Tascam 388 DBX Card Calibration

ninnghizidda

New member
I am in the process of completely overhauling my 388 and have finally discovered the very low and very distorted levels on tracks 3 and 4 were due to the DBX cards. Swapped them around, and now tracks 7 and 8 are bad, regardless of engaging the DBX or not. Everything Deox'd and still consistently follows the card. The contacts still look plenty springy. I tried to adjust the trim pots for those channels, and it didn't really make any difference, so I put them back where they were. I would like to try to calibrate the DBX cards properly now—if for no other reason than a true complete overhaul would not be complete without this calibration—to signal trace as I calibrate should things not look how they ought, and hopefully repair my card instead of chucking it for a new one from ebay. My searches on this forum, Google and Yahoo, calls to Tascam, New Jersey Factory Service, and DBX themselves has all led nowhere.

So, does anybody have any information that might be useful? Maybe somebody here has attempted/succeeded this already?
 
I’m assuming in all your research you looked at the manual first? In case you missed it, it’s section 1-6 of the service manual.

But slow your roll...the dbx circuitry does not typically need adjustment, and unless you fully understand how to do it should not be messed with.

Let’s step back a little...the fact the problem occurs with and without dbx engaged narrows the possible problem cause(s)...it indicates the dbx companding circuitry is not likely at fault. I would take the time to hone in on the cause before tearing into the whole card.
 
If my issue follows one of the DBX cards, whether engaged or not, that suggests to me the issue is somewhere on the DBX card.
 
Correct. That’s a reasonable assumption. But that doesn’t mean you have to dive into the whole dbx circuitry. If the problem exists when the dbx is bypassed, there is minimal circuitry on that board that is involved vs when the dbx is engaged. I’m suggesting you focus on the minimal involved circuitry rather than messing with stuff that isn’t even involved when your problem manifests. Does that make sense?
 
I'm assuming it's an LC4966 as I've seen somewhat similar issues with switching IC's in the past, but the prudent and professional approach would be to verify that by signal tracing until things look wrong instead of shotgunning some fresh parts at it and hoping for the best. Is there really a better time to do that signal tracing than when the cards are being calibrated, or at the very least, during the verification of their 30+ year old calibration? It very well may pass the ear test, but if somebody asks me if the machine has been calibrated I can say "Yes," instead of "Should be, it hasn't been touched in 30 years."

That's not the issue I am having, however. What I would like to know is peoples experiences with calibrating the DBX boards. You need a Teac FG-1 to create the manuals seemingly-arbitrary staircase wave signal. Unfortunately, Tascam, New Jersey Factory Service, DBX, and Google (you might be better at Google than me, it's not impossible and I am human) all have never heard of such a unit, and nobody knows exactly what it does. Does it indeed generate functions? Would it still generate functions on its own without an oscillator attached to the front as shown in the manual? Maybe it's actually more of a function synthesizer, taking a sine and creating pulses, triangles, staircases, etc? If I knew exactly what it was the FG-1 was doing, in contrast to the generic descriptions of what it is doing as shown in the manual (which also has a discouraging note about calibrating the DBX cards, if I remember correctly), I have no doubts that this would be a rather straightforward calibration, assuming I sort out the (switch IC?) issue mentioned above. Again, I'm looking for peoples experiences calibrating the cards.
 
Okay. Thanks for that reply. I think I understand you better now.

I still don’t entirely agree with your ideology to calibrate the dbx encode/decode circuitry to troubleshoot your signal issue simply because I’ve learned from my own experience it’s better to start with the most basic reasonable cause of the problem and drill down from there rather than jump to the more complex and less likely cause to see if that captures the problem. But you also seem to be very intent on calibrating the dbx circuitry regardless of your signal problem, and want answers to your questions about the dbx calibration. Just proceed with caution because the dbx calibration is a much more sensitive and detail intensive procedure than other “calibration” procedures like record and playback levels and frequency response, and on the off-chance there *is* a dbx calibration issue (which, again, is uncommon), any tapes recorded with the off-spec dbx circuitry may no longer reproduce properly once the dbx circuitry is calibrated...that’s often one of the key drivers behind a calibration procedure; ensuring recordings made on your machine reproduce properly on another, but if your machine is the only one likely being used to record and reproduce then it’s more of an exercise than a necessity. And the procedure requires low-distortion precision tone generation equipment and, if you are not using the right spec of test gear, out-of-spec test results may simply appear so because the test equipment isn’t up-to-snuff and pretty soon you are tweaking things that either don’t need tweaked or tweaking blind because your test gear can’t produce reliable results. I like things being “just so” or “correct” though too, so I understand your impetus. FWIW I have never calibrated dbx noise reduction circuitry...I’ve only discussed the topic with more experienced technicians, and as a result decided to stay away from messing with it.

