DAC datasheet - how to read if the DAC can drive my circuit?

Thread Starter

tysseng

Joined Apr 14, 2013
1
Hi! I am trying to select a DAC that will be connected to a filter circuit. I would prefer NOT to use an additional op amp to buffer the DAC output. The DAC output will be 0-5V. The first cell of the filter looks like this:

1600024323692.png

Now, looking at two different DACS (one 12bit and one 16bit, this is not really important to the question), I can only find the following about driving capabilities:

MCP4921: http://ww1.microchip.com/downloads/en/DeviceDoc/22248a.pdf

In this datasheet it says that all tests have been done with Rload = 5kOhm and output of 4.096V. I am having a hard time to understand exactly how to interpret Rload and what this means - as far as I can read, Rload means a (theoretical/equivalient) resistor from output to GND.

I understand this as "any Rload HIGHER than 5kOhm will be fine", is this correct?

DAC8830: https://www.ti.com/lit/ds/symlink/dac8830.pdf
This datasheet says:
"These DACs are capable of driving unbuffered loads of 60 kΩ".

Does this mean that anything LOWER than 60kOhm is not possible without additional buffering? And does this mean that this DAC is unsuitable for directly driving my filter?

Sorry if these questions are stupid...
 

OBW0549

Joined Mar 2, 2015
3,545
Now, looking at two different DACS (one 12bit and one 16bit, this is not really important to the question), I can only find the following about driving capabilities:

MCP4921: http://ww1.microchip.com/downloads/en/DeviceDoc/22248a.pdf

In this datasheet it says that all tests have been done with Rload = 5kOhm and output of 4.096V. I am having a hard time to understand exactly how to interpret Rload and what this means - as far as I can read, Rload means a (theoretical/equivalient) resistor from output to GND.

I understand this as "any Rload HIGHER than 5kOhm will be fine", is this correct?

As I read it, the 5 kΩ is simply their test condition under which the various parameters are measured for the data sheet. Certainly, any load resistance higher than that (including your filter, with the component values shown) should be fine.

DAC8830: https://www.ti.com/lit/ds/symlink/dac8830.pdf
This datasheet says:
"These DACs are capable of driving unbuffered loads of 60 kΩ".

Does this mean that anything LOWER than 60kOhm is not possible without additional buffering? And does this mean that this DAC is unsuitable for directly driving my filter?
Yes, to both questions. The output of this DAC appears to be directly from its R-2R ladder, without buffering. An external buffer is necessary.
 

Papabravo

Joined Feb 24, 2006
14,393
Your opamp circuit does not show the supply pins so I don't know if you intend for this to be a single or dual supply application. This circuit is going to give you trouble because there is no DC path to Ground on the high impedance inputs. The ability to drive a load depends on the nature of the DAC output. I'll take a closer look and get back to you.

Never mind. I'll go with what Obie says, but I still think you need a DC path to ground for dual supply operation or to Vcc/2 (including a DC path to ground) for single supply operation.
#1 google hit for "do opamp circuits need a DC path to ground"

It is important to note that, just as we found for transistor circuits, one should always provide a DC path to ground for op-amp inputs. Otherwise, charge will build up on the effective capacitance of the inputs and the large gain will convert this voltage (= Q/C) into a large and uncontrolled output voltage offset.

Pay heed young Jedi to the wisdom of the masters.
 
Last edited:

Irving

Joined Jan 30, 2016
1,006
Your interpretation is largely correct.

You can think of the DAC output as a pure voltage source in series with a resistor. The resistor and the input impedance of the 'load' act as a voltage divider so the lower the load impedance the lower, proportionately, the DAC output voltage. The DAC's output voltage swing is only guaranteed at the spec'd load impedance.

The MCP4921 datasheet simply states that : "...the input impedance of the output amplifier should be as high as possible. ", which implies buffering. Unfortunately it doesn't state what the output impedance is, only the load for the test condition which is 5k. In reality you should go much higher than that, with a simple unity gain voltage follower.

The DAC8830 gives the output impedance as 6.25kOhm. This implies it is an unbuffered output and therefore needs buffering, in line with the minimum 60kOhm suggested. Again a simple unity gain voltage follower would suffice.
 
Top