| broodis |
Hi guys,
Currently I'm having some problems with the National Semiconductor LM4651 driver. I followed the guideline given by National to debug the chips (no input test), however I wasn't able to get a duty cycle of exactly 50% but it was close enough.
When I tried to connect an DC input signal of greater than 1V the amplifier output seems to be overdriven (100% duty cycle). Is there a way of finding out the amplitude of the input signal triangular waveform?
Thanks. |
|
|
| SpittinLLama |
What was the duty cycle with no input? It should not be off by more than about 1-2% unless running without gain, then more off. This just translates into DC offset at the output. A perfect 50% duty cycle will give 0V, right between the Vcc and Vee voltages.
1V input should not over drive, what is the gain set to? If possible, it is best to use low gain to reduce noise. Are you making a DC supply or why putting a DC voltage into the chip?
-SL |
|
|
| broodis |
| When the input is 0V we get around 55% duty cycle and I am using a gain of 10. We are planning to use the LM4651 and LM4652 as a current amplifier to drive 8 coils for a magnetic bearing. |
|
|
| SpittinLLama |
Seems a little high for duty cycle. Checking the LM4651 datasheet it shows a maximum error of 3% and typical of 1%. Are you using the Roffset resistor? Sometimes that value needs to be changed depending on the application and the supply voltages. You can try it without this resistor and see what you duty cycle error is then adjust the value of this resistor until you get it where you want it. You can see this resistor on the typical appliation on page 4. Be sure to read the section on adjusting DC offset on page 12. You will need an input capacitor to block the DC voltage. The amount of voltage at the input from the Roffset resistor is very small so it can seem touchy. You could try a POT and just adjust it to get it where you need it to be.
-SL |
|
|
| broodis |
Finally got some sensible results. The problem is I'm not getting a consistent gain for different input signals. Is it because of the Roffset?? What is the formula for Roffset?? Is there another way of getting a consistent gain which is based on the gain equation on the datasheet?
Thanks. |
|
|
| SpittinLLama |
The gain equation in the DS might be a little hoky. Gain is also affected by the frequency you have set for the modulation. The supply voltage also affects gain. What is the schematic you are using? Where are you seeing gain differences? According to the datasheet the gain v frequency should be flat out to where ever you have chosen as the output filter point. The graph on page 9 shows a subwoofer type of solution so LC filter sits around 5kHz. The graph on page 23 shows the other demo board which is for full range audio. It seems to be flat out to way beyond the typical 20kHz of audio. So it dure seems possible, if the datasheet is to be believed, to have flat frequency response. More information about your board, schematic, and what you're trying to do might be helpful.
-SL |
|
|
|
