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.

DAC7811: Obtaining 0 to 5.5V swing with only a positive power supply/reference

Part Number: DAC7811
Other Parts Discussed in Thread: OPA350, DAC902, DAC80501, DAC60501

Tool/software:

My team is looking to use a SPI-controlled DAC to generate 0 to 5.5V analog signals with maximum frequencies of 1-2 MHz. We think the DAC7811 fits these requirements but are open to any other suggestions if you think there's a better part.

In any case, we are trying to figure out the circuitry necessary to achieve a 0 to 5.5V swing from this DAC without a negative supply or reference to pull from. I was looking at the below topology from this appnote:

I had two thoughts here:

  1. Use a 5.5V LDO to provide the reference into IOUT1. This would prove less accuracy (1% vs ~0.05%) but 1% accuracy should be fine for our application. Then use an OPA350 on the REFIN to provide a fast output voltage swing.
  2. Use a REF5050 5V voltage reference, then a precision op amp to boost the IOUT1 reference voltage up to 5.5V. 

Would either of these solutions work, making it mostly question of how much accuracy I want from the DAC? Also, are there any other considerations I might be overlooking when pulling the output from REFIN instead of IOUT1? 

Best,

Malcolm

  • Hi Malcom, 

    What type of 2MHz signal are you trying to create? The max SPI clock is 50MHz, and each DAC update takes 16 SPI clocks. You could get a 1.5MHz square wave as there are only 2 voltages per period in a square wave. To get something like a 24-point sine wave the max output frequency would be around 130kHz.

    Best,

    Katlynne Joens 

  • Hi Katlynne,

    We are looking mostly at sine waves and then some AWG. I've spent a good amount of time looking through the portfolio and if a 130kHz sine wave is the best we can do on the precision DAC end then that's fine. I'm looking at the DAC902 as well, but we will need 4-8 total output channels, and I don't think we have the bandwidth to incorporate such a large amount of separate inputs as demanded by parallel CMOS.

    Best,

    Malcolm

  • Hi Malcom, 

    The max will be more like 130kHz per SPI bus. If you have two channels on the same 50MHz SPI bus both with 24-points per period, then the max frequency gets cut in half. Basically, you can only make one DAC update on any channel every 480ns. 

    You could use 4-8 single channel DACs with their own SPI busses. 

    Either way, I don't think you need to use an MDAC like the DAC7811 and in that case you don't have to worry about the DAC output being inverted with respect to the reference. You can use something like DAC80501.

    What is this for? 

    Best,

    Katlynne 

  • Hi Katlynne,

    We are making an EVM that will allow customers to test devices in real time using various waveforms. The vast majority of these are going to be square waves/digital patterns, but some will be sine/triangle/sawtooth/etc. analog inputs mainly for comparators, which is what we need the DACs for.

    We have a 12-bit ADC elsewhere on the board so probably the DAC60501 makes the most sense here.

    However I still have the same question as before: if I want a 0 to 5.5V voltage swing, do I either have to sacrifice precision using an LDO to get a 5.5V reference or introduce another amplifier to increase a 5V external reference to 5.5V?

    Best,
    Malcolm

  • Hi Malcom,

    If you're fine with the accuracy of the LDO then it is fine to use that as the reference to save having to use another part. If you just use the supply as the reference, you won't need any other external components because this device has an internal output buffer. 

    Do you need any current from the output? If you do need more than a few mA of current, then the DAC will have a tough time operating at full-scale given that the full-scale is equal to your rail. If you do need to source current, then I would recommend the external op amp method with the internal reference. 

    Here's our headroom plot:

    Best,

    Katlynne Jones

  • Hi Katlynne,

    No, I think this should be good; we don't require any current drive.

    Just checking my math/understanding here but this device would require 24 SPI clocks/update meaning that our max SINE frequency @ 24 points would be about 86kHz?

    Best,

    Malcolm

  • Hi Malcom, 

    Yes. The original device you were considering had a 16-bit SPI frame (many of our older devices do). But DAC60501 does have a 24-bit SPI frame so your calculation is correct.

    Best,

    Katlynne Jones  

  • Hi Katlynne,

    Thanks so much you've been super helpful.

    Just one last question, given what you know about our application (i.e. trying to keep the analog waveform frontend relatively simple) if we needed a device that could generate 1MHz sine waves, what device would you suggest?

    Best,

    Malcolm

  • Hi Malcom, 

    I don't have a good answer for that. We don't have any precision DACs that would get up to a 1MHz sine wave. We'll max out at 100-130kHz. 

    You can check with the high speed team for suggestions on the simplest communication protocol for whatever MCU you are using. Otherwise, I know there are ways to get a sine wave out of a 555 timer, or you can explore more discrete sine wave generation options. 

    Paul from the amps team shared some suggestions in this post: TLV3501: 6 MHz Sine Wave signal Generator Circuit - Amplifiers forum - Amplifiers - TI E2E support forums

    Best,

    Katlynne