Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Make a manufactured SMPS to actually deliver 12V

Status
Not open for further replies.

daokj

Newbie level 6
Joined
Jan 18, 2021
Messages
14
Helped
0
Reputation
0
Reaction score
0
Trophy points
1
Activity points
172
Hello,

I am trying to reuse a SMPS coming from an old DVD player with internal hard drive. It has among other lines a 12V line (when not feeding anything, multimeter says it actually outputs 11.3V) and a 5.4V (multimeter says 5.2V).
In the original device, those two lines are used to feed a 3.5" HDD - 5V for the HDD PCB and 12V for its motors.
However I did not keep the original PCB to connect a HDD and in my use case I am willing to use the 12V line to feed such **broken link removed**) which is meant to be plugged to 3,5" HDD (for test purposes I use the one of the original device) and a Raspberry Pi.

Now, when I use the power supply which came with the converter (multimeter says it outputs 12,2V), the HDD can be recognised by the RPi.
But when I use the above mentionned SMPS (basically through a modified jack connector), the drive is not properly recognised (and its motors do no seem to work even though the LED of the SATA to USB converter is lighting up).

The converter should get a 12V input, from which it manages to feed the drive in 5V and 12V via the SATA connector.
So I guess the actual voltage of the current supplied by the SMPS' 12V line is too low. From here I wonder (novice questions):
  1. How comes that it managed to successfully feed the same hard drive in the original device (even though the 5V was supplied by a separate line, and not converted from the 12V as it is now with my SATA to USB converter) ?
  2. When not feeding anything, the 12V line outputs 11,3V. Once the SATA to USB converter is plugged (with nothing on the other end), i measure 10,9V. Once the hard drive is plugged to the converter, I measure 10,3V. Is the tension supposed to drop like this ?!
  3. You will find here attached the schematics of the SMPS in 2 parts. I'd rather use that SMPS than the power supply shipped with the converter (which works fine) since I need other lines from the SMPS (-30V, VF+, VF-) to feed other modules from the original device. What would be the best way to get a tension closer to actual 12V ? Creating an additional rectifying circuit to actually feed 12V to 12,2V ? From the 12V line (named PC_12V where the 12V is made from 15V using 278R12, but currently outputting less than 12V)? From the 33V line?
Thank you for reading
 

Attachments

  • SMPS1.png
    SMPS1.png
    248.9 KB · Views: 86
  • SMPS2.png
    SMPS2.png
    75.4 KB · Views: 85
Last edited:

What do you have connected to the '_CTL' lines?
It isn't unusual for voltages to drop as the load on the outputs increases, it may simply be that you are trying to draw more from the PSU than it can manage. Consider that the manufaturer made it for a specific purpose so they wouldn't have allowed for extra loads.

Brian.
 

So far I do not have anything anything connected to the _CTL lines. Now you mention them, not sure if this is normal but when trying to measure the voltage with a multimeter between GND and PC_CTL it makes a light crackling noise (among all the output pins of the PSU, it only happens with PC_CTL).

Good to hear that voltage drop isn't unusual - that does not mean the PSU is broken.
I measured the pins again, this time also disconnecting the fan (FANCN1) which goes with the SMPS (and which is fed by line PC_12V as well). Then only I can measure a 12V sharp output (but if then I plug my HDD, it does not work).
With the same set up, I have also measured the voltage on the pins of the 278R12 (12V 2A output low drop voltage regulator) of which I provided a link to the datasheet in my previous message. With reference to pin 3 (GND), I get I measure pin 1 (IN) or 2 (OUT) with reference to pin 3 (GND), I get 12V for output and 12,3V as input (when it is supposed to be 15V as per the schematics, but I guess it is not a big issue as long as the regulator can actually outputs 12V). In addition, I double checked the schematics up to the transformer for the 15V line and I noted a difference - on my PSU there are no L1S32 ("R-METAL OXIDE) and C1S34 (330uf capacitor). The rest is as per the schematics.

I am puzzled regarding your point on the load. So far I actually tried to draw less than at the time of the original device (only HDD and fan when in the original device the same PSU was feeding in addition to that several boards, a DVD-ROM reader, etc.). What is more, the voltage regulator used to provide the 12V (278R12), is supposed to output 2A, when the PSU shipped with the SATA to USB converter (with which I don't have such issue) is also rated 12V 2A.
 
