search

rusefi / hw_microRusEfi

microRusEfi board for rusEfi ECU
66 stars 36 forks source link

loading caps on the crystal #58

Closed rusefi closed 5 years ago

rusefi commented 5 years ago

@mck1117 says "loading caps on the crystal should be 8pF instead of 20pF"

mck1117 commented 5 years ago

confirmed - 8pf caps function properly

jharvey commented 5 years ago

mck1117, have you seen the note on page 5 of Frankenso? I have some notes and a reference note when we were trying to figure out this intermittent clock issue. See below snippet which is on the Frankenso schematic. http://www.crystek.com/documents/appnotes/Pierce-GateIntroduction.pdf PCB per predictions with SaturnPCB has less then 3.5pF traces, TPIC pins assumed 5pF ESR = 80ohms max Rf = 2meg could be between 1meg and 10meg. Cload should be 18pF per XTAL datasheet Cload = ([Cin+C1][C2+Cout])/(Cin+C1+C2_Cout)+PCBstray Cload = ([5+24][24+5])/(5+24+24+5)+3.5= 18.0pF C1=C2=C166=C167 = 24pF Rs = 1/(2piFC2) = 1/(2pi8MHz*24pF) = 829ohms, 820ohms is close enough = R177

I think we technically for this one, we want 4pF not 8pF. See my calcs from here https://docs.google.com/spreadsheets/d/1O6-u6T_O083pT9SEm70DGKKy6YZJ89tGeXWsZTZnmM4/edit?usp=sharing

mck1117 commented 5 years ago

I hadn’t seen that yet, no. The boards I’ve assembled so far have 8pF caps, and seem to work fine, fwiw. Is there a way to empirically measure whether the crystal is over or under loaded? I have a 200mhz scope at my disposal, and a 4ch/2ghz one at work if I need it.

jharvey commented 5 years ago

I don't know a way to measure if the loading is proper. I understand the problem is mostly a failure to start issue. If you change your pF's you can wander the 8MHz slightly, but I don't think we care about the slight warping. I guess we could measure it bey building 1k boards and see if they all start at the min and max temperatures. If yours is starting, with 8pF then it's probably fine. I see 4.7pF is available, so I updated the BOM and schematic to call for that.