THS3491: Gain = 1 V/V use case

Well you need to use a higher Rf, but why use this part for a buffer, look at the BUF602

Hello Michael-san,

I introduced BUF602 but the buffer amp needs to handle 0 to 15V input signal. BUF602 is 12.6V.

I have some questions for the amplifier.

1) Your simulation, gain starts to increase from ~10MHz. I think this shows THS3491 is getting unstable. The signal frequency is 100MHz. Can THS3491 use 100MHz signal in stable?

2) Is there any good buffer amplifier to support 0 to 15V input signal?

3) The feedback value is 1.2kohm in your simulation How to determine the value?

Please advise us what amplifier is the best for this usage.

Best regards,

Toshihiro Watanabe

Hello Watanabe-san, 

Well yes, a 100MHz 15V sine wave will require 4.7V/nsec slew rate

The current feedback THS3491 needs higher Rf at lower gains as this table shows, they don't like to be buffers and running non-inverting puts the full slew rate and bandwidth requirement across the input buffer, 

Here I did a small attenuation (0.75V/V) at the input to 

1. reduce demand on the buffer

2. bandlimit using input C at about 300MHz

and then gained that back up 1.33X

This 1nsec input step is getting slew limited, but doesn't look too bad for a 100MHz 0 to15V input

Here is the file, 

THS3491 buffer.TSC

Hello Michael-san,

I checked this option but Hi-Z buffer is required. BUF634A is +/-18V buffer. Do you think BUF634A can use this usage?

Best regards,

Toshihiro Watanabe

what slew rate does the signal require

Hello Michael-san,

Required slew rate is 4.7V/nsec.

Best regards,

Toshihiro Watanabe

That was my calculation for a 100MHz sine wave, the BUF634 will not make that. 

Hello Michael-san,

I could understand BUF634A slew rate is not enough. In non-inverted buffer configuration, I think THS3491 may be less phase margin. The customer does not like resister divider because power consumption is big.

I could not find a good solution. Do you have any idea to receive 0 to 15V, 100MHz signal?

Best regards,

Toshihiro Watanabe 

Well what I had shown is 7.5mA into the input attenuator at +15V input, what amount can they tolerate?

The problem with CFA at gain of 1 is normally you would just keep increasing the feedback Rf to get flatness, but you run into a problem with the pole introduced with that resistor and the inverting input capacitance causing reduced phase margin. 

Hello Michael-san,

The input current is required uA range. x1000 input impedance is required.

THS3491 requires 4.3V head room for input and 1.7V head room for output. My idea is to use -5V, +25V power supply and THS3491 gain x.1.5.

If gain 1.5 non-inverting amplifier is acceptable, is this help for stability?

Best regards,

Toshihiro Watanabe

Gain of 1.5V/V should be workable.

Hello Michael-san,

I tried gain of 1.5V/V but slew rate was not good. I guess BW got narrowed.

I tried gain of 1.333V/V as you simulated but input attenuator removed. I think it is the best under the condition. What do you think?

The waveforms of input and output/1.333 on the same plot. bottom is output waveform.

THS3491IRGT_gain1.333.TSC

Best regards,

Toshihiro Watanabe

Incidentally, to get this run I had to go into Analysis options and change the matrix solver to Davis and integration to trapezoidal,  here is that file, not sure the options come through with the file, saved as V9

5086.THS3491IRGT_gain1.333.TSC