Power Supply DC-DC Step Up Converter Problem

Tough to say without a schematic. It looks like you would spend as much replacing parts as the cost of another one, unless time is critical or you just want to troubleshoot.

I agree. I do have one piece of added advice.

When the OP replaces or fixes this supply, they should change the cabling connectors so that it is impossible to miswire power again.

Recycle it and buy another one.

I believe that the words, "faulty" and "defective" are improperly applied here.

"Blown" and "operator error" seem more appropriate here.

As suggested, purchase a new one and connect it properly and have a lesson learned.

Will you please post the actual amps drawn from your power source when the light bulb is connected? The 5W lamp at 13V would draw less than .4 Amps at the output. It does in fact seem that your power source is insufficient. Considering the typical design of most DC-DC converters, applying 11 or 12 volts to the output should not blow it out. My estimate is that with the output load of your lamp, the input current shoul be less than 1 amp.

Thanks for all the comments. I appreciate the feedback!! With regards to the above, the input current drawn with the 5W-lamp is around .5amp. I'll check what current is being drawn at the output and post it. My power supply is not the problem, as it can and will deliver 5 amps. The power supply is a Voltcraft DIGI40 with adjustable current and voltage. No worries when the lamp is connected directly to the power supply. Also I have connected a H4 halogen 12v 55w. I'm getting the same behavior, the input voltage drops and the amount of amps being drawn from the power supply is similar to what it should be like 4.5amps. The voltage at the input should not drop with the amount of amps I have available from my power supply. It seems the DC to DC step up converter is influencing the power source. Also what I'm noting is that the output light seems to be blinking means fast on of visible for the eye. The same I'm seeing on the led's of my power supply for both the voltage and amp-LED. In normal behavior, the voltage led on the PS should be on when enough amp is given and the amp led is off. Whatever I do on the DIGI40 power supply with the current-adjuster resistor the amp led will not go off and will together with the volt led start very fast blinking. I have also a step down converter which acts properly with the same lamp and Power supply. I do understand buying a new converter is the most easy thing, but also this will take time and I have still an unused part. Besides that I want to know what is the problem. I could also post later a couple of pictures of the pcb if that will help anyone. Maybe the particular behavior described rings a bell related to a certain component.

What gives you the idea that increasing your load should not cause voltage drops. None of your wiring is identified or made of superconducting material.

Check your cables and connectors from your power supply to the DC-DC converter input. If you have 12 volts at the Power Supply, and 8 volts at the DC-DC Converter, then you have a bad cable or improperly sized cable.

Thanks all for the provided feedback. What I'm trying to say is that once the DC to DC converter is connected to the Power Supply, the Power Supply is acting strange provided there is a load connected to the DC to DC converter. What happens is that I can't control the current at the Power supply and the voltage at the Power Supply drops with around 3 volt. The reading is coming from the reading on the Power Supply itself. Also the LED's at the power supply are flashing rapidly. Looks to be the same frequency as the load-bulb. The connecting leads from PS to converter and from converter to load are proper and there is no issue there. Somewhere on the PCB of the DC to DC converter there is a problem.

It sounds like it's oscillating. Do you have a capacitor 470uf or larger, that you can connect right at the input to the DC-DC Converter? I'd use an AM radio to see if it's broadcasting a lot of noise. If the external cap solves it, check the ones on the DC-DC Converter, especially near the input.

Don't forget, you can download the data sheet. It gives pinout of the PWM IC, On Semiconductor's data sheet explained theory. There are also example circuits, but none for ste-up converter.

This is your DC source ?......

...and the LEDs next to the output voltage and current limit knobs are blinking when you load it with a lamp, right?

"Bite the Bullet" and admit you ruined it. Go and get another one and as suggested change the input and output connections so they are not a match.

Be thankful that it was an inexpensive power supply and not a $6,000 Tektronix scope or something like that. Experience is the best teacher! Be thankful you were working with low voltages, it could have been you that got fried besides the power supply. You might end up getting the biggest headaches from trying to rejuvenate your problem unit.

Better luck next time.

