This structure is suitable for such high-power occasions. In fact, I have been exposed to this topology since I studied the design of switching power supplies. I have made several high-power power supplies using this structure. The performance is excellent. This kind of structure has a special liking, that is, the structure is simple and the performance is reliable. In this post, we mainly take this work as an example to share the process and experience in the design. If the analysis fails, please understand and correct me.

Here are some parameters of this work:

Name: High-power adjustable power supply (a good helper in the laboratory);

Input voltage: 220V;

Output voltage range: 0-50V;

Output current range: 0-50A;

Power: 1000W.

At present, this is the second version, and the performance has reached the design requirements. Unfortunately, flying a few devices is not beautiful enough. It does not matter. My third version of the PCB has been sent to proofing, so this post will use the third version. Focus on explanation and analysis.

The purpose of designing this adjustable power supply at the beginning was to be used as an experimental power supply to facilitate debugging of various circuit boards. To be honest, a high-power power supply is necessary for inverter development. Usually, it is directly used with several large batteries, connected in series or parallel. However, the current is not adjustable, and I always feel afraid. If I buy a high-power adjustable power supply, the price is often relatively high, so I decided to design one myself to meet various debugging requirements and use it handy.

designing process:

After determining the topology, there are several more important places that need to be designed and considered.

Auxiliary power supply experience: The two previous versions have encountered the same problem. The auxiliary power supply is stable and works in a discontinuous state. The interference generated at this time is enough to affect your current and voltage loops. .

Upload several power supplies of various versions:

This is the first version of high power adjustable. The problem is that there is no auxiliary power supply, and the auxiliary winding of the transformer is directly used, so that there is no way to output a low voltage. The first failed product was made in October 2015.

The second version adds a TNY276 auxiliary power supply to improve the loop, but the working of this auxiliary power supply really sads me. The duty cycle is very small, the waveform is discontinuous, and various transformers are rewound. There is no way to solve it. In this way, the interference from the auxiliary power supply is very large, which makes the loop oscillate seriously when the power supply reaches 300W. Finally give up November 30, 2015.

This is the fourth version, because the third version is slightly improved, so I will not stick out the third version, mainly from a single panel to a double-sided, the main control has become a patch, so save A large part of the PCB area, the entire power supply is much more compact, the third version is the first photo of the real thing, the fourth version is still on the way to proofing … mainly if the power buffer is changed In addition, the relay is delayed and closed, and a short circuit protection.

Today I shared with you the experience of making this power supply. The road to power supply is really bumpy and painful. In fact, doing so many versions is not a project for others. It is mainly an interest, and you must do a good job when you decide to do it. I also hope that everyone will support me, and then slowly update …

The last one of my main structure diagram, because it is a wide range of adjustment, so I also learned the linear power supply switching transformer winding to achieve a wide range.

This relay determines the status according to the output voltage, which pulls in above 25V and releases below 20V. The PCB is back again, and a quick close-up:

In one breath, try two.

The large PQ5050 transformer used is common to the previous two versions.

Three 5mR constantan wires are used for current sampling. The front stage does not collect current. A 7P terminal is used to adjust voltage, current, ADC-I, ADI-U and VREF + 5V + 12V. Two separate 2P terminals are connected to the digital meter.

Continue to install:

It’s almost installed. Let’s try it last time, but the process of testing the new machine must be cautious. It is not just to take 220V into it.

Let me share the test process below:
First, use the experimental power supply to adjust it to 50V (the same as the DC power supply, the bridge is adjusted anyway). The current is adjusted to 1A. After powering on, the auxiliary power supply does not start. According to previous experience, this VIPer12a auxiliary power supply should be Just start. Check the power off and find that the problem is that the capacitor of the power IC VCC is reversed, the starting voltage cannot be established, and it feels not very comfortable. I have a feeling that if the first time you do something, if it starts relatively smoothly, That means that this work must be relatively smooth throughout the debugging process, I feel! feel! Quickly change the capacitor position. Power on again, and the power indicator turns on inadvertently.
Well, it ’s not a big problem, and then the value of the auxiliary voltage measured by the meter is normal. After all, the auxiliary power transformer is wound later, +13.5, -13.6, +13.4, the three-way voltage is normal and stable, use an oscilloscope In view of the waveform of VIPer12A, the duty ratio is relatively large, because the input voltage is relatively low, but it is relatively stable, and it is ok. Then look at the working condition of TL494, measured the waveform of the 9-pin driving pin, and I saw that it has already waved, but the frequency is not correct, 100K found that the 5-pin capacitor is installed again, and I feel that something big is happening behind ~ change back to 102 After normal, then measure the driving waveform of MOS GS:

The upper tube and the lower tube are measured with the same waveform. This is a step closer. There is no problem with the drive. I immediately thought that since the wave has been emitted, then the post stage must have voltage. When the full duty cycle is reached, the post stage Normally, the output voltage is about 14V. Immediately use a multimeter to measure the output voltage at the back. Full duty cycle, not according to experience, just hear a bang …
The capacitor was installed backwards, responding to the hunch and big event just now. This experience tells me to be careful and not rough. Well, since this is the case, change the capacitor and test it, but always remind yourself to be careful.

Replace the capacitor, wash the points on the PCB with thinner, and then directly enter the voltage loop. The 9.7V waveform is very stable. Then adjust the potentiometer, and the direct voltage changes accordingly.

The minimum voltage can be as low as 0.11V, but it is meaningless. Because the input voltage is only 40V, the maximum output is 14V. The voltage is normal and the loop is stable.

Then it is necessary to test the current with a little load!

This is to lead two long lines to short-circuit the output end, and then adjust the current potentiometer. I saw that the ammeter walked smoothly, and it was relatively smooth ~~

The minimum current is around 0.2A, and now the duty cycle is up to 30A, and the high voltage should be normal. 30A does not dare for a long time, and no radiator is installed. The tube felt hot within 10 seconds.

Well, at this step, it almost means that the PCB circuit is no problem, look forward to high voltage …