Tascam 34B vs TEAC A3440

[noisegeek]

I’ve been refurbishing old decks for a few years now. Recently picked up a 34B and a 32 to replace the A3440 and I gotta say there’s no level of tuning I can do to get the Tascams sound to stack up to the old A3440. The old TEAC just sounds sweet and smooth. Everything coming out sounds better than what went in. Even with a spectrum analyzer I can’t get the frequency and transient response as good as the TEAC. Even went as far as upgrading the OP amps and still no improvement. Is there something electronic wizardry going on in the A3440 that’s missing in newer ones? The heads are different than the A3440 but one wouldn’t expect a downgrade on a pro-sumer model 10 years newer. Anyone have experience with the circuitry on these?

 

[sweetbeats]

I can’t say for sure…both machines use the 4558 and 4559 opamp…what I’ve learned from experience is the playback amp plays a key role in the sound of a machine, and what I do see on a somewhat macro level that’s different between the two is on the 34 the playback signal goes through a signal transformer first, presumably to mitigate DC in the signal path, which is not a bad idea, but the quality of the transformer and the circuit design is crucial to how well signal is handled dynamically. And then it goes straight into a 4558. On the A-3440 the first stage of the playback amp is a discrete transistor-based circuit block…no transformer. This may have something to do with the sound difference. In my experience there is nothing to rave about the signal transformers Teac used in their designs. So that’s one suspect area that might account for what you are experiencing. My first multitrack open-reel machine was a 3340S…the entire signal path, except for the headphone amp, is 100% discrete transistor-based…not a chip in sight. It’s one of the best sounding “prosumer” tape machines I’ve ever encountered…really sounded great. So maybe that discrete front-end on the playback side is a good thing on the A-3440.

 

[noisegeek]

Great information sir I appreciate it! And good catch! The swap from the 4558 to 5532 made no measurable difference. Then tried LM4562 still no difference so I think you might have something there with the transformers. It’s just weird how they nailed it back in the 70s but still felt the need to change it up. The 34 has a noticeable lower noise floor so maybe the engineers had to compromise a bit to achieve that?. Looks like I have some diagrams to study. I’ll post back results If/when I find a fix

 

[sweetbeats]

Noise: don’t assume it’s a design issue. The A-3440 is many years older than the 34B…have either been recapped? My hunch is a power supply recap would help in either case, and the A-3440 may be in a worse state. Both units feature audio power supplies with discrete regulation adequate global and local filtration, etc. Sure it’s quite possible the discrete stage on the A-3440 is noisier than a 4558 or 4559, but, for sure, if the A-3440 is running all original caps, that machine is about 50 years old now…well due for a recap considering the type of caps used (vs. like the computer-grade stuff you might find in a bonafide “professional” machine like an Ampex MM or 3M M series machine). And they also used larger local filters in the A-3440 (100uF), which is more expensive and can have benefits, but the larger the capacitance the trade-off is HF performance, and in the case of a filter cap that means decreased performance passing HF noise to ground. The factory parts used likely have higher ESR than the equivalent period parts for the 34B, and certainly compared to anything modern. So that may also contribute to higher noise between the two, but I would start with recapping the power supply. Hm. It also looks like the secondary filter in the A-3440 could benefit from a parallel small value film cap for better HF bypass. They didn’t include that at least on the schematic.

Opamps: it’s such a wormhole. And you have to remember any solid state opamp whether discrete or integrated in a chip is an amplifier…it’s not a generator…it’s not a color box…it, in concert with the necessary surrounding components, either buffers signal, or provides gain to signal. That’s it. Yes different opamps have different noise, distortion, slew, bandwidth and open loop gain characteristics, among other things, but if the input signal is garbage, electronically speaking, they won’t fix that. They buffer or amplify the garbage. And they are highly dependent on their surrounding components. Sure it is possible to hear differences in noise level or sound character depending on the application and especially if you are replacing many opamps in a signal chain, but it is not a silver bullet. And I would never, ever, ever replace an opamp without measuring the changes…checking for changes in DC offset…checking for oscillation…checking how it handles a square wave input…that 5532 might be super unhappy in that application where you used it. But you won’t know it if you don’t check things like I mentioned above. Never ever consider one opamp as a “drop-in” for another just because it’s pin-compatible. Ever. Unless there is evidence-base that supports that is true (in other words somebody tried the same thing in the same device for the same amp stage and measured to verify the new part was happy). And THEN you use your ears and compare A to B to determine if the change is beneficial. It might not be. It’s been about 50/50 in my personal experience. So I think first you need to recap the supply and see if that helps your noise issues, and then from there replace the local filter caps, and then start replacing your coupling caps.

