If you're going with SMD components, you can also try using a SMD crystal to simplify the build process some more. I used this one for example (although currently out of stock, but there are probably similar ones) https://www.digikey.com/en/products/detail/fox-electronics/FK135EIHM0-032768-T3/1024594
Also for future reference, sometimes it's actually the RTC that is faulty. I order a lot of components from aliexpress and have had no issues with going cheap, except in the case with RTCs. With cheap RTCs they are sometimes the ones that failed a self test off the production line and they'd be way off on time. I have a testing rig set up to make the RTC output a test frequency square wave that I can measure with a multimeter before I solder them. On Monday, 19 July 2021 at 08:27:43 UTC-4 MrThe50sanchez wrote: > Hello! > > It looks like I was in a Mistake with the datasheet of the crystal, > thinking that the 6.0 to 12.5 pf capacitors would be perfect for the time > running perfect, But with this configuration, I have like 2 mins of delay > in 24H. > > I really appreciate the help, Just took a quick look and you are right, > 22-27Pf and 220K ohm seems that is going to fix the delay I have. > > I will have a good reading of the links you added, will try, and will let > you know. > > Thanks a lot! > > El lun, 19 jul 2021 a las 1:24, gregebert (<greg...@hotmail.com>) > escribió: > >> It is gaining time, or losing time ? >> >> If it's losing time, could it be caused by execution delays in your >> software ? I've seen some really bizarre things happening with my Raspberry >> Pi based clocks because of the chaotic behavior of Linux, and I have even >> seen metastability-related failures due to the way I had software scanning >> hardware for asynchronous events. They are fixable, but they definitely >> challenge your debugging skills. >> >> BTW, the last time I saw such a thing was *yesterday*....I'm testing my >> RZ568m tubes for a new clock, and saw my time jump by 2 seconds, instead of >> the usual 1 second, because of the way in which I update the displayed >> time. These are often subtle bugs you dont catch while typing-in your code. >> >> On Sunday, July 18, 2021 at 12:19:21 PM UTC-7 Dekatron42 wrote: >> >>> The capacitance you cite from the datasheet is not the recommended >>> capacitance for the external capacitors, it is the load capacitance of the >>> crystal - you must check your ordering document to see which crystal you >>> got unless you can read the information on the crystal itself if it is >>> laser printed there. The load capacitance can then be used to calculate >>> what capacitance value you need for the two other capacitors - do a quick >>> test with two 22pF or 27pF capacitors if you have the crystal with CL = >>> 12.5pF - twice the load capacitance from the formula in the application >>> note: http://ww1.microchip.com/downloads/en/appnotes/91097a.pdf is >>> usually a good start, so if you have a lower CL value use twice that for >>> tests. >>> >>> Neither is the series resistance in the datasheet the value you need for >>> the external resistor, it is the internal resistance of the crystal itself. >>> You should use the value in the application note to start with, 220k - it >>> might be that you have damaged the small crystal by over driving it if you >>> used a lower value, it is explained in the application note why you need to >>> protect the crystal from over driving. You can read some on series >>> resistance and crystal selection here: >>> https://www.maximintegrated.com/en/design/technical-documents/tutorials/7/726.html >>> >>> . The datasheets and product documentation found on the manufacturer >>> website: >>> https://www.microcrystal.com/en/products/khz-quartz-crystals/ms3v-t1r/ >>> and >>> https://www.microcrystal.com/en/products/khz-quartz-crystals/ms1v-t1k/ >>> do contain the equivalent circuit for the crystal similar to the one above. >>> >>> So try with twice the load capacitance for both capacitors and use the >>> recommended 220k resistor as in the application note and see how it behaves >>> - also use the pins recommended in the application note, ie CLKIN & CLKOUT. >>> >>> /Martin >>> On Sunday, 18 July 2021 at 20:53:32 UTC+2 MrThe50sanchez wrote: >>> >>>> Hello Again group! >>>> >>>> >>>> I've been testing this for a while, and I still have some delay >>>> minutes. This is the PCB design I'm using and components. >>>> >>>> *Components: * >>>> ·PIC16F1936 >>>> ·CRYSTAL MS3V-T1R CRYSTAL 32.768 Khz. >>>> ·Capacitors ( used 6.0 ; 7.0 ; 9.0 & 12.5) pF as datasheet from crystal >>>> recommend. >>>> ·Resistor value between 50/65 K ohm. >>>> >>>> *PCB Design: * >>>> >>>> This design is done following the document you guys added called: >>>> Microship TB097 which give's good instructions on how install Crystals to >>>> our pics. >>>> >>>> >>>> Guys, My head is gonna explote because I don't find the problem that is >>>> making minutes delays in 24H (About 3 mins per 24H). >>>> >>>> I'm questioning two fact's that can be making this delay: >>>> >>>> *1.