I've been adjusting tape speed and azimuth on Aiwa HS-J9/09/600/36/360/10 by ear. I understand this is inferior to instrumented adjustment. My method was to adjust speed on a commercially recorded cassette so the music sounded "right". Adjusting the head comprised looking for the "brightest" tone while also trying to get the same volume from both channels. So prerecorded cassettes played in stereo, using earphones, would sound centered. Presumably commercial cassettes are properly balanced, and typically the lead vocalist is centered. I realized Audacity could be used to check channel balance. Channel balance should equate to head alignment. I obtained a frequency test tape and installed an application with frequency display, on a pc. The application I used has a pro version with specialized frequency analysis tools, but I used the basic free version. I tested several of the players I've repaired and adjusted manually. I plugged the players' earphone jack into the computer's microphone jack and played the test tape. I boosted the volume to broaden the channel tracks for Audacity to see if the volume of the two channels was equal. They were. Turning to the tape speed and the frequency analyzer, all of them were bang on the 1000, 2000 and 3000 frequencies. (The application wouldn't let me center the cursor on the maximum peaks between 996 and 1004 hz.) I tested what was going on by pressing on a pulley to slow the tape. The peaks displaced to lower frequencies as would be expected. So. Is adjusting head positioning and tape speed by ear as effectively good as using fancy instruments? Is using the software as good as using fancy instruments? In fact I've found that just positioning the head in line with whatever it attaches to is pretty accurate. It seems odd these players don't need instrumented tape speed adjustment after an initial manual adjustment. Why are these players moving the tape at the right speed whatever their history? I find it hard to believe my perception music playback speed is that accurate. Or does the tape drive electronics in these players have a speed regulation function?
As far as speed adjustment goes, if you use a song you know well, you can get within +/-10Hz just by ear. But there's another variable that has to be taken into consideration: the pre-recorded tape itself. How do we know the speed on that is correct ? After all, what we're doing is tune our device to that tape, whether that's a good reference or not. In my opinion, the best scenario for recording such a test tape is to use a quartz lock deck, since those don't have speed adjustment and the quartz itself is very stable and has minimal long-term drift. Of course, at the end of day, the by ear adjustment on a commercial tape can be good enough, but not perfect. As far as stability/drift, once the speed is set correctly wouldn't expect it to drift significantly since all these motor speed regulators use precision references and feedback. Azimuth indeed can be adjusted accurately just by making head parallel to the head bridge. That's what they did at the factory with all those non-adjustable units and with some adjustable ones too. Also the capstan plane has to be perfectly perpedincular to the head bridge plane. Which is not the case in center gear DDs for example, hence why on most of these the factory adjustment is wrong. Then the method of using a pre-recorded tape also can render good results, but has the same potential problem as in the speed case. Yes and no, depends on the frequency this balance is measured at. You can have perfect channel balance at 400Hz and singificant imbalance at 10kHz. When it comes to azimuth, the higher the frequency, the more precision in the adjustment. The minimum I would suggest is 6.3kHz for such adjustment.
It is possible prerecorded tapes could have been recorded at the wrong speed, but there was a huge brand image disincentive to allow that. I've never heard of it happening. And presumably the cassette manufacturers would have sophisticated equipment and checks to prevent it. More likely would be speed problems caused by degraded cassettes. Which all of us have run into. Recorded music might not be centered, and a vocalist track may not be center stage. I've been using a Graceland cassette. A factor I considered is whether the test tape was recorded at the right speed. There are websites where you can generate frequencies you can choose. Besides buying a frequency test tape, I recorded a 1000hz tone onto blank cassette and compared it to the test cassette. They were exactly the same frequency. While a player running at the wrong speed will replay a frequency recorded on it at the right frequency, the professional test tape won't match. And the frequency deviation of the cassette recorded on the defective player will show up when the cassettes are compared on a known-good player. It was the consistency in these results that made me think these players have a speed regulation system. Which makes instrumented speed testing less important. Seems to me heads can be aligned with the tape in 3 axis. But the few player models I've worked on have adjustment for just plane. Which might be called "skew" but because of the mounts also has a degree of lateral positioning. This limitation of adjustability seems to be adequate relative to the precision of the manufacturing. It reminds me of when plastic body SLR cameras came out. Conventional wisdom was that bodies had to be metal to be sufficiently precise. Good point that speed accuracy is better adjusted at higher frequencies than lower frequencies. Here's a question. Is 500hz compared to 505hz, more or less audible than 5000hz compared to 5050hz? Or, by percentage, is a .5% difference of 500hz compared to 505hz, more or less audible than 5000hz compared to 5250hz? I suppose I could check this by generating those frequencies. I was thinking that for that reason, I should add, say, a 7000hz tone to my own test tape. The question then would be the accuracy and sensitivity of the test instrument. The frequency tests I considered claimed to sample at 44,100hz. Good point about channel balance possibly varying by frequency. It's possible test tapes are louder on one channel than the other. The only way I can think of to test this is by comparing it on multiple players, or on a known-good player. Typically people have better overall hearing in one ear than the other, and that difference can also be a factor of frequency. Unbalanced hearing could be checked by reversing the earphones on our ears. One could also make an adapter that crosses over the channels. Hearing can be professionally tested, of course. The player's amplifier may have stronger output on one side than the other. I was thinking that once the test tape is verified for equal channel strength, the earphone wires can be reversed to test this, with or without using an instrument display. And if all that weren't enough, earphones may have unbalanced output. Obviously wow&flutter is best checked on professional equipment. But an amplified pure frequency should expose w&f as the cassette plays. Another question is whether the tape speed changes due to varying torque in the takeup reels as the amount of tape wound on the reels changes. This would be why healthy pinch roller and capstan surfaces are important. Do any high end players change applied takeup torque as the amount of wound tape changes? As players evolved, was that found to be unnecessary, or should tape speed be checked near both ends of a tape? I notice the most professional people who restore players, do instrumented testing of takeup torque. Even if a variation is detected but is not obvious due to eating tape, what can be done about takeup torque anyway? I neglected to mention that speed must be checked in both tape directions.
