke as a finished component. When installing the fuse clips, note that they each have little lugs on one end which stop the fuse from moving. If you install the clips the wrong way, you will not be able to fit the fuses. The 220Ω 5W wirewound resistors 20 100 Fig.2: THD versus frequency for both channels driven into 8Ω loads at 90W. can also be installed at this stage; they are wired to PC stakes adjacent to each fuseholder and are used during the setting of quiescent current. Next, mount the smaller transistors; Fig.4: the component overlay for the PC board. Note that the resistor feeding ZD1 has been changed to 2.7kΩ 5W wirewound. 20  Silicon Chip ) 28 MAR 100 20:19:46 AUDIO PRECISION SCCRSTK XTALK(dBr) & XTALK(dBr) vs FREQ(Hz) 0.0 28 MAR 100 20:45:48 -20.00 -40.00 -60.00 -80.00 -100.0 -120.0 1k 10k 20k 20 100 1k 10k 20k Fig.3: separation between channels across the frequency range from 20Hz to 20kHz. When you’ve finished assembling the first PC board and mounted it to the heatsink (see overleaf), it should look exactly like this! Now repeat the assembly procedure for the other channel . . . May 2000  21 Fig.5: the drilling and tapping details for the tunnel heatsink extrusions (not to scale). All holes above left are tapped for M3 screws while the base (above right) is tapped for M4 screws. Note these are not same size! Fig.6: these diagrams show how the transistors are mounted to their respective heatsinks. 22  Silicon Chip Fig.7: this the component overlay for the regulator PC board. Make sure you don’t inadvertently swap REG1 and REG2. ie, BC546, BC556, BF469 and BF470. The transistor pairs Q1 & Q2 and Q5 and Q6 are mounted so that their flat faces actually touch each other. Since we want each pair to thermally track each other, put a dab of heatsink compound on the flat faces and squeeze them together. Both Q8 & Q9 need to be fitted with U-shaped heatsinks, as shown in Fig.4. The four output transistors, the driver transistors (Q11 & Q12) and the Vbe multiplier Q10 are mounted vertically on one side of the board and are secured to one section of the tunnel heatsink with M3 and M4 machine screws. The heatsink needs to be drilled and tapped to take the screws. Fig.5 (not full-scale) shows the drilling and tapping details for mounting the transistors to the heatsink and the heatsink to the chassis. Alternatively, if you are building this amplifier from a kit, the heatsink may already be drilled and tapped. At this stage, you can temporarily attach the transistors to the heatsink but don’t bother with heatsink compound or washers at this stage. This done, poke all the transistor leads through their corresponding holes in the board and line up the board so that its bottom edge is 15mm above the bottom edge of the heatsink. This ensures that the board will be horizontal when fitted with 15mm tapped spacers at its front corners. Note that you will have to bend out all the transistor leads by about 20°, to poke them through the PC board. You can now solder all the transistor leads to the PC board. Having done that, undo the screws attaching the transistors to the heatsink and then fit mica washers and apply heatsink compound to the transistor mounting surfaces and the heatsink areas covered by the mica washers. The This view and the inset above shows how the two transformers were stacked and their primaries and secondaries terminated to an insulated terminal block (as shown in Fig.8). After this photo was taken we rewired the speaker terminals with much heavier figure-8 cables (2 x 79 strands 0.2mm). This made a significant difference to the power output and damping factor. Note that all power and output wiring to the amplifier modules is tightly twisted to provide maximum AC field cancellation. May 2000  23 PARTS LIST Amplifier Case 1 ATX tower PC case (available from CAM1 Computers; phone 02 9975 2919) 2 225mm tunnel heatsink extrusions (Jaycar Cat. HH-8530) 1 12V 80mm fan (see text) 1 toroidal power transformer, 300VA, 2 x 35V and 2 x 50V secondaries OR 1 300VA toroidal power transformer with 2 x 35V secondaries (Altronics Cat. M-5535) and 1 20VA or 30VA toroidal power transformer with 2 x 12V secondaries (Altronics Cat M-4912 or Jaycar Cat. MT-2112) 1 long bolt, nut, and washers to suit transformers 1 pushbutton DPST 250VAC switch to suit case (Jaycar Cat. SP-0746) 4 insulated female spade connectors (to suit DPST switch) 1 3-pole, 4-position rotary switch (adjust to 3 positions) 1 10kΩ dual-ganged log potentiometer 2 knobs, to suit rotary switch and potentiometer 1 IEC male power socket 1 IEC female power socket 2 insulating boots, to suit IEC power sockets 1 panel-mount 3AG safety fuse-holder (Jaycar Cat. SZ-2025 or equiv.) 