* Pins connections between Crystal + Pic. In some other designs, >>>> crystal is attached to other pins. For example. >>>> >>>> Example A) Crystal pins connected to *PIN 9: A7/CLKIN* & *PIN 10: >>>> A6/CLKOUT.* >>>> >>>> Example B) Crystal pins connected to *PIN 11: RC0/T1OSO *& >>>> *PIN 12: RC1/T1OSI. * >>>> >>>> this make's me think that I'm connectin the crystal to wrong pins? >>>> what's the difference between them? >>>> >>>> *2*. Maybe I need to try better with other cap values? >>>> >>>> *3.* Software getting the count wrong? >>>> >>>> Thanks a lot guys, I really appreciate your time and help. >>>> >>>> Christopher.. >>>> >>>> >>>> >>>> >>>> >>>> El jueves, 20 de mayo de 2021 a las 21:09:01 UTC+2, Dekatron42 escribió: >>>> >>>>> Since this is a Tuning Fork type of crystal they are sensitive to to >>>>> much energy and might become damaged if you use them incorrectly. You >>>>> should read this article: >>>>> http://ww1.microchip.com/downloads/en/appnotes/91097a.pdf and also >>>>> read the Microchip application note above that Bill van Dijk suggested as >>>>> they both explain/discuss how to use this type of crystal. >>>>> >>>>> In this document for the PIC-processor in question: >>>>> http://ww1.microchip.com/downloads/en/devicedoc/41364e.pdf they also >>>>> mention the 91097a.pdf above as a document to read on page 71. >>>>> >>>>> /Martin >>>>> >>>>> On Thursday, 20 May 2021 at 20:40:49 UTC+2 MrThe50sanchez wrote: >>>>> >>>>>> Ei Martin! >>>>>> >>>>>> I'm using PIC16F1936 (16F family!) >>>>>> >>>>>> my connections are the following. >>>>>> >>>>>> Crystal pin 1 - RA7/OSC1/CLKIN/SEG2 >>>>>> Crystal pin 2 - RA6/OSC2/CLKOUT/VCAP/SEG1 >>>>>> >>>>>> >>>>>> I add you my pcb design (The old one) and the new one! >>>>>> >>>>>> (OLD ONE) >>>>>> [image: old ddesign.png] >>>>>> >>>>>> >>>>>> >>>>>> NEW DESIGN >>>>>> [image: new design.png] >>>>>> >>>>>> >>>>>> Thanks a lot! >>>>>> >>>>>> El jue, 20 may 2021 a las 16:28, Dekatron42 (<martin....@gmail.com>) >>>>>> escribió: >>>>>> >>>>>>> What PIC-controller is the crystal hooked up to? Did you follow the >>>>>>> design specifications for using the crystal you chose? >>>>>>> >>>>>>> /Martin >>>>>>> >>>>>>> On Thursday, 20 May 2021 at 16:03:55 UTC+2 jf...@my-deja.com wrote: >>>>>>> >>>>>>>> All the cheap watches and clock movements seem to use the same >>>>>>>> small cylindrical crystal. I got many of mine by taking them out of >>>>>>>> old >>>>>>>> watches or clocks that had stopped working properly (like worn out >>>>>>>> gears in >>>>>>>> a clock). If you do not already have any in your junk box, you should >>>>>>>> be >>>>>>>> able to buy a working one from a street vendor or flea market for >>>>>>>> under >>>>>>>> € 5, and probably less than € 1. Also, new crystals will cost >>>>>>>> less than US$1 each (plus shipping and handling), so you could buy >>>>>>>> several >>>>>>>> and select the one that works the best. >>>>>>>> >>>>>>>> On Wednesday, May 19, 2021 at 11:55:42 PM UTC-7 MrThe50sanchez >>>>>>>> wrote: >>>>>>>> >>>>>>>>> ? Any recommendations of where can I get a good quality cristal >>>>>>>>> for my watch? >>>>>>>>> >>>>>>>> -- >>>>>>> >>>>>> You received this message because you are subscribed to the Google >>>>>>> Groups "neonixie-l" group. >>>>>>> >>>>>> To unsubscribe from this group and stop receiving emails from it, >>>>>>> send an email to neonixie-l+...@googlegroups.com. >>>>>>> To view this discussion on the web, visit >>>>>>> https://groups.google.com/d/msgid/neonixie-l/b9d34649-4fef-4fe8-a44b-8544b51dee8cn%40googlegroups.com >>>>>>> >>>>>>> <https://groups.google.com/d/msgid/neonixie-l/b9d34649-4fef-4fe8-a44b-8544b51dee8cn%40googlegroups.com?utm_medium=email&utm_source=footer> >>>>>>> . >>>>>>> >>>>>> -- >> You received this message because you are subscribed to the Google Groups >> "neonixie-l" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to neonixie-l+...@googlegroups.com. >> > To view this discussion on the web, visit >> https://groups.google.com/d/msgid/neonixie-l/b0456e09-3348-44d4-a2d5-5ec466983b82n%40googlegroups.com >> >> <https://groups.google.com/d/msgid/neonixie-l/b0456e09-3348-44d4-a2d5-5ec466983b82n%40googlegroups.com?utm_medium=email&utm_source=footer> >> . >> > -- You received this message because you are subscribed to the Google Groups "neonixie-l" group. To unsubscribe from this group and stop receiving emails from it, send an email to neonixie-l+unsubscr...@googlegroups.com. To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/eff98cda-5b58-45ee-a1de-db9de3d345a8n%40googlegroups.com.