STE V1.5 32MHz BOOSTER - STATUS

[exxos]

The 36 MHz oscillators came today.. So swapped out the current booster's 40MHz one... Its now on test.. will see what happens...

This particular board seems to want to pop out of the socket for some reason.. But I am not really surprised about this....

My first 3-D object sort of like this...

999.jpg

The idea roughly is to have this fit around the outside the CPU, and the angle at the top will match the slant of the shielding... So the shielding will push on top of this block, and the block will push down on top of the CPU holding it in place in the socket.

Of course it would help if I was actually any good at cad :lol: I think it will be easy to do, I just need to actually work out the proper dimensions for this :)

[PhilC]

That should be pretty easy in Fusion 360 exxos, I've only just started using it, so I'm sure you'll manage to learn it quicker than me :-)

[exxos]

So 53mins loop 17 :roll: so double the run time from before, but still its tripping up after a hour... will re-test.. but these CPU's don't seem to overclock well at all :(

[troed]

I don't think this makes much sense actually. Constant load for almost an hour should be plenty to show any overclock issues. My other guess was that it's heat related (when I overclocked the Shifter to 64MHz a simple heatsink made a huge difference) but even then it should show up quite quickly.

Are they stable at 32 ... ?

[exxos]

So after some measuring, thankfully the CPU is directly under the flat part of the metal shielding.. So only really needs to gap padding out couple of millimetres..

Idea of this object is it will sit on top of the CPU, grip it around its edges by about 1.5mm, which will be all it really needs so it does not move.. And the plastic will sit 1.5mm above the CPU and push up to the shielding.. I also put a hole in the middle cos it looks cool :lol:

I need to print one out, but it will have to wait :)

yay.jpg

[exxos]

I don't think this makes much sense actually. Constant load for almost an hour should be plenty to show any overclock issues. My other guess was that it's heat related (when I overclocked the Shifter to 64MHz a simple heatsink made a huge difference) but even then it should show up quite quickly.

Are they stable at 32 ... ?

The slower I run it, the longer it will run for so it is tripping up quicker with the faster clock..

I have not tried 32MHz yet, I do not think I even have a SMT osc that speed.

Of course there are 3 boosters out there which have run at 40MHz for some time now. Even my original one I ran for days at 40MHz. So it is not like the design has any issues.

There could be a bit of noise on some signal somewhere causing it, or a iffy connection, but I need to do more testing yet..

[exxos]

So its up to loop 36 so far.. got to 17 before ( 1 hour) ... so its been going about 2 hours now... maybe just was a bad connection...

[exxos]

I turned it off last night (needed sleep for some reason) It got to 182mins loop 56, so basically 3 hours run!

I need to try the machine "cold" but I'm out all day today, so might have to wait until tomorrow..


EDIT:

Had it running this morning for a little bit, probably half hour and it quit to desktop :roll:

[exxos]

Only just got home.. but gave the MB a "hard calibration" and set it running again...Will try to get a hour run before I pass out for the night..

Also just tweaked the spacer... (previous ideas were here https://www.exxosforum.co.uk/forum/viewt ... 102#p13097 )

This one has 2 sides missing to allow air flow.. also a inner spacer so there will be a vent gap between the spacer and top of CPU... Also made 0.2mm wider as a bit "tight" fit on the CPU. Also made the outer walls 0.5mm longer... will print tomorrow ...

stev2.jpg

[exxos]

I had the booster on test last night when I got home, it ran for 63mins then quit to desktop..

First run this morning, it ran for 100mins, then locked up... Second time it ran for about 5mins and locked up... They just started the run to see what happens..

It is difficult to say this is a bad connection, or if something is tripping up somewhere..

EDIT:

OK 27mins that time.. this looks like to me the CPU just can't run at 36MHz.. I will do some more testing...