ISO124: Signal delay with ISO124

Bhupendra Suthar

Part Number: ISO124
Other Parts Discussed in Thread: ISO224, TINA-TI

Hello All, I am using ISO124 with Analog Signal Isolation. Signal is analog , 5Vpp , 82Hz & 2V DC Offset. As per application requirement i am switching this signal (On/OFF) using Mosfet. Mosfet OFF signal must be zero ( gradually ) , Mosfet ON signal must be appear as it is (gradually) . there is passive low pass filter @ o/p of ISO124U.

Problem is : on MOSFET OFF, signal takes much time to become zero (350ms more) , if ISO124U is used. bypassing ISO124U will result less time to become signal zero.

Anyidea what is happening in this case ?

Any suggestion for ISO124U replacement part ?

Thanks,

over 3 years ago

0 kai klaas69 over 3 years ago

Guru 140200 points

Hi Bhupendra,

I would need a schematic :-)

Kai

0 Bhupendra Suthar over 3 years ago

Prodigy 70 points

Sch attached here:

SCH

0 Alexander Smith over 3 years ago

TI__Mastermind 47625 points

Hi Bhupendra,

As per the ISO124U datasheet, Frequency Response, the time to fully settle the signal takes approximately 350uS for the device itself with +/-10V input. 350mS is significantly longer than this, perhaps there are some additional filters or components that account for the additional delay? As Kai mentioned, please provide a schematic.

May I ask you about your application? What are you using the ISO124U to measure and what type of end application is this measurement supporting? Any addtional details will be helpful.

If the wide-body package of the ISO124U is not required, you may want to consider the ISO224 which is a newer device with a shorter delay time of ~3uS.

0 Bhupendra Suthar over 3 years ago in reply to Alexander Smith

Prodigy 70 points

Here Full Sch Sheet :

Signal Waveform is this:

Let me know , if anything more you need.

ISO224 has Differential Output , so any example to make it Single ended ?

0 Alexander Smith over 3 years ago in reply to Bhupendra Suthar

TI__Mastermind 47625 points

Hi Bhupendra,

Looks like there is some significant ringing in the circuit. Kai and I will take a closer look to help identify the source.

In the mean time. yes - this tech note shows how to perform a differential to single-ended conversion: https://www.ti.com/lit/an/sbaa229/sbaa229.pdf

We also have a few circuit "cookbooks" for the ISO224. Here are a couple of examples that I think may help. The first one implements a differential to single-ended conversion and provides the calculations necessary to tailor the voltages to your application:

https://www.ti.com/lit/an/sbaa317/sbaa317.pdf

https://www.ti.com/lit/an/sbaa299/sbaa299.pdf

0 Bhupendra Suthar over 3 years ago in reply to Alexander Smith

Prodigy 70 points

Appreciate your quick support,

Can i simulate ISO124U or equivalent ( I m using SOIC) in TINA-TI Simulator ?

will simulation help me out considering this issue ?

thanks,

0 kai klaas69 over 3 years ago in reply to Bhupendra Suthar

Guru 140200 points

Hi Bhupenra,

the schematic is hard to decipher. And I cannot see the MOSFET section.

Also, add electrolytic caps of 10...100µF in parallel to the 100nF caps in the Pi-filtes. Otherwise the Pi-filters might generate huge ringing. What series resistance do the 10µH chokes provide?

Kai

0 Bhupendra Suthar over 3 years ago in reply to kai klaas69

Prodigy 70 points

Hello Team,

Here is confusion , what you see is not ringing....it is because mosfet off time takes longer hence signal appears for a while...see attached signal image for more clarification. Signal is measured on Pin 2 (CH0) of U13.

Green Signal (CH1) - Original input Signal

Blue Signal (CH2) - MOSFET Switching Signal (OFF = Low Signal).

MOSFET is on another board not part of this schematic /board. It will just On/OFF input Signal. MOSFET OFF = Input Signal OFF.

In same circuitry if bypass ISO124U signal takes 1.4 Sec down to zero (as per image).....but with ISO124U it takes 1.75 Sec to become zero.