1 5A 3AG fuse 2 gold-plated binding post terminal pairs (Jaycar Cat. PT-3008) 1 6-way RCA phono terminal panel (Jaycar Cat. PS-0265) 1 stereo headphone socket 1 12-way insulated terminal block 16 adhesive cable twist-ties 2 solder lugs 1 400V 35A bridge rectifier (BR1) 1 1N4001 1A silicon diode (D1) 4 8000µF 63VW chassis-mount electrolytic capacitors 1 470µF 25VW electrolytic capacitor 2 8.2kΩ 1W resistors 1 1kΩ 0.25W resistor 2 330Ω 1W resistors (to connect headphone socket) 24  Silicon Chip 1 120Ω 5W resistor (to suit fan; see text) Cable & Hardware 1m 250VAC 7.5A figure-8 flex 2m 2 x 79/0.2mm heavy-duty figure-8 speaker cable 2m red 7.5A hook-up wire 2m white 7.5A hook-up wire 2m black 7.5A hook-up wire 1m green 7.5A hook-up wire 1m rainbow cable 1m figure-8 shielded cable 2m red light-duty hook-wire 2m black light-duty hook-up wire 8 15mm tapped spacers 16 M3 x 6mm screws 35 M3 x 10mm screws 2 M3 x 15mm screws 24 M3 nuts 45 M3 flat washers 8 M4 x 10mm screws 1 M4 x 15mm screw 1 M4 nut 9 M4 flat washers 4 No.6 x 15mm self-tappers Amplifier Boards 2 PC boards, code 01103001, 105mm x 176mm 8 M205 PC mounting fuse clips 4 M205 5A fuses 2 coil formers, 24mm OD x 13.7mm ID x 12.8mm long, Philips 4322 021 30362 2 200Ω multi-turn trimpot Bourns 3296W series (VR1) 3 metres 1mm diameter enamelled copper wire 26 PC board pins 4 TO-126 heatsinks, Altronics Cat. H-0504 or equivalent 8 TO-3P insulating washers (for output transistors – see text) 6 TO-126 insulating washers Miscellaneous Heatshrink sleeving, heatsink compound, tinned copper wire, solder, insulation tape Semiconductors 4 MJL1302A PNP power transistors (Q13, Q14) 4 MJL3281A NPN power transistors (Q15, Q16) 2 MJE15030 NPN transistors (Q11) 2 MJE15031 PNP transistors (Q12) 2 MJE340 NPN power transistors (Q10) 2 BF469 NPN transistors (Q8) 2 BF470 PNP transistors (Q9) 6 BC546 NPN transistors (Q5-Q7) 8 BC556 PNP transistors (Q1-Q4) 2 3.3V 0.5W zener diodes (ZD1) Capacitors 4 1000µF 63VW electrolytic 4 100µF 63VW electrolytic 2 100µF 16VW electrolytic 2 2.2µF 25VW electrolytic 2 0.15µF 400VW MKC, Philips 2222 344 51154 or Wima MKC 4 10 0.1µF 63V MKT polyester 2 .0012µF 63V MKT polyester 2 100pF 100V ceramic Resistors (0.25W, 1%) 4 18kΩ 2 330Ω 2 12kΩ 1W 4 150Ω 2 3.3kΩ 6 120Ω 2 2.7kΩ 5W 8 100Ω 2 1.2kΩ 4 47Ω 2 1kΩ 2 6.8Ω 1W 2 390Ω 16 1.5Ω 1W 4 220Ω 5W (for current setting) Regulator Board 1 PC board, code 01103002, 61 x 92mm 6 PC pins 2 2kΩ multi-turn trimpots Bourns 3296W series (VR2,VR3) Semiconductors 2 TIP33B NPN power transistors (Q17, Q18) 1 LM317 adjustable positive 3-terminal regulator (REG1) 1 LM337 adjustable negative 3-terminal regulator (REG2) 1 BR610 bridge rectifier (BR2) 2 1N4004 silicon diodes (D1,D2) 2 33V 5W zener diodes (ZD2, ZD3) Capacitors 2 470µF 100VW electrolytics 1 220µF 63VW electrolytic 1 100µF 63VW electrolytic Resistors (0.25W, 1%) 2 6.8kΩ 2 47Ω 2 180Ω 6 15Ω 1W Fig.8: this diagram shows the details of the mains wiring and all the transformer secondary terminations at the insulated terminal block. details for mounting these transistors are shown in Fig.6. Alternatively, you can dispense with mica washers and heatsink compound and use silicone impregnated thermal washers instead, as can be seen in the photos. Whichever method you use, do not over-tighten the mounting screws. Now check with your multimeter, switched to a high Ohms range, that there are no shorts between the heatsink and any of the transistor collector leads. If you do find a short, undo each transistor mounting screw until the short disappears. It is then a matter of locating the cause of the short and remounting the offending transistor. Double-check all your soldering and assembly work against the circuit published last month and the component layout diagram of Fig.4. Set trimpot VR1 fully anticlockwise so that it is at minimum resistance. Remove both fuses and ensure that the 220Ω 5W resistors are wired across both fuse-holders, as described above. Power supply & case Assuming that you have built two amplifier modules you can now set them aside and proceed to build the regulated power supply board. Its component overlay is shown in While we elected to wire both IEC sockets and switch the female socket, most builders will probably take the simpler approach and not wire the female socket. May 2000  25 options for wiring