LMH6552: About Differential output + and output - DC error.

Hard to say, this current feedback based FDA appears to have neglected to specify offset voltage. Normally CFA devices are poor DC offset, the only spec here is input bias current offset which will get tothe output times the feedback R values,

Hard to say, this current feedback based FDA appears to have neglected to specify offset voltage. Normally CFA devices are poor DC offset, the only spec here is input bias current offset which will get tothe output times the feedback R values,

Hi Xinjian,

what is driving the OPAmp? What looks the circuit like at the unshown left of your schematic?

Kai

Hello Xinjian,

This offset seems high to me assuming the input offset voltage of this device is similar to that of the LMH6555. The sim file for the device also seems to indicate an input offset of around 1mV, although it doesn't specify the differential offset was added in the sim parameters.  My assumption is that there is some mismatch in your circuit that might be adding to your offset. Can you provide more of you schematic?

Best,

Hasan Babiker

Really not going to be able to answer this question without a min/max input offset spec on the LMH6552

Not sure where you got that 1mV typ number fromthe LMH6555, that fixed gain of about 5 device only quotes a min/max output of 50mV or about +/-10mV input, 

Hey Michael,

I agree its hard to determine without any min/max specs. For the 1mV offset I was referring to spice file for the LMH6552.. however it does look like the LMH6555 input offset can get to around 10mV so it might not be uncharacteristic for the LMH6552 to get this high. 

Best,

Hasan Babiker

This seems to be the offset listed in the parametric search as well, however I think the common-mode offset voltage was used for this field since VosCM is listed as 16.5 mV max in the datasheet. I checked the original datasheet of the device to confirm and the older version also has no indication of differential offset voltage.

Best,

Hasan Babiker

So I checked that line item and I entered that data, not a mistake I would have made, the other CFA FDA in this table is the LMH6554 - it shows 16mV input offset -CFA typically do have poor offset voltages. 

So I did find a 2007 LMH6552 datasheet, no input offset voltage spec there either.Frankly,I think I need to not support FDA's that fail to provide basic specs like this. 

I will say (by way of clue) most parts have the same person doing test development - no self respecting test guy would do a test flow without an input offset voltage test - it is possible the production data is out there if you wanted to fill in this egregious gap. 

Hi, Michael，

    thanks your answers. 

    this LMH6552 is used for AWG driver.   signal path is   DAC output to buffer to driver. 

   our requirement is signal bandwidth 80MHz, output differential level 16Vpp, SNR>75dB, THD>85dB. 

    please review if any recommendations for better performance differential device? 

   thanks. 

Hi, Kai;

    this is awg circuit,  ADC output to buffer, then connect to OPAmp, I just screenshot the driver circuit.

    usually we pay atteation OCM input level and output DC level error.  out+ and out-  usually no DC error.    LMH6552 have DC error form out+ and out-,  I don't know this is device performance or may circuit mistake.

    thanks.

Well Xinjian, if you eventually need a single ended output, look at the THS3215 or THS3217 - specifically intended for this purpose. 

I don't think you have a ckt mistake, I think the unspecified LMH6552 input offset voltage is likely pretty high with a wide +/- range. 

THD of 85dB at 80MHz 16Vpp likely into a 100ohm load is a stretch, but  the THS3491 was done to approach that.

Great posts from you, Michael :-)

Kai

Thanks Kai, The guys doing those LMH CFA based FDA's were actually the original Comlinear Designers - really kind of slippery to just not spec what is probably a poor input offset voltage. 

After doing some digging regarding differential offset voltage of the LMH6552, it looks like there was some early measured data collected for the part. From what I found, the device seems to have a typical offset of 5mV although not sure what the min/max of this spec is. Either way, it seems that the 10mV offset measured by Xinjian is definitely not outside of the scope of this device.

Best,

Hasan Babiker

Interesting Hasan, 

It would be far more interesting if there was a test line for this (normally, if you are doing Ios you have to do Vos  to solve for it). I bet it is in the production test flow and those folks have years of histograms (and probably a limit to screen bad parts). 

Hey Michael,

Looks like I spoke too soon.. characterization for the device looks to have speced min/max Vos at +/- 40mV. Not sure if this is what is being used as limit for screening however.

Best,

Hasan Babiker