Last edited:

So in the meantime I have been told that those kind of multi voltages SMPS have a of mechanism which makes them use the 5V to "compensate" the 12V (I can't explain, but would be interested to understand if someone can). Meaning that I needed to draw a certain amount on the 5V line to find the 12V properly regulated.

I used a 2N2222 NPN transistor (hfE 75) as in the schematics below to draw about 430mA on the 5V line... and the hard drive starts and gets recognised by the RPi!
 

Attachments

  • PSU 5V load to compensate 12V regulation.png
    PSU 5V load to compensate 12V regulation.png
    5.3 KB · Views: 76

That will likely destroy the transistor! Have you swapped the resistor values by chance?

The idea behind 5V 'compensating' the other supplies is that as they are all derived from the same transformer they will all rise and fall simultaneously. That means that is any single one is used to monitor the voltage and provide correction, the rest should fall into line. It isn't a perfect system as each output has it's own individual drops from rectifier and filter components but it is a good compromise, giving a degree of regulation without having to fit a regulator on each output.

Brian.
 

    daokj

    Points: 2
    Helpful Answer Positive Rating
That will likely destroy the transistor! Have you swapped the resistor values by chance?

No. Hem... indeed I read the 2N2222 datasheet too quick - I can now see peak base current shall be max 200mA (fortunately the transistor did not seem to compain receiving about the double during my test which lasted a few minutes). For the final implementation I shall find either a higher resistance (like 22ohm -> 193mA) for the the base pin, or another transistor if this amount of current drawn on the 5V line is too low to have the 12V line correctly regulated.

The idea behind 5V 'compensating' the other supplies is that as they are all derived from the same transformer they will all rise and fall simultaneously. That means that is any single one is used to monitor the voltage and provide correction, the rest should fall into line. It isn't a perfect system as each output has it's own individual drops from rectifier and filter components but it is a good compromise, giving a degree of regulation without having to fit a regulator on each output.

Understood, so I believe that is the point of bringing the 5V to the ICE3B2065J that we see on the first page of the SMPS' schematics.

Thank you for your follow-up.
 

Cheap supplies will regulate one output and depend on
"cross regulation" for ones deemed less critical. This is
not great, but often acceptable when cost is king and
nobody's looking.

But 11.2V is more than 5% out of regulation and maybe
some element of the assembly does an undervoltage
lockout or a logical check of some "PGOOD" signal
indicating in-regulation.

It ought to be easy to substitute a bench power supply
and determine the "not-dead envelope" of the 12V input.

If you want more voltage just on the 12V output you
might consider putting some power Schottky rectifiers
in parallel with the ones currently in the 12V output path,
this ought to net you a few hundred mV under load.
But you need to watch the Schottky I-V curve as they
do not have the conductivity modulation and Vf will
cross over at some point of If.

You could also adjust the main feedback, let 5V rise
to "whatever makes 12V centered" and add a good
6V-input LDO to drop the "5V" back to 5.00V with
not too much added inefficiency.
 

Yes, thats right. You are basically asking why your output cross regulation is not perfect, and it never will be.
You can do stuff to improve it....eg, you can use shared proportinal feedback from the 12v rail and the 5v rail.......basically have an upper divider resistor coming from each rail. Power integrations have a good app note on it.
But in your top schem, i dont see pin 11's connection (transformer)...it doesnt look connected.
Also, the bias coil doesnt seem to be connected to ground.
 

Thank you for the feedbacks.

@dick_freebird , what is the reference of the Schottky diodes you see on the 12V output line currently (I can't see them)?
Beside the waste of current, the second solution you mention sounds easier to me but I don't get what would be the role of the LDO here? I actually need a low amount of 5V current, most of the amount to be drawn would probably be a dummy load to have the 12V "centered".
Additionally it is difficult to me to fully understand what mean in this case since the voltage I measure on the lines depends on what is connected to them. For instance, during the dangerous test for my 2N2222 transistor, I measured the voltage while it was connected, and got something lower than what it is when empty (not sure how much anymore but of the like of 0,4V less).

@cupoftea
in your top schem, i dont see pin 11's connection (transformer)...it doesnt look connected.
Also, the bias coil doesnt seem to be connected to ground.
Should I worry about that ? I can't see the consequences of it.
 

Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top