Continue to Site

Eng-Tips is the largest engineering community on the Internet

Intelligent Work Forums for Engineering Professionals

  • Congratulations KootK on being selected by the Eng-Tips community for having the most helpful posts in the forums last week. Way to Go!

stablility of dc to dc converter

Status
Not open for further replies.

UtrasoundGuy

Bioengineer
Nov 18, 2013
21
we have the following DC to DC converter.
It has been in use in our company for around 10 years without any major problems Im aware of.

The converter is capable of converting 15V DC to -200 to -400V DC (depending on trimpot turning)
Since Ive been working here for the last 2 years, we have been trying to get this device produced by 2 different vendors (one 2 different times) and the PCBs keep failing our incoming inspection.
The failure is that the output voltage is not stable at -220 and -400V which are test points in our test procedure.
An example of the output failure over the converter is shown below (AC coupling on oscilloscope is used to remove the -400V DC), the time is in seconds and volts are actually 10x what is shown and are 20-25V peak to peak fluctuation

After the tests are performed, and if everything checks out, the device then gets set to -300V where it gets used.
We are hesitant to use these PCBs if they are not passing inspection though,.
We are not sure why there are now difficulties with stability or what has changed.
Maybe the transformer is different although its speced the same but Im not sure how this would cause oscillation in the output waveform?

I believe part of the difficultly in debugging this is that it is a feedback system
So some symptom shows up throughout the board, its hard to track to just one part.
The current sense (pin 3) of the pwm controller appears as follows when it is stable
but when the waveform is oscillating, it appears as follows
you can see the blurriness in the waveform shows it is changing size and amplitude.
the pwm output from the pulse width modulation chip also changes as a result.
This is an in area where the voltage is changing rapidly and you can see the waveform of the pwm is changing on the output

The transformer end of the primary (attached to mosfet) appears as following the oscillation during the changing cycles

Any ideas on what to do or check or even where to start looking?
Can I start cutting traces and if so, what one is the best, all of the feedback?

any help is really appreciated
Jim
 
Replies continue below

Recommended for you

Try a little load and see what happens. I have seen circuits that became stable when loaded.


Bill
--------------------
"Why not the best?"
Jimmy Carter
 
If your design has worked in the past and is not any more, then what has changed? I suspect it's PCB layout. How much specification of layout do you provide the manufacturers?

I'd actually be looking at the voltage feedback signal first. The current sense might actually be correct for the switching signal provided. The feedback network is fairly high impedance - could any voltage be getting coupled on to it? Then maybe a closer look at the current sense network. How close is it to the sense pin? How far from the transformer? Have these metrics changed?
 
UC3843 is a current controller and for good stability need slope compensation when Duty > 50%. See Also, I recommand to add a resistor in OP feedback loop(pins 1-2) of about 100K, reduce R7 to 47-100ohm, add a fast diode between pin 6 (UC output) and ground, reversed biased, reduce R8 to 1K and connect it directly to source MOSFET not to ground.
 
Status
Not open for further replies.

Part and Inventory Search

Sponsor