This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

VCA810: Non-linear issue of VCA810

Part Number: VCA810
Other Parts Discussed in Thread: ADC3424, VCA820, LMH6609, VCA821

Hi team,

   My customer is using VCA810 to control the sine waveform amplitude, see schematic below, but they find that our device will cause a non-linear issue to the input signal.  See test below, this table shows that the VCA810 will affect the input signal and reduce the Vpeak value. This effect is hard to be calibrated. Do you have any idea about how to avoid this effect? Thanks. 

Input signal

Vpp

Data in ADC3424

Gain=0dB

Data in ADC3424 without VCA810
1V 783 843
0.5V 411 422
0.25V 211 212

  • PRobably slew rate compression, what frequency?

  • Hi Michael,

        The max frequency is 30M. But customer is using 1MHz signal will also have such problem. Please help share some suggestion for us to find the root cause. Thanks.

  • Well a 1MHz test will not have a slew limit, but a 30MHz is probably too high for this part. Not sure why it is not working at 1Vpp, but you should be looking at the VCA820 type part 

  • What do you mean by 30MHz is too high for this device? VCA820 is not p2p with VCA810, it is difficult for customer evaluate our device again. Have we done some test in TI lab? We really need some professional suggestions from team to locate the root cause. Thanks.

  • Well, I did all these parts and the VCA810 was a process migration of the earlier VCA610 - I re-positioned the VCA810 at that time as a pulse oriented VCA since it has abysmal HD performance, If you are headed to 30Mhz, you are headed to dissappointment with the VCA810, a quick review of the datasheet shows, 

    All of these curves are starting to roll off at 30Mhz, normally you do not want to operate in the rolloff region, 

    And the distortion is amazingly poor vs The VCA820 series (4 parts). We stopped this plot at 10Mhz out of embarrassment. 

    and then another look at this, if you don't care if you have a highly distorted waveform - sure, but the VCA820 series will be far better - sorry they are not pin compatible. Again, I don't know why you are seeing the 1Mhz 1Vpp compression, could be lots of things,but you have more serious issues at 30Mhz. 

  • Thank you for your response here. Could you please help recommend the right device to support below requirement from customer's side,

    1. Max input signal frequency 30MHz;
    2. bipolar signal input;
    3. Gain Adjust Range 40dB;
    4. Max Vpp is 2V;

     Will VCA820 support this case? Or any other device can support this case for good linearity? Thanks.

  • Thanks for the detail, '

    When you say good linearity, can you specify a harmonic distortion target - I think this is driving an high pass filter in the LMH6609? That is a 900MHz part for a 30MHz signal? Seems kind of fast. What filter cutoff you are you looking for?

  • Also do you want a dB/V gain adjust (useful for AGC loops) or a linear V/V gain adjust ?

  • Hi,Michael,

    Thank you for your response.

     Use VCA810 to control the  waveform amplitude,then drive an high pass filter in the LMH6609(filter cutoff is 300kHz), then drive an low pass filter in the LMH6609(filter cutoff is 30MHz).

    When VCA810  is at a fixed gain,I do not want the waveform  was distorted.

    I used 1MHz 1Vpp  sine waveform,but  within  the valid range   I  found  the waveform's  peak  was  not linear.

    Thanks

  • Hi Michael,

       Thanks. The output of the VCA part will be sent to bandpass filter (300kHz-30MHz), For the 900MHz LMH6609 is enough for 30MHz signal, let us focus on choosing a VCA part in this application. We do not have a specific target for harmonic distortion, the smaller the better. Thanks.

  • Well let's assume a dB/V gain adjust is preferred, the two choices you have are the 150MHz VCA820 or the 720Mhz VCA821 - here is an HD plot for the VCA820. If you are going to follow this with a 30Mhz LP, that will attenuate harmonics, so lets focus on 20Mhz HD and note your loading is light - so HD2 will be dominant, and is about -63dBc for the VCA820

    Then the VCA821 is spread across two pages, here is the first plot, where again the HD2 is dominant at 20Mhz, 

    and then the lighter load HD - looks to be about -74dBc at 20Mhz on the HD2. So, which would you prefer?? Both have plenty of slew rate for a 2Vpp at 30Mhz although oddly the VCA821 does not show a large signal BW plot? 

  • Hey Lin, 30Mhz active LP filter is a bit of challenge, what filter specs. Gain, Q (I am guessing Butterworth?) I would do these as an inverting MFB where these designs come easily out of tool I developed. If I assume a gain of 1 butterworth at 30MHz, Filterpro says you need a 2.1GHz GBP - not really, but you do need a pretty fast part - my most recent updated tool says 1.7GHz - so, you might want to consider this low pass to be an RLC filter followed by an amplifier buffer. 

  • Hi Michael,

       From the HD2 spec, I think VCA821 is a good choice for customer. I will penetrate this device to them, BTW, you mentioned filter pro, could you please share this tools for customer to design their Butterworth filter? I want to recommend a better circuit for them. Thanks.  

  • Well Lin, I think Filterpro is gone,but that code is in the newer TI online active filter designer, however, it stops accepting cutoff frequencies above 10Mhz. Like I said, might be that an RLC followed by an amplifier will be the way to go - here is a possible solution using a low insertion loss design flow from this application note, 

    http://www.ti.com/lit/an/sbaa108a/sbaa108a.pdf

    you would then buffer this with amplifier stage - this is hitting 29.8Mhz F-3dB, pretty close.