these and we will come to those in a moment. We also elected to use the front panel power switch and if you are using an older computer case you can use the standard DPST (double-pole, single-throw) switch. However, if you are using a newer ATX case, its power switch will be a momentary contact type which is not suitable. If that is the case, you will be need a push-on push-off DPST switch to mate with the pushbutton on the front panel of the case. Again, you may able to obtain that from an older PC or you can purchase a suitable replacement from Jaycar (Cat. SP-0746). Now to the 240VAC mains wiring options. As far as the female IEC power (output) socket is concerned, you can either leave it unwired (and just use it to blank off the hole) or wire it in parallel with the male IEC power (input) socket. Alternatively, if you decide to switch the IEC female socket, you will need to run two lengths of figure-8 250VAC cable (to run from the IEC sockets to the switch and back). We took this approach but the wiring diagram of Fig.8 shows the simpler approach with the IEC female socket unswitched and just one length of figure-8 250VAC cable running from the IEC male socket to the DPST switch and then to the multi-way insulated terminal block. Note also that a panel-mount safety fuseholder is required and its contacts should be sleeved with heat­ shrink tubing. Similarly, the two IEC sockets should have insulating boots fitted over them to prevent accidental contact with the wiring terminals. Another point to note is that the IEC female socket is larger than its male counterpart and therefore requires a larger boot. Drilling the case Before you can start doing any wiring on the case, all the holes must be drilled for the hardware and any cutouts made. For simplicity, we won’t mention all the holes that are required and we’ll only talk about specific hardware as we discuss the wiring but you have to do all drilling and metal-bashing first. For example, you have to drill all the holes to mount the tunnel The rear panel has an extra cutout for the tunnel fan and has gold plated speaker terminals as well as a 6-way RCA input socket panel. Fig.7. This is quite straightforward to assemble but don’t make the mistake of inadvertently swapping REG1 & REG2, the positive and negative regulators. And make sure that the zener diodes and electrolytic capacitors are inserted the right way around. The next step is to work on the tower case for the amplifier. As stated previously, we purchased a new ATX tower computer case for $66 (from CAM1 Computer Wholesale Pty Ltd; phone (02) 9975 2919. This came with a power supply which we removed and that leaves quite a few metal working details to be sorted out. First, the opening where the power supply was needs to be filled in and to do that we cut off the rear panel of a non-working PC power supply. That gave us a panel with a fan cut-out (for ventilation) and two IEC power sockets – male and female. There are two 26  Silicon Chip An advantage of this case is that you can run some of the wiring between the chassis and one of the side panels. This improves shielding as well as giving a neater result. Note the ribbon cable for the input signal wiring – this is much easier to run than shielded cable. Fig.9: this is the alternative power supply arrangement using two power transformers with their secondaries added together. heatsink and the amplifier modules, regulated power supply board, the power transformers, the multi-way insulated terminal block, bridge rectifier, chassis-mount electrolytic capacitors and the chassis-mount fuseholder. You also have to make the cutouts in the rear panel for the 12V fan, loudspeaker terminals and RCA phono terminal panel. Make sure that all holes and cutouts are de-burred and that the chassis is completely clean of all metal swarf. It is also a good idea to wipe the entire chassis clean with a cloth moistened with methylated spirits or kerosene. This will remove grease and finger-pints which eventually become a site for surface corrosion in these (normally) bright zinc-plated chassis. We also had to remove the 3.5-inch disk drive cage but elected to leave the 5.25-inch drive cage where it was as it was spot-welded in place. Power transformer wiring There are two options for the power transformer. The power supply circuit on page 22 of the March 2000 issue shows a single transformer with two 35V windings and two 50V windings. The prototype 225VA transformer was made by Harbuch Transformers Pty Ltd (phone 02 9476 5854) and they will no doubt be able to supply a 300VA version for this stereo amplifier. As an alternative, we decided to power our prototype with two off-theshelf toroidal transformers: a 300VA unit with two 35V windings and a 30VA unit with two 12V windings. These are wired so that they effec- tively provide two 35V AC windings (from the 300VA unit) and two 47V windings, with the 35V and 12V windings added together, as shown in the circuit of Fig.9. One of the 12V windings is also used to power the front panel LED and the 12V DC fan; more on that in a moment. The two transformers were stacked, with a neoprene washer under the 300VA transformer, a washer between the two transformers and another neoprene washer underneath the steel cup washer for the 30VA transformer. One bolt passes through both transformers and secures them to the case, as can be seen in the photos. So the first steps in wiring the power supply are to stack the transformers together and terminate their primary May 2000  27 This underside view of the finished modules mounted on the heatsink shows how the extrusions have been drilled for M4 screws to secure it in the case. and secondary windings to the multi-way insulated terminal block, as shown in the diagram of Fig.8. Run the 240VAC mains wiring around the top and sides of the case, as shown. Do not connect any of the other power supply components yet until the phasing of the two transformers is confirmed as correct. To do this you connect the unit to the 240VAC mains, switch on and use your multimeter (switched to a 100VAC range or higher) and check that you have the two 35V windings delivering around 37VAC (unloaded) and the summed windings delivering around 50VAC unloaded. If the phasing is incorrect, you may find that the summed wind- Resistor Colour Codes                   No. 4 2 2 2 2 2 2 2 2 2 2 4 6 8 6 6 2 16 Value 18kΩ 12kΩ 8.2kΩ 6.8kΩ 3.3kΩ 2.7kΩ 1.2kΩ 1kΩ 390Ω 330Ω 180Ω 150Ω 120Ω 100Ω 47Ω 15Ω 6.8Ω 1.5Ω 28  Silicon Chip 4-Band Code (1%) brown grey orange brown brown red orange brown grey red red brown blue grey red brown orange orange red brown red purple red brown brown red red brown brown black red brown orange white brown brown orange orange brown brown brown grey brown brown brown green brown brown brown red brown brown brown black brown brown yellow purple black brown brown green black brown blue grey gold brown brown green gold brown 5-Band Code (1%) brown grey black red brown brown red black red brown grey red black brown brown blue grey black brown brown orange orange black brown brown red purple black brown brown brown red black brown brown brown black black brown brown orange white black black brown orange orange black black brown brown grey black black brown brown green black black brown brown red black black brown brown black black black brown yellow purple black goldbrown brown green black gold brown blue grey black silver brown brown green black silver brown ings actually deliver around 24VAC. If this happens, you will need to swap the connections from the two 12V windings. Note that while the wiring diagram of Fig.8 shows the colour-coding of the transformer wires to acheive the circuit shown in Fig.9, you will still have to check the output voltages, as noted above. In fact, while our prototype was wired as shown in Fig.8, we still had to swap one of the transformers secondaries to achieve the correct result; so don’t take it for granted. Next, install the four chassis-mount 8000µF filter capacitors and the bridge rectifier and the regulated power supply board and complete the wiring, as shown in Fig.8. Then apply power again and check the resulting ±55V regulated rails and the unregulated ±52.5V rails. For the regulated supply rails you will need to adjust trimpots VR2 & VR3 to obtain exactly ±55V DC. As far as the main unregulated supply rails are concerned, they will probably deliver around ±53V as they are completely unloaded. These measurements were made with an AC supply voltage of 240VAC. If your mains voltage is higher, and this will normally be the case, then the amplifier supply rails will be increased accordingly. Note that when you switch the unit off, the 8000µF capacitors will take a very long time to discharge. Hence, you should use a resistor of, say, 470Ω 5W to safely discharge each supply rail after your initial tests have been done. Testing the amplifier modules Before the amplifier modules are installed in the case, they must be tested. To do this, you need a steel