Fullscreen HTML5Med Res (??MB)HTML4

Handsome is as handsome does. The 500W amplifier is a big brute but is very-well finished throughout. Feedback on the 500W power amplifier Reader reaction to the 500W amplifier featured in the August, September & October 1997 issues have been very enthusias­tic and a good number have now been successfully built. But it is only recently that Dick Smith Electronics have released their version of the kit and we were interested to have a good look at it. By LEO SIMPSON & BOB FLYNN It’s one thing for us to labour to produce a complex design for publi­ cation and often quite another to see the versions built by readers from kitsets. Sometimes they’re not what we expect and sometimes they are very good. 62  Silicon Chip So we were very interested to fi­ nally see the Dick Smith Electronics kit version of the 500W amplifier, as built by one of their staff members. In brief, while the kit has been produced to a very high standard, it had not been assembled to get the very best performance of which it was capable. We thought we might pass on what we discovered. The external appearance of the kit­ set amplifier is quite professional and pleasing to the eye. All metalwork, except the front panel, has a black, powder-coat finish. The front panel is finished a wet sprayed charcoal and both front are rear panel legends are silk screened in white paint which is very easy to read, even in subdued light. All the cabinet panels are made of steel. The base plate is 1.5mm thick. The front sub-panel, back, sides and top are 1mm plate and the front panel is 2.5mm material. The turned over edges of the vertical panels are fitted with 4mm captive nuts and the top and bottom panels have matching 4mm clearance holes for the 4mm round head assembly screws. The panel holes and nuts are in good alignment, ensuring easy assembly. The complete amplifier is very rigid. Lifting at any corner does not flex the case. Five large rubber feet are fitted to the case to provide good support for what is a heavy unit. Removing the case top revealed a well constructed amplifier with good quality components throughout. Both PC boards have an epoxy substrate and the copper tracks are tinned. The large toroidal transformer is made in Australia by Harbuch Electronics Pty Ltd. The chance of accidental contact with the power supply electrolytics has been eliminated by the provision of a screen of Elephantide insulating material. Instead of an XLR socket, this am­ plifier had an insulated gold-plated RCA socket for the signal input but apart from that all the componentry is pretty much as we specified. Naturally, we were anxious to put it to the test and while we did not expect it to be below par, it was good to find that it equalled the power output of our prototype on both 8Ω and 4Ω loads. Nor did it become too hot. We understand that Dick Smith Electronics also subjected it to full power testing for long periods with­ out any significant problems showing up. There was some evidence of over­ heating but we will come to that later. Where the performance was below par was with respect to noise and distortion. For example, while our prototype gave a signal-to-noise ratio of 117dB unweighted, the DSE ver­ sion was only -100dB. “Only -100dB!” you might say but while that’s pretty respectable it is not as good as it might have been. Similarly, the distortion was not as low as it should have been. Wiring layout is important There are several possible reasons for these differences. First, the power transformer orientation probably was not quite optimum to obtain lowest possible noise. While toroidal trans­ formers do have low hum radiation compared to conventional E-I trans­ formers, their secondary winding ter­ minations are still responsible for the radiation of hum and rectifier buzz. Inevitably there will be some dif­ ferences in this regard between our prototype transformer and the pro­ duction transformers finally used in the Dick Smith Electronics version. So it is difficult for us to nominate the precise orientation. Anyhow, rotating the transformer for minimum hum radiation can give a reduction of several dB in noise. To find the best orientation for the trans­ former, you need an oscilloscope and better still, a sensitive AC millivolt­ meter or as we have, a distortion measurement test set. If you have a quiet location and very keen hearing, it is also possible to do it by ear. Doing the adjustment is not easy be­ cause you need the amplifier powered up and the transformer bolt loosened off so you can rotate it while watching the rectifier buzz on the oscillo­scope. The transformer leads are very stiff and they make it hard to obtain the optimum position. At the same time you must be ex­tremely careful not to come into contact with any high voltage yourself of cause any shorts. In other words, you need to do it very carefully. The pragmatic approach Putting ourselves in the shoes of a typical hobbyist con­structing this amplifier, we would not be inclined to worry about getting the absolute minimum noise out of the amplifier unless rectifier buzz was audible in the loudspeakers. As we said above, -100dB is pretty respectable. Second, and much more important to obtaining minimum dis­tortion and noise, the power supply leads to the PC board were not run in the same way as in our prototype. They looked neat and tidy but they were not right. On page 59 of the September 1997 issue we were quite spe­cific in stat­ ing that “it is important to follow the wiring details of Fig.5 quite closely” and we went on to give details of the transformer wiring. Well, perhaps we should have been even more specific and should have stated that the sup­ ply wiring layout should be exactly as shown in the photos. Why? Because the orientation of the supply leads and output wiring does have a major effect on the harmonic distortion. What happens is that the heavy cur­ rents in the supply leads and output wiring have an associated magnetic