Regarding pre-recorded cassettes speed, yes it's indeed unlikely to be off by a significant amount. It's just that (unlike a test tape) we don't know what that amount is (can be 0.1%, 0.5%, 1%). It's important to be mentioned "wrong speed" can be 3005Hz, 3030Hz or 3085Hz. All 3 are technically wrong, but the first 2 will fall within specification for most players. Third one however is something that will sound obviously fast, even to someone with "untrained" ears. A 0.5% difference at 500Hz vs 5kHz won't be perceived that differently because the ear perceives frequency in chunks rather than individual frequencies. Those chunks become larger the higher the frequency. Also in music we don't listen to a single tone at a fixed frequency, rather what we hear is a complex waveform with many harmonics. Hence we hear how these frequencies mix toghether rather than pure tones. That's why things that may be audible with test tones may not be audible with music. For example, a very narrow dip in frequency response (at higher freq) of a headphone will be audible with a frequency sweep, but won't be audible with music. A wide dip though, will be audible with music also. Output can (and usually is) indeed not perfectly balanced at the headphones out, that's because a stereo volume potentiometer has its own imbalance. However when doing azimuth adjustment, point is to have both outputs at max level, not necessarily equal. Because azimuth only affects the higher frequencies, adjusting for the brightest sound should render a good result even with uneven hearing. Takeup torque will affect both speed and W&F, but this effect should fall within a certain tolerance. The torque iteself also has a tolerance, which can be quite large like between 25 and 43 g x cm. While checking torque is important, in most cases being grossly out of spec means a serious problem like frozen clutch rather than some adjustment needed. Very few walkmans have the takeup torque adjustable. As far as changing the torque based on position of tape, that can only be done electronically on decks that have a separate reel motor and no clutch. And yes speed/W&F tests should be done in both directions for autoreverse players. It's not uncommon for only one direction to have problems. Another thing I want to mention is in many walkmans there can be a difference in speed between FWD and REV. That doesn't necessarily mean there's a problem if difference is small.
Because what you say makes perfect sense, I'm going to go back and redo the testing using 7000hz. instead of 1000. I was thinking about why high frequencies are more sensitive to head position. I know recording onto tape involves very briefly applying varying charges to the head "gap"(s), creating a magnetic field which aligns the fields of the magnetic particles in the tape coating. The "gap" is a certain size. Presumably a signal for a low frequency is applied for more time than a signal for a low frequency. Does this mean a low frequency signal has a larger "footprint" than a low frequency signal? The footprint of the low frequency signal would affect an area of tape wider than the signal for a high frequency. And in turn a misaligned head would be more likely to detect magnetism for a low frequency signal. (Come to think of it, examining the edges of data tracks when trying to restore a fully reformatted hard drive would depend on the magnetized data signals spilling beyond the actual head width. And why the most effective means of erasing a hard drive (besides physically destroying it) comprises multiple iterations of writing 0's and 1's while repositioning the heads laterally.) My earlier photos for frequency do indeed show a whole spectrum of frequencies besides the target 1000hz frequency. Indeed, one can see the test tape I bought has two other test frequencies besides 1000hz. Yet, as you said, we hear the mix as a single tone. I'll make a tape with just 7000hz. The test tape also seems to have some frequencies with very low presence. Of course that was a snapshot and the amplitudes of the lines were continually moving.
Yes that is correct, higher frequencies have a smaller "footprint", that's the reason head azimuth and tape path alignment in general affect the higher frequencies way more. It's also why the head gap has to be fine in order to be able to "read" higher frequencies. This is even visible with naked eye: a head with 15kHz response head has the gap visible, an 18kHz signigicanly less so, while on a 20kHz one barely so. Yes, it's also why on higher capacity SMR drives data would become unrecoverable after a format. SMR drives write the data in chunks (many adjacent sectors at once), as writing one sector at a time would compromise the adjacent ones due to spilling of magnetic field on adjacent sectors. It's also why on tape we have guard bands between the tracks, to minimize crosstalk.
