LM5007: How to reduce a very high Pk-Pk voltage on the output?

While it is possible to make a working dc/dc converter on a breadboard circuit, generally you will need a well designed actual printed circuit board to make your design work correctly.

Thanks for the reply John!

However at the moment it's not feasible to make a separate PCB specifically for the power supply of my PD.
Is there any other indication that comes to mind which I haven't tried yet? Or are some calculated values at fault here?

I realize that I might/will not get a perfect output via the breadbord, but I'd go for something to fix the Peak voltage down to at least 2-3V. This way I can confirm my circuit's functionality and know where/how to tackle the problem.

Flint

We would have to know all your inputs (Vin, Vout, Iout) to recreate this design in Webench. I do not directly support LM5007, but I see these type of posts with some regularity, where users cannot get their circuit to function properly with point to point hand wired circuits. If you build this on a double sided copper board with the bottom as a dedicated GND plane, and pay close attention to the component placement and wiring, you may possibly get your circuit to work. Another option would be to purchase the official TI EVM, and modify it to match your schematic.

As far as my input and output goes:

Input: (at the moment) Power supply EST 150, giving 36-50V (but should be 36-57V)
Desired Output: 12V, 0.500A

Thanks for your other suggestions, I think that I might try the double copper board if possible.
After thinking a bit I came across the idea of placing a zener diode at the end to limit the peak value. This could also work right?

Flint

Actually what you said about recreating the design in WEBENCH got me thinking (didn't know it was possible to change the values), so I tried it myself and entered almost all components with nearly the same characteristics as the TO-components that I use on my breadbord. After simulating a few times and comparing with the default values that WEBENCH itself calculates based on my inputs and outputs, I found that 1 major factor influencing the peak voltage is the ESR.

I played with the ESR of Cout and noticed that I had a TO-capacitor with the correct capacitance but with an ESR of 13 Ohm instead of ~4m Ohm (which WEBENCH default uses). Changing only the Cout based on the ESR gave me a peak voltage of almost 2V with 13 Ohm ESR compared to 100mV with ~4m Ohm ESR.

In the WEBENCH altered schematic above I've circled the Cout. In this screenshot I had the 13 Ohm ESR capacitor selected. You can see the effects in the plots below when I use a 13 Ohm and ~4m Ohm ESR. 

First up is the 13 Ohm ESR voor Cout:

In Second place we get the ~4m Ohm ESR voor Cout:

These results made me realise that I have a bunch of other capacitors/components which have a different or unknown ESR and all together could be the main (or one of the) culprit to cause such an abnormal Peak value. So i'm going to try out some SMD components with low ESR, which WEBENCH suggests, on a double sided copper board. 

I'll post my findings once again when it comes to that. If there happen to be any other suggestions, I'd be happy to read them. 

Flint

After some time I managed to build the WEBENCH suggested circuit on a single sided copper board with Through-hole and SMD components combined (only resistors are TO, de rest are SMD). 

WEBENCH suggested circuit with a 36-57V input (PoE) and a desired output of 12 V & 500mA:

With a 100 Ohm load:
Output: ~4.38 V
Output Pk-Pk: ~2.24 V

WIth a 4.7 kOhm load:
Output: ~7.3 V
Output Pk-Pk: ~2.08 V

It seems that somehow I'm not getting the ~13V Pk-Pk value anymore. However it still there with a lower value of 2.5 V. Any idea's to fix this?

Furthermore, I've designed this circuit with WEBENCH to produce 12 V on the output (and 500mA). I'm confused to see that the output voltage seems to be depended on the load resistance (as you can see in my results above in this post)). Shouldn't it be giving a steady output value of 12 V??

Any suggestions??

Flint

Hi Flint,

 

I strongly suggest you get this design on a PCB, wiring the circuit on a breadboard will not work.  Even on a PCB please be careful to follow closely the layout guidelines in the datasheet, this is also strongly advised.  Failure to do so will likely end in a design not working correctly. 

Thanks for your reply David.

I guess there's no way around it then, however since time is of the essence...I'm considering to buy an EVAL board of the lm5007. However, it's configured for 10V 400mA output. Is it allowed/meant to change/replace parts on the board to get and test my desired output?

Plus, I think that the LM5007 might not be able to supply enough power for my application after all....but I'll make a new topic for that.

Hello Flint,

You can certainly adjust the output to 12V by changing the Feedback resistors, you should be ok for 500mA too.