field and this is radiated into the early stages of the amplifier. This leads to higher harmonic distortion. In our prototype, the power supply LOUDSPEAKER SALE Limited stocks Prices include sales tax Australian Audio Consultants PO Box 11 Stockport SA 5410 Phone or Fax 08 85 282 201 Vifa D25AG-35-06 Vifa D19SD-05-08 Vifa D26SG-05-06 SEAS ExcelT25-001 Vifa M11WG-09-08 Vifa P13WH-00-08 Vifa P17WJ-00-08 Vifa P17SJ-00-08 Vifa M18WO-08-08 Vifa M18WN-19-04 Vifa M22WR-09-08 Vifa M22WR-19-04 Vifa M22WR-29-04 DVC Vifa M26WR-09-08 Vifa M26WR-19-04 Vifa M26WR-29-04 DVC Dynaudio 24W100 Scanspeak D3806/8200 Dynaudio 20W75 Dynaudio 15W75 Call for full specs. $69.00 $45.00 $49.00 $34.00 $62.00 $49.00 $194.00 $179.00 $99.00 $79.00 $99.00 $75.00 $99.00 $75.00 $109.00 $75.00 $145.00 $90.00 $149.00 $90.00 $195.00 $110.00 $195.00 $120.00 $210.00 $125.00 $229.00 $135.00 $229.00 $150.00 $249.00 $159.00 $399.00 $199.00 $181.00 $90.00 $296.00 $170.00 $257.00 $140.00 leads were run close to the bottom plate because this minimises harmon­ ic radiation. But on the Dick Smith Electronics amplifier the same leads are run about 50mm above the metal­ work and make a right-angled down turn to the PC pins. Re-routing those wires so that they can lie flat on the chassis will have a very worthwhile effect on the harmonic distortion. Some readers may have wondered why we go to the trouble of provid­ ing large colour photographs of our prototypes. It is not just to show off the project or to use up more space. We do it so that constructors can see precisely what we have done. In most cases, they should try to duplicate wiring and other details as closely as possible, unless they have the skills and measurement equipment to check the performance of the finished project. This applies particularly to audio equipment where wiring layout is extremely important. Even the best audio circuit will be below par if the wiring layout is not optimised. Mind you, there is a down side to having large clear photos of projects in the magazines and that is that it lets March 1998  63 needs to be found. Winding the coil with 1.25mm diam­ eter enamelled copper wire instead of the 1mm diameter wire listed in the parts list will also reduce the coil dissipation. Note: while we used and spec­ified a Philips 0.15µF, 275VAC capacitor at the output of the amplifier, this device is not rated for the continuous applica­ tion of the high­ er audio frequen­ cies at full power. A better choice of device for this position is a Wima MKC10 0.15µF 630VDC poly­ carbonate or if avail­ able a Philips MKP378 or MKP379 0.15µF 630VDC polypropylene capacitor. Alternatively, you could consider the Wima MP3-X2 range. These are metallised paper 250VAC or 275VAC RF interference suppressors for class X2 service. DC offset error We also noted that a 30Ω resistor was shunted across the emitter resistor of Q2, one of the input differential pair tran­ sistors. This had apparently been done in order to get the output offset voltage down to an acceptable level and should not have been necessary and indicated that one of the devices must be leaky or faulty in some other way. We pulled the pair out and meas­ ured them. One was made by National Semiconductor and had a beta of about 100 while the other was unbranded and give a beta reading of over 1000 and was probably leaky as well. No wonder the DC offset could not be adjusted to zero! We replaced the input pair with Mo­ torola (again!) devices and the shunt resistor was removed. The output offset voltage was then easily set to zero with VR1’s wiper only slightly off-centre. That is how it should be. All told though, we were very grat­ ified to see this ampli­fier. It is well presented and anyone who built it would be pleased with its perfor­ mance and finish. The complete kit is very good value at $799 (Cat K-5582) and is available from all Dick Smith SC Electronics stores. The interior of the assembled 500W amplifier from Dick Smith Electronics. While all the componentry was the best quali­ty, the wiring itself did not duplicate the exact layout shown in the photos of our prototype. keen-eyed readers sometimes see that we have made a mistake. Oh well . . . of the mounting; a smear of heatsink compound is all that is required. Voltage gain stage Heat buildup There is another factor which can be significant in the distortion perfor­ mance of an amplifier such as this and this concerns the driver and output transistors. The output and driver transistors can only be obtained from Motorola so there is no problem with substitutions – there aren’t any that we know of. However, we specified MJE340/350 transistors in the voltage gain stage and in our experience, none are as good as those made by Motorola; they may be rated the same but their distor­ tion is worse. For the record, we did not specify Motorola MJE340/350s in the parts list but perhaps we should have. There was one other point concern­ ing the MJE340/350s in the review amplifier. They were mounted on their heatsinks with interposing mica wash­ ers. These insulators are not necessary and only add to the thermal resistance One of the tests Dick Smith Elec­ tronics performed on their amplifier was to run it at more than 500W out­ put into a 4Ω load for a full working day. This is a severe test of the output coupling network but does not run the output transistors at their maximum dissipation. Maximum power dissipa­ tion actually occurs at around 40% of power output. During the full-power test, the for­ mer of the output choke (L1) melted and the three 18Ω 1W resistors also overheated. By any normal standard, a full power test for 8 hours is a long way from typical operating conditions. During extensive testing of our prototype, we had no prob­lems with individual components overheating. If the amplifier is intended to be run continuously at full power or at very high levels with program material of a small dynamic range then a coil former with a higher melting point 64  Silicon Chip