0
\$\begingroup\$

I have assembled a power supply from five HP dps-750rb common slot power supplies. All five units are in series. I have the first grounded electrically and have the other four units modified and floating (with all cases earthed).

Each of the power supplies contain a pin between the control board and the power board which supplies a consistent 2.500V. By varying it, it is possible to control the output voltage of the supply at the expense of the inbuilt current limiting (I have overcurrent protection on the 0V low side sorted separately), from just under 1V-13V per supply. (Aiming for the assembly to be variable upto 60V or 60A or 3000W max output before my set cutout, which is operational)

I can do this quick and dirty by having five separate pots and controlling each of the five units separately. I HAVE to use separate isolated pots because each of the 2.500V signals is relative to the low side of each of the floating supplies, so at 12V output per supply, the signal is 2.500v for first supply, 14.500V for second, 26.5V, and so on...

From looking, 5 gang linear pots seem impossible to come by...

So I am looking for ideas please, on how best to approach controlling all of these together. I would welcome using a DAC and uC, however I am still unsure what might be the easiest way to achieve my feat?

Thank you in advance. I would provide more details on the supplies but it seems unnecessary as I only wish to achieve the signals required, rather than be concerned with reverse engineering the unit any further. If it would help I can try to provide further information.

\$\endgroup\$
  • 1
    \$\begingroup\$ Your question is missing a link to the datasheet for the power supply. Key to solving this will be the input circuit of the voltage regulation: input impedance, capacitance, is there a voltage source available for the isolated receiver, will it take PWM, etc. Add the info into the question rather than scattered through the comments. \$\endgroup\$ – Transistor Sep 23 at 20:21
  • \$\begingroup\$ Look at digital potentiometers as a starting point, plenty on the interweb on them. You can control them all from a central point, be it push switches or an MCU. \$\endgroup\$ – TonyM Sep 23 at 20:23
  • \$\begingroup\$ Would the logic lines not be raised too high to command from a single point to be able to use separate DACs? If I had a datasheet, or one was even availiable online, I would have linked it, but there isn't. cannot take PWM, I have tried this much. Each board has a voltage supply that can be used to supply any DAC or otherwise I would use on each stage. \$\endgroup\$ – Rendeverance Sep 23 at 20:25
  • \$\begingroup\$ does this pin sink or source current, how is the potentiometer connected, what are the voltage limits on it. there may be some way to do this using simple analogue parts and a single potentiometer \$\endgroup\$ – Jasen Sep 23 at 20:44
  • \$\begingroup\$ Sinks current, at the moment I have it connected between a 12v supply and a gnd on each of the supplies control boards, and the signal pin of the power board. The pot I ended up using is 1k, with a voltage divider resistor to limit the upper range. The swing is 0-5v, though I will limit it somewhere within that range since strange (bad) things happen at the limits. \$\endgroup\$ – Rendeverance Sep 23 at 20:50
1
\$\begingroup\$

build a current mirror with 5 outputs connect one to each supply control pin add resistors to pull the control pin in the opposite diorection.

\$\endgroup\$
  • \$\begingroup\$ I think I am responsible for confusing matters, since the pin on one board obviously sinks and the one on the other sources, it's my mistake if I am confused as to which we referred to. I do not think a current mirror on the low side is possible because they need to see a voltage that is referenced to the supply before. The power board seems to give feedback to the CTL board by means of providing 2.5v with respect to the 0v it sees, which is floating. I presume that the control board therefore is a current source. I am really sorry I appreciate the potential elegance of the proposed solution. \$\endgroup\$ – Rendeverance Sep 23 at 22:12
  • \$\begingroup\$ perhaps put a current mirror on the high side an pull down to "ground" of each supply. \$\endgroup\$ – Jasen Sep 24 at 2:04
  • \$\begingroup\$ I do not mean to sound naive but I have absolutely no idea how I would achieve that? Surely the load is then after the transistor / fet, and the currents would not be the same, only the gate voltage and supply voltage, and would require P type switching? Every practical implementation of a current mirror I can find as a schematic seems to place the load before the switching? Perhaps I am missing something crucial in either my description or understanding of how a current mirror functions, but I cannot visualise it. \$\endgroup\$ – Rendeverance Sep 24 at 8:29
  • \$\begingroup\$ you can use a PNP current mirror to source current or an NPN current mirror to sink current if you can describe the devices in more detail I can explain in more detail \$\endgroup\$ – Jasen Sep 24 at 21:24
  • \$\begingroup\$ I did not realise I could do it high side with PNP and it would still work. This is great I will try and get back to you once I've received a few units. I have hundreds if not thousands of transistors and fets but few PNP that would appear suitable. I shall also update the question if I have further details, which I should in due course. I can't seem to add picture on phone so shall use pc. Thanks again. \$\endgroup\$ – Rendeverance Sep 24 at 22:29

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.