 

[BeatleFred]

Interesting discussion, as I have wondered how the two models compared to each other.

I always liked the A-3440 better, though I dont have any hard evidence to tell anyone why a 34 would be inferior.
I own the 3440, never had a 34, but I would not think they would sound all that much different from each other.

As far as timeframe's go, they are not spread out very far. My older brother bought his A-3440 new in June '79,
(still has it), and I believe they were already available in late '78. Even when the 34 came out in '82, the A-3440 was
not discontinued, which I find interesting- if someone were looking to buy a new 4 track at that time, which of
the two to pick, assuming their $ price was close to each other. The A-3440 ran thru '83- the serial # date code verifies
that, as well as Teac brochures, one of my two 3440's is an '83 model, though despite atleast a solid 5 year run, I dont think
there were any service bulletins or design changes to the 3440. The matching 2A mixer/MB-20 had a few minor changes- the
knobs, switches- colors, etc..

If the A-3440 you had performed so well and sounded so appealing, why did you not just keep it?

ps: The 34 was reviewed in Modern Recording- loved that magazine (Tape Op nowadays not quite the same...) if you scroll
down here in the link, but the mag messed up the photo in the review!

https://www.worldradiohistory.com/Archive-All-Audio/Archive-Modern-Recording/80s/Modern-Recording-1982-08.pdf

 

[sweetbeats]

I never had an A-3440. I referred to a 3340S.

 

[noisegeek]

Thanks sweetbeats! Thats great in-depth stuff there. Honestly I’m just an automotive professional but making music has been a lifelong hobby. So just kinda blending my automotive electronic skills with my hobby haha. I.e. I picked up an MX-80 24trk (now my primary tracking machine) with a dead remote and used my Snap-on Zeus to find all the bad chips, then used the same tool to align and calibrate the machine after it was up and running. These machines and 90s era automotive electronics have a lot in common surprisingly. But audio path circuitry has proven to be quite an art and I need all the education I can get haha. Still wrapping my head around bypassing and filtration. Typically with op-amps, I’ll try a newer “better spec” one, then run tests, if there’s no improvement I end up putting the original back in and go back to the drawing board until I learn more about circuit compatibility. Trying to get million dollar studio sound out of dumpster finds can be as challenging as it is rewarding imo. You’re definitely right about the caps on the A-3440. I recapped the ch1 rec/play pcb a couple years ago after it started dropping audio. Ch1 would be dead on a cold power up. Cranking the output gain would pop it to life until the next power up. Recapping fixed it and if there’s one, I don’t doubt there’s many others in similar condition.

 

[noisegeek]

BeatleFred the A-3440 is going to a comfy place in my living room for home use. The 34B and 32 are to replace it in the studio and for mixdown from the MX-80. Just want the Tascams to be on-par with the A-3440 first. But for sure that old A-3440 is honestly the best and clearest sounding deck I’ve ever heard. Even the MX-80 is on-par at best and that things a beast. Noise floor is the only area in can be beat it seems.

 

[BeatleFred]

Ps- An interesting comparison would be with the 40-4 as well. 40-4, A-3440 and 34.

You may know the 40-4 is the four track version of the 80-8, but it looks like it has alot in common with the A-3440,
perhaps different motor (ac vs DC). I would think they would sound alot alike, but I have not owned a 40-4 to know.

