Server Powersupply mod

VK3ZYZ

Moderator
Staff member
For the last couple of days I've been trying to increase these server power supplies from 12V to 13.8V.
M1Z-5500V_1.jpg
M1Z-5500V_2.jpg

M1Z-5500V_3.jpg


You can see from the ratings, these would make a pretty good radio power supply.
12V at 40Amps is quite nice, in a 240x100x40 case.
M1Z-5500V_4.jpg

M1Z-5500V_connections.jpg

But, so far, the max volts I can get out is 12.4V

The control PCB has the trim pots marked, VR12V, VR12C, VR3V...... so that was a help.
Here you can see some of my attempts at raising the voltage.
M1Z-5500V_pots.JPG


The circuit is like this...

M1Z-5500V_Rs.jpg

As there are 2 places for R42, and it is 12V, I'd hoped changing the 12K to 13.5K (2 x 27K) it would magically give me close to the sought after voltage.
But, no.
So, the next try was to just drop the 12V into the control board by cutting the 12V feed and adding 2 series diodes inline.
M1Z-5500V_2diodes.jpg

Still, no :(
So far, that is the end.
 

VK2RK

Member
I modified an HP PL18 unit for 13.8 V, the mod was to link out a 100 Ohms resistors in the sensing network.
The normal assumption is to raise the feedback voltage, I think these units work the other way, so I would try lowering the voltage using the R41.

One way to test my theory is to measure the pot value from one extreme to the other noting the higher voltage, if I am correct the highest voltage will be achieved with the lowest pot value.
IMAG4143-20220412-154119488.jpg
 
Last edited:

VK3ZYZ

Moderator
Staff member
Yes Rob.
I have tried both increasing the top resistor and decreasing the bottom one in the voltage sensing string.
Either way, something else is limiting the max volts.
In other power supplies I've played with, the over voltage limit shuts the power supply down, but maybe this one is different.
More testing and circuit tracing is needed.
 

VK3ZYZ

Moderator
Staff member
At the moment, a front is being printed for the PS.
Here is one, but the USB sockets do not fit.
M1Z-5500V_front.jpg

M1Z_5500V_front2.jpg
PS_Front_printing.png

It will have 3 x sets of the 12V outputs and 4 x USB sockets. Also, a power switch and LED.
 
Last edited:

BillC

Active member
The printed case makes a really professional looking job. Good on the pot, I hope you can get the voltage up to where you want, looks as though it will be a beaut. power supply. Is the PSU. to be mounted under the desk? , could make a good legwarmer??//.
 

VK3ZYZ

Moderator
Staff member
M1Z-5500V_controlBoard.png

A small win..
The control board unplugs :)
So, as shown by the sheet the board is on, I've printed it out over 3 x A4 sheets for each side, so maybe I can see to trace the circuit!

And yet another cover design (#5???) is to be printed.
This will have bonding posts for the voltages.
M1Z-5500V_cover_5.png
 

VK3ZYZ

Moderator
Staff member
M1Z-5500V_cover_5_printed.jpg

This is what the printed one looks like.

Upon fetching another PS to see how it fits, I find they are different.
R2Z-6500P-R_specs.jpg


No edge connector on this one.
R2Z-6500P-R_connector.jpg


The outside cases look the same.
 

VK3ZYZ

Moderator
Staff member
I am really having trouble tracing this board out. It is so compact and tightly loaded. Also, the parts are quite small.
The biggest trouble is trying to probe from one side to the other.


So, maybe this will help?
TestProbe1.jpg

TestProbe2.jpg
 

VK3ZYZ

Moderator
Staff member
This is not a lot to show, but here is the control circuit so far.
M1Z_5500V_control20220418.png


There are still connections I have not found.
 

VK3YNV

Administrator
Staff member
Interesting CCT, would it be worth trying a couple of diodes in series with D6, so that the output has to go higher before D6 turns off.
 

VK3ZYZ

Moderator
Staff member
I've not tried adding diodes to the outputs as yet.
One thing that I am puzzled about is the OpAmp outputs need to go low to turn the diodes on. So the OpAmps may have a -ve supply.
What the OpAmps drive via the main board is still a mystery too. This is not the easiest circuit I've traced out, that is for sure. There are still things that go to points unknown!
The mods I've tried is to increase R42, decrease R41 and add a couple of diodes in line to drop the sensed voltage from pin 39.
And I have not measured the voltages on the board so I do not know Vref or the OpAmp supply voltages as yet.
You can see IC1 in some of the photos. It is at the bottom left and looks like a surface mount power transistor. The type number is not known.
A few wires will need to be soldered on to get access to measure them. Maybe get that done today?
 
Last edited:

VK2RK

Member
Interesting CCT, would it be worth trying a couple of diodes in series with D6, so that the output has to go higher before D6 turns off.
D6 is on all the time, this is part of the analogue feedback the voltage at point 38 will raise and fall in accordance with the required regulation
The swing of the opamp will be limited by the reference set point, not knowing the actual Vref voltage is a bit difficult to say in what direction to go, however since the voltage will not raise above a voltage slightly greater that 12 Volts I am confident to say that the reference is the issue so designed to also act as an overvoltage protection.
Altering R47 is not a good thing to do as this will alter the gain of the opamp that is set at 4.25, finding the source of Vref and seeing what goes on there in my opinion is the best path in this rabbit hole
 
Top