Good Luck, Old Salt

Maybe I shouldn't be too quick in my praise for this circuit. The problem is its advertized 600W rating. What a joke! It's based on an assumption you'll run with 60V input (overheating R17) and 10A average current. Assuming the inductor, electrolytic, and rectifier can handle the dissipation issues, this the converter will still need to be in CCM (continuous conduction mode) to run at this high current at this high voltage. But the circuit does not have the necessary pole-zero compensation scheme to deal with right-half plane 2-pole operation. Nor does it have the slope compensation it needs for stable operation at duty cycles past 50%. The only takes a few parts, but they aren't present. It also doesn't have the parts to implement soft start, which might have saved out O.P. some pain. Modern controller ICs make it easy to implement these features, plus they have other protective features. But they cost more than 57 cents (Mouser, qty 1k).

Here's one of my designs, for a 14V to 42V converter, which I conservatively rated at 200W, although it can do 300-350W at room temperature (it's fused for 420W). It uses an LT3782 controller IC, which costs $5. The converter is two phase, so each half gets to work at 1/2 power, and the current-ripple stress on the electrolytic caps is far less. As you can see, much more is invested in components and features. But it can't be manufactured and sold for $17 with free shipping. :-)

Hello all, Thanks for all the feedback!! Great forum. I also appreciate the detailed feedback of Winfield Hill with a schematic of the converter as well as the explanation. Currently I have connected a 12v battery as input source, just to rule out any current source or other PS issues. I'm getting the same behavior. Below are some measurements I did. With no load and R14 set to 95k: Uin 12,8v Uout 13,7v R14 Variable resistor for current is set to 95K Then pin 3 of UC3843 is 24mv and pin 6 of UC3843 is 6mv pin 8 of lm358=10,1v With load H4 55w lamp Uin 11,7v (so voltage drops with 1 volt uout 13,7v R14 for current is 95K Iout =4.9amp Then pin 3 of 3843 is 166mv and pin 6 of 3843 is 1.9v pin 8 of lm358=9,4v Now with r14 variable r resistor is set to 3K With no load: Uin 12,8v Uout 13,7v R14 Variable resistor for current is set to 3K Then pin 3 of UC3843 is 11mv and pin 6 of UC3843 is 7mv pin 8 of lm358=10,2v With load H4 55w lamp Uin 11,7v (so voltage drops with 1 volt uout 14.2v R14 for current is 3K Iout =4.9amp Then pin 3 of 3843 is 119mv and pin 6 of 3843 is 1.7v pin 8 of lm358=9,44v So I conclude that changing the current resistor from 95 to 3k will only give a slight voltage change on pin 3 of the UC3843 when a load is applied. Changes only from 166 to 119mv. The output current in both settings will not change and stays at 4.9Amp. Funny enough I see also that the output voltage raises when put under load and R16 is changed from 95k to 3K. The voltage increases from 13.7 to 14.2Volts. In my opinion the above behavior is not correct. As mentioned on the diagram. The voltage on pin 3 of UN3483 should change from 0 to 1v. The input voltage for the lm358 and the uc3843 will be around 9,4 volt under load. Any idea what is going on and how to solve this, so I can adjust the output current.

From your observations described above alone, I'm not sure your unit is malfunctioning. It's being called upon to provide barely any boost, and is skipping cycles and generally blipping along, probably with very high output ripple. To evaluate it thoroughly you'd need to use a 'scope. For example, the current-sensed signal on pin 3 is a peak value. A couple of blips of Q1 charging the inductor can send an unloaded output beyond the setpoint voltage, which then makes the error integrating amplifier turn off the switching entirely until the output voltage sinks back below where it belongs.

For testing purposes, I suggest you run the converter at higher output voltages (higher step-up boost ratios) and with heavier loads. Be aware this design is not a precision, low-ripple regulated converter, but it should keep the output more or less where it belongs, within an error band of a volt or two.

If your use is as a low-voltage low-load modest-ratio boost converter, and you want tighter output-voltage control, you may want to remove the inductor and rewind it with 2 to 3x as many turns.

Also, redfred has speculated that you reversed the input power voltage polarity when you first mistakenly connected the power to the output. If so, the output electrolytic caps might have been damaged. You can test them by adjusting the converter in boost mode to a moderately high voltage, say 50V, and applying a light load, like 5k ohms. Then remove the input and see how long the output takes to decay.

Another excellent explanation of how a circuit should work. Bravo. However, I disagree with your comment that wiring the input power to the output terminals should not cause damage. If the polarity of the input power is also backwards then that blocking diode is now forward biased and several polarized electrolytic capacitors are reverse biased. A reverse biased electrolytic capacitor will be quickly damaged but may not exhibit visible damage. This miswiring agrees with results stated by the OP.

