Integrating the Trimming & Satisfying the Customer’s Needs, Part 2

In the first part of this blog (Integrating the Trimming & Satisfying the Customer’s Needs, Part 1), we looked at a typical switching regulator and considered some ways to make the set-point voltages programmable. The method for setting the voltage uses a one-time programmable device that uses a 16 V signal to force open the [fuse-type] programming link. We'll continue that examination of the testing and programming techniques. We'll look at what can go wrong.

The switching DC-to-DC converters, for portable consumer applications, could integrate some important features that can be adjusted by software through the testing program, during the starting phase of the mass production. I illustrated this concept in Part 1 of this blog. There I described the engineering process of a step-up switching DC-DC converter, focusing my attention on the integration of a special block that I asked the design team to work on. That block was to perform the regulation of the UVLO and ENABLE thresholds through the trimming of a reference voltage which is inside the converter. My request resulted in a trimming table that was the result of the designers' simulation, and it was a theoretical table. See Table 1:

Table 1

Trimming table (theoretical).

Trimming table (theoretical).

Utilizing the trimming table in Figure 1, I expected to have an effective trimming procedure based on that table. I started to debug my testing program by supplying each logical sequence of each row of Table 1 and by measuring the reference voltage after this trimming procedure. Here follows an example of the programming code of the test program implementing the trimming procedure. See Figure 1:

Figure 1

The trimming code implementing the trimming procedure of Table 1.

The trimming code implementing the trimming procedure of Table 1.

To check the effectiveness of the logical sequence corresponding to (for example) the +4 percent step of Table 1 (i.e. bit 0=0, bit 1=1, bit 2=0, bit 3=1), I tried forcing the value of the variable var_perc to +4 percent. I then measured the value of the voltage reference parameter after the execution of the related trimming procedure step (bit 0=0, bit 1=1, bit 2=0, bit 3=1). If the measured value of the parameter after trimming changed by an amount in the range of 4 percent to 4.8 percent (the next step), I concluded that the test was passed and the procedure worked well. Otherwise, it failed, and the trimming sequence was ineffective.

I verified that only some of the logic sequences (corresponding to one row of Table 1) worked well. The time available for the debug of the testing code was expiring. But not everything worked as it was supposed to. I decided to fix this bug during the starting phase of the mass production of the unit at my company's power supply facility.

In Part 3 of this blog (next week), we'll look at what happened with my testing as we moved from the prototyping stage to large-scale production. Also, we'll what happened when I used an automated testing and programming system to try to set the correct voltages for UVLO and ENABLE thresholds.

Related posts:

2 comments on “Integrating the Trimming & Satisfying the Customer’s Needs, Part 2

  1. etnapowers
    November 12, 2013

    The software code of Figure 1 implements a procedure to simulate the trimming procedure (bit_seq_sim function call) and then to trim the resistor bridge (bit_seq_trim function call) . The trimming procedure is processed only in case of positive result of the simulation.

  2. etnapowers
    November 12, 2013

    It happens so many times to have a  theoretical indication by the design team and to verify that the procedure works but it has to be fixed because of reasons that you cannot simulate. Has anyone experienced this?

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.