[rescue] 501-5365 in E3000 causes "Overtemp detected on board 16" during POST

[Ethan Hawke]

The EXX00 don't use I2C, it looks like the clock board is implemented 
almost entirely in discrete logic, which seems pretty crazy to me. I 
believe (though not 100% sure) that the temperature is through an 8bit 
LTC1099 ADC mounted on the clock board, I may try wiring the measure pin 
through a low resistance resistor to ground, there doesn't seem to be an 
undertemp alarm, so in theory this will work (with the caveat of no 
temperature monitoring on the clock board).

Cheers,
Ethan

On 2/19/23 00:39, Dave McGuire wrote:
>   I assume those sensors are I2C.  If this is the case, and if the 
> firmware accessing them is poorly-written (i.e. not properly handling 
> the lack of a presence pulse when the sensor is accessed), and if the 
> sensor is dead or missing, 0xff would be the expected result.
>
>          -Dave
>
> On 2/18/23 07:00, Ethan Hawke via rescue wrote:
>> An update, I entered the extended POST tests as that is before the 
>> overheat checks, and probing the temperature register on the clock 
>> board reads '0xFF', or in other words it's stuck high.
>>
>> Not sure where on the board the temperature probe is, I did find a 
>> ADC so that's probably a good start.
>>
>> Cheers,
>> Ethan
>>
>> On 2/18/23 21:42, Ethan Hawke via rescue wrote:
>>> Board 16 is the clock board, boards 0 to 15 are the CPU or IO boards.
>>>
>>> I am carefully looking over it now, to see if I can see a scratch or 
>>> other damage.
>>>
>>> I am running OpenBSD on it at the moment (with the original clock 
>>> board) and it can display the clock board temperature, so it does 
>>> contain some sort of "smart" temperature probe.
>>>
>>> Cheers,
>>> Ethan
>>>
>>> On 2/18/23 21:38, Jonathan Katz via rescue wrote:
>>>> Ugh. Not sure of a silver-bullet answer. The x500 boards should work
>>>> in x000 and vice-versa, just at the lower backplane clock speed (50Mhz
>>>> vs. 83Mhz, IIRC.)
>>>>
>>>> E3x00 don't have a board 16 (right?) so I'm not sure where it would
>>>> get the temp from. Almost sounds like a resistor or surface-mount
>>>> capacitor broke off and it's confused, thinking it's in an E6500 or
>>>> something. Any bent pins anywhere?
>>>>
>>>> On Sat, Feb 18, 2023 at 10:32 AM Ethan Hawke via rescue
>>>> <rescue at sunhelp.org> wrote:
>>>>> Hey all,
>>>>>
>>>>> I picked up a 501-5365 Clock board recently and when installed in 
>>>>> one of
>>>>> my E3000s it causes an "Overtemp detected on board 16" and then shuts
>>>>> off the machine (happens in both my E3000s). All the fans are running
>>>>> and it always occurs at the same point, during the "Environmental 
>>>>> Probe
>>>>> Test". The board is definitely not overheating (and there isn't 
>>>>> really
>>>>> anything on it to get that hot). I know that the 501-5365 actually
>>>>> shipped with the EX500 systems and not the original EX000 systems, 
>>>>> but
>>>>> the handbook does mention that it is supported in the EX000 
>>>>> machines. Is
>>>>> this not true? Is mine faulty? Is there a jumper or modification I 
>>>>> need
>>>>> to make?
>>>>>
>>>>> If anyone has any suggestions, I'm all ears!
>>>>>
>>>>> Cheers,
>>>>> Ethan
>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> rescue list - http://sunhelp.org/mailman/listinfo/rescue_sunhelp.org
>>>>
>>>>
>>>
>>> _______________________________________________
>>> rescue list - http://sunhelp.org/mailman/listinfo/rescue_sunhelp.org
>>
>> _______________________________________________
>> rescue list - http://sunhelp.org/mailman/listinfo/rescue_sunhelp.org
>