Seems like alot more A-3440's were sold and still out there these days vs 40-4 for whatever reasons.....

 

[noisegeek]

Ps- An interesting comparison would be with the 40-4 as well. 40-4, A-3440 and 34.

You may know the 40-4 is the four track version of the 80-8, but it looks like it has alot in common with the A-3440,
perhaps different motor (ac vs DC). I would think they would sound alot alike, but I have not owned a 40-4 to know.

Seems like alot more A-3440's were sold and still out there these days vs 40-4 for whatever reasons.....

I haven’t been inside a 40-4, but if the design is based on the 80-8, it’ll have input transformers from the play head. I don’t have a schematic but I suspect the 40 and 80 series are brothers and the A-3440 is the cousin. I did bypass the transformers along with supporting mods a 32B (same audio card as 34B). And it did improve the clarity to be more on par with the 3440. I believe the 3440 has a better playback head than the 34B. Likely would explain the wider frequency response provided by the A-3440. The transformers found in the 40,80, and later Tasman series, lower the noise floor but sacrifice some sound transparency and frequency response based on my findings. The 40,80,32B and 34B should sound virtually identical assuming properly aligned and calibrated. A-3440 is in its own league. Clearer sound but DBX unit required if noise is a concern imo.

 

[sweetbeats]

The signal path of the 30 series is indeed very similar to the earlier generation 80-8 and 40-4; the latter does center around the later generation 455x opamp (4559 vs 4558), and I’ll generally say the signal path is a bit “value engineered” on the 30 series compared to the earlier generation machines. The 30 series has later generation head tech, but indeed retains the transformer as first stop in the reproduce circuit (from the sync or repro head). Also note the Tascam 44 has two generations, the early one that is identical to the 40-4 except for cosmetics, and the later generation that was introduced with the 42 and 48; completely different machines than the 40-8/80-8. But the signal path of the A3440 is actually very similar to the 40-4/80-8…4558-based, similar circuit design…no transformer though. None of them are like the 3340, which features an all-discrete signal path. That doesn’t always mean “better” necessarily, but as I’ve worked with more and more tape machines over the decades, my very first open-reel machine, a 3340S, remains one of the better sounding multitrack machines I’ve used, and I think it’s related to the signal electronics. Maybe the A3440 is somewhere in between.

 

[noisegeek]

The signal path of the 30 series is indeed very similar to the earlier generation 80-8 and 40-4; the latter does center around the later generation 455x opamp (4559 vs 4558), and I’ll generally say the signal path is a bit “value engineered” on the 30 series compared to the earlier generation machines. The 30 series has later generation head tech, but indeed retains the transformer as first stop in the reproduce circuit (from the sync or repro head). Also note the Tascam 44 has two generations, the early one that is identical to the 40-4 except for cosmetics, and the later generation that was introduced with the 42 and 48; completely different machines than the 40-8/80-8. But the signal path of the A3440 is actually very similar to the 40-4/80-8…4558-based, similar circuit design…no transformer though. None of them are like the 3340, which features an all-discrete signal path. That doesn’t always mean “better” necessarily, but as I’ve worked with more and more tape machines over the decades, my very first open-reel machine, a 3340S, remains one of the better sounding multitrack machines I’ve used, and I think it’s related to the signal electronics. Maybe the A3440 is somewhere in between.

I think you’re spot on with all that. Always helpful to have someone that knows what they’re talking about on these forums . I firmly believe that “value engineering” later models like the 30 series is definitely a factor. The older units seem packed with more robust components. Like everything, newer isn’t always better I suppose. I just wish I could find a NOS playback head for the A-3440 and put it in the 34B. Everything I pull up for replacement brings up the same part for 34B. The A-3440 uses the same record head as 34B but playback is different. I suspect by the way it performs, the gap is a bit tighter on the original A-3440 head. On the A-3440, I get flat response all the way out to 20khz at a mere 7.5ips. Same speed on 34B, rolls off at 16.5khz.

 

[RRuskin]

Ps- An interesting comparison would be with the 40-4 as well. 40-4, A-3440 and 34.

