24 MHz Crystal Specifications ?

[ChrisNielsen]

Could you please consider posting specifications for the 24 MHz crystal so that I can make sure to purchase a crystal that meets the requirements?

Datasheet Table 63 lists Cload and typ ESR for the 32 KHz crystal, but no similar table elements are given for 24 MHz.  I do see from the Tower board BOM that it used an 18 pF Cload, but I want to make sure this meets the DS requirements -- is 18 pF the datasheet required Cload?

Of more concern is the unknown typ or max ESR required.  For 32 KHz, the ESR is listed as typ, but it is MUCH preferred if you could specifiy a max ESR value since this is by far the most common specification found on crystal datasheets.

Thanks, Chris

[naoumgitnik]

Dear Chris,

• The 24MHz crystal's characteristics are identical to those recommended for the i.MX6 series in their datasheets (planned to be added into the Vybrid datasheet, too).
• The finalized recommended capacitance value is 12pF.
• Please remember to apply the relevant Erratum about the 2.2M resistor form XTAL to GND.

Sincerely, Naoum Gitnik.

[ChrisNielsen]

To clarify, the i.MX6 recommends:

- 80 Ω typical ESR crystal rated for a maximum drive level of 250 μW
OR

- 50 Ω typical ESR crystal rated for a maximum drive level of 200 μW

To clarify the 12 pF value:

- I think you are referring to the Rev G Tower schematic where the external load caps changed from 27 to 12 pF?

- Could you confirm that when purchasing a crystal, we still need to ask for crystal Cload = 18 pF? (as used on Tower)

Regarding the 12 pF, could you confirm the math?  I think it is:

- 12 pF = 2 * ( 18 pF Cload - ( 3.0 pF pcb stray cap + 9 pF Cin E/XTAL pins ) )

Did I get it right?

Thanks, Chris

[ChrisNielsen]

Also, if the above assumptions on the specs are correct, I'm having a hard time finding a crystal datasheet that can meet the specs.  Could you share a few example crystal part numbers that are know to have their datasheets meet these specs?  Could you share the part number that is soldered on the Tower boards? (the BOM xls file only lists as generic CADpart place holder)  Thanks, Chris

[ChrisNielsen]

Amazingly, the iMX6 eval board (MCIMX6DLAICPU1) uses the best candidate I was able to find on the market during my 1/2 day web hunt:

Epson Toyocom #TSX-3225 24.000 MHZ

The issue is that the Epson datasheet lists:

- Epson = 40 ohms max ESR, MISMATCH, Vybrid = 50 ohms typ ESR (~=75 ohms max ESR?)

- Epson = 200 uW max, SUCCESS, Vybrid = 200 uW max

I haven't gone hunting yet for standard crystal typ/max ESR ratios but in general I use 1.5X in a lot of typ/max BSEE areas -- it's a rule of thumb but decent in the absence of hard data.

Using 1.5X, I get an estimate of 27 ohms typ ESR for the Epson.

If this is the design point, could someone at the factory study it and consider changing the official datasheet language to include "40 ohms max ESR" as the spec?  If we can't do that, I suppose I could keep hunting for a crystal that does meet the specs, but it might be difficult.

If the oscillator design will accomodate it, it would be nice to extend the spec to 60 ohms max ESR since a lot of crystals are speced there (and 40 ohms max is less common).  Similarly, it would be helpful if the max uW drive could be reduced since a lot of parts use 100 uW max.

Thanks for considering it, Chris

[naoumgitnik · Accepted solution]

Dear Chris,

You are right – for the layout similar to that of our Tower board, please, select a crystal with CL = 18...20pF and use two 12...15pF capacitors on the board.

Also provided below are crystal types used on several Vybrid and i.MX6 boards, with their specific parasitic capacitances:

Manufacturer	Manufacturer PN	CL in datasheet, pF	ESR max. in datasheet, Ω	Processor / Board	Capacitance used on board, pF
FOX ELECTRONICS	FOXSDLF/240F-20	20	30	Vybrid / Automotive	12
PERICOM SEMICONDUCTOR	FL2400041	18	60	Vybrid / Tower Module	12
EPSON TOYOCOM CORPORATION	TSX-3225 24.0000MF15X-AC3	9	60	i.MX6DQ / SABRE_AI	16
SIWARD INTERNATIONAL INC	XTL571300LLI24.000-10TR	10	80	i.MX6DQ / Sabre_SD	18

Our documentation is still being worked on, and we will take proper care to provide all the required data in an as clear way as possible, but until the Vybrid materials are ready, feel free to use the i.MX6 ones to select the proper crystal type.

Sincerely, Naoum Gitnik.