The FG-1 appears to be a function generator...a sine wave generator in this case and low distortion at that...less than 0.02% based on Figures 1-6-2 through 1-6-4 in the Service Manual. But then steps 1-6-4E and 1-6-5E call for a pulse generator to produce those tone bursts with control over the number of cycles on and off. So any piece of equipment or combination of devices that can produce low distortion sine waves over the audio band and generation cycle specific pulses will do.

U104, that LC4966, is a good catch...if I was going to shotgun anything it would be that part because I have worked on or with a fair number of 388s that needed that part replaced in the monitor section. The problem you’re describing sounds similar to bad signal switching relays. We don’t have mechanical relays here, but we DO have the logic equivalent in the form of that LC4966. If it was me, I’d probably jump the gun and replace it...could be a quick solution, or a good preventative measure since there something of a trend with that part in the 388 (and other period Tascam gear) going bad.
 
I'm definitely not above shotgunning parts. This 388, though, was bought as a beat up parts unit on ebay a few years back not working. It had bad line and mic in's due to some bad card connections, cold solder joints, and/or shorted op amps, the solenoid was rattling around inside and needed a new roll pin, new belt, of course, shorted transistor on the power supply board, all jacks a nice green oxide colour, a couple bad capacitors, I could go on. At this point, the only thing that seems wrong with it is the head could use a relapping, and this DBX card distortion issue. Given so many things were wrong with this particular unit and have already had to do so much to it, it doesn't really seem that big of a deal to me to go ahead and at least verify the cal of the cards. The only tapes I own are a blank roll of LPG35 and an MRL 21T204, so no issues there. My thinking was that the calibration calls for a signal to be injected, but I couldn't trace the signal easily unless I had jumpers (which I don't) or the card is powered up otherwise, as it would be during a calibration, and from the sounds of the procedure as outlined in the manual, as long as everything measures okay I shouldn't have to adjust anything anyway. Maybe this is where my lack of experience with logic circuits comes in? Either way, yes, I am set on calibrating the cards, if for no other reason than as a learning exercise.

I have access to the equivalent or better test equipment as called for in the manual, so no issues there, either. Except that damned FG-1. I interpret the images as an external oscillator and attenuator feeding the FG-1, but it doesn't seem to matter either way based on the instructions, unless the FG-1's frequency depends on an external oscillator. It doesn't give much detail about what sort of staircase signal or triangle wave other than I need to apply them. It really seems to me based on the pictures that the FG-1 is taking a sine wave and converting and splitting it out as a staircase and triangle. Does it matter as long as the shapes are fine? It looks like they are the same frequency, if that even matters? Maybe it can be assumed they're 1k since that's mostly what you're using anyway? Or maybe 100hz? I'm beginning to feel like I might need to try it at a couple different settings and figure out which one falls best within (or beyond) the factory specs for running it with DBX? I wouldn't expect much adjusting, whether it needs it or not, so maybe I ought to experiment with my known good card and try to replicate the good DBX output signals by adjusting the input frequencies?

That's good information to know about those IC's, I was unaware of them being potential trouble spots. I once worked on an Oberheim 2 Voice that did make sound, but had all kinds of wacky problems due to what I've later read were probably bad CMOS chips. I've read that older CMOS chips have a tendency to degrade over time? Any insight into this, and could these LC4966's suffer the same problem?
 
I don’t know what types of functions the FG-1 can produce (i.e. waveforms), but if you read the procedure in the manual, you only need a low-distortion sine wave generator or oscillator. If you can find a generator that meets the distortion spec, then fine. Otherwise you’ll need an oscillator. The 0.02% distortion spec is what is called for. The attenuator must be there specific to the FG-1 (i.e. maybe the FG-1 nominal output level is much greater than -10dBV). I don’t know where you are seeing the need to generate any other waveform than sine? Read the procedure. Maybe you are referring to the waveforms you should see on the scope? And don’t forget you need some way to generate the low-distortion sine wave for 8 cycles on and 128 off. That’s necessary to verify the response of the encoding/decoding circuits. Looks like the procedure is done with the card in the cardbay and interfacing with test points. And you’ll want to make sure you have a low-impedance RMS measuring AC voltmeter accurate across audio bandwidth. So, low-distortion sine wave generator/oscillator, pulse generator, oscilloscope and AC voltmeter.
 