I redid the testing using 7000hz, and indeed there was room for more precise speed adjustment. I can't get the graph to register exactly 7000hz only because the granularity of the software jumps by ~15 hz at those frequencies . I figure 6996+hz is good enough. That also undermines my speculation these players have a built-in speed regulator.
3khz should be good enough for speed and w&f. Azimuth requires better test tapes at 6.3khz or 10khz. One thing I found about my J9 (maybe just this unit) is that the tape speed may drift up ~30hz or 1% on both directions, when the player is fully assembled. I guess it’s some heat stability issue when back cover is on. A little bit odd since J9 has a dedicated regulator circuit for each direction.
Yes, 3kHz is good enough for speed ajustments. At a higher frequency, despite there is more resolution available there's also more variation. In regard to speed regulators, all cassette machines have speed regulators, but they're not all equal. 1. The most basic one present in very early portable recorders and decks consists only of a very big flywheel and very elastic belt, while motor speed relies on AC mains frequency. It's good enough, but big, heavy and speed cannot be adjusted being tied to the mains frequency. 2. Most tape machines that use DC motors or 3-phase brushless motors use a regulator that uses the motor current as input for speed regulation. While significantly better than previous version, the speed will still have variations with temperature, load and some long-term drift. It's also unable to sense very small variations in speed. Another disadvantage is the fact this system regulates the speed of the motor, not the speed of the capstan. So it's sensitive to belt uniformity, wear in the bushings, etc. Most belt walkmans will use this system. It's good enough if well implemented correctly but has its drawbacks. 3. DC motor + FG coil on the motor. This is what the TPS-L2 and some late 70s portables used. Main advantage is it can read the speed of the motor rotor directly, instead on relying on current measurement. For this reason, this type can sense and correct very small variations in speed. While generally more stable, this sytem also regulates the speed of the motor, not the speed of the capstan. Still senstitive to mechanics. 4. Disc Drive system with capstan servo. Used in DD walkmans, TC-D5 portable, etc. This system does not use a belt, motor driving the flywheel via a rubber disc. Main advantage is this system can sense the speed of the capstan directly, instead on relying on motor speed regulation. For this reason it's way more accurate and way less prone to mechanical wear, being very stable even with significant load variation (like poor cassette shells). 5. Direct Drive system. Used in Panasonic RX-S40/S41 walkmans and in decks. Here the capstan is the motor rotor itself, having the advantage that any corrections done on motor are tied directly to capstan speed. 4 and 5 can also use a quartz oscillator as the speed reference, making the speed extremely stable with temperature variations and having almost negligible long-term drift.
[QUOTE="novrain2012, post: 90728, member: One thing I found about my J9 (maybe just this unit) is that the tape speed may drift up ~30hz or 1% on both directions, when the player is fully assembled. I guess it’s some heat stability issue when back cover is on. A little bit odd since J9 has a dedicated regulator circuit for each direction.[/QUOTE] These J9/09/600 and similar have minimal clearance between the rather flexible main board and the pulleys. But probably if something is rubbing because of fitting the back cover, there would be a scraping noise. Maybe not if it's a wire rubbing. I've been testing mine with the back cover off, so I can access the speed adjustment screws on the motor board. But given your observation I'll repeat the test with the cover installed. I have a bunch of these models. They are all slated for speed etc. testing. So I'll be able to determine if the change in speed with the cover installed is common. All of mine sound good to me, but it adds to their value if I can say they have been tested and fine tuned like this. Is the motor speed regulation circuitry on the motor board? Although the same board is used in many Aiwa's with the 2ME-7 mechanism, unfortunately there are no circuit diagrams for it.
Yes, the regulation circuitry is on the motor board. Seen that before motor + servo being treated as a single unit by the service manual. One possible reason for that is they got that manufactured by a 3rd party and they bought it as a complete assembly. These AIWA slim motors are made by Namiki and it's possible they designed the servo too based on AIWAs specifications. However these servos are pretty simple, using a single IC, sometimes even with the drive transistor integrated and some passives around. Should be easy to identify the chip itself.
It's an LA5524, seen this IC in a couple of other walkmans although not the most common. I think the drive transistor is also integrated into the chip, as the 3 discrete ones are a bit small.
Replacement bare motor boards like this are being made in China. A lot of the Aiwa's that use this board mount it right next to the internal battery compartment. In fact the batteries rest against the board. So when the batteries leak the board is the first thing to be ruined. Which in turn would be common enough to make it worthwhile to reproduce the boards. They are available as a bare board, or with all the components except the two ic's. I've ordered two of the populated boards and hope the chips on my corroded boards are ok and that my soldering skills are up to swapping them over. The chips, or substitutes, are also advertised separately. It's surprising, so far as I can determine, that none of these populated motor boards are available with the ic's mounted.
