The original "A300" power supply is rated mere 22 Watts and it is just enough for the basic system. Theoretically it should handle memory expansion, 2.5" HDD and one or two external floppy drives. 3.5" HDD and CD-ROM will be too much to handle.
Disclaimer
All information on these pages is provided "as is". The information may be
misleading, inaccurate or completely incorrect. I take no responsibility if
you break, fry (as I did) or in other ways demolish your assets after following
these instructions. These pages contain graphic images of violence towards
valued Amiga hardware. Viewer discretion is advised.
Estimated
power consumption
of A600 Dlx (as of 2018):
| Part | +5V | +12V |
|---|---|---|
| Core | 1.7A | 0.3A |
| Vampire | ? | ? |
| DVD | 1.5A | - |
| Ext FDD | 0.7A | - |
| TOTAL |
Cisco 2514 PSU is quite compact switching power supply. It has handy detachable mains cable e.g. for UPS use.
For the safety regulations the PSU should be enclosed into appropriate casing with active cooling,
because it has live voltage exposed thru ventilation holes.

The PCB has three +5V "lines", but I am going to use only one wire of 5x0.75 mm2 cable and 5-pin DIN connector.
DIN connector intended for audio may not be rated for 5 amperes, but I am hoping it will suffice in occasional use.
Someone may wonder why I am not simply using complete cable from the original Amiga PSU. I have cut the power connectors
for my previous tower project, so I do not have any more complete cables.

Because I was little worried about throughput of 5-pin DIN connector, I decided to double its performance by using 8-pin connector.

The GND, +5V and +12V are now connected via double pins. Shield and -12V are using just one pin as before.
After seeing some videos on the Tube about assembling modern switching PSUs inside the original C64 PSU cases, I fancied marrying the Cisco PSU and original Amiga PSU case together.

Looking good so far. Sadly I forgot to insert the thin film insulator between heatsinks, so I short circuited the mains and low volt sides. -> Kaputt!
After the previous mishap, I needed a new supply of sweet power for my loved one. Many have had success stories with Mean Wells. PT-65B should fit nicely inside.

For secure mounting, I recycled some leftover motherboard standoffs.
Drilled 2.5mm hole and forcibly screw in the standoffs.
Then I put a blob of super glue and tighten the standoffs to the final position.
I also cut off one of the cover screw pole thingies and covered the hole with piece of genuine Commodore plastic from a C64 PSU.

Fits nicely inside. The PT-65B has 2 pins for both +5V and GND and I used only one pair.
The Molex SPOX is rated at 11A, so theoretically it should be enought.
As an halloweenish aftertought I added bright red led to the spare pins to give me some spooky glow when the PSU is powered on.

Suitable for Vampire powered Amiga.
I have had some intermittent problems about system freezing randomly and keyboard not recognised on some boots.
It was just putting caps lock on.
After I swapped current tray model DVD-ROM to slot-in DVD-burner, the caps lock led started flickering simultaneously with ODD seeking.
And the drive was making horrible noises and was super slow.
This sure was indicating some power problems.
I hooked The Mother of all Amiga PSUs (from Cisco 4000) outputting beefy +5.2V, 24A --> no problems with keyboard and ODD was fast and silent.
I connected the Mean Well back and measured voltages around places.
| Point | +5V | +12V |
|---|---|---|
| Mean Well | 5.0 (idle) 4.9 (load) | . |
| Floppy/DVD | 4.6 (idle) 4.4 (in-use) | 13.0 |
| Keyb IC | 3.8 | . |


| Activity | Amps |
|---|---|
| Complete idle System | 2.6 |
| Floppy/DVD inserted | 2.9 |
| XCopy formatting | +0.3 (int) +0.4 (ext) |
| Creating OS3.9 EBD | 3.8 |
I found my unaltered A600 light-weight PSU, which still had complete cables.
So I was dreaming of less destructive retrofit of the internals.
The A300 PSU has lightly different layout than A500 PSU and PT-65B was just about 2mm too wide.
I dremeled some bits off from the PCB and top cover screw poles - nice and tidy fit.
Please note those brilliant mounting brackets made with Artesan Anti3D Pro printer (Dremel).

For safety reason, I changed the SPST layout mains switch to DPST one.
Because of nonpolarized Schuko plug, the single pole switch does not always cut the Line wire depending the orientation of the plug.
Thus it is better to cut the both L and N.
Some cheaper PSU have some leakage or design faults causing unwanted voltage if Line and Earth are live and Neutral is cut off.
To minimise voltage losses, I cut the output cable to half lenght.
Also overclocked the +5V output to 5.2V to compensate, which is within the tolerance with lighter loads.
As a side-effect, this caused the +12V output rise to +13.3V - way out of the specs.
The PT-65x datasheet states current range of 0.2~3.2A for the +12V line.
AFAIK the A600 has basically no load on ±12V, just audio amp and serial port, so probably the minimum 200mA load is not achieved.
I got to the conclusion, that some kind of dummy load is needed.
I started with moving the LED from the +5V rail to the +12V rail, increasing its load by 10mA, and soldering a 180R 5W resistor.
The resistor should cause about 70mA extra load (and 0.8W extra heating).


Sadly this was not enough and voltage was still +13.0V.
I experimented with diffrent loads including few car light bulbs and the best case that I got was about +12.3V.
At the end I tamed the +5V rail overboost a notch and added second 180R resistive 70mA load, totaling 150mA.
For a sidenote: Do not put hot resistors next to each other, because you will get a 1.6W spot heater.
After a few minutes power on I measured 47'C surface temperature outside bottom case even with the top cover removed.
Please give those resistors some air to breath and the plastic will be thankful.
I settled to Amy getting about +4.8V, +12.6V and -12.1V when she is idling.
These are inside reasonable limits.

Greenlighted 100R for the OG.
To my opinion, the PT-65B is not the perfect solution for this kind of +5V biased load.
According to the datasheet's block diagram, all voltage regulation is based on the +5V rail.
When the load on +5V increase and voltage drops, the PSU compensates this by increasing ALL voltages, as the other rails do not have their own PWM/switching circuit.
Similar behavior can be seen with the infamous Topfield STB-PVR PSUs when their crappy caps glued to heatsink go bad after 2-3 years.
The PSU cannot keep up the +5V under a stress like recording and start increasing the voltages, including the +12V.
My box output +16V to the HDD, which was not a good thing on a long run.
AFAIK the PT-65B is not designed to be powered on without load.
Just for curiosity, when there are no load on the +5V rail and some (dummy) load on the +12V rail, the +5V output is excatly at adjusted +5.0V and the +12V is below designed, like +11.9V.
When a load (2.5A) is introduced to the +5V rail, it drops to +4.9V and the +12V rail swings up many many decimals, up to 13V.
Maybe a better option would be using a more expensive PSU with invidually regulated output, or one with less powerfull +12V rail.
Or really DIY persons could overdrive the +12V rail and put inline a 7812 linear regulator?
Jump to
Part 1: Base system
Part 2: ATA Connections