ATX PSU in original Amiga PSU casing

TTSAddict · 25 February 2018, 21:37

Using an ATX PSU to power Amigas is propably old news for you folks, but I'd like to share this specific one because it uses the original PSU casing.

I wanted to do this for a long time and finally mustered up enough energy to tackle this.

When I was browsing my junk, I came across this specific PSU from Zalman, a ZM350-FX. I don't know why I bought this in the first place, I only remember it being a mispurchase originally. Later I powered a router via its 12V rail, it didn't even get lukewarm; the fan is temperature regulated, and it uses a rather small scale design. Looks like its an ideal candidate.

Removing the guts from its casing didn't proof as straightforward as one might imagine. There are parts for interference supression soldered directly to the kettle jack, and they also physically prevent removal of the board.
Since I don't need the casing anymore, I just bent that side down to make life easier while unsoldering, which was very tricky. The little legs of the parts were intertwined with the jack before soldering and I didn't want to break or cut them too short, so this was a hot job, which also damaged my solder tip, because I'm stupid. (Used the tip to push out the pins, which scratched the coating). Additionally, I had to cut out a part of the shielding around the jack, which was soldered to the earth/ground pin.

Testing the size

I hot glued the fan to the cover, with the pushing side up. Don't know if that can squeeze anything through the small vent holes, but I don't think it will need the fan at all anyway.

Originally I intended to use the original switch as a breaker of the mains side and an additional switch for the PS_ON signal. Also instead of hooking the red LED to 3,3v, I wanted to hook it to 3,3v_standby, a green LED to the PWR_GOOD signal and a third one directly to the fan PWM as a... kind of temperature warning, but I scrapped all that and just hooked one red LED to the active 3,3v rail. Also the original switch now just operates the PS_ON signal. The ZM350-FX originally had no power switch either. I should propably measure the standby power consumption NOW and revert to my original plan but hey...

This is the current state:

So I still need to solder the mains side. Note that I rotated mains and secondary, since the switch operates a 5V signal now. I also needed that GN/Y wire for the shield pin, which was actually connected to ground in the original commodore design.
I will update this thread once when I'm done. I will not test this contraption on my A1200 for obvious reasons, but I have two A500 guineapigs that can put this PSU to the test.

Oh btw. I forgot to mention, the board is held down by hot glue as well.
Also, here is the complete imgur-album

Final state of the mains side:

Closed lid:

The thing powers up and gives out the expected voltages without a load/Amiga connected. Now I'm getting a reading of 80V AC from all rails to the shield pin/earth which stumps me a bit. Need to think about how to proceed before connecting anything to that.

~
I think I might need to connect the secondary side ground to earth as well, I presume it was like that when it was housed in its original casing through the chassis. Measuring any rail of my PC PSU to ground gives the intended DC voltage without ripple, so I think thats the norm. The current reading of the Zalman is 0V DC to earth along with the 80v AC.
~
Yes, looks like you need to manually connect the secondary side GND to earth, when running it outside its metal case.
Final modification:

Voltages are okay now. Now for a test run.

daxb · 26 February 2018, 11:50

The PSU is completely oversized for an A500. Around 450W with 350W on 12V line. With some luck it will work but in a very bad operating range.

TTSAddict · 26 February 2018, 13:52

Yeah, I was anticipating that, but its quite hard to find modern ATX PSUs in a lower output range.
A pico ATX with an additional 12v source would have been an option, I didn't do my research.
Will finish this nonetheless, because *after* the test, it is supposed to power my A1200 with a 030 card, so it shouldn't be as critical as with a 500.

Sinphaltimus · 26 February 2018, 13:56

I like the mod. Nicely done. You can throw a load on it if you don't have a lot of things running off the PSU. I've read about attaching large resistors to it someplace to create a big enough load in order to stabilize the PSU.

daxb · 26 February 2018, 14:14

Quote:

Originally Posted by TTSAddict

Will finish this nonetheless, because *after* the test, it is supposed to power my A1200 with a 030 card, so it shouldn't be as critical as with a 500.

I fear that doesn't change much. Maybe 10-20W more power consumption but good luck that it will work for you.