or aluminium baseplate which can be earth­ed back to the tower case. This becomes a temporary chassis for the amplifiers. Place a piece of cardboard over the base-plate to reduce the Capacitor Codes     Value IEC Code EIA Code 0.15µF  150n  154 0.1µF  100n  104 .0012µF   12n  121 100pF 100pF  100 Fig.10: use this diagram when running all the signal wiring and power wiring to the amplifier modules. Note that the routing of the ±52.5V wiring is critical if you want to obtain the very best harmonic distortion performance. May 2000  29 Fig.11: the full-size PC board pattern for the amplifier power supply. Only one of these boards is required. chance of any shorts from the modules. Test one module at a time. You need to run the five power supply leads from the tower case to the amplifier module: ±52.5V (unregulated), ±55V (regulated) and 0V. Now apply power. No loudspeaker or resistive load should be connected at this stage. Now measure the voltage at the output of the amplifier module. It should be less than ±30mV of 0V. If it is not close to zero, switch off the power as you have a fault. Check over your work very carefully. Check the base-emitter voltages of each transistor; they should all be in the range of 0.6V to 0.7V. Also check for missed solder connections, solder splashes between tracks, incorrectly connected transistors, incorrect transistor types, parts in the wrong way around, etc. Check the voltage across the 3.3V zener diode. Our examples proved to be low as they were 1W types and they needed more current through them. Accordingly we changed the 8.2kΩ 1W bias resistor to 2.7kΩ 5W to increase the zener current to around 20mA. We recommend this change. Now monitor the voltage across one of the 220Ω 5W resistors. With VR1 fully anticlockwise, the voltage should be close to zero since there is no quiescent current in the output stage. Now slowly wind VR1 clockwise until the voltage starts to rise. Set VR1 for a voltage of 4.4V across the 220Ω resistor. This is equivalent to a quiescent current of 20mA or 10mA through each output transistor. You can check this by measuring the voltage drop across any of the eight 1.5Ω 1W emitter resistors. The average value across the resistors should be 7.5mV. Leave the amplifier to run for 10 minutes or so and then retouch the setting of VR1 if necessary. Finally, fit the 5A fuses and the module is finished. Repeat the procedure for the second amplifier module. Wiring up There remains quite a bit of wiring to be done. First, you need to run the rainbow cabling between the RCA phono sockets and the selector switch on the front panel. Note that the white phono sockets are for the left channel; red for the right channel. The wiring from the selector switch to the volume control and then to amplifier inputs is run in figure-8 shielded cable. The details are shown in the wiring diagram of Fig.10. Using ribbon cable for the signal wiring is much easier than running Fig.12: the full-size etching pattern for the amplifier PC board. Two boards are required for the stereo amplifier. 30  Silicon Chip shielded cable. It must be laid flat on the chassis and kept away, as much as possible, from power wiring. Using our tower case, we were able to run the ribbon cable between the chassis and one of the side covers to improve the shielding. The cables for the power supplies must be run exactly as shown in the diagram of Fig.10. First, run three leads, using 7.5A-rated hookup wire, from the regulator board to both modules. These leads must be tightly twisted as shown. Second, run three leads, again using 7.5A hookup wire, from the unregulated ±52.5V rails to both modules and again, tightly twisted. Particularly critical is the way in which these three leads are routed underneath the centre of both amplifier boards and then having the positive and negative leads radiating out to the respective PC stakes on the boards. The routing shown is critical because the heavy class-B currents produce a magnetic field which partially cancels the fields produced by the same class-B currents in the PC board tracks. Note the positioning of these wires carefully; see how they align with the tracks carrying the class-B currents from the paralleled 1.5Ω resistors on each side of the board. A change in position by as little as 5mm can make quite a significant difference to the resulting high frequency distortion performance of the amplifier. This photo shows the detail of wiring to the headphone socket (top left) and power switch. Note the