I agree with Sweetbeat's "K.I.S.Sir philosophy and would be very averse to touching those pre sets IF for no other reason than they are low grade open tracks and get dusty. Once moved they will NEVER go back to the original set value and will Jodrell all over the shop.

The test signal generation problem is an interesting one. It would take SOME work but you should be able to generate any form of tone burst in a DAW? I once spent a wet weekend building some multiple frequency-various level tones with a view to speeding up amp tests. Unfortunately the Big C got me* before I could develop the idea! Samplitude seem excellent for that but Audacity might work as well.

If you can sort the distortion I think a test over a range of levels should give a good indication of the state of the machine's calibration? Slowly catchee monkey?

One thing is for sure? ANYTHING you buy off Evil BAY ain't gonna be aligned to ***t!

I REALLY don't know why you guys do it!

*Oh! I am not dead! Just had to jack the job for a bit. (and they paid me ALL the time. Did not HAVE to!)

Dave.
 
And if done on the computer the resolution would have to be at 192kHz/24-bit, and the distortion on the outputs meet the 0.02% distortion spec.
 
And if done on the computer the resolution would have to be at 192kHz/24-bit, and the distortion on the outputs meet the 0.02% distortion spec.

Hmm? 0.02% THD? Easy, nay bother there.
24bit word length? Would not dream of using anything else!
192kHz sample rate WTF for? 48kHz give you a 22-23kHz bandwidth, good enough for cassette surely?
Or am I missing something?
 
I don’t know what types of functions the FG-1 can produce (i.e. waveforms), but if you read the procedure in the manual, you only need a low-distortion sine wave generator or oscillator. If you can find a generator that meets the distortion spec, then fine. Otherwise you’ll need an oscillator. The 0.02% distortion spec is what is called for. The attenuator must be there specific to the FG-1 (i.e. maybe the FG-1 nominal output level is much greater than -10dBV). I don’t know where you are seeing the need to generate any other waveform than sine? Read the procedure. Maybe you are referring to the waveforms you should see on the scope? And don’t forget you need some way to generate the low-distortion sine wave for 8 cycles on and 128 off. That’s necessary to verify the response of the encoding/decoding circuits. Looks like the procedure is done with the card in the cardbay and interfacing with test points. And you’ll want to make sure you have a low-impedance RMS measuring AC voltmeter accurate across audio bandwidth. So, low-distortion sine wave generator/oscillator, pulse generator, oscilloscope and AC voltmeter.

It very strongly seems to be nothing like that to me. The manual consistently shows different pieces of equipment connected by simple lines, suggesting that even when using the low distortion signal, it's still running through the FG-1 first. In my experiences, if you want that low distortion of a signal you would want the straightest path to where you're sending it. Funneling it through the FG-1 does what other than potentially adding distortion to the signal? When you're not using the low distortion signal and are using the functions on the FG-1, that low distortion signal is still connected to the input of the FG-1. Does it matter? I can only make out "Burst" on the FG-1's functions on the PDF, but it clearly shows in Fig. 1-6-4 VCA Symmetry Adj. the FG-1 sending a staircase wave to TP-01 on the DBX card, which is then hooked up the scope, as well as sending a triangle wave to the horizontal input of the scope. The staircase and triangle are probably linked somehow. Decoder calibration step B.1 confirms this. Step B.2 says to short J110-1 and J110-2, which are not test points but pins on the card connectors. I can't make out the text on the PDF but it definitely looks to me that step B.2 and the L and H signals are being connected directly to the pins of a removed card, as well as the + and - 12V. It doesn't look like "J" or "TP" for the DBX card to scope out, but again, I can't really make out the letters.

My test equipment is quality, no worries there. I can do the triangles and pulses no problem. I actually did just notice it looks like the width of the triangle is measured on Fig. 1-6-4 but I can't make it out really. I wish there was a high quality scan of the service manual. That being said, I really haven't encountered anything that will do a staircase wave aside from an arbitrary waveform generator, and even then most of the staircases look nothing like the one depicted in the manual. And again, it really seems to be that the triangle and staircase are somehow related on the FG-1.

