• Updated 2023-07-12: Hello, Guest! Welcome back, and be sure to check out this follow-up post about our outage a week or so ago.

Modern PSU for the SE/30

JDW

Well-known member
I mention the model number at 10:20 in my video. My model matches yours.

Regardless of what the specs say, I am hearing the fan turn on after the Mac has been on for a little while.  I've not measured the current consumption of my entire Mac, but I highly doubt it draws substantially less current when I first switch it on, and then 5 minutes later suddenly starts to draw massive amounts of current such that the fan is triggered by current.  In light of that, I made my assumption that heat is triggering the fan rather than current.

 

Crutch

Well-known member
Lots of possible explanations there - it could be measuring average load over X amount of time.

Would be easy to test if anyone cares - plug a Seasonic PSU out of the box with no load into the wall, then point a hairdryer at it.

 
Last edited by a moderator:

JDW

Well-known member
Well, I'm not going to take out the SEASONIC from my SONY PSU chassis only to blast it under no load with a hairdryer.  But back in 2018 I made the video below which serves to better explain what I said in my previous post, showing very clearly that "high current" does not appear to trigger the fan.



 
Last edited by a moderator:

Crutch

Well-known member
Good and clear video as usual.  Did anyone respond confirming similar behavior?  The statement on the spec sheet that the fan doesn’t run under a 30% load is so specific it’s hard for me to believe they just completely made it up and the fan really is just temperature controlled.  I have a spare Seasonic PSU, maybe I will try it this weekend.

 
Last edited by a moderator:

JDW

Well-known member
Did anyone respond confirming similar behavior?  I have a spare Seasonic PSU, maybe I will try it this weekend.
No one responded, so I look forward to your test results this weekend! :)   No doubt you'll confirm what I show in my video.  I doubt the unit I have is unique.

 

Bolle

Well-known member
I just installed a SSP-250SUB in one of my SE/30s and I am not getting good results at all.

The 5V line is at around 4.8V with just the logicboard and some RAM and no PDS cards. It drops to 4.6V if I add a network card and down to 4.2V (resulting in a non-booting Mac) if I add an accelerator on top of that.

Exchanged the power supply for another unit just to make sure it is not faulty - same results.

I am running a spinning hard drive so there is at least some added load on the 12V rail.

I can not recommend doing this unless some other modification is done to get a stable 5V output... I am going to experiment with additional dummy loads on the 12V and 3.3V rails. This kind of defeats the purpose of a modern power supply though because adding a dummy load to the 12V line is a thing I already have to do on some of my stock Sony power supplies to get the 12V rail back into spec.

 
Last edited by a moderator:

JDW

Well-known member
I just installed a SSP-250SUB in one of my SE/30s and I am not getting good results at all.

The 5V line is at around 4.8V with just the logicboard and some RAM and no PDS cards. It drops to 4.6V if I add a network card and down to 4.2V (resulting in a non-booting Mac) if I add an accelerator on top of that.


I had the same problem until I checked the main wire harness that connects the Analog board to the motherboard.  A couple wires were higher resistance than they should have been.  Since I didn't have the required terminals to merely swap out the wires, I soldered on a wire in parallel with each of the bad wires and the voltage then tested right at about 5 volts at the external floppy connector.  So it could be a high resistance wire (due to age) or corrosion on the terminals inside the connector or both.

 

CircuitBored

Well-known member
I had the same problem until I checked the main wire harness that connects the Analog board to the motherboard.  A couple wires were higher resistance than they should have been.  Since I didn't have the required terminals to merely swap out the wires, I soldered on a wire in parallel with each of the bad wires and the voltage then tested right at about 5 volts at the external floppy connector.  So it could be a high resistance wire (due to age) or corrosion on the terminals inside the connector or both.
This is a problem I've encountered in some late 70s/early 80s micros so there's potential for it to be at least partly to blame here. A complete harness transplant fixed all the issues on my old Heathkit, which seemingly couldn't hold voltage even after a total PSU rebuild. I'm very curious to see how things develop on the Mac side of this issue.
 