Quote:

I've read about attaching large resistors to it someplace to create a big enough load in order to stabilize the PSU.

Only the stupid do this. Sorry to all the stupid.

Really, if you start thinking about a PSU replacement there are better solutions then ATX.

demolition · 26 February 2018, 15:10

Quote:

Originally Posted by TTSAddict

Yeah, I was anticipating that, but its quite hard to find modern ATX PSUs in a lower output range.

Doesn't have to be ATX? This industrial supply is very suitable:
https://www.distrelec.de/en/switched...50b/p/16905972

Sinphaltimus · 26 February 2018, 16:16

Quote:

Originally Posted by daxb

I fear that doesn't change much. Maybe 10-20W more power consumption but good luck that it will work for you.

Only the stupid do this. Sorry to all the stupid.

Really, if you start thinking about a PSU replacement there are better solutions then ATX.

I dunno why you would say that. ATX PSU (especially older ones) are pretty well suited. I've been running on an ATX PSU for about 6 months now without any issues. And I have not added any resistors. I have a collection of ATX PSUs so not only is it A solution, but it's no additional cost and with plenty of spares that I never have to worry about it ever. Better solutions out there, sure I bet there is. However, for the folks who have done this AND used the resistor as a stabilizing method as needed, what makes them stupid? Pretty harsh statement there based on opinion.

TTSAddict · 27 February 2018, 14:00

Final state of the mains side:

Closed lid:

The thing powers up and gives out the expected voltages without a load/Amiga connected. Now I'm getting a reading of 80V AC from all rails to the shield pin/earth which stumps me a bit. Need to think about how to proceed before connecting anything to that.

~
I think I might need to connect the secondary side ground to earth as well, I presume it was like that when it was housed in its original casing through the chassis. Measuring any rail of my PC PSU to ground gives the intended DC voltage without ripple, so I think thats the norm. The current reading of the Zalman is 0V DC to earth along with the 80v AC.
~
Yes, looks like you need to manually connect the secondary side GND to earth, when running it outside its metal case. Voltages are okay now. Now for a test run.

TTSAddict · 15 April 2018, 17:01

So I have this in operation for some time now without any problems on my A1200 and on three different 500 models during their maintenance. It even saved one of the Amigas life twice due to its short protection, since I had a faulty floppy power cable.

25 February 2018, 21:37	#1
TTSAddict Influence Device Join Date: Nov 2005 Location: Germany Age: 44 Posts: 75	ATX PSU in original Amiga PSU casing Using an ATX PSU to power Amigas is propably old news for you folks, but I'd like to share this specific one because it uses the original PSU casing. I wanted to do this for a long time and finally mustered up enough energy to tackle this. When I was browsing my junk, I came across this specific PSU from Zalman, a ZM350-FX. I don't know why I bought this in the first place, I only remember it being a mispurchase originally. Later I powered a router via its 12V rail, it didn't even get lukewarm; the fan is temperature regulated, and it uses a rather small scale design. Looks like its an ideal candidate. Removing the guts from its casing didn't proof as straightforward as one might imagine. There are parts for interference supression soldered directly to the kettle jack, and they also physically prevent removal of the board. Since I don't need the casing anymore, I just bent that side down to make life easier while unsoldering, which was very tricky. The little legs of the parts were intertwined with the jack before soldering and I didn't want to break or cut them too short, so this was a hot job, which also damaged my solder tip, because I'm stupid. (Used the tip to push out the pins, which scratched the coating). Additionally, I had to cut out a part of the shielding around the jack, which was soldered to the earth/ground pin. Testing the size I hot glued the fan to the cover, with the pushing side up. Don't know if that can squeeze anything through the small vent holes, but I don't think it will need the fan at all anyway. Originally I intended to use the original switch as a breaker of the mains side and an additional switch for the PS_ON signal. Also instead of hooking the red LED to 3,3v, I wanted to hook it to 3,3v_standby, a green LED to the PWR_GOOD signal and a third one directly to the fan PWM as a... kind of temperature warning, but I scrapped all that and just hooked one red LED to the active 3,3v rail. Also the original switch now just operates the PS_ON signal. The ZM350-FX originally had no power switch either. I should propably measure the standby power consumption NOW and revert to my original plan but hey... This is the current state: So I still need to solder the mains side. Note that I rotated mains and secondary, since the switch operates a 5V signal now. I also needed that GN/Y wire for the shield pin, which was actually connected to ground in the original commodore design. I will update this thread once when I'm done. I will not test this contraption on my A1200 for obvious reasons, but I have two A500 guineapigs that can put this PSU to the test. Oh btw. I forgot to mention, the board is held down by hot glue as well. Also, here is the complete imgur-album Final state of the mains side: Closed lid: The thing powers up and gives out the expected voltages without a load/Amiga connected. Now I'm getting a reading of 80V AC from all rails to the shield pin/earth which stumps me a bit. Need to think about how to proceed before connecting anything to that. ~ I think I might need to connect the secondary side ground to earth as well, I presume it was like that when it was housed in its original casing through the chassis. Measuring any rail of my PC PSU to ground gives the intended DC voltage without ripple, so I think thats the norm. The current reading of the Zalman is 0V DC to earth along with the 80v AC. ~ Yes, looks like you need to manually connect the secondary side GND to earth, when running it outside its metal case. Final modification: Voltages are okay now. Now for a test run. Last edited by TTSAddict; 27 February 2018 at 18:45.