I agree with Sweetbeat's "K.I.S.Sir philosophy and would be very averse to touching those pre sets IF for no other reason than they are low grade open tracks and get dusty. Once moved they will NEVER go back to the original set value and will Jodrell all over the shop.

The test signal generation problem is an interesting one. It would take SOME work but you should be able to generate any form of tone burst in a DAW? I once spent a wet weekend building some multiple frequency-various level tones with a view to speeding up amp tests. Unfortunately the Big C got me* before I could develop the idea! Samplitude seem excellent for that but Audacity might work as well.

If you can sort the distortion I think a test over a range of levels should give a good indication of the state of the machine's calibration? Slowly catchee monkey?

It's definitely Jodrelled all over the shop, then, because although I marked where they were, they did get cleaned and turned. Only the ones on the distorted channels, though. The channels that were fine have factory DBX calibration, or wherever they were set when it was purchased by me. I can't imagine it being a make or break issue, though, especially if I plan on doing a full calibration later anyway.

That's what I was thinking, as well. The rest of the machine is calibrated. If I hook the DBX up to calibrate it and everything looks how it should, I will confirm the DBX is fine, or I will find the trouble spot and hopefully correct it.

I was never even planning on even using the DBX. I was always under the impression that most people thought it didn't sound very good on the 388 and it wasn't worth using. Then I started to read about people saying how great it works and how wonderful the 388 sounds with it. I know that some people may just not like what DBX does to the sound overall, whether it works or not, and I know some people like it as an effect. These are old pieces now, and who knows the history of the units people are making these claims about. I would like to ensure my machine is mechanically and electrically within factory specifications and come to my own conclusion.
 
Is that what I think it means...the Big C...? :(
So...are you doing OK...?
Don't mean to pry...just hate to hear that kind of stuff.

Yes, pretty good (but...) It was bladder cancer. Peed blood at work and like a twat ignored it. Week or so later 'in retention' with clots. Very successful op (but they take out thru the Ole Feller and that smarts when you wake up!) In five years and many 'probings' no return.

The but: Last but one scan showed a small tumour on left kidney. biopsy prove it malignant but of a very slow growing form. I wait the results of my scans last month.

All this might seem a bit dire but honestly at 72 I think I am pretty lucky to function as well as I do. Bit O preaching?

DO NOT ignore 'man tap' issues. The treatments are not as bad as you think. I also have ARMD. Left eye buggered (still legal to drive but not at night) having injections in the right. AGAIN do NOT ignore sight problem especially past 50. Plenty of 'coloured veggy, toms, beets, peppers and a regular eye test. Make sure both eyes 'match' from time to time.

Dave.
 
It very strongly seems to be nothing like that to me. The manual consistently shows different pieces of equipment connected by simple lines, suggesting that even when using the low distortion signal, it's still running through the FG-1 first. In my experiences, if you want that low distortion of a signal you would want the straightest path to where you're sending it. Funneling it through the FG-1 does what other than potentially adding distortion to the signal? When you're not using the low distortion signal and are using the functions on the FG-1, that low distortion signal is still connected to the input of the FG-1. Does it matter? I can only make out "Burst" on the FG-1's functions on the PDF, but it clearly shows in Fig. 1-6-4 VCA Symmetry Adj. the FG-1 sending a staircase wave to TP-01 on the DBX card, which is then hooked up the scope, as well as sending a triangle wave to the horizontal input of the scope. The staircase and triangle are probably linked somehow. Decoder calibration step B.1 confirms this. Step B.2 says to short J110-1 and J110-2, which are not test points but pins on the card connectors. I can't make out the text on the PDF but it definitely looks to me that step B.2 and the L and H signals are being connected directly to the pins of a removed card, as well as the + and - 12V. It doesn't look like "J" or "TP" for the DBX card to scope out, but again, I can't really make out the letters.

My test equipment is quality, no worries there. I can do the triangles and pulses no problem. I actually did just notice it looks like the width of the triangle is measured on Fig. 1-6-4 but I can't make it out really. I wish there was a high quality scan of the service manual. That being said, I really haven't encountered anything that will do a staircase wave aside from an arbitrary waveform generator, and even then most of the staircases look nothing like the one depicted in the manual. And again, it really seems to be that the triangle and staircase are somehow related on the FG-1.



It's definitely Jodrelled all over the shop, then, because although I marked where they were, they did get cleaned and turned. Only the ones on the distorted channels, though. The channels that were fine have factory DBX calibration, or wherever they were set when it was purchased by me. I can't imagine it being a make or break issue, though, especially if I plan on doing a full calibration later anyway.

