The information in this document is the result of my experience with an Akai 1730D-SS 4-track, 4-channel deck. This particular deck, which was a flagship of the Quadraphonic era, utilizes removable/replaceable Playback and Record amplifier boards (four of them). It is my understanding that these boards are similar, if not identical, to those employed in some other Akai models of the same vintage. Therefore, much of the information in this document may also apply to those units. If you have information concerning which Akai models fall within this category, or a schematic/service manual you'd like to contribute, please contact me so that I can revise this document.
Possibly similar models: 202D-SS 4400D 4000D GX-M11 GX-220D GX-280D
The Sanyo LD3141 is a preamp IC in a ZIP package. It is a very simple chip, basically comprising a two-transistor amplifier. In Akai's application it is powered by a single 24V supply. This IC requires no input bias, and it's output is biased internally. It also has a reputation for being noisy. However, as you will see elsewhere in this document, that reputation may be unfounded. There are at least two versions of the LD3141, differentiated by their sub-code, 10 or 20. I have no idea how they may differ otherwise, although I believe the 10 contains three transistors (in the 1730D-SS 20's are used on the P.B. Amp boards, 10's on the Rec.).
The (CS)C458 is an NPN silicon transistor suitable for audio frequency signal amplification. Vintage versions of these transistors have a reputation for going noisy that I can personally vouch for. The the noise is produced at the base of the transistor and is temperature sensitive. In Akai's design, noise produced at the base of the first C458 in the two-transistor power amp stage is injected back into the negative feedback loop of the LD3141 through the latter's output. This results in a somewhat non-intuitive behavior as the noise level grows when the LD3141's gain is increased, which would normally imply that the noise source was at the LD3141's input, or, after eliminating that possibility, within the LD3141 itself. The implication is misleading: the LD3141 is only amplifying the noise produced by the faulty transistor at its output.
In the case of my unit, the noise was the fault of these transistors. Of course, thanks to the law of probabilities, I only discovered this after having bothered to devise a way to substitute an Op-Amp for the LD3141. If you are reading this because your Akai unit makes a rumbling noise in one or more channels which increases in volume as the unit heats up, resembles the sound of someone blowing on a microphone, and tracks in intensity as you adjust the gain pot on the playback board, then you can save yourself a lot of trouble by simply replacing these troublesome components.
Only the first C458 in a given channel, whose base is connected to the output of the LD3141 (although this isn't apparent when looking at the circuit), needs to be replaced in order to eliminate this annoying noise. If you have enough parts, however, you might as well replace them all. For this purpose I employed C1815-GR's, though C945's or any similar general purpose silicon NPN transistor should work--perhaps the most obvious substitute would be C458's of more recent manufacture. In any case, replacing these is easy. Take care, however, not to install the replacements backwards; the part number is on the BACK of the old C458's, whereas it will almost certainly be printed on the front of your replacements.
There are also C458's on the Rec. Amp boards (one per channel), but these don't appear to feed back any noise. Replace them if you feel like it.
Fig. 1 illustrates the location of these transistors (actually, their replacements) on the Akai LF-5022 P.B. Amp board.
|
Since this part has long been obsolete, it can be difficult to obtain. There are some people selling these on the Internet at outrageous prices. But if the LD3141 is really so prone to failure as some individuals claim, why would one replace it with a NOS part when a modern component could be substituted? The LD3141 is noisy by modern standards, but this is only broad-spectrum, white noise, and is probably insignificant in the presence of tape hiss. But if you're like me, and you actually use your deck for recording or dubbing (esp. to digital), rather than simply listening to those old Sinatra tapes, then you might benefit from replacing even a perfectly good LD3141 with a low-noise Op-Amp. At least that's what I've convinced myself of in order to justify all the trouble I went to figuring out how.
The substitution of an Op-Amp for the LD3141 is complicated by several factors. First, the LD3141 has 9 pins, eight of them used, the function of most of which is at best poorly defined. Second, since the input of the LD3141 requires no bias voltage, its feedback loop is AC coupled in Akai's circuit. This is incompatible with Op-Amps, which require a bias voltage at their input and throughout their external feedback loop (in situations such as this). For these reasons, substituting an Op-Amp requires the addition and replacement of a handful of passive components, as well as the exclusion of some existing connections.
|
Fig. 2 is a schematic diagram of the LF-5022 P.B. Amp circuit modified to accept an Op-Amp (IC1) instead of an LD3141. The named components are additions or replacements, the unnamed shown for reference. The pin numbers refer to the LD3141 (see Fig. 3), not the pins of the Op-Amp (for information on which you should check your data-sheet). The parts list is as follows:
| name | value |
|---|---|
| RA | 10kΩ |
| RB | 10kΩ |
| Rin | 100kΩ |
| R1 | 100Ω |
| C1 | 25V 47µF |
| C2 | 16V 100µF |
| IC1 | 1/2 TL082 |
|
Obviously, an 8-pin DIP Op-Amp won't fit in the LD3141's footprint. You'll have to install it with wires. I wired the TL082 on the top the board, with enough slack to allow it to reach the midpoint, secured it with a dab of hot-glue, and installed the other components on the solder-side of the board as shown in Figs. 4a and 4b.
 |
|
The highlighted areas in Fig. 4b point out where a trace was cut and jumpered around. This is necessary to exclude the output coupling capacitor from the Op-Amp's feedback loop and corresponds in the schematic to moving the feedback connection from one side of the 10µF output capacitor to the other. Without this measure, the gain of the Op-Amp at DC will be, for all practical purposes, infinite, resulting in the output being railed to Vcc (24V), which might destroy your Op-Amp and would in any case result in extreme distortion the AC waveform.
