1
\$\begingroup\$

I am using the LT3845 to obtain 1.3V of output. From the formula given for calculating the resistors to obtain 1.3V, I conclude that R5(the one tapping into the Vout) should be approx5.69Ohms and the other resistor be 100KOhms. The test schematic simulation in LTSPICE is as shown below -

Simulation circuit.

But, the Vout is almost 0V. The waveform is as shown below -

Waveform

Why so?

\$\endgroup\$
2
  • \$\begingroup\$ Try connecting the circuit to the smaller (i.e. with higher R) load. 2 Ohms might be just too few. \$\endgroup\$
    – Mike Spark
    Dec 9, 2015 at 13:15
  • \$\begingroup\$ I have put 100K. Still result looks the same. \$\endgroup\$
    – Board-Man
    Dec 9, 2015 at 13:16

4 Answers 4

1
\$\begingroup\$

In the datasheet, the minimum start voltage (see page 3, second parameter) is listed as 7.5V max.

There is also this note: Note 4: VIN voltages below the start-up threshold (7.5V) are only supported when the VCC is externally driven above 6.5V.

As you are not satisfying the startup voltage requirement with your 6V source, I do not expect your circuit to properly simulate (or operate, should you build it this way).

Either supply a minimum Vin of 7.5V, or externally drive Vcc (pin 13) with at least 6.5V.

You could always use a different converter, of course.

\$\endgroup\$
0
\$\begingroup\$

You ASSUME that it is the feedback network that is giving you an unexpected result. But is that the problem ? I think not.

Simulating a buck converter is not so simple. Have you checked the basic operation of the switching converter ? Do the outputs of the IC do anything ? The model of the LT3845, does it even support a full switching mode transient simulation ? Have a look inside the modelfile and see what is there. You're trying to simulate 200ms, for a DCDC converter switching at 100 kHz that is a lot of switching. That simulation could take hours.

\$\endgroup\$
6
  • \$\begingroup\$ Even with 2mSecs. The result is the same. \$\endgroup\$
    – Board-Man
    Dec 9, 2015 at 13:16
  • \$\begingroup\$ You can change every parameter and hope it "magically" starts to work but trust me, it will not. You should try to find out WHY the simulation is not behaving like you want. If you cannot then trying to simulate this is a waste of your time. Trust me, I have worked with circuit simulators for 30 years, I know what I'm talking about. \$\endgroup\$ Dec 9, 2015 at 13:22
  • \$\begingroup\$ But how can I open the model file. Issnt it proprietory info ? \$\endgroup\$
    – Board-Man
    Dec 9, 2015 at 15:01
  • \$\begingroup\$ Depends on the model, but proprietary or not is irrelevant. What matters if it is encrypted or not. I have no experience with LTSpice nor LT's models and what they contain. But before I use a model I like to know WHAT is modelled. In electronics there are no models that model everything. A model is always a simplified representation of reality. \$\endgroup\$ Dec 9, 2015 at 15:46
  • \$\begingroup\$ "All models are wrong. Some are useful. " to paraphrase George Box. \$\endgroup\$ Dec 9, 2015 at 15:57
0
\$\begingroup\$

I am using the LT3845 to obtain 1.3V of output. From the formula given for calculating the resistors to obtain 1.3V, I conclude that R5(the one tapping into the Vout) should be approx5.69Ohms and the other resistor be 100KOhms.

Incorrect. The Vfb voltage reference is 1.231 volts therefore, if R6 is a 1kohm resistor, then the current thru that resistor is 1.231 mA. That leaves 1.3 volts minus 1.231 volts across R5 (69 mV) and, at a current of 1.231 mA, the resistance should be 56 ohms.

Bizarrely you have R6 as 100k and R5 as 5.69 ohms - this is plain wrong.

\$\endgroup\$
0
\$\begingroup\$

Well, it all turned out to be a silly issue. Now, when the Vin is > 20V all was fine and dandy. The moment I kept reducing it (to a lowest of about 10V or so), it started to malfunction. Finally found out that it was the SHDN pin that was playing spoil sport. The buck was getting shut down as the Vin was reduced and there by the whole thing did not work.

The moment I ensured decent values of the voltage divider to make sure the SHDN was in the 5V range all worked fine.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.