I'm wondering if anyone has any idea why I see a 50hz spike when using a ADuM7440 SPI isolator IC between a micro-controller and a DAC. I don't have much experience with isolators but just read the datasheet and designed the PCB myself. I have two 5V PSUs where I intended to use one for the DAC IC and one for the Arduino microcontroller with the ADuM7440 isolating them.
These are my observations:
1x 5V PSU with ADuM7440 bypassed = Good
2x 5V PSU with ADuM7440 for isolation = 50hz spike.
Here is the schematic:
http://www.diyaudio.com/forums/atta...g-switches-multiplying-dac-mdac-schematic.jpg
To bypass the ADuM7440 IC I just connected the VDD/5V and GNDs from either side of the IC.
Attached is the layout and two graphs showing the 50hz spike.
Did I make some mistake? Poor layout? It just seems weird that either PSU on its own measure fine but as soon as I use the IC to isolate I get a 50hz spike
These are my observations:
1x 5V PSU with ADuM7440 bypassed = Good
2x 5V PSU with ADuM7440 for isolation = 50hz spike.
Here is the schematic:
http://www.diyaudio.com/forums/atta...g-switches-multiplying-dac-mdac-schematic.jpg
To bypass the ADuM7440 IC I just connected the VDD/5V and GNDs from either side of the IC.
Attached is the layout and two graphs showing the 50hz spike.
Did I make some mistake? Poor layout? It just seems weird that either PSU on its own measure fine but as soon as I use the IC to isolate I get a 50hz spike
Attachments
I would have to guess the problem is with the power supply.
If it was a problem with the board I would expect the spike to
be present all the time.
Try swapping the two supplies and see if the spike goes away.
What power supplies are you using?
Are you able to reverse the phase of the AC input for one of
the supplies?
If it was a problem with the board I would expect the spike to
be present all the time.
Try swapping the two supplies and see if the spike goes away.
What power supplies are you using?
Are you able to reverse the phase of the AC input for one of
the supplies?
I tried this. Either power supply on its own doesn't produce the spike. I also tried a battery supplying the Arduino and my 5V PSU powering the DAC but still get the same spike.
Two of these.
You mean just swapping L and N on the mains side? I just tried this, still a spike!
Very interesting, i found exaclty the same problem with an ADUM4160 USB isolator, i have two different boards and one of them has a 50hz spike, the other doesn't. I assumed the issue was due to the 12" of cable between the power supply and isolator module. The board without the issue has an onboard rectifier and gets AC power directly from a wall wart, where as the noisy one has a power supply in a separate enclosure.
Yeah. It's on and connected for all measurements. It controls the DAC so is required.
Hi,
I asked the question because when you send the data from the micro to the dac you have to clock it. You are sending 16 bits to the dac. The clock would radiate noise when clocking the data to the dac. You can do a check to see if it is coming from the communication between the dac and the micro. Connect everything and program the micro to halt and see what happen.
I asked the question because when you send the data from the micro to the dac you have to clock it. You are sending 16 bits to the dac. The clock would radiate noise when clocking the data to the dac. You can do a check to see if it is coming from the communication between the dac and the micro. Connect everything and program the micro to halt and see what happen.
I'm guessing I had something to do with one of those...
Yup, yours would be the quite one
The default SPI clock for the Arduino is 4Mhz
I could try disconnecting the SPI lines after all the parameters have been set and then see how it measures though?Yeah I can power the Arduino with a battery and I still get the spike.
Not sure! This is my rectifier: DFL1508S-E3/45 - VISHAY - BRIDGE RECTIFIER, 1.5A, 800V | Farnell UK
The spike is present in both of your plots, just the level changes. The first one has around -119dB relative to FS, the second around -134dB. Assuming we're looking at 1VRMS for the 0dB reference point these signals are in the region of 1uV and 200nVRMS respectively.
Given the absence of harmonics my first guess would be capacitive coupling from the 50Hz mains. Let's get some handle on the numbers involved - at 50Hz a 1pF capacitor looks like 3Gohms - if your DAC has a 100ohm output impedance this makes a 30,000,000:1 potential divider. So with 240VRMS on the input to this, the output would be 8uV. It seems therefore that your spike at -119dB could be due to just (very roughly) 0.1pF of stray capacitance to mains.
Given the absence of harmonics my first guess would be capacitive coupling from the 50Hz mains. Let's get some handle on the numbers involved - at 50Hz a 1pF capacitor looks like 3Gohms - if your DAC has a 100ohm output impedance this makes a 30,000,000:1 potential divider. So with 240VRMS on the input to this, the output would be 8uV. It seems therefore that your spike at -119dB could be due to just (very roughly) 0.1pF of stray capacitance to mains.
With all due respect, but what do you want to achieve... a perfect measurement? The -119db is inaudible and with the sensitvity of the human ear at that frequency you'll probably never hear it no matter how much you cranck the volume up..
Now ta further reducing the peak.
Stray capacitance is usually caused by the construction of your setup. So you could resort to srparate the psu from the DAC if possible.
Also you could be plagued by a minor grounding issue.
So perhaps a pic or two might be helpfull assessing the situation.
I know -119 is quite low but sometimes it measures near -100, not sure why it changes though. To be honest my main reason for persuing this issue is because I'd like to learn why it happens out of curiosity, not because I'm obsessed with measurements. And I also thought I'd made some mistake with my layout or something else obvious since the difference in measurements was so noticeable.
Thanks! My setup wiring is a bit of a mess so perhaps I'll fix that and put some internal shielding in place between the analog circuits and the mains/PSUs and then redo the measurements.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.