Capacitor C2 in the schematic is actually a replacement of the existing 6.3V 100µF electrolytic on the board. The purpose of this capacitor is to decouple the bias voltage from power supply variation and eliminate the noise introduced by resistors RA and RB. Since our input bias voltage is 12V, this capacitor must be upgraded to a 16V version. Fig. 5 shows the location of this capacitor for each channel. It is interesting to note that the capacitor for my front right channel was installed backwards at the factory, severely impairing it, but apparently without noticeably affecting that channel's operation. When you replace this capacitor, be sure to double check its orientation; the side with the minus or stripe should face the more negative voltage, in this case ground.
|
As you saw in Fig. 4b, the rest of the components, four resistors and a capacitor, are wired point-to-point on the solder-side of the board. This is actually pretty easy if you use resistors having one long and one short lead, but it is still a tight fit. Be sure you check your work for solder-bridges before you try plugging it in. The wiring should be apparent from the pin-numbers in the schematic and Fig. 3, but I shall nonetheless attempt to put it into words:
Since an 8-pin dual Op-Amp package like the TL082 has only two power connections, you can skip steps 4 and 5 for one of the channels on each board.
If everything is connected properly, then the board will function as normal when it is reinstalled; ie. the gain trim-pot will behave as it did with the LD3141 installed and the frequency response should be just as close to (or far from) the NAB curve as Akai designed it to be. The gain, however, will be slightly higher than it was before the substitution; I find this makes calibrating the playback levels easier. You may substitute a 150Ω resistor for R1 if you decide you don't need the extra gain.
Once you've converted all the channels you like--I did them all--you'll have to recalibrate the unit for proper playback and record levels. If you have the manual, I'm sure it will tell you the proper way to do this. I didn't have the luxury of manufacturer's instructions or an MRL tape, so I just connected a 40mV 1kHz sine-wave to the line-in and adjusted the gain until the meters read 0vu while recording the tone, in both TAPE and SOURCE modes (to adjust the gain in SOURCE mode you must fiddle the pots on the Rec. boards, which are inter-dependant), for each channel. While you're at it you might as well adjust the tape bias voltage levels for minimum distortion of low-frequencies, but explaining that procedure is beyond the scope of this document. Be aware, however, that you will not be able to get exactly the same record level in 4-channel mode as in 2-channel mode; this because some aspect of the switching process reduces the bias voltage and, to a lesser extent, frequency when in 2-channel mode. This effect may or may not be a peculiarity of my unit. Perhaps someone with the service manual could offer more insight.
Unfortunately the above procedure does not apply exactly to the record circuitry; in order to substitute the LD3141 with our Op-Amp here we must modify our approach. Fig. 6 is a schematic diagram of the relevant portions of the record circuit. The previously stated rules of notation also apply to this diagram and the pin layout on the board is identical. Also, recall that the LD3141-10 used on the record boards is subtly and mysteriously different from the LD3141-20, although I believe the that versatility of an Op-Amp allows us to safely ignore those subtleties. I should state that I found nothing wrong with the record boards in my unit and undertook to upgrade them only out of a sense of completeness. It is also entirely possible that the design of the record circuit makes it immune to noise feedback from any following stage.
|
Take note that the inverting input of IC1 is connected to hole 3 on the Rec. board, while we had it going to hole 2 on the P.B. board--this is due to differences in the circuits. Otherwise, the Op-Amp connections are identical and the above wiring instructions should apply.
Here is the parts list:
| name | value |
|---|---|
| RA | 10kΩ |
| RB | 10kΩ |
| Rin | 100kΩ |
| R1 | 154Ω |
| C1 | 25V 47µF |
| C2 | 16V 100µF |
| IC1 | 1/2 TL082 |
The record circuit, unlike the playback, is a fixed-gain configuration. The gain required (of the LD3141 stage) is approximately 144 (the excess power being used elsewhere in the circuit). The schematic specifies an ideal R1 value of 154Ω; this is not going to be a resistor you just have lying around. I solved this problem by measuring a handful of 150Ω resistors and picking out the 154Ω's. If you don't care whether or not the deck is in spec. as regards its input/output levels, then it doesn't really matter what value you use, within reason, as long as it's the same in each channel. I arrived at the value of 154Ω by measuring the gain of the unmodified circuit, not calculating from scratch based on the desired input/output ratio, which I didn't think it wise to attempt without a service manual.
|
Fig. 7 shows the locations for C2 on the Rec. Amp board. As in the playback circuit, these must be replaced with 16V 100µF electrolytic capacitors.
To remove the output capacitor from the feedback loop, instead of having to cut a trace, we can simply lift one leg of the 22kΩ resistor seen in the schematic and bend it over to the back of the board. Fig. 8a shows this part's location on the unmodified board. In Fig. 8b you can see where the free leg of this part has been connected on the solder-side of the board. For the right channel you'll have to use a jumper wire to reach your target.
|
 |
I sincerely hope someone will find this information useful. I searched high and low for something like it before deciding to dive into this project on my own. Perhaps these instructions will put to rest a few rumors surrounding the LD3141 and the Akai decks containing it. Don't let some fool tell you your deck is worthless just because it uses a now obsolete part--a normal person would consider any reel-to-reel machine obsolete and worthless in this iPod era anyway. These and other simple modifications and repairs have made my 1730D-SS a quiet, effective, and reliable--not to mention beautiful--part of my audio system.
October 2006
Jonathan Moore Liles -- j dot liles at unix dot net
 |