27 February 2018, 14:00	#8
TTSAddict Influence Device Join Date: Nov 2005 Location: Germany Age: 44 Posts: 75	Final state of the mains side: Closed lid: The thing powers up and gives out the expected voltages without a load/Amiga connected. Now I'm getting a reading of 80V AC from all rails to the shield pin/earth which stumps me a bit. Need to think about how to proceed before connecting anything to that. ~ I think I might need to connect the secondary side ground to earth as well, I presume it was like that when it was housed in its original casing through the chassis. Measuring any rail of my PC PSU to ground gives the intended DC voltage without ripple, so I think thats the norm. The current reading of the Zalman is 0V DC to earth along with the 80v AC. ~ Yes, looks like you need to manually connect the secondary side GND to earth, when running it outside its metal case. Voltages are okay now. Now for a test run. Last edited by TTSAddict; 27 February 2018 at 17:52.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
A4000 PSU -> Mini-ATX PSU mod	keropi	Hardware mods	53	09 December 2019 03:52
Best ATX PSU for Amiga?	PaulG	support.Hardware	7	28 January 2017 12:18
ATX PSU gradually lowering voltages to my amiga when turned off!!!	R-Typer	support.Hardware	6	29 November 2014 18:19
Decent ATX PSU for Amiga 4000	CU_AMiGA	support.Hardware	1	31 May 2011 22:27
ATX PSU 2 Amiga 500 case tutorial now up	r6stu	Hardware mods	12	20 January 2008 19:14

26 February 2018, 11:50	#2
daxb Registered User Join Date: Oct 2009 Location: Germany Posts: 3,303	The PSU is completely oversized for an A500. Around 450W with 350W on 12V line. With some luck it will work but in a very bad operating range.

26 February 2018, 13:52	#3
TTSAddict Influence Device Join Date: Nov 2005 Location: Germany Age: 44 Posts: 75	Yeah, I was anticipating that, but its quite hard to find modern ATX PSUs in a lower output range. A pico ATX with an additional 12v source would have been an option, I didn't do my research. Will finish this nonetheless, because after the test, it is supposed to power my A1200 with a 030 card, so it shouldn't be as critical as with a 500.

26 February 2018, 13:56	#4
Sinphaltimus Registered User Join Date: Aug 2016 Location: Cresco, PA, USA Age: 53 Posts: 1,126	I like the mod. Nicely done. You can throw a load on it if you don't have a lot of things running off the PSU. I've read about attaching large resistors to it someplace to create a big enough load in order to stabilize the PSU.

15 April 2018, 17:01	#9
TTSAddict Influence Device Join Date: Nov 2005 Location: Germany Age: 44 Posts: 75	So I have this in operation for some time now without any problems on my A1200 and on three different 500 models during their maintenance. It even saved one of the Amigas life twice due to its short protection, since I had a faulty floppy power cable.

Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)