I have obtained (cheaply!) an old ADC Sound Shaper One Equalizer (schematic attached) and I wish to "make it all that it can be!!". I have an LM317 power supply regulator board w/6800uF filtering that will fit; and I am replacing the coupling and filter circuit capacitors with polypropylenes. My question is:
Are there better (lower noise/distortion) transistors that can be substituted for the current 2SC1222 and 2SA539 that are on board?
Are there better (lower noise/distortion) transistors that can be substituted for the current 2SC1222 and 2SA539 that are on board?
Last edited:
Transistors are transistors. Inherent distortion is all the same.
The circuit is not of the best; but the 42V supply for 1V signals takes much of the curse off.
The unbalance in the long-tail makes 2nd harmonic distortion predominate. Exact balance would null the 2nd and may put the 3rd well down. One resistor added makes it "better". More elaborate mirroring makes it "more perfect". However there is still 2nd in the input buffer. My matchbook says it may be 0.06% at 10V RMS, and much less at typical 1V signal level.
I don't know how a regulated supply helps?
42V is rather high for a '317 chip.
The "real upgrade" would be to drop the rail to 36V, and replace both stages with any good audio op-amp chip.
The circuit is not of the best; but the 42V supply for 1V signals takes much of the curse off.
The unbalance in the long-tail makes 2nd harmonic distortion predominate. Exact balance would null the 2nd and may put the 3rd well down. One resistor added makes it "better". More elaborate mirroring makes it "more perfect". However there is still 2nd in the input buffer. My matchbook says it may be 0.06% at 10V RMS, and much less at typical 1V signal level.
I don't know how a regulated supply helps?
42V is rather high for a '317 chip.
The "real upgrade" would be to drop the rail to 36V, and replace both stages with any good audio op-amp chip.
Attachments
Replace all the electrolytic capacitors in the audio path, including the gain control pcb,
with fresh bipolar types. The two 330uF and the 47uF in the power supply also will
need refreshing. Maybe delete the tape monitor switch, and run the inputs directly to the pcb.
Clean the pots, of course.
.
Last edited:
Thanks for your suggestions! I think I will go with all polypropylene capacitors in the signal path, and I will replace the power supply with a neat little 2x2" LM317 regulator board that I found here on the Forum. I'm going to modify the 'Monitor' switch to be a 'Bypass' switch instead; this will allow easy comparison between EQ'd signal and straight. I think the trick with equalization is to use it very moderately.
Last edited:
Thanks for your suggestions also! I am still trying to grasp this 'imbalance": TR103 has 27.1v on its base, whilst TR105 has 27.2v; TR103 has 41.3 v on its collector, TR105 has 42.1v; Both share a common emitter resistor with 26.6 v. How does this result in the WILDLY different (9x) current difference you have shown?
I'm not sure the regulation will help tremendously, but the board does allow putting a 6800uF 63v filter cap on the supply, which certainly will improve the humm performance over the ~660uF filtering it currently has. The LM317 datasheet says that its limitation is 40 volts between input and output; I will have just over 18 volts, so I should be good to go there---there is no mention of a MAXIMUM output voltage that I can see. I may indeed end up replacing the whole circuit board with an IC-based design, but that is beyond my pocketbook at the moment!!
Many of the signal path capacitors are too large in value to be pp types,
so they'll have to be electrolytics.
> How does this result in the WILDLY different (9x) current difference you have shown?
Do the arithmetic. TR107 Base needs only maybe 0.12mA, and that will BE the current in TR103 when the NFB loop forces everything to its happy-point. Meanwhile the shared emitter resistor is clearly flowing over 1mA. Anything TR103 does not need, TR105 must be taking.
> TR103 has 27.1v on its base, whilst TR105 has 27.2v
The theoretical Vbe difference for 10:1 diff of current is about 0.060V. In old-production parts the device variation may be similar due to process-slop. So quick reading on a needle meter will not tell a lot about balance. The noted 0.1V difference does back-up the theoretical ~~0.060V difference within the width of a needle meter or 2.5 digit DVM.
It is not a bad design. Since then many authors have pointed out that even slight imbalance raises THD enough to detect, maybe hear. My added 1.2K resistor forces TR103 to pass a half-mA before it wakes-up TR107, getting the unbalance down. It will also improve loop gain by improving hIE of TR103 more than hIE of TR105 is spoiled. A more exact balance would be a current mirror up top. But there are many other "improvements" possible. All of them built-in on many chips, better than you can hand-hack.
Short of chipifying it, I absolutely agree with Rayma. e-Caps go bad, switches can turn sour, pots do go sour. I do not think this unit "sucked" when new. (I recall it could be distressed in PA use where signals can run much higher, and loads sometimes lower, than hi-fi duty.)
Do the arithmetic. TR107 Base needs only maybe 0.12mA, and that will BE the current in TR103 when the NFB loop forces everything to its happy-point. Meanwhile the shared emitter resistor is clearly flowing over 1mA. Anything TR103 does not need, TR105 must be taking.
> TR103 has 27.1v on its base, whilst TR105 has 27.2v
The theoretical Vbe difference for 10:1 diff of current is about 0.060V. In old-production parts the device variation may be similar due to process-slop. So quick reading on a needle meter will not tell a lot about balance. The noted 0.1V difference does back-up the theoretical ~~0.060V difference within the width of a needle meter or 2.5 digit DVM.
It is not a bad design. Since then many authors have pointed out that even slight imbalance raises THD enough to detect, maybe hear. My added 1.2K resistor forces TR103 to pass a half-mA before it wakes-up TR107, getting the unbalance down. It will also improve loop gain by improving hIE of TR103 more than hIE of TR105 is spoiled. A more exact balance would be a current mirror up top. But there are many other "improvements" possible. All of them built-in on many chips, better than you can hand-hack.
Short of chipifying it, I absolutely agree with Rayma. e-Caps go bad, switches can turn sour, pots do go sour. I do not think this unit "sucked" when new. (I recall it could be distressed in PA use where signals can run much higher, and loads sometimes lower, than hi-fi duty.)
> into a lowish load
When new, some kids confused it for Pro-type gear.
It will drive 600 Ohms; just not to Pro levels with point-oh THD. Savy sound engineers knew how to interface it.
I really have no qualms about it in hi-fi. Unless it has been abused, I don't think its small self-hiss will approach the noise of any recording. It is as you say getting old enough for a cap and switch/pot job.
When new, some kids confused it for Pro-type gear.
It will drive 600 Ohms; just not to Pro levels with point-oh THD. Savy sound engineers knew how to interface it.
I really have no qualms about it in hi-fi. Unless it has been abused, I don't think its small self-hiss will approach the noise of any recording. It is as you say getting old enough for a cap and switch/pot job.
Well, the ADC Sound Shaper One arrived today. Fortunately, it works!! Unfortunately, the input coupling and first two filter stages sport the DREADED tantalum electrolytic capacitors---yeech!! Fortunately, there is plenty of room for polypropylenes to fit. Also fortunate is the fact that this device is pretty damn quiet. I finally found a spec sheet for the Sound Shaper Two, which has the same exact circuit, except for a regulated supply and metering circuitry added on. They spec noise as -85 db below a 1 volt rms input, and from listening, I believe 'em---there is no audible excessive 'hiss" to worry about. So, off to ordering my polypropylenes and installing my little LM317 regulator board. for which there is also plenty of room.
- Status
- Not open for further replies.