Question on ISO124

Hi,

The high-voltage isolation characteristics of the ISO124 are defined across the barrier from one side to the other, input and output. Should the input side ground (earth) rise a few kilovolts relative to the output ground, then the rated isolation  barrier voltages would apply; 1.5kV continuous, 2.4kV for 1sec. Anything above that is likely to damage the isolation barrier.

If it is a matter of the input level rising to a few kilovolts, then without any protection in place the amplifier would be damaged. However, the schematic provided does show a transient voltage suppressor (TVS) and a MOV between the input and the input ground. Those protection devices, along with some added series resistance, should provide a current-limiting clamp, effective against any large applied voltage. The ISO124 datasheet states an absolute maximum input voltage (Vin) of 100V. Although I would never recommend testing the input to that level, clamping it to a much lower voltage should keep it from being damaged.

Regards, Thomas

Applications Engineering

HPA - Linear Products

Texas Instruments

Thanks for your reply Thomas!

Your post has cleared up some question I had.

However, I would also like to ask as to whether I can use some charge pump ICs for the +/- rail supply of the ISO124. The datasheet of the ISO124 seems to suggest the use of isolated unregulated dc/dc converters and some LDO regulators.

Is there something fundamentally wrong with using only charge pump ICs?

Thanks and regards

Alex

I am under the impression I have thought of things the wrong way round.

I realize now, that I do indeed need to use an isolated unregulated dc/dc converter to protect the rest of my circuitry from earth potential rise.

However, I do have a question regarding the input voltage swing with regards to the input supply voltages. Figure 9 of the ISO124 suggest that from a +/-5V supply one can get a +/-10V input voltage swing. Is this correct? does this require that the +Vs2 and -Vs2 need be at +/-15V?

Also, the input resistance of 200kOhm is quite low for my needs. Most high voltage dividers (kV range, ratio ranging from 1000-3000:1) require 1Mohm termination. Does TI offer a similar IC to the ISO124 that can do the job?

I should point out here that the high voltage divider would provide the input to my data acquisition system.

Thanks and regards

Alex

Hello Alex,

I don't see anything specifically in the ISO124 data-sheet regrading charge-pump issues. The ISO124 is a legacy product that has been available for a few decades. I suspect that when the data-sheet was written that charge-pumps were not as commonly applied as they are today. Most of the data-sheet applications show transformer coupled power sources that were designed to provide a level of high-voltage isolation. No matter what the power supply sources used they would have to be able to sustain the required level of isolation voltage.

Another issue with a charge-pump, or even a DC-DC transformer solution, is the switching frequency involved with them. The ISO124 modulator and demodulator uses a switching frequency of about 500kHz. It doesn't have high power-supply rejection and it would be easy for power supply-related switching noise to couple into the ISO124 through the power supply lines. The switching noise could mix with the ISO124 switching frequency to produce unexpected frequency spectra in the pass-band. That may be avoided, or reduced, by the power supply filtering shown in Figure 4.

Regards, Thomas 

Alex,

As stated in the data-sheet you can operate the input stage with +/-5V supplies and still drive the input +/-10V (Page 6, Basic Operation - Signal and Supply connections), but the output section must be powered by +/-15V.

The only analog isolation products that are offered are the ISO122 and ISO124. The 200k resistance internal value is set for specific input integrator/ switched current source circuit operation.

Regards, Thomas

Thansk for your swift reply Thomas.

I have quickly created two basic block diagrams to try and explain the problems I am facing.

Components for the figure above, I have found are much cheaper. The three input resistances along with the isoaltion amplfier, should provide a non-inverting  positive offset attenuation to the input signal. However, I face the problem of impedance isolating (buffering) the input impedance (call it Zin) from the source, as well as requiring the isoaltion amplifier to have a large input impedance (which they generally do).

So my question would be how do I buffer the input impedance with components that will withstand high voltage surges.

 Components for the figure above, I have found are much more expensive. However, I still require a bipolar isolation amplifier that has a large input impedance (1Mohm and above).

hope you can provide some light to my questions.

Thanks and regards

Alex

Alex,

Earlier you mentioned that the 200k input impedance of the ISO124 would load your input divider. An operational-amplifier connected as either a unity-gain buffer, or non-inverting amplifier presents a very high impedance to the source that drives it. If you are concerned about possible input transients an input protection circuit can be added. I've attached a PowerPoit slide showing how this is accomplished. If you drive the non-inverting input the proetction would be added to that input and the power supply pins. The PowerPoint slide has viewer notes that explain the protection circuits in detail. 

Regards, Thomas

Hi Thomas and thanks for your reply.

I am grateful to you for the slides you added. I had never though/seen transient protection on each input/output/power pin of an op-amp before, though it makes perfect sense.

Could I ask whether these slides were taken from a book, application note or website? I would like to read more on the topic, especially if they include some 'real life' test results.

I do not think I have seen the op-amp transient protection circuit you showed me, in TI's "Op-amps for everyone" book.

Back to more of my questions....

The more I think about it the more bipolar operation up to the ADC seems way too expensive. I would also have to calculate the possible degradation the ISO124 has on the input signal and I am not quite sure about all the factors involved. The input offset of the ISO124 seems HUGE.