That's what I was thinking, as well. The rest of the machine is calibrated. If I hook the DBX up to calibrate it and everything looks how it should, I will confirm the DBX is fine, or I will find the trouble spot and hopefully correct it.

I was never even planning on even using the DBX. I was always under the impression that most people thought it didn't sound very good on the 388 and it wasn't worth using. Then I started to read about people saying how great it works and how wonderful the 388 sounds with it. I know that some people may just not like what DBX does to the sound overall, whether it works or not, and I know some people like it as an effect. These are old pieces now, and who knows the history of the units people are making these claims about. I would like to ensure my machine is mechanically and electrically within factory specifications and come to my own conclusion.

I’m not sure what it is people are going for by not using the dbx on a 388, but each to his/her own.

I see now what you are saying about the different waveforms. My bad for skimming instructions. I *think* the FG-1 must just modify a sine wave right? So forget the oscillator and the attenuator...but I don’t know how to generate a stair-step wave either.

This is all going to a much further “nth” degree than I’ve even really considered. I wish I could be of more help.
 
You did help by forcing me to more carefully read and digest the schematic and manual, so thanks for that! I did post here hoping that maybe you had some experiences with it. Your 388 Story thread is the best (maybe only?) 388 restoration guide out there.

I found this link to a staircase wave generator which is literally the only reference I've seen online to a wave that looks similar to the one in the 388 manual. I am going to build that and hope for the best. Hopefully I can trigger it with the same triangle wave I need to send to the scope to align it with the staircase, assuming that's what's going on with the FG-1. I'm terrible at taking notes, but I'll try to document the process once everything is together, if anybody cares. For fun, assuming this all works out, I'd like to compare my "factory" board with the one I've messed with and see if I can hear a difference.
 
Yes, pretty good (but...) It was bladder cancer. Peed blood at work and like a twat ignored it. Week or so later 'in retention' with clots. Very successful op (but they take out thru the Ole Feller and that smarts when you wake up!) In five years and many 'probings' no return.

The but: Last but one scan showed a small tumour on left kidney. biopsy prove it malignant but of a very slow growing form. I wait the results of my scans last month.

All this might seem a bit dire but honestly at 72 I think I am pretty lucky to function as well as I do. Bit O preaching?

DO NOT ignore 'man tap' issues. The treatments are not as bad as you think. I also have ARMD. Left eye buggered (still legal to drive but not at night) having injections in the right. AGAIN do NOT ignore sight problem especially past 50. Plenty of 'coloured veggy, toms, beets, peppers and a regular eye test. Make sure both eyes 'match' from time to time.

Dave.

Well...best of luck.
Even at 72, I'm sure you don't want to give in too soon.
 
U104, that LC4966, is a good catch...if I was going to shotgun anything it would be that part because I have worked on or with a fair number of 388s that needed that part replaced in the monitor section. The problem you’re describing sounds similar to bad signal switching relays. We don’t have mechanical relays here, but we DO have the logic equivalent in the form of that LC4966. If it was me, I’d probably jump the gun and replace it...could be a quick solution, or a good preventative measure since there something of a trend with that part in the 388 (and other period Tascam gear) going bad.
If I may, what symptoms does a LC4966 display when they are failing?
I am having trouble with some distortion in 312B monitor section - when I switch monitor sources, I see a perfect sine wave test signal for 2 seconds and then followed by clipping of the negative side of the waveform -
replaced all the biasing resistors around the transisitors feeding the LC4966 - i never suspected a switching IC would have impact on the audio quality and could cause distortion

Any ideas?
 
The dBx cards on most tape deck should only be adjusted for Nominal encode or decode levels. The rest of the pots will just send you into trouble with a lot more specialized equipment needed. At Teac we only saw the special generator one time and we had to send it back. To build one does not work as it is lacking as one guy who tried found out but he was connected to a EE so he was able to get one to work. dBx usually does not fail but solder joints can and then there are the guys who tamper with the controls- like head azimuth by ear this never turns out good, The LC4966 is like a CD4066 only better part. I think they have higher voltage capability. You can trouble shoot those by trying a new one in a IC socket which should be put in these old boards anyway. There are some tape deck I had found out that the signal goes through the dBx board all the time and taking the connectors out of the dBx even if off cuts the signal- I think I found that on a X1000R. Dolby chips in cassette decks also use Dolby IC for gain even when no Dolby is engaged.
 
Back
Top