ark (11/10/82)
I said I'd report my experiences on the Heathkit AA-1800 power amplifier. I am finally in a position to do so. Since it has been a while since my last comments on it, let me first review the amp's basic characteristics. The AA-1800 is a basic power amplifier, rated at 250 watts per channel by the (very conservative) FTC method. Heath claims about 3 dB of dynamic headroom -- in other words it will put out 500 watts for short periods without clipping. Distortion is rated at 0.025 percent. The amplifier has three circuit boards: a fairly elaborate protection circuit and one board for each amplifier channel. Although the protection circuit is quite complicated, the amplifier boards are still more complicated. Each amplifier board has 27 transistors, eight of which are power transistors in TO-3 cases. Said power transistors mount to a huge heat sink, one for each channel. The amplifier has several nice features. First, it is designed to be put somewhere out of the way and forgotten. Thus, the front panel has no controls, not even a power switch. Because most preamps do not have switched outlets that will handle the 1200 watts the amp draws at full power, the amp comes with two (!) AC cords. The thick one supplies power to the amplifier itself; the thin one drives a relay that turns the thick one on and off. This keeps you from melting the power switch in your preamp. There are input level controls, but they are on the back panel. I think this is where they should be -- use them to get the overall gain of your system right and then forget about them. I have a nice technique for this: set your preamp on "mono," and then connect a sensitive voltmeter between the hot leads on each channel. Adjust the gain on one channel for the level you want, and then adjust the other one for minimum reading on the meter. This will compensate for gain differences between your preamp channels. The only things on the front panel are carrying handles and four LEDs. The LEDs are: power, protection, L and R full power. The latter two come on whenever a channel is getting close to clipping. The protection LED is on whenever an internal relay has disconnected the speakers from the amplifier output. This happens: (a) during the first 8 seconds of operation, (b) when either heat sink gets too warm (about 200 degrees F), or (c) when excessive DC is present at the output. This latter thing is determined cleverly: whenever the output exceeds about 2.5 volts magnitude, it starts a timer. If the voltage has not found its way below 2.5 volts within 50 ms, the speakers are disconnected. The 50 ms figure means that not even a full-power signal at 15 Hz will trip it, but even a moderate amount of DC will. The protection board contains 12 transistors, the aforementioned four LEDs, an integrated circuit (four xor gates), a relay, and other components. Chassis wiring includes a huge (about 25 pounds) power transformer, two 13,000 microfarad (125V) capacitors almost as large as beer cans, two relays, a bridge rectifier, level controls, a fuse, switches, sockets, and a thermistor, all interconnected by two pre-cut wiring harnesses. The thermistor serves an interesting purpose. It is rated at 3.3 ohms cold, and it is in series with the power transformer primary. When the amplifier is first switched on, the thermistor limits the primary current so that charging the filter capacitors will not blow the household wiring or the 10 amp line fuse. As the capacitors charge, the thermistor heats up and (presumably) decreases its resistance to let the capacitors charge more. When things have gotten close enough to full charge, a voltage taken off a different secondary transformer winding closes a relay that shorts the thermistor and couples the primary directly to the wall socket. The worst part of the physical design is the mounting of the LEDs on the protection circuit board. They protrude from, and are soldered to, the foil side of the board. The directions say to make sure that the LEDs are exactly a specific distance from the board, and when they say "exactly" they mean it. When the board is finally mounted, behind the front panel, any small variations in how high the LEDs sit shows up as a variation in how far the LEDs peek out of their little holes in the front panel. Even a single millimeter of variation is unsightly. I reheated and slightly shifted the LEDs so many times that I began to damage the circuit board foils. These LEDs should be in sockets. When I first turned on the power supply circuits, the power LED did not come on. This turned out to be due to a dead power LED. Maybe it was defective and maybe I resoldered it too many times. At any rate, neither of the Heath Electronic Centers in my neighborhood have these LEDs in stock, so I obtained a replacement locally. Before I did that, though, I tried to check out the protection circuits. The instructions tell you to connect the thermal circuit breaker from one of the boards into the protection circuit and to short the other circuit breaker socket with a wire jumper, and then the protection LED should go off after eight seconds of operation. It didn't. Two hours later, I discovered that if the power LED is broken, the protection LED will never turn off. Sigh. After fixing that problem, I started testing the amplifier boards. The first board was uneventful. The second time I applied power to the second board, the result was a shower of sparks and smoke. After I disconnected everything and went trouble-shooting, I found that one of the power transistors had a leaky insulating gasket. It wasn't a dead short, so the resistance tests I had previously performed on the board didn't turn up any problem. There was enough leakage, though, to blow out two resistors that ordinarily should have almost no current through them. I unsoldered and