Edit: Oops! I didn't pay enough attention. The CPU rotation does look correct after all.

for some reason my answers are not getting posted. my board has pads for jp7 but nothing soldered. the golden triangle i thought was to go to pin 1-a not 1-p because of the shape of the socket. it seems like it's pointing to pin 1-a

I edited my original post and deleted it.

AlaricD: I believe the other side of the CPU is also 1 thru 14. The triangle points to matric location 1-A, which I believe is correct.

if you look at the picture in one of my latest posts where the socket is unpopulated it has number 1 printed on both sides. i oriented the cpu to pin 1-a

balthazor4ever, measure the solder pads to see if they go to FLT# on the CPU and to GND.

I wish I knew about the FLT trick earlier.

You can probably even bend back the QFP lead once or twice before it falls off if you think you want to go back to the onboard CPU. Another approach would be to use JB Weld to affix a jumper to connect and disconnect GND to the FLT# pin.

feipoa wrote:

I do not understand why your board contains a a PGA132 socket, yet no jumper to set FLT# to GND. My board was similar.

Unless installing a CPU in that socket then trips a relay to close the FLT#-to-GND.

if i install the 486dlc in the socket then both cpus get warm and the 386 even gets really hot (a lot more than usual) after a few seconds.

balthazor4ever wrote:

if i install the 486dlc in the socket then both cpus get warm and the 386 even gets really hot (a lot more than usual) after a few seconds.

Maybe you should try putting jumoers on your motherboard (particularly at JP7) and see if that doesn't help.

AlaricD wrote:

feipoa wrote:

I do not understand why your board contains a a PGA132 socket, yet no jumper to set FLT# to GND. My board was similar.

Unless installing a CPU in that socket then trips a relay to close the FLT#-to-GND.

I haven't come across such an advancement on these 386 boards.

balthazor4ever, are you able to set the FLT# pin to GND? It is not necessary to bend the QFP pin over as I did, you can merely solder a wire to the FLT# and the other end to GND. The other end can even be poked into the GND on the molex cable off the PSU.

i could try with a wire because i wouldn't like to accidentally destroy the 386dx

I attached an image to ensure you get the right pin.

belated thanks for the tip feiopa, I hope the cpu on my Unichip 367C will be suitable 😀
any idea where #FLT is normally tied to? I assume VCC?

Thank you!! I hope it works!!

keropi wrote:

belated thanks for the tip feiopa, I hope the cpu on my Unichip 367C will be suitable 😀
any idea where #FLT is normally tied to? I assume VCC?

That is a good question. If it is tied to directly Vcc then wiring it to GND will cause a short. In my recomendation to wire FLT# to GND, I am under the assumption that FLT# was left floating. You best ensure that FLT# is not connected to Vcc by using the continuity test function or resistance function on your multi-meter. I don't recall if my FLT# pin was connected to anything or not.

Edit: looking at the photos, it appears that FLT# on my board was indeed connected to Vcc. Not sure why they did that. Check yours.

yeah better lift+connecto to gnd rather than risk anything

Things kinda went quiet. Balthazor4ever, how did it work out for ya?

I'm still unpacking some stuff in my new house and haven't found my soldering equipment. i'v tried tin foil to set jp 7, 9, 10, 11 but no luck. it might just be that tin foil isn't making enough contact so i'll have to find my soldering iron to check them again and the flt pin as well.

I prefer copper tape with electrically conductive adhesive.