insulating sleeves on the power switch: they’re essential! Output connections Again, this is a critical aspect. For the output leads from the amplifier modules to the output terminals we used a heavy-duty figure-8 speaker cable (Jaycar Cat. WB-1712 or WB1713; 2 x 79/0.2mm). Do not use lighter gauge cables as they do have a significant effect on the ultimate performance. These cables must be tightly twisted for effective field cancellation. The speaker terminals themselves should be heavy-duty solid metal units such as the gold-plated types from Jaycar (Cat. PT-3008). These are another essential item – do not use cheaper plastic or spring-loaded speaker terminals; they do not make reliable low resistance connections and they can make a large difference (like 10 times worse) to the distortion. The recommended terminals will The selector switch and the 10kΩ dual ganged log volume control are mounted on one of the plastic in-fill panels. also take the largest of jumbo speaker cables. You also need to run light duty hookup wire for the wiring to the headphone socket although you may decide to dispense with the head­ phone facility altogether. Assuming that you do wire the headphone socket, you need to run the twisted-wire pair from both channel outputs to the socket but only one earth return is connected while the other remains unconnected, as shown on Fig.8. Two more points about the head­ phone socket: first, do not earth the headphone socket, otherwise you will end up with an earth loop (they cause hum and distortion). Hence, the head­phone socket is mounted on one of the plastic infill panels, as shown. Second, do not use the headphone socket to switch the loudspeakers on and off. While we did this in the above-mentioned Class-A amplifier, the much higher speaker currents involved in this 100W amplifier are too much for the headphone switch contacts to handle and give a low distortion result. In a future issue, we will address the May 2000  31 Fig.13: the existing badge can be removed from the front panel of the case and this one used instead. Finally, here is the whole rear panel of the assembled amplifier. As previously noted, we replaced the cabling to the speaker terminals with much heavier wire – with very worthwhile results. problem of speaker protection, muting and headphone switching. Fan control As mentioned above, the fan is run at low speed and it runs continuously. We were able to salvage an 80mm 32  Silicon Chip fan from a defunct computer supply although they are readily available from electronic parts retailers. If you are buying a fan, purchase the one with the lowest noise rating. These days such 12V fans are brushless (ie, electronically commutated) which means that they are polarised; if you connect them the wrong way around they won’t run. The fan we used is rated at 12V <at> 200mA but we throttled it right back to around 5.8V by using a 120Ω 5W resistor in series with the half-wave rectified DC supply (see Fig.9). While you may be able to run your fan at lower than 5.8V and thereby make it even quieter, you will need to check that it runs properly; if the voltage to the fan is too low, it may not start reliably. Note that while the fan will run much quieter than if it was being powered by the full 12V, it will still make a low level hum which may still be a problem, depending on your listening room and how close you are to the amplifier. In our situation, we found that while ever music was playing, even at very low levels, the fan was not audible but when the program stopped, the fan could be heard as a very muted hum. We’ve taken this approach for simplicity. If fan noise is a problem in your situation, you may need to position the amplifier well away from your listening position or even put it in a cupboard. The DC supply for the fan also runs the front panel LED. Wiring for this LED will already be present in the computer case and it is simply a matter of connecting the wires to the DC supply at the multi-way insulated terminal block. When all your wiring is complete, you need to check all your work very carefully. Then apply power and recheck the voltages on the amplifier. Readjust trimpot VR1 on the amplifier modules if necessary. Finally, place the covers on the case, connect your CD player and loudspeakers. Have a listen close to the loudspeakers without any music playing. There should be only a very low level hiss coming from the speakers. Now place your favourite CD in the machine and sit back to enjoy the SC sound.