My customer has seen the following issues using the LMH6321.
I am evaluating the LMH6321 buffer using the LMH6321 evaluation board. I am using the buffer in a loop with a THS4031 op amp. I have a 10K resistor with a 1 nF cap across it in between the op amp output and the buffer as recommended in the evaluation board schematic. I have a 100k resistor from the current limit pin to the V+ supply. The supplies are +/-15V. I found that the current limit works properly if I short the output when driving the circuit with a frequency of 10 KHz or less. At higher frequencies the current limiting doesn't work. Why doesn't the current limiting work at higher frequencies?
The load is capacitive which can be as high as 4500pf. I was using 100K for Rext for initial testing. The final value will be 50K to limit the current to 120 mA. The resistor is much larger than the data sheet because I am connecting the resistor to +15V.
I am not having any problems driving the load. The circuit drives the load up to 1 MHz which is the highest frequency we are using. The problem is when the output of the buffer is shorted the short circuit current limit is much higher than expected at 1MHz. The short circuit current limit at frequencies below 50 KHz is what is expected.
The circuit is the schematic I sent you
2) Input voltage = 1 Vrms Sine wave
3) The output current was measured by shorting the output with a 0.1 ohm resistor, measuring the rms voltage across it with a scope and calculating the current
4) The current limit set resistor is 50K to +15V
The expected current limit is 400 * 15V/50k = 120 ma per the data sheet.
The data shows the current limit is as expected when the frequency is below 20 KHz. As the input frequency is increased the current limit increases until there appears to be no current limiting at all at 1 MHz. Please let me know why the device behaves this way and if there is a way to limit the current at high frequencies.
Is what the customer seeing normal?
I can send their schematics if you give me an email address.
Thanks for your help with this issue.
Richard Elmquist