As for how or why old wiring becomes more resistive I have no idea. Oxidisation? 

 

jessenator

Well-known member
A complete harness transplant fixed all the issues on my old Heathkit, which seemingly couldn't hold voltage even after a total PSU rebuild. I'm very curious to see how things develop on the Mac side of this issue.
I wonder if this is just good due diligence with machines this age; i.e. if replacing the wiring harnesses all around should be required. I mean, it's all Mini-Fit Jr. (thanks, @cheesestraws) receptacles, right? Even though I really dislike crimping, perhaps this will avoid the problems?

Here are PNs I've found (checked against an AB image and molex'sCAD viewer):
2x5 receptacle: https://www.molex.com/molex/products/part-detail/crimp_housings/0039012100
2x6 receptacle: https://www.molex.com/molex/products/part-detail/crimp_housings/0039012120

2x7 receptacle: https://www.molex.com/molex/products/part-detail/crimp_housings/0039012140

I think JDW references the CRT receptacle and board plug in his analog board re-cap video, because I don't know it off hand :lol:  if that's another one to do out of due diligence.

What wire gauge should one use, if they were to tackle this challenge?
 

Which invites another question I had: the OE harnesses that come with the SeaSonic PSUs have 4-pin molex power coming out of them. Is there any harm in splicing off of them (fan power mainly) instead of using the analog board's connector? My gut says "no problem" because some folks aren't installing a spinner to draw from anyway, however my gut is also not an electrical engineer ;)  

This is all very timely, as I was considering a swap, as well as implementing Bolle's adapter+NIC with an accelerator.

 

cheesestraws

Well-known member
it's all Mini-Fit Jr. (thanks, @cheesestraws) receptacles, right?


Both the AB <-> LB and PSU <-> AB connections are mini-fit jr. on both ends.  I don't have an SE AB out right now to check what connectors the other AB cables are. 

For people who are good at crimping and cable-building, replacing those cables might not be a bad plan.  For people like me who are terrible at it, it is less likely to help ;-) .

 

JDW

Well-known member
What wire gauge should one use, if they were to tackle this challenge?


Answer: The thickest wire that can be used to fit the terminals in a compatible connector.  As such, I do not believe you can go thicker than the stock 18AWG wire.  If you could, that would be great.  The thicker the better, especially for an upgraded SE/30.

 

CircuitBored

Well-known member
I wonder if this is just good due diligence with machines this age
The folks over at the Society of Eight Bit Heathkit Computers certainly seemed to think so. It's fairly normal for cars to need new wiring once they approach 30-40 years old so it figures that computers of that age do too. The fact that those kit computers sometimes required the builder to make their own wiring harness compounds the issue. I've read a few accounts of machines with wiring that was two thirds or even half the gauge it should have been – scary stuff when you're working with old-school CRT boards. 

What wire gauge should one use, if they were to tackle this challenge?
I would strip some of the old wiring and measure the gauge with a micrometer. Obviously you'll want to match both the gauge and the type of wiring with your replacements. Thicker wiring would be advantageous for upgraded systems.

I sure am jealous of the upgradeability of other compact Macs. I have a lowly Classic so the upgrade path is somewhat more complex than any other compact.

 
Last edited by a moderator:

cheesestraws

Well-known member
As such, I do not believe you can go thicker than the stock 18AWG wire.


Yup.  Per Molex's website, the biggest is 18AWG (c. 0.8 mm² for those of us in metric lands) in the normal mini-fit terminals.  You can get terminals to push into the mini-fit jr shells that go down to 16AWG, but I think those require matching high current terminals at the other ends.  (I may be wrong; I am not an expert, just someone who worries too much about connectors).

 

Bolle

Well-known member
until I checked the main wire harness
You sir are the hero of the day.

I was running this on my testbench setup with an extension cable and believe it or not one of the two 5V pins slipped out of its seat in the connector housing and wasn’t making good contact.

After crimping on a new contact everything is now running well. Voltage on the 5V line still is a little bit on the low side with 4.7V, but at least it’s staying there no matter if there’s no card installed at all or if I am running 2 PDS cards + accelerator.

