Page 2 of 2

Re: Knock detection

Posted: Fri Oct 29, 2010 11:50 am
by STC
10 cycles is for redundant initialization and checking, previously it was only 1 time. I did it for experiment and left for now. I don't know anything about diagnostic mode, because data sheet has no proper information, so that is why I don't use diagnostic mode and leave ~TEST in high state. SO terminal immediately repeats state of SI, I think it is directly multiplexed without any logic involved :-)
I tried to find more information about diagnostic mode, but without success. My chip is without 'Z'. Only 'AB', but it does not matter I think. IMHO - datasheet for HIP9011 is full of shit.

Re: Knock detection

Posted: Fri Oct 29, 2010 11:58 am
by STC
Interested thing. One of users tried to connect 10k resistor between SI and SO lines. He told me that in this case all works OK ! I don't understand :-)

Re: Knock detection

Posted: Sat Oct 30, 2010 12:53 am
by jharvey
I just posted a note I got from Intersil found here.

viewtopic.php?f=9&t=480&p=11261#p11261

From what I see the in the noted reference chip, it does not appear to have the same diagnostics layout. I see the last 2 bits appear to be used, and the table shows only analog features, where the 9011 appears to include digital features. Perhaps the TPIC chip is better supported. Looks like the TPIC may have the same diagnostics feature. I wonder if the Intersil folks simply got bought out by TI. Would appear the TPIC is better supported and has a better datasheet, that exactly matches the Intersil chip.

Re: Knock detection

Posted: Sun Oct 31, 2010 10:30 am
by STC
Thank you for the link! I will read data sheet for HIP9010. As I understood, business of Intersil is slightly bad :-)
Yes. I also don't see information how to read registers via SO pin.

By the way. I've fixed bug. I've inserted initialization of prescaler before initialization block. I don't fully understand how, but it works.
See my last commit.