AMC3336: Using AMC3336 for DC-link measurement. How to get full scale for DC-voltage 10%...90%?

Marco Beldermann

Part Number: AMC3336
Other Parts Discussed in Thread: AMC3330

Hello,

I want to measure a DC voltage and connect it like the schematic on the datasheet AMC3336-Q1 p.26 (typi. application).

The HGND is in the middle of the resistor divider and INP/ INN inputs are connected upper and below the shunt resistor. But DOUT shows me a duty cycle of 50% at 0V, 90% at 1V and 10% at -1V input voltage (like the ac-measurement connection below => HGND connected to INN).

So if I change the connection with HGND from INN to the middle of the resistor divider, I don´t notice any changes.

Is the only change to get a higher accuracy and linearity, but the measure range is unchanged? I thought I can measure from +1V INP to -1V INN means 2VDC instead of 1VAC (peak), but I think I am wrong?

over 2 years ago

0 Alexander Smith over 2 years ago

TI__Guru 54122 points

Hi Marco,

If I'm following correctly, you have your schematic similar to the configuration as shown in this cookbook circuit with the split-tap HGND connection?: https://www.ti.com/lit/an/sbaa553/sbaa553.pdf

Are you connected to a middle point and measuring from -DC to +DC or from 0V to +DC? Attaching your schematic(s) would help.

If only measuring from 0V to +DC, then the input signal would only only swing from 0V to +1V, which would result in losing some of your resolution, however it very well may still meet your accuracy requirement.

0 Marco Beldermann over 2 years ago in reply to Alexander Smith

Prodigy 10 points

Hi Alex,

the schematic is similar to your link with the split-tap HGND connection.

I am measuring from +DC to -DC = 1V (+DC to HGND = 500mV; HGND to -DC = -500mV). And the result is getting a duty cycle of 90%. I thought I can have a absolute input voltage of 2V (input voltage to the amplifier INP to INN) with this configuration. But when the duty cycle is 90% at 1V, it is not possible to increase the input voltage to 2V. In this kind I can only use the half of the resolution, from 50% to 90% duty cycle.

Attached my schematic. As you can see i have changed my schematic with the split-tap HGND connection (red board). It is the first prototype and I have note that I also need some small RC-Filter for CLKIN and DOUT. I suggest at 10MHz CLK I am using 47Ohm and 33pF (Fg = 102.6MHz).

+1 Alexander Smith over 2 years ago in reply to Marco Beldermann

TI__Guru 54122 points

Hi Marco,

When HGND = VINN, then VINP can swing +/-1V with respect to HGND = VINN.

So not only would the device be able to measure from +DC to -DC = 1V (+DC to HGND = 500mV; HGND to -DC = -500mV), but also +DC to -DC = -1V (+DC to HGND = -500mV; HGND to -DC = +500mV). "1AC peak" as you wrote in the original post is correct.

The split tap connection helps to increase linearity in the measurement and depending on your accuracy requirement, may very well be able to meet your needs. Comparing accuracy with the AMC3330 (amplifier with similar +/-1V input structure), only using half of the range results in a very minimal degradation in output referred accuracy. https://www.ti.com/lit/zip/sbar013