Thinking the same here.

As an aside, I recall many decades ago, when switch mode PS started to appear on the scene, that they had to have a load connected when energised or they would release smoke.

Does anyone else remember this from way back?

Greetings,

I just registered to this site in order to get trough to ask you one particular detail: Concerning eBay "600W" schematic which you posted here, I am about to use this same converter to boost output from 24V (32V MPPT) 200W solar panel into 48V lead-acid battery bank. Two issues are to be concerned: (1) It is high impedance source which I have to harvest to the last coulomb, (2) energy-optimal conversion can only happen around MPPT point of the PV, so I have to maintain quasi-constant input voltage.

To get to the point: I will have to enable and disable this converter by external circuitry in 1Hz (order of magnitude) intervals by sampling input voltage, switch ON when input raises to 33V and OFF when it falls to 30V. Very large input capacitors will slow down input voltage ramps. Where would you "tap" to the circuit in order to achieve that? What point to control in this particular design to start/stop the converter? My idea is to force down Vcc of UC3842 below UVLO treshold, but that method would be rather aggressive to R2. Any better thoughts?

Many thanks in advance for considering this post,

Koča Vrančić

Thanks again for all the input. Have still a couple of things to try. I do have a oscilloscope as well so when I find some time will do some testing. Actually I want to use this converter for my ebike for testing purposes. The ebike has a 24v input which actually delivers around 25.5V. Just want to boost the voltage to the motor to gain a bit more topspeed. For this I need to figure out how much extra voltage I need for let's say 30km/h. Later I then can adapt the output voltage with for instance batteries, rather then losing watts in the converter. Another option I'm considering is to use the converter to charge batteries. of the bicycle (if still available on the bike and doesn't consume much of my battery power) So have an old laptop PS connected to the DC to DC converter and provide 28.8v to 2 SLA batteries in series. Does anybody have experience with the efficiency of this converter. They claim up to 95% but not sure if this is the case. Before doing any further testing on the bike,I need to make sure the converter works fine.

My boost-converter board draws 25mA, or 0.5 to 1W, even when it has no load. There are improved replacements for the UC3843A controller which draw much less power. For example the CMOS technology UCC38C43 says it draws only 100uA standby and 2.3mA operating. If you change the controller to CMOS (direct replacement) that will help improve the low-load efficiency. If you're going to use input voltages above say 40V, you can also increase the value of R17, say by 5x, to reduce the U1 current (and you can also get low-current versions of the '431 IC). If appropriate, you can also down-scale the converter's switching-current operating region to lower currents by increasing L2 and R9.

I have had a sequence of power electronics course back in the early 1990's in B.S.EE program. In fact designed, built, and tested one power supply as a group (team) project. The course sequence had buck, boost, buck-boost etc. topology, Matlab simulation and C & C++ programming projects...It has been very long ever since seeing such a circuit diagram. :) I am into electrical power NOW (switched major) than semiconductor (electronics or fiber optic electronics...).

Dear Winfield

I have purchased the 600W converters and I am having some problems.

I cannot get the current limit to work.

I have supplied the unit with 24V DC and adjusted the output to 33.75V.

I require a current of 3A. i have connected a dummy resistive load to the unit and the output current just increases even if I turn I-ADJ to minimum.

Please advise.

Regards

Thego

Changing controller to UCC38C43 will improve low-load efficiency only in low power converters, in high power converters it will have negligible effect due current needed to drive MOSFET, which for high current devices will be much higher than current needed by controller. RMS value of current needed to drive MOSFET will also change with operating frequency, so for high power high frequency converters total effect of changing current consumption of controller will be negligible (it will still affect total efficiency but effect will be extremely small)

The power side of the step-up circuit on its own should NOT have been damaged by connecting power at the output (if polarity was not reversed of course), but the PWM controller voltage feedback circuitry could, especially if output voltage adjustment was set very low at the time (near the IC reference) and that situation was not taken in account in the circuit design, so no external resistance was used there. (I don't blame the board designer for that, as it's not in the IC reference designs, and its only use would be to protect it from circuit board... Terminators, like the OP.) S.M.

I bought the same dc-dc step up converter but I have a problem: the output is always at 0.60V and doesn't change with the potentiometer.
The converter is alimented with a 12V pc power supply (that works perfectly).
What could the problem be?