I'm attempting to design an LM3409HV circuit for linear IADJ operation down to 100 mA.
According to the math, 24 KOhm Roff, 470pf Coff, and 330 uH inductor should produce what I want in the 800 KHz range (depending greatly on efficiency).
I have one channel of our production board setup with these values. I removed the output cap we currently use. We use a ZXMP10A18KTC and I've populated Rsense = 0.22 Ohm.
I derived these values from this spread sheet:
| Vin | 57.5 | Volts | ||||||||
| iled | 1.053 | Amps | ||||||||
| DeltI | 0.05 | Amps | ||||||||
| eff | 0.9 | |||||||||
| Coff | 4.7E-10 | Farads | ||||||||
| Vo | 10 | 15 | 20 | 25 | 30 | 35 | 40 | 45 | 50 | 54 |
| Roff | 15545.6 | 20995.62 | 24452.94 | 25920.64 | 25399.69 | 22890.51 | 18393.3 | 11908.16 | 3435.17 | -4774.45 |
| Toff | 1.01E-06 | 8.88E-07 | 7.67E-07 | 6.46E-07 | 5.25E-07 | 4.05E-07 | 2.84E-07 | 1.63E-07 | 4.23E-08 | -5.4E-08 |
| fsw | 800000 | 800000 | 800000 | 800000 | 800000 | 800000 | 800000 | 800000 | 800000 | 800000 |
| Lind (uH) | 201.6908 | 266.3043 | 306.7633 | 323.0676 | 315.2174 | 283.2126 | 227.0531 | 146.7391 | 42.27053 | -58.6957 |
| iL(max) | 1.078 | Amps | ||||||||
| Rsense | 0.230056 | Ohms | ||||||||
| Pick Values and derive actuals | ||||||||||
| Resense | 0.22 | Ohms | ||||||||
| Roff | 24000 | Ohms | ||||||||
| Lind | 330 | uH | ||||||||
| Eff (Guess) | 0.6 | 0.78 | 0.85 | 0.87 | 0.89 | 0.91 | 0.92 | 0.93 | 0.94 | 0.95 |
| Toff | 1.56E-06 | 1.01E-06 | 7.53E-07 | 5.98E-07 | 4.96E-07 | 4.24E-07 | 3.7E-07 | 3.29E-07 | 2.95E-07 | 2.73E-07 |
| fsw | 456128.4 | 655907.3 | 784895.4 | 836176.6 | 833535.1 | 780537.5 | 658483.4 | 482301.9 | 253722.3 | 41881.15 |
| DeltaI | 0.047179 | 0.046123 | 0.045618 | 0.045323 | 0.045128 | 0.044991 | 0.044889 | 0.044809 | 0.044746 | 0.044704 |
However, when I run these values, the behavior is not really matching the predictions. And there is a strange behavior which I don't understand.
Here is a picture of the PMOS gate signal from the LM3409HV:
Here is the input to the 330 uH inductor:
You can see that the gate is charging and then shutting off. Making Ton vary greatly. This causes the part to be running at twice the frequency, taking a lot of energy to charge the gate, and then not actually accomplish much half the time.
I understand that noise can cause this. What else can cause this and how can I get control over this behavior and get the part to generate more predictable outputs?
Thanks in advance.
