| Author |
Message |
jegandren
Joined: 13 Dec 2007
Posts: 30
Helped: 1
|
 29 Sep 2008 8:45 help needed on current control SMPS |
|
|
|
hi all,
attached is a current controller circuitry that i have build using UC 3846, but with no success..
the problem here is i guess with the current sensing part as for an open loop circuit without feedback i would get +/- 40V with no noisy coming from the switchimg transformer, with a nice pulse.
but when the feedback is given im getting only +/- 17V with a noisy sound coming from the switching transformer and the mosfet gets heated up quite fast.
Is it the rite way to sense the current. please advise.
|
|
| Back to top |
|
 |
ghydda
Joined: 02 Oct 2008
Posts: 24
Helped: 4
Location: Denmark
|
| 04 Oct 2008 22:43 help needed on current control SMPS |
|
|
|
There are most likely several things which needs to be corrected.
1) What is the point of isolating the voltage control loop between primary and secondary side when the same technique is not applied to the current loop?
Surely the secondary side is meant to be completely isolated from the primary side (mains) - and the direct connection made between the controller current measure inputs (3+4) and the sense resistor at the secondary really spoils this.
2) Current sense is only applied to the positive rail of the secondary. There is no protection of the negative rail, thus any current overload will not be detected at the controller and thus the converter is at high risk of being destroyed.
3) Due to only measuring the positive rail current there is no signal available for the controller to tell it when the desired current level even has occurred, and thus the negative pulse will stay on until the clock terminates it. This means there is no control over the voltage on the negative rail either. Quite a problem.
4) The 1uF capacitor in series with the transformer is redundant as the other end of the transformer is also DC-decoupled. It should be removed.
5) The midpoint of the input smooting capacitors should not be connected the transformer. It is good practise to use dedicated small value capacitors for this (values between 220n to 2u2 is commonly used in offline half bridge designs).
6) Topologically it troubles me that the current control scheme is mixed with DC-decoupling of the transformer. Should you correct problems 1), 2), 3), 4) and 5) this one will knock you circuit off its feet for sure.
The slightest unbalance in the voltage between the the 330uF capacitors or different Rds_on of the FETs or an imbalance in the current sense circuitry between the the positive and negative pulses will cause the midpoint between the 330uF capacitors to slide to either side and the full rectified voltage of the mains will be applied onto one of the 2 capacitors, possibly violating their voltage rating.
I'm very sorry to say that I'm not the least bit surprised it works really badly.
In fact if the concerns listed above is not corrected you will not end up with a reliable power supply.
Best wishes,
/Ghydda
|
|
| Back to top |
|
|
ghydda
Joined: 02 Oct 2008
Posts: 24
Helped: 4
Location: Denmark
|
| 07 Oct 2008 10:15 Re: help needed on current control SMPS |
|
|
|
Hi,
The attached picture is a principal schematic only!
It shows a way to solve the otherwise unsolvable problem of having current mode control with a DC-decoupled primary transformer winding.
The balancing inductor will correct any mismatch in voltage division between the two capacitors used to create the center point to which the transformer is connected. The power exchanged from one capacitor to the other should be very little, thus it is not necessary to have high power fets driving the inductor.
I hope it will help you get your converter fully operational.
Cheers,
/Ghydda
|
|
| Back to top |
|
|
DrBob13
Joined: 26 Mar 2008
Posts: 133
Helped: 36
Location: UK
|
| 07 Oct 2008 13:54 Re: help needed on current control SMPS |
|
|
|
All good stuff, alternatively you could transformer couple your mosfet drive signals and then only the swiching transistors and the switching transformer primary will be mains live. You do require some clipping zener diodes to shape up the pulses a bit and an RD network is also required to rapidly discharge the gate capacitance on switch off. In this case the mosfet driver device is not required. Design your driver transformer with a 1:1 ratio with enough turns to generate the required output voltage without overloading the controller chip, I have not had any trouble driving normal sized mosfets directly using this method but it is necessary to include all the required protection components on the gate drive. If you are worried about the driver chip, include complementary NPN/PNP emitter followers (configured like transistor radio af amp outputs used to be) to drive the transformer inputs, they do not need to be big, 2n2222a/2n3907a will do it, the driver signals are already available in complimentary form.
WARNING, as you have got it, it is a potentially lethal power supply, make sure you remove the mains coupling to the DC outputs, no matter what else you do.
The heating and the noise is probably due to the control loop being destabilized.
Good luck,
Bob.
|
|
| Back to top |
|
|
