Osc. ckt for DAC chip

[whatever]

In a 1987 Kenwood dp-460 CD  player's dac circuit, I measured 55.234 Mhz on X_o and X_in (pins 12 and 10) on the DAC chip (Toshiba td6709N). Note that the KDS xtal can has "16.9344".:Mhz printed on it. ,Not sure what that other sliver can is (T 7E1 JPN, F62SM )? It may be some kind of freq. multiplier.
That 30-pin Toshiba td6907N dac iis used n similar CD players, including other Kenwoods, which often use a 67.xxx MHz xtal . And I do measure exactly that on the Xout and Xin pins.But Kenwood has chosen a different strategy in the dp-460 .

Anyone know what that  sliver can is (T 7E1 JPN, F62SM ) ?


.(Note that the Toshiba td6709n dac seems to be designed for high clock rates as normally, one sees audio dacs of this vintage in the 8-16Mhz range) ).


BELOW: Kenwood dp-460





===
===

Below Kenwood dp-750 (uses 67.xxx Mhz xtal; td6907N has heatsink in it)

[RES]

Cannot find diagram, only two datasheets of different Toshiba DACs; TD6710 and TD6720.

https://tvsat.com.pl/PDF/T/td6710an_tos.pdf
https://vasiltech.narod.ru/files/TD6720N.pdf

[wasedadoc]

I doubt it is a coincidence that 16.9344 multiplied by 4 is 67.xxx.

[whatever]

I doubt it is a coincidence that 16.9344 multiplied by 4 is 67.xxx.

Right ... perhaps because Kenwood had issues obtaining the orig 67.737 Mhz by the time the dp-460 came out (mid 1987); the dp-750 (with  67.737 Mhz) is from mid 1986.
Not sure whether the higher Mhz xtals are harder to cut for low-jitter applications.
About jitter, that plastic "jitter jacket" -- shown  in photo of KDS 16.93444 -- is new to me. In the serv. manual, the jacket itself has dedicated part no.!!

The mystery remains about my measurement of  55.234 Mhz on X_o and X_in (pins 12 and 10) on the dp-460  DAC chip. In my dp-750, I measure (with a Siglent o'scope) a pure 67.737 Mhz sine wave.

Cannot find diagram, only two datasheets of different Toshiba DACs; TD6710 and TD6720.

Yup ... I have those already, and they are not very detailed, maybe deliberately by OEM to protect IP. Philips and Burr-Brown produced much better datasheets.
IAC ...
If you go to
https://www.dutchaudioclassics.nl/the_complete_d_a_dac_converter_list

... and look for any of those TD chips -- TD6709, TD6710 and TD6720 are identical -- you'll note CDP they are used in. Then you can look for a serv. mnl thru the usual free portals like Archive.org.
Here are screen captures from a Sansui CDe70 schematic, that uses that same "OEM" circuit (which may have been Toshiba???). Note the td6709 with the 67.737 Mhz xtal.

[CaptDon]

Are you talking about the gray shrink wrap sleeve on the crystal with the part number and frequency printed on it?? It serves no function at all other than a place to print the crystal part number and frequency. Jitter Jacket?  That's a new one!! Have audiophools invented a name for a crystal with shrink wrap? I have dozens of old crystals many of the same frequency, some with the gray wrap (or even black) and some without. It makes no difference in stability.

[CaptDon]

Most crystals won't run very well on the 4th overtone. They like fundamental, 2nd, 3rd and 5th overtone. Maybe the crystal is running at the 2nd overtone and the other device is a resonator or S.A.W. device to act as a doubler circuit to the 2nd overtone which would create the 4th overtone and exactly the frequency needed, Don't know why your scope shows a totally unrelated frequency unless it is pulling the crystal some kind of terrible!! Does the unit still play while you are measuring the crystal? I am guessing it may sound whacko.

[whatever]

Don't know why your scope shows a totally unrelated frequency unless it is pulling the crystal some kind of terrible!!

Well, my Siglent sds-1202x-e dso is very finicky to use. So I'm going to renege  on what I noted earlier. With certain osc. circuits, I can measure just fine with the Siglent. For example, right now, I'm looking at 11.289Mhz signal from a dedicated ckt (not connected to anything but the o'scope) . I want to make sure the Siglent can measure clean Mhz freqs. And the 11.2889 Mhz osc. looks and measures normally on both Siglent and my old Tek analog 'scope.
In a CD player or D/A processor, the Tek 465 behaves but the Siglent is almost always weird. I use the idiot button "Auto Setup" and sometimes (rarely), it'll lock in, but most often not. I  wish I could get to the bottom of the Siglent. After all, it is 200Mhz and the Tek 465 is only 100.
I assume the xtal is at its schematic (serv. mnl) spec. of 67.xxx Mhz. Otherwise, the CDP would not work or work well. It works well.

[CaptDon]

You would be surprised at how many crystals oscillate fundamentally at one third of the frequency marked on them. Many of the 27MHz C.B. crystals are actually cut at 9MHz. It is very easy to get the third overtone of a crystal. When the forth overtone is needed it seems they oscillate on the 2nd overtone and then feed a doubler  or phase lock a P.L.L. to the 4th overtone of a crystal which is oscillating at its fundamental cut. Without a schematic it is hard to know what they are doing with that 16MHz crystal but the 4th overtone is exactly correct when compared to the 67MHz frequency. I suspect the smaller metal can is also a resonator of some sort and probably resonant at the 67MHz frequency. Perhaps the DAC has a built in frequency multiplier circuit like most microprocessors.

[MarkT]

I doubt it is a coincidence that 16.9344 multiplied by 4 is 67.xxx.

Or that 16.9344MHz is 384 x 44.1kHz (ie 3 x 48 x 44100).

In these days of I2S busses you see (multiples of) 6.144MHz crystals a lot (for 48/96/192kSPS), and for CD audio multiples of 5.6448MHz

[whatever]

I doubt it is a coincidence that 16.9344 multiplied by 4 is 67.xxx.

Or that 16.9344MHz is 384 x 44.1kHz (ie 3 x 48 x 44100).

In these days of I2S busses you see (multiples of) 6.144MHz crystals a lot (for 48/96/192kSPS), and for CD audio multiples of 5.6448MHz

I'm assuming that Kenwood had supply issues for that original 67.737 Mhz , so for later models of the same topology, they opted for that weird 16.9344MHx x 4 option. I also assume that original pure 1x xtal/osc. is the cleanest option. Yes?