It also seems that while the stock PSU would not be able to run the Turbo040 stably at anything below 4.8V the replacement PSU handles this just fine.

Might still see if I can adjust the voltage a little bit closer to 5V.

 

Crutch

Well-known member
Great tip @JDW.  I have a recapped SE/30 analog board with a Seasonic PSU that isn’t getting the floppy port voltage anywhere near 5V ... definitely going to give replacing the harness a try.  I would not have thought of that, obvious as it is.

(Still haven’t tested out if I can trigger my Seasonic PSU fan with heat .... one day soon!)

 

sos_nz

Well-known member
Loooong time no post, but just to add my thanks for the instructions on this thread and JDW's youtube video.

Had my brother come and stay and so the ol' Mac SE/30 got fired up for some retro games .... she mostly worked, but often nothing at all would happen when the power switch was flicked. Eventually wouldn't turn on at all. Quick multimeter check revealed the old Astec power supply had up and died.

I contemplated recapping, but chose to go with a Seasonic SSP-250SUB (cheapest price for me was RS electronics, $98 NZD, shipped).

I wired up an IEC plug (https://www.jaycar.com.au/iec320-240v-female-line-power-socket/p/PS4006), taking the feeds from the original switch. This results in a secure connection from the switch into the new PS (sure, could have soldered it too). Drilled one hole to screw the power supply to the "floor" of the original cage. Secured the other end with one zip tie.

As others will know, the longest part was removing the many pins and undoing the birds nest of wires to rewire the original 10-pin connector. So far so good. Grounded the new ATX connector's green wire so the new PSU will power on from the original SE/30's power switch. Voltages all working as expected with the PSU on the bench. Time to reassemble everything.

Put it all together and.... nothing. Hmmmm :unsure:. Tested the new PS - no voltage anywhere now. What the???

Turned out I'd mangled a couple of traces just under the new PSU's ATX connector when reassembling the cage. This effectively bricked the new PSU :cry:
Sigh. Big breath.

Reordered another PSU, luckily didn't have to redo any wiring this time, simply moved the new PSU further forward in the original cage. Secured more carefully this time. Reassembled.....

...."BONG" - ALL WORKING 100% :D
 
Last edited:
For testing your voltage levels, make sure you have it under some decent load. I thought my test was OK with my Sony factory Mac SE power supply but I was getting some weird voltages. I had used a random load resistor on each output. Then I changed to a bigger load on +5 and voila, things suddenly looked good so don't kill yourself trying to fix something that isn't actually broken!
compare.PNG
 

s_pupp

Well-known member
I purchased a Seasonic SSP 250SUB. Unfortunately, it has no harness. Does anyone have a pinout of the Seasonic?
 

s_pupp

Well-known member
As it turns out, Crutch’s post from 11/26/18 provides the info I need.

I am not sure, though, of the connector I need. Would this one from Digikey - a 26 pin Molex - work? Are there different 26 pin connectors that are keyed differently? I have always found locating Molex connectors that I need to be challenging.
 

zigzagjoe

Well-known member
As it turns out, Crutch’s post from 11/26/18 provides the info I need.

I am not sure, though, of the connector I need. Would this one from Digikey - a 26 pin Molex - work? Are there different 26 pin connectors that are keyed differently? I have always found locating Molex connectors that I need to be challenging.
I have a SSP-300SUB. It also has a 26 pin connector, I think it is likely the same pinout as the 250. It appears to be a standard Molex Minifit JR as an 24 pin ATX connector fits in just fine. The connector you linked isn't correct, I don't actually see it on digikey on a quick look.

If you have a sacrificial PSU or ATX extension, you could just cut its connector off and use it, the missing pins seem to be an extra ground and +5VSB (standby power). Also a redundant +5v line in the middle of the connector. It does not latch due to lack of tab to catch on, but it fits very firmly so I wouldn't worry about it coming out. This would be the easiest way to proceed IMO, buying all the stuff to crimp for a single connector is annoying. Colors are incorrect, but you can make a tool to remove pins out of a paperclip and re-pin it if that bothers you.

1688859357893.png
 
Top