tested every transistor on the board after that, and found one leaky power transistor. I replaced that and the two fried resistors, and everything was almost normal, but that "almost" had me worried. The output level would wander, apparently randomly, with an amplitude of about 10 mV and a frequency of about 2 Hz. A friend who knows more about these things than I do said that these symptoms look very much like a transistor that is about to give up the ghost, and I sure didn't want that to happen with the amplifier connected to my loudspeakers. Thus, on September 9, I dispatched the ailing board to Heath in Benton Harbor. They told me it would be about three weeks in their shop. Unfortunately, they didn't know at the time that they didn't have any power transistors in stock. The first thing they did when the board came in was to pull all the power transistors and run them through their transistor tester. All eight showed up leaky, so they wanted to replace them. It wasn't until almost the end of October that they were finally able to return the board. When the board came back, I tried the same test I had the last time: putting power on it and measuring the DC output level with my sensitive voltmeter. I was dismayed to find that it showed exactly the same symptoms as before. I shipped the board back to them yet again, expecting to have to fight with them before they would do any more work on it. After all, they claimed that they had measured the board and it met their power and distortion specs. But I described the symptom again, in detail, and asked them to call me when they had gotten the board and verified that it was behaving as I claimed. The sequel was a pleasant surprise. They called back and said, in effect, "We got your board and measured it carefully. It does indeed misbehave as you described. It should not do this. We have no idea what is wrong. Rather than waste a whole lot of time trying to fix it, we are going to build you an entire new board, test it thoroughly, and then send it back to you." Total time in their shop this time was about two days, and no charge for any of the service. Their decision to do this may have been influenced by the fact that my solder joints are at least as neat as theirs, and I tried to sound like I knew what I was talking about on the phone. The new board arrived last night. I put it through its paces, and it appeared to work, so I buttoned the whole amp up and tried, not without trepidation, connecting it to my speakers (through fuses, of course). To my delight, it worked! It may be my imagination, and I haven't done A-B tests, but it sure seems to be cleaner than the 60 W/ch receiver it replaced (a Heathkit AR-1500). At the end I was running my new preamp through the power amp section of the receiver, so I know that the only difference was the power amp. The receiver is going to be the nucleus of a second system for the bedroom. Meanwhile, my first impressions. Construction: when it is good it is very very good, but when it is bad it is horrid. The sanity checks recommended during construction are pretty good, but I would add one other: just before you apply power to a board for the first time, look at the four screws that secure the board to the aluminum bracket on which the driver transistors are mounted. Exactly one of these four screws makes electrical contact with the board foils. If you remove this screw, there should be no electrical contact at all between the heat sink and any components on the board. Measure the resistance between the heat sink and the case of each power transistor. When you have removed this screw, this resistance should be many megohms (it won't be truly infinite, because the thermal glop on the transistors will leak a tiny bit). Now replace the screw and continue with the published instructions. If I had thought to do this before, rather than after, I applied power, I would have had the amp working in August, not November. The Heath Electronic Centers are sympathetic, but seem to stock only about half of the parts I needed while building the kits. I had to order the rest from Benton Harbor. Voltages inside the main power supply can be lethal. During testing, make sure you follow all their instructions EXACTLY and DO NOT WORK ALONE. Operation: This amp has exactly the features I think it should. You use the level controls to get the overall gain of your system right, and then put it in a corner and forget it. The arrangement with the two power cords is wonderful. Do be careful when you install it -- it is heavy enough that you can injure yourself trying to lift it. It runs warm, so make sure it is in ventilated quarters. Performance: It sounds good. There is NO audible hum or noise, even when I walk up to the speaker and put my ear against it. I have not gotten the clipping LEDs to go on yet, though I haven't really tried very hard. When I've had a chance to run lab tests, I'll report the results.
logo (11/11/82)
At the recent AES show, I spoke with an engineer from Sounstream and a product manager (ex-engineer?) from Sony. Both of them claimed that the low pass filters that they used in the input and output stages of their digital recorders were minimum phase (in what freq range? hmmmm) and were associated with appropriate all pass filters to make the phase shift (nearly?) zero for all frequencies. As discussed in a recent article in JAES, a constant phase shift is probably not acceptable for audio reproduction (especially if the shift is 180 degrees). For those who are not familiar with the terminology, minimum phase means that the circuit has a linear phase shift vs frequency curve. Also, sorry if this article has a funny name. They have changed the name of my MAILER program and readnews binds at invocation. David (Reisner) uucp : ...!ucbvax!sdcsvax!logo arpanet : sdcsvax!logo@nosc