What is wrong with my DAC

[didibabawu]

I have a DAC taped out and test. The following is some of the test result. in the first one, the Fs=100MHz Fin=100k and the result is ok. In the second one, the Fs=125MHz Fin=100k and something is wrong with the DAC.

The capture instrument I used must work below 50MHz, so when I did my test, I reduced the capture frequency to Fs/M, and I used M=8. So when the Fs=100MHz, the capture freqency was 12.5MHz ,and 15.625MHz for 125MHz.

Can anyone give me some advise about my DAC. Thank you very much.

 

[keith1200rs]

I'm not convinced by the test method. If the sample clock of the measuring system and the DAC are not synchronous you could get strange alias problems and that could be what you are seeing. Look on a proper spectrum analyser if you can. Try different sample frequencies - if the high frequency harmonics move then they are most likely caused by the sampling.

Keith

 

[didibabawu]

I'm not convinced by the test method. If the sample clock of the measuring system and the DAC are not synchronous you could get strange alias problems and that could be what you are seeing. Look on a proper spectrum analyser if you can. Try different sample frequencies - if the high frequency harmonics move then they are most likely caused by the sampling.

Keith

Hi Keith,

Thank you very much for your reply, I also do some test when Fs=160MHz, the frequency harmonics did not move when Fs is larger.

 

[keith1200rs]

What about if you keep Fs the same and change your capture sample rate? Also, check your sampling system with a bench sine generator.

Can you measure the exact frequency of the two harmonics? There must be some significance. Have you looked at the actual waveform output?

What type of DAC is it?

Have you tried to reproduce your test system in simulations?

Keith

 

[JoannesPaulus]

It seems that those two tones are at about 56MHz+-100k in both cases. What happens if you increase/decrease the input frequency?

 

[didibabawu]

What about if you keep Fs the same and change your capture sample rate? Also, check your sampling system with a bench sine generator.

Can you measure the exact frequency of the two harmonics? There must be some significance. Have you looked at the actual waveform output?

What type of DAC is it?

Have you tried to reproduce your test system in simulations?

Keith

It's a current steering DAC, I also do some test when Fs=100MHz, Fin=1M. The result is in the pic.

I'm sorry the pic is large,but you can see it more clearly.

 

[keith1200rs]

I definitely think you should test a bench sine generator and look at the DAC output directly on an oscilloscope.

Keith.

 

[JoannesPaulus]

So this tone moves with the input frequency but does not move with the clock frequency. Moreover, it always shows skirts. Do you have an oscillator/PLL on chip? What frequency? Do you have a switching power regulator? I believe, you need to find a 56MHz signal that mixes with your input.