Inductor PART is : AIML-1206-100K-T, (DC Resistance :1 Ohm)

Sensor Signal

Thanks,

0 Alexander Smith over 3 years ago in reply to Bhupendra Suthar

TI__Mastermind 47625 points

Hi Bhupendra,

We don't have a model for the ISO124U as it is an older device. However, simulating the passive components may help you identify the source of the issue.

I can't read the schematic so I'm not sure if ISO124U is U13 or not, can you please re-post it?

I don't see CH0 in the waveform capture.

The input of the ISO124U seems to be ringing back and forth constantly. Can you please provide additional detail on what is between the MOSFET and the input of the ISO124U? How are the boards connected if they are separate?

Can you provide a schematic for the MOSFET board?

0 Bhupendra Suthar over 3 years ago in reply to Alexander Smith

Prodigy 70 points

SCH ISO124U SCH.pdfattached again here : IDD_IN is Original i/P signal.

0 Bhupendra Suthar over 3 years ago in reply to Bhupendra Suthar

Prodigy 70 points

Thanks for your help Alex,

Simulating Passive component will not solve purpose, as if ISO124U is bypassed then everything works fine in place.

About ISO224 , it has internal gain of 1/3 instead of our requirement of unity 1 gain.

thanks,

0 kai klaas69 over 3 years ago in reply to Bhupendra Suthar

Guru 140200 points

Hi Bhupendra,

does the issue persist when taking a fresh ISO124?

Kai

0 Bhupendra Suthar over 3 years ago in reply to kai klaas69

Prodigy 70 points

Yes , Kai.
I have three boards, same problem with all.
I am guessing , may be sample hold / buffer ckt inside ISO124U may causing this delay.

Thanks,

0 kai klaas69 over 3 years ago in reply to Bhupendra Suthar

Guru 140200 points

Hi Bhupendra,

are you overdriving the input? The guaranteed input voltage range is +/-10V. But your first scope plot shows a much higher input signal. Please reduce the input signal amplitude to a value not higher than 10V.

Also, the usual input signal of ISO124 is a smooth analog signal without being chooped so heaviliy as in your application. So you might overdrive the ISO124 even dynamically with respect to the finite slew rate of ISO124.

What is the slew rate of your input signal?

Kai

0 Bhupendra Suthar over 3 years ago in reply to kai klaas69

Prodigy 70 points

Hello Kai,

Even signal is lower than 10V ( second scope plot) same issue present.

I am not sure about Slew rate. Maximum signal freq is : 1Khz.

Is it possible to measure slew rate with oscilloscope ?

-Bhupendra

0 kai klaas69 over 3 years ago in reply to Bhupendra Suthar

Guru 140200 points

Hi Bhupendra,

you could post a scope plot with small horizontal deflection time allowing to see the shape of edges :-)

Kai

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

TI__Mastermind 47625 points

Hi Bhupendra,

Thank you for considering the ISO224. I understand that the internal gain of 1/3 is not ideal, especially when working with only a +/-5V input signal. The output of the ISO224 is also differential, so in order to maintain a single-ended output of the same magnitude from the isolated measurement, a differential to single-ended conversion with a gain of 3 would be necessary.

Following Kai's points, I see 4 cycles in 50mS, which corresponds to an 80kHz input signal while the small signal bandwidth of the ISO124U is typ 50kHz. This calculation is very rough, would be best to have a properly zoomed in picture to confirm. Figure 9 from the ISO124U datasheet shows the gain response to input signals of higher frequencies, but I am not able to find anything related to change signal delay vs input frequency. Comparatively, the ISO224 bandwidth is 185kHz for the A version or 275kHz for the B version.

0 Bhupendra Suthar over 3 years ago in reply to Alexander Smith

Prodigy 70 points

Thanks Alex,

Can you help me with ISO224 reference design having 1 to 1 gain i/p and o/p ? Instead of 1/3. Single ended i/P to single ended o/p ?

Moreover, I heard about eFuse for load switching from TI newsletter , (apart from protection ) can we use that instead of mosfet load switching , Switching Freq. in my case is 1Hz ? (www.ti.com/.../slva811a.pdf