Things are a bit nuts right now but I did finally find all my notes on this subject and am working on compiling them and screeding out the irrelevant.
High-level points:
The 58/MS16 amplifier circuit is pretty neat. It’s a DC servo design which eliminates almost all the coupling caps you’d normally have with a more garden variety design, and this is better. This same circuit concept carried over to the 48 as well as the ATR60 series. I don’t know if it carried all the way forward to the ATR80 series…don’t have a service manual for that. If somebody knows or can share tech docs from an ATR80 manual I’d love to know. But it’s a cool design. Not “boutique” per se, but definitely not “garden variety”. I’ve always wondered why it is I read some subjective comparisons between the sound and noise levels of the MS16 compared to the ATR60-16, that the MS16 sounds better and is quieter, but the heads are the same, and the amplifier circuitry is conceptually the same. I wonder if those impressions were machine dependent; if the ATR60 machine was not setup correctly, etc.
Some comments on Jim Williams’ work on this subject and in general: I don’t know the man at all, he is respected and his advice widely heeded. My personal opinion, and with all due respect because I’m just little people in these matters, is some of what he advises is over-the-top…over-the-top for me. Maybe not for you. I’m going to speak about this with regard to me and my opinions. My perception is that part of what categorizes a device as sounding good to Jim Williams is transparency…input equals output. And that doesn’t interest me as much. And I believe some of what he advises is in that pursuit. Again, these are entirely my opinions and perceptions. That doesn’t mean at all he’s wrong or bad, certain things just don’t strike me as coincident to what sounds “good” to me. I think the main facet in this regard is the practice of bypassing all the electrolytics with a small value film cap. For me there might be certain applications where I want to do this but only for specific circumstances. I don’t feel for me it’s necessary to do this wholesale bypass modification. Clearly a lot of people feel otherwise and maybe you do too, but for me, generally, if it is an older device with general purpose electrolytics installed, replacing the factory caps with modern low ESR parts is an improvement. I typically use “audio-grade” parts for whatever that’s worth. I’m about 50/50 wondering if I’m only paying extra for marketing hot air. But I do it anyway. The other thing I always look for is how the audio power supply rail or rails are filtered, if there are secondary filters either in the supply or local to the devices being powered and if not make a decision if there’s something to add there. Recap the power supply if it’s older. All of this is to help lower the noise floor. Lastly a general thing to look at is the value of any coupling caps, and to make sure they are not too low in value either capacitance or voltage. If the capacitance is too small that can create an audible HPF. I ran across this in a mixing console I was working years ago. The manufacturer used a lot of a certain value of cap throughout the console, probably because that helped drive the cost down with higher volume purchasing of the caps, but that cap at the input created a HPF and I don’t recall the value of the cap but I think the -3dB knee was in the low 20Hz range. Well that’s below human hearing you say…but that’s just the -3dB point. Depending on the slope of the curve that could easily encroach into audible territory and even so I don’t want an input coupling deciding every time for me that I want to filter out everything below 20-some Hz. So I increased the value so the -3dB was more like 4 or 5Hz. Understand that the higher the capacitance of an electrolytic cap comes the potential for an impact on HF performance. That’s the trade-off with an electrolytic. As well as inherent non-linearities. But with more modern caps this is much less of an issue, and the advantages of the electrolytic cap remain low cost and small packages relative to capacitance and voltage tolerance values. Now, back to the 58/MS16 amp cards, as is usually the case with 80s era Teac stuff, there’s no reason to change any values on the amp card. “They did their homework.” And sure you can go bonkers and bypass the few electrolytics in the signal path or be super obsessive like some and try and replace the electrolytics and cram polypropylene equivalents on there, which is expensive and troublesome because they are relatively giant physically compared to electrolytics, or you can just get modern low ESR ‘lytics, know it’s an improvement, and remember that countless hits and generally great sounding recordings were made using unmodified equipment with general purpose caps. So it’s a recap, and if your cards still have the original 4558 opamps on them then it might be beneficial to replace those. You mentioned 5532. I’m a fan of how they sound. But that’s me. But I can’t recall if the 5532 works in that application, and I haven’t found my notes yet for what I was going to do with my cards as far as opamps…you can’t always replace a JFET part like the 455x, TL072, etc. with a 5532. If you have reference of it being successfully done on the 58/MS16 amp cards then go for it. If not you would just want to measure the DC offset at the output of either side of the opamp before and after the replacement, and if the offset significantly increases after the replacement then the 5532 is not suitable. I also like to use the OPA2134, which is also JFET and a good drop-in for parts like the 455x, TL072, etc. If you’re going to change out opamps then I recommend installing good quality machine-pin sockets. I also recommend bypassing the power rails with a small value cap at each opamp to arrest the possibility of the replacement opamp going into oscillation. Be aware not all the opamps on the card are audio path opamps. Resistors…here again, lots and lots of sought after gear is loaded with carbon film resistors. But if you are interested in exploring this, the only resistors I would replace are the ones that are in amplifier stage feedback loops, particularly those with gain. That’s where the noise stemming from the higher thermal coefficient of the carbon film parts will be more prevalent if at all. Lastly, specific to the 58/MS16 cards, the early 58 cards were fitted with polystyrene caps in the feedback loops instead of Mylar or ceramic. You can see this in the 58 parts list. There is a general opinion that the polystyrene caps sound better. YMMV. They are harder to find, more expensive, and more fragile…they don’t handle heat as well such as when being soldered in. It’s interesting to me that in a lot of earlier Teac stuff I see them everywhere in these applications…late 70s and early 80s, but it didn’t take long before they were value engineered out. My 58 cards all have the Polystyrene caps. So that’s something to consider. But probably unnecessary. It’s super valuable to have a spare card you can modify against which you can do A/B comparison to a stock card. And to do mods in stages. That’s how I would approach it because some of this stuff is not likely to be audibly different, and in some case might make it sound worse. I upgraded a bunch of opamps in a couple modules on my early 80s Tascam prototype console…glad I used sockets. I liked the stock channels better. Now, I didn’t know enough at the time to do some analysis to maybe identify why I didn’t like it and if the cause was only 1 or 2 of the upgraded stages…I’d like to experiment some more someday and I have done some since then and like the changes, but my point is don’t assume any change is worthwhile or beneficial. Try stuff. Just position yourself so you can compare. Best thing is to have a buddy switch back and forth when you are doing A/B comparisons so it’s a blind test. It’s super easy to fool ourselves into thinking different is better because we *want* our changes to be magical.
Oh! Last but not least, relays. I’d buy a full set of new relays. Signal passes through them. And if you’re not yet experiencing maddening intermittent signal issues that seem to randomly move from card to card, sometimes follow troubleshooting card swaps and other times not, you will at some point. And relays are typically suspect with that set of symptoms.
I can still post specifics about what my punch list is as far as preventative maintenance and possible part replacements…those are the notes I’m compiling.