So upgrading the PSU *isn’t* recommended?
-
Hello MLAers! We've re-enabled auto-approval for accounts. If you are still waiting on account approval, please check this thread for more information.
SE/30 Sony CR-44 PSU unstable voltage output
- Thread starter Bolle
- Start date
This is what my thinking was as well. My Brother's SE30 has a Dimmo 50mhz, and an ethernet card. Executive decision recently by me was to pico it. It still works but is proving more cost effective/efficient when refurbing the machine.This is why I ended up ATX upgrading my SONY SE/30 power supplies, only to now find that the Seasonic units that are commonly used for this purpose don't seem to handle lots of SE/30 upgrades well (accelerator + Ethernet card + a 2nd PDS card).
Not a huge fan of using PicoPSU in heavily upgraded machines: the PicoPSU datasheets are quite insistent that forced air cooling is required when pushing 3.3V/5V heavily, but I'm not sure that anyone has actually measured the temperatures of a PicoPSU in a closed up SE/30. There's some airflow, but not sure it's sufficient given the small size of the PicoPSU.
Either way, the actual issue remains: When upgrading your SE/30, you load the 5v line heavily, and regardless of what PSU you're using the stock harness (or Xceed Grayscale harness) only uses two 22 gauge wires to carry *all* the 5v to the logic board. This is a recipe for voltage drop. With the stock PSU, you can at least crank up the output voltage to combat it to some degree, but with ATX and other modern retrofits you don't have this option.
I strongly recommend replacing those 22 gauge wires with 18/16ga which can be purchased pre-crimped from Digikey for cheap. Some more details in this thread.
Relating to the original topic, though, I'm curious what that symptom looked like as I've experienced a slight horizontal wobble to the screen on HDD access which I always assumed was just the way things were. Suppose I'll need to pop open my Sony PSUs and have a check. I'd think it'd be doable to retrofit a modern LDO in its place, if dead: IC252 is already mounted in a way that insulates the back from chassis ground (thermal pad insulates rear of IC, and a clip is used to hold it against the PSU chassis) so just some jumpers would be needed to adapt the pinout of whichever modern LDO is used.
My seasonic refitted unit also has the same issue, but that's expected as the 12v rails in ATX PSUs aren't independently regulated, just independently current monitored. Putting in a filter of a schottky diode into an inductor and some caps to ground on the 12 sweep line did help. Or, just don't use an actual HDD may be the easier fix
Fan /airflow is also being upgraded. Theres a BlueSCSI and no HD. It is also having a full recap as is the Accelerator expansion card.Not a huge fan of using PicoPSU in heavily upgraded machines: the PicoPSU datasheets are quite insistent that forced air cooling is required when pushing 3.3V/5V heavily, but I'm not sure that anyone has actually measured the temperatures of a PicoPSU in a closed up SE/30. There's some airflow, but not sure it's sufficient given the small size of the PicoPSU.
Either way, the actual issue remains: When upgrading your SE/30, you load the 5v line heavily, and regardless of what PSU you're using the stock harness (or Xceed Grayscale harness) only uses two 22 gauge wires to carry *all* the 5v to the logic board. This is a recipe for voltage drop. With the stock PSU, you can at least crank up the output voltage to combat it to some degree, but with ATX and other modern retrofits you don't have this option.
I strongly recommend replacing those 22 gauge wires with 18/16ga which can be purchased pre-crimped from Digikey for cheap. Some more details in this thread.
Relating to the original topic, though, I'm curious what that symptom looked like as I've experienced a slight horizontal wobble to the screen on HDD access which I always assumed was just the way things were. Suppose I'll need to pop open my Sony PSUs and have a check. I'd think it'd be doable to retrofit a modern LDO in its place, if dead: IC252 is already mounted in a way that insulates the back from chassis ground (thermal pad insulates rear of IC, and a clip is used to hold it against the PSU chassis) so just some jumpers would be needed to adapt the pinout of whichever modern LDO is used.
My seasonic refitted unit also has the same issue, but that's expected as the 12v rails in ATX PSUs aren't independently regulated, just independently current monitored. Putting in a filter of a schottky diode into an inductor and some caps to ground on the 12 sweep line did help. Or, just don't use an actual HDD may be the easier fix
~But will bear in mind replacing the wires.
Last edited:
I haven't looked at the transistor yet, but I have a strong suspicion that my symptoms were due to the reset line going low due to heavy voltage drop on the +5V line. I've upgraded both my harnesses with 16AWG wires and am seeing more reliable operation.