You may know the 40-4 is the four track version of the 80-8, but it looks like it has alot in common with the A-3440,
perhaps different motor (ac vs DC). I would think they would sound alot alike, but I have not owned a 40-4 to know.

Seems like alot more A-3440's were sold and still out there these days vs 40-4 for whatever reasons.....

When I was involved with TASCAM, I was told that the 40-4 was the same as the 80-8 but with 1/4" heads and guides. Furthermore, the motors were designed to be used on a 1" transport so an 80-8 inspired 1" 16 track would have been possible. Of the machines mentioned above, I'd rank them in the following order - 40-4, A3440, 34. I put the 3440 ahead of the 34 because its transport is much sturdier.

 

[sweetbeats]

I wouldn’t automatically assume you can just drop an A3440 reproduce head on a 34 and get the same frequency response performance out of it…you *may*, but it’s important to understand that record and reproduce amplifier circuits are designed in such a way that they are tuned to the head coils to which they are connected, and there may need to be resonance trimming in the reproduce circuit of the 34/34B to take advantage of the specific differences in the A3440 head that provide for that extended frequency range. Again, maybe, maybe not, but also consider the likelihood it will be impossible to drop a different head in there with upsetting the pre-existing wear geometry present on your 34…to do it right, if you change any component of a headblock tape path you need to have freshly lapped heads and an optical alignment. Otherwise if the wear pattern on the “new” head is substantially different than what’s already there in the tape path, the tape will no longer have the same contact geometry with the record head. Sure you can throw it together, but it’s pretty much impossible to get everything to match up when you bring something into a headblock from outside that headblock’s wear history, and you might end up not gaining anything in frequency response, or even worsening the situation without doing the above. I’d leave it alone. I suspect the reason the HF response is rolling off in the 34 is its time to recap the coupling caps in the audio path. Same for the A3440, because you should be -3dB well past 20K on that machine. Not that 20K isn’t good. Way past my hearing range, that’s for sure.

On the subject of value-engineering, it gets bad as you progress through the late 80s and into the 90s, specifically with the Portastudio machines. The open reel machines continued to get better in many regards in terms of features and transport performance…it was the amplifier circuits that were value engineered. The 58 and MS-16 are kind of the sweet-spot, though I’m not sure about the last generation ATR80. Would still love to get my hands on the service manual for that machine. But there was some innovative stuff they did leading up to the 58 amplifier circuit, which is the same as the MS-16, but then it was value-engineered in the 48 and then further in the ATR60 series. The early Tascam open-reel multitrack machines were not very well built IMO, specifically the first generation of them, the Series 70 machines. And then the first 1” 16 track the 90-16…not nearly as bad of build quality issues as the Series 70, which were particularly not good in the area of the signal electronics assembly, both the structures and the circuit design…not good quality components and SUPER thin low-quality PCB material…not impressed. To my surprise the 85-16 1” 16-track machines are actually by far the most impressive highest quality build of any Tascam 1” 16-track. The audio specs aren’t as good as the later machines, but that’s due to earlier generation heads, and regardless I suspect they might sound the best of all of them…completely different animal in terms of the engineering design throughout that machine than anything else they did in a 1” machine by far…all discrete DC servo signal path with almost no coupling caps at all…not needed because of the thoughtful design…glass fiber PCBs with 0.156” pitch cardedge contacts on all the cards including transport logic…proper thick cast webbed transport plate, reeling motors have relatively large motor shafts, at least twice the diameter of of the MS-16 and ATR60 series…like I said it’s just a completely different animal. The signal electronics are the same concept as the 58 and MS-16, but they moved away from discrete signal path to opamp-based with the 58 and MS-16 using the 4558. But I digress.

 

[Monkey Allen]

I just bought a TEAC A-3440. Should arrive in a week.

 

[BeatleFred]

I just bought a TEAC A-3440. Should arrive in a week.

Congrats, its such a nice-looking unit. I love all the things that went into it, even just things like the color-contrast between the top and bottom half of the front display, and the wood-style side panels. A little more modern looking than the previous 3340.

