LMH6554 Fully DC-Coupled Configuration with Shifted Dual Supplies

Hi Masa,

It seems that your input swings 2Vpp centered at 1.9V dc (Vdcsource_cm) from 3.9V to -0.1V. So, I would think that you don't need the Rpu resistor.

You only need the Rpu resistor when the Vdcsource_cm and output common mode voltages (Vocm) are different which happens to be the case in the application note AN-2177 with Vdcsource_cm = 0V and Vocm = 1.25V. In your case, the Vdcsource_cm equals Vocm to 1.9V and the Vicm will by default be set to 1.9V. However, one thing that you need to do is tie the Rt and Rm inputs to the same dc voltage as the Vdcsource_cm to maintain input voltage symmetry as shown in the below diagram.

Following is the general equation which calculates what is the Vicm set to with your Vdcsource_cm and Vocm:

 where Rin = Rs||Rt + Rg

For your case with Vdcsource_cm = Vocm = 1.9V, the Vicm will be equal to 1.9V (mid-supply) providing +/-1.3V room to swing for the Vicm pin. But, in the application note AN-2177, the Vdcsource_cm = 0V and Vocm =1.25V which set the Vicm to a different voltage other than 1.25V (mid-supply). Hence, there was the need of Rpu to maintian the Vicm close to 1.25V or mid-supply.

To answer your questions, I think your circuit configuration of LMH6554 should work. Two things I noticed was that: 1. Connect the Rt and Rm to the same dc voltage as the source dc voltage (Vdcsource_cm) shown in the above diagram. 2. Inorder to get 2Vpp(diff) from a 4V(p-p) single ended input, you would need a 0dB gain in the LMH6554 because of the -6dB loss due to voltage division of Rs/ (Rs + (Rt||Rg)).

"About Vdcsource_cm(DC input voltage), is it voltage connected through Rt? If so, is it okay to use same V+(4.4V)?"

Ans: The voltage connected to Rt and Rm (where Rm = Rs||Rt) should be the same as the dc source voltage (Vdcsource_cm).

"I referred equation (2) on page 6 in AN-2177. Is this as same as the equation you gave in your response?"

Ans: It is a slightly different equation for Rpu which would end up giving you the same equation as in the AN-2177 if you set Vdcsource_cm = 0V and Vocm = Vicm. The AN-2177 document shows you how to set Vicm with Vdcsource_cm = 0V but depending upon application the Vdcsource_cm might be different than 0V, which you have to then consider in your calculation. In the below equation, the Vdcsource_cm is already 0V and if you set Vicm = Vocm, you end up with the formula Rpu = (V+ - Vicm)*Rin/Vicm.

"Also may I have recommended Rg, Rt, Rm at -6dB? I am referring Table 1 on page 3 in AN-2177 and assume Rf=200ohms (can increase if need)."

Ans: As I explained in my 2nd point above, you would need a 0dB gain in the LMH6554 because of the -6dB loss due to voltage division of Rs/ (Rs + (Rt||Rg)). I have attached TINA-TI simulation for your reference. I would choose the Rg, Rt and Rm to be the ones in Table 1 on page 3 in AN-2177 for 0dB gain.

6758.LMH6554_schematic.TSC

Best Regards,

Rohit

Hi Masa,

One last comment about increasing Rf is that the bandwidth of LMH6554 would decrease with increase in Rf because of the LMH6554 being a current feedback amplifier. I have attached a plot of bandwidth vs Rf which you could consider while choosing the Rf and the excel spreadsheet with the measurement.

6013.4786.8524.LMH6554%20Freq%20Response%20for%20various%20RF6_14_12(1).xls

Best Regards,

Rohit

Hi Masa,

I think the fourier series simulation is correct. With the fourier series output format set to C*exp(j*(kwt + phi)), the first coefficient with k = 0 is the dc and if you calculate the fourier series expression for k=0, the output is 1.89 with phase set to 0degrees. However, this is not the fundamental output signal. The k = 1 is the fundamental output signal and the HD2 and HD3 must be calculated from this reference amplitude(C1). As a result, the HD2 and HD3 values come out to be 102dBc and 113.8dBc which is closer to Figure 13 on page 7 in the datasheet.

Some things to note however is that the datasheet plots are specified with a gain of 2V/V which is different than the 0dB gain configuration you are planning on implementing. Also, the harmonics specially the HD2 will be highly dependent upon the layout of the board and matching of external resistors. To get very good distortion performance, I suggest to use 0.1% or better tolerance external resistors. The LMH6554 models for TINA-TI simulator are mainly derived from the datasheet and it is difficult to include matching or circuit parasitics thoroughly into the simulator. So, the simulation results will be better than the actual results to a certain degree.

A suggestion could also be to characterize an LMH6554 evaluation board in the desired configuration and check true performance of the device.

Hi Rohit,

Thank you very much for all of your helps.

Best regards,

Masa