ISO224: Output ripple issue

Zhou Dong

Part Number: ISO224
Other Parts Discussed in Thread: , ISO124

Hello,

I have an issue with using ISO224. The thing is that I measured an output ripple of around 110kHz and 700mV at the output of ISO224(OUTP to OUTN).

The voltage for Vdd1 and Vdd2 are both 5V and measured with negligible ripple. The input is a DC voltage around 7.2 V DC.

I tried to disconnect everything connected OUTP and OUTN, which does not help solve it. Also, I tried another ISO224 chip, which show the same results.

Any idea about the issue?

Thanks

over 3 years ago

0 kai klaas69 over 3 years ago

Guru 140200 points

Hi Zhou,

the ISO224 provides a differential output. So have you added a differential amplifier to the output as shown in the ISO224EVM?

Kai

0 Zhou Dong over 3 years ago in reply to kai klaas69

Prodigy 100 points

Yes, I also tested the differential amplifier but it still has the ripple.

kai klaas69 said:

Hi Zhou,

the ISO224 provides a differential output. So have you added a differential amplifier to the output as shown in the ISO224EVM?

Kai

0 Alexander Smith over 3 years ago in reply to kai klaas69

TI__Mastermind 47625 points

Hi Zhou,

In addition to Kai's comments, have you tried measuring this with a differential probe?

What kind of filters do you have? Can you share your schematic?

0 Zhou Dong over 3 years ago in reply to Alexander Smith

Prodigy 100 points

Hello Alex,

I didn't try with a differential probe but the circuit is isolated from the ground. So it should be fine even with a passive probe.

Below please find my schematics

Thanks,

Zhou

0 Zhou Dong over 3 years ago in reply to Zhou Dong

Prodigy 100 points

Sorry, the figure seems not to work.

0 kai klaas69 over 3 years ago in reply to Zhou Dong

Guru 140200 points

Hi Zhou,

please connect GND to AGND.

Also, add caps in parallel to R14 and R31.

Why not exactly adopting the circuit (or at least the passive components) shown in the EVM?

Kai

0 Zhou Dong over 3 years ago in reply to kai klaas69

Prodigy 100 points

Hello Kai,

GND and AGND are connected.

I just added the caps (10pF) in parallel to R14 and R31. I also changed all the resistance to be 105kohm. The results did not change. Also, the ripple is large, that's why I think these passives should not influence that much. Below is the tested voltage across Voutp and Voutn.

By the way, the chip has a wide power supply voltage of the high side. What could be the difference between using a 5V power supply and an 18V power supply except the power consumption? Should I make sure the input voltage (VIN) to be smaller than VDD1? Based on the datasheet, it should be fine. I just want to double check it.

Thanks,

Zhou

0 Alexander Smith over 3 years ago in reply to Zhou Dong

TI__Mastermind 47625 points

Hi Zhou,

There should be no problem with Vin > VDD, as there is an internal charge pump for this device. I see on the input that you only have a capacitor, there is no series resistor. Try adding a series resistor to the input such that an RC filter with 1MHz cutoff fc is achieved. I would recommend 10-20ohms, and reducing the size of the capacitor on the input. It is possible that a high frequency signal is aliasing into the sample band.

0 kai klaas69 over 3 years ago in reply to Zhou Dong

Guru 140200 points

Hi Zhou,

eventually the ripple can be made smaller. What is the signal bandwidth you want to transmit?

Kai

0 Zhou Dong over 3 years ago in reply to kai klaas69

Prodigy 100 points

Hello Kai,

The signal transferred is just DC. And I tested the input, it is DC. That's why I feel weird the output can have a large ripple.

Thanks,

Zhou

0 kai klaas69 over 3 years ago in reply to Zhou Dong

Guru 140200 points

Hi Zhou,

how do you generate the many supply voltages? Are you using any switchers (switch mode power supplies, DC/DC-convertors)? Can you show scope plots of the supply voltages? Can you measure DG_GND against GND and publish the scope plot?

Can you connect all the grounds to each other for a test?

Kai

0 Zhou Dong over 3 years ago in reply to kai klaas69

Prodigy 100 points

Hello Kai,

Thank you for your reply. I'm using linear power supply for the high end while a switching power module for low end. But both should have small ripples and they are shown below.

The voltage is measured between two GNDs. I'm using AC measurement to look at the ripple and the DC bias is around 6V. There's some ripple but the frequency is around 3 times higher than the ripple of voltage across Voutp and Voutn. Also, I tried to short circuit two GNDs, but the ripple across Voutp and Voutn seems not to change.

0 Zhou Dong over 3 years ago in reply to Alexander Smith

Prodigy 100 points

Hello Alex,

I've also tried adding a resistor, but did not help. Any other possible issues? I'm just curious, no one ever met similar issues when using this chip?

Thanks,

Zhou

0 Alexander Smith over 3 years ago in reply to Zhou Dong

TI__Mastermind 47625 points

Hi Zhou,

Can you share your layout?

Ripple on the output is most often caused by some disturbance in the power supplies. Is there something you have that is injecting noise into one of the device grounds? Switches, big motor, etc?

Do you have an EVM that you can try a similar setup to try and eliminate a few of the variables?

0 kai klaas69 over 3 years ago in reply to Alexander Smith

Guru 140200 points

Hi Zhou,

and please post a photo of your setup.

Kai

0 Alexander Smith over 3 years ago in reply to kai klaas69

TI__Mastermind 47625 points

Hi Zhou,

Just checking to see if this is resolved?

0 Gerhard Kreuzer over 3 years ago

Expert 2580 points

.... you wrote:

"The voltage for Vdd1 and Vdd2 are both 5V and measured with negligible ripple. The input is a DC voltage around 7.2 V DC."

If VIn is higher than Vdd1 (supply on the input), there maybe a problem.

Is this really possible, I thought, you have to use 9V or more to get correct results. The gain of the device is 1/3, so using a 5V supply on the output side is quite ok.

With best regards

Gerhard

0 Alexander Smith over 3 years ago in reply to Gerhard Kreuzer

TI__Mastermind 47625 points

Hi Gerhard,

I think you're remembering ISO124. ISO224 is the newer generation device with an internal charge pump that allows for +/-12V input with a 5V high side supply.

0 Zhou Dong over 3 years ago in reply to Alexander Smith

Prodigy 100 points

Hello Alex,

Sorry I didn't reply in time.

The issue has been solved. I didn't notice that there's a 1 uF capacitor soldered between ISO224 and the following differential amplifier in the back of the PCB board.

After removing it, the low-frequency ripple (~100kHz) becomes MHz ripple and easily been filtered by the RC filter after the differential amplifier.

Not sure about the reason. Please let me know if you have any idea about it.

Thank you so much for your help.

Best,

Zhou

0 kai klaas69 over 3 years ago in reply to Zhou Dong

Guru 140200 points

Hi Zhou,

that's really good news :-)

Soldering a 1µF cap directly at the output of ISO224 erodes the phase margin and makes the output stage of ISO224 become instable and oscillate. See the maximum capacitive load drive capability of ISO224:

Kai

0 Alexander Smith over 3 years ago in reply to Zhou Dong

TI__Mastermind 47625 points

Hi Zhou,

Great to hear! Please let us know if you have additional questions!

Amplifiers

Amplifiers forum

ISO224: Output ripple issue