Hopefully, the seller kept it decently maintained. What mixing board are you planning to use with it? The matching unit is the 2A with MB-20 meter bridge, though you can really use many other mixers, whether Teac/Tascam or others designed for four track recording/overdubbing. The matching noise reduction unit was the RX-9, though some folks might find it unnecessary. I heard the Tascam DX-4D is better, and can work with a 3440 if a soldering mod is done to make it compatible.

 

[Monkey Allen]

Congrats, its such a nice-looking unit. I love all the things that went into it, even just things like the color-contrast between the top and bottom half of the front display, and the wood-style side panels. A little more modern looking than the previous 3340.

Hopefully, the seller kept it decently maintained. What mixing board are you planning to use with it? The matching unit is the 2A with MB-20 meter bridge, though you can really use many other mixers, whether Teac/Tascam or others designed for four track recording/overdubbing. The matching noise reduction unit was the RX-9, though some folks might find it unnecessary. I heard the Tascam DX-4D is better, and can work with a 3440 if a soldering mod is done to make it compatible.

Good question about what mixer. Finding a 2A or something like that in Australia could take many, many years. And prices would be insane. Something like a Tascam M208 are over a thousand AUD. And they just don't come up much. I was wondering if a modern mixer like a Yamaha MG16 (or 12 or whatever) would work. I really don't know. If you have any suggestions let me know, thanks.

 

[RRuskin]

Good question about what mixer. Finding a 2A or something like that in Australia could take many, many years. And prices would be insane. Something like a Tascam M208 are over a thousand AUD. And they just don't come up much. I was wondering if a modern mixer like a Yamaha MG16 (or 12 or whatever) would work. I really don't know. If you have any suggestions let me know, thanks.

Of all the TEAC/TASCAM mixers, the Models 2 and 2A are the worst. They are extremely limited in function and are RF magnets.

 

[BeatleFred]

@MA: I overlooked your location, so I do understand about availability & prices when it comes to these vintage units.

Sure, you can use modern mixers- they would be convenient to hook-up your equipment to. You can download just about any Operating manual, whether its for a Yamaha, Mackie and so forth and look at the connections diagram that are typically shown within them to see what-gets-connected-where, etc..

However, modern mixers are usually more-suited to recording onto laptops/music software and only just a stereo mix. A vintage mixer, like the 2A/MB-20 was designed to record onto a 4 track multitrack, because that was the technology of its time. Thus, its has 4 channel assign buttons. Everything is arranged very neatly, and easy on the eye, the 2A/MB-20 has a sloped panel with 4 large upright VU meters, and those channel assign buttons are also seen on other larger Teac/Tascam mixers, such as the M-30, which does not have the sloped shape, but thats not really a big deal, other than basic ergonomics. The M-208 you mentioned was made later after the 2A was discontinued, and considered to be a 'higher' model, the one above the M-106 in the product line, but neither one has Upright meters, so unless someone likes that, I usually have to prop a book under my M-106 to be able to see the meters as easily as they are on the 2A/MB-20.

Just to clarify, I simply mentioned the 2A/MB-20- being that you acquired an A-3440, to point out, its the actual matching mixer. Unless you intend to get into
very complicated recording schemes, the 2A would be a very decent mixer to start out with. Its not like you are locked into using it forever. As I said, the 4 channel assigns
make things very straightforward, and the MB-20 manual shows how to make the connections between the mixer, meter, and recorder.

I have not had any issues with my 2A that have soured me as it apparently did to RR. I have no interest in any ongoing negative debate with someone of an opposing opinion, so I leave it at that, and you can decide for yourself when you're ready on which mixer would work best for you & your A-3440.

 

[Monkey Allen]

Of all the TEAC/TASCAM mixers, the Models 2 and 2A are the worst. They are extremely limited in function and are RF magnets.

Fair enough. I really don't know much about them...or about mixers in general. I've only done DAW recording really. Apart from some 424 cassette stuff here and there.