In order to make a more informed decision could you please point out for me an unipolar isolation amplifier for data acquisition systems. The isolation amplifiers needs to have  a bandwidth of at least 10kHz and would work with a 12-bit ADC. The supply voltage of the isolation amplifier should be 3.3V.

Thanks and regards

Alex

Hello Alex,

An isolation amplifier such as the ISO124 must provide reasonable AC accuracy while withstanding a high-voltage (kV) applied across the barrier. It is a difficult feat to achieve very high accuracy performance under this condition. If you review the specifications for the various capacitively-coupled, magnetically-coupled and optically-coupled analog isolation amplifiers available you will find they each have their areas of electrical performance where they shine, or don't shine. The TI analog isolation amplifiers are limited to the ISO120 through ISO124 which are all capacitvely-coupled amplifiers. Some other electronics companies do offer their analog HV solutions and it may be worthwhile to see if one of their products can provide the performance you require.

Possibly, a mixed-signal solution would take care of your requirements. You would follow your input sensor/ transducer with any necessary analog signal conditioning. Then, convert the analog signal to digital using a serial output A-D converter. The serial output stream would be transmitted across the HV barrier via a digital isolator such as TI's ISO721. It has a 2.5kV RMS isolation voltage rating. Then if needed, the serial stream could be converted back to analog using a D-A converter.

I sent my colleague in the Data Converter group a message asking about suitable converters that would have the resolution/ accuracy, speed and supply voltage necessary for your application. Here's an excerpt from what he replied:

I suspect we’re looking at some sort of SPI interface, so the ISO7321 may work out.  He would need another ISO721 as well – that way he’s got isolated SPI (SDI, SDO and SCLK) plus a /CS on the quad so that he can talk to either the ADC or the DAC.  If he’s only got one /CS, he could use an inverter on the isolated side of the ISO7321.  If he’s got two, he could use the IS0721 to control /CS to the DAC for instance.  The SCLK, SDI and SDO can be shared between the two.  For the A/D, look over the ADS8326 and its associated sisters – all are capable of running through at least 100KSPS sample rates and have pin compatible footprints. Resolution ranges from 12/14/16-bits

Here's the list of A-D converters he suggested reviewing:

ADS7816 - ADS7818, ADS7822, ADS7834, ADS7835

And the following D-A converters (if required):

ADS7311, ADS8311 (12/16 bit options)

I hope this is of help to you.

Regards, Thomas

Thanks for your reply Thomas and all the work you are putting into my questions.

I think a mixed signal solution is definitely something much more complicated than I am looking for. However, I am using an external bipolar ADC with SPI interface (on a PCB I have made). The idea of using four ISO721 is, I guess feasible. I am guessing with this option, I will have to surge protect all the analog and power pins being use by the external ADC. I would rather not reconvert the now digital signals back to analog. However, this means adding a low pass active filter before the ADC and surge protecting that as well.

Difficult choices…..I will have to think long and hard about the top approach.

I am contemplating the idea of making a PCB with a unity gain non inverting op-amp (with electrical overstress protection as you suggested), the ISO124 next and finally a low pass active filter.

The PCB should give me a good indication as to whether the ISO124 is acceptable or not.

Could I ask you to point me in the right direction, with regards to more information on EOS protection (books/websites/application notes), like the slides you posted on a previous post?

Thanks and regards

Alex

Alex,

TI has a line of ESD device protection solutions.  You can visit www.ti.com/esd and find a selection guide and other colateral.  Many devices were tailored to certain applications, but can definitely be used in other applications. 

Best Regards,

Dave Yaeger

HPA Product Marketing

Texas Instruments

Hi Alex,

The diagram I posted showing external EOS protection of an operational-amplifier is a slide from a TI Tech Day presentation that I developed a few years ago. I have a pdf version of this presentation I can email to you. I will need your email address. You can call me at (520) 750-2169 if you do not want to list your email address on the forum.

Also, one of my colleagues sent me the circuit diagram of an optically isolated, isolation amplifier design. It uses three, TI wide band OPA350 operational-amplifiers and an Avago Technologies HCNR201 analog opto-coupler. I've attached the circuit for your review.

Regards, Thomas 

Hi Thomas,

I have been reading more on TI’s TechDays seminars on signal conditioning and isolation.

The presentation on EOS-ESD you attached, shows how to surge protect an Op-amp. However, if were to isolate my circuit directly after the ADC (isolate the SPI interface), how would I protect the ADC from earth potential rises on its GND pin?

The figure below shows in very basic block diagrams the signal chain of the circuit.

Is there a relative seminar/case study showing how to surge protect a Op-amp and ADC and then isolate the analog part, as you suggested a couples of posts above?

Thanks and regards

Alex

Alex,

"How would I protect the ADC from earth potential rises on its GND pin?" Are you referring to a rise in the ground potential in conjunction with a current increase in the analog ground circuit? If so, the rise would be a function of the ground impedance. Ideally, the the ground impedance is zero; zero ohms resistance, zero ohms reactance. That cannot be achieved but with correct mixed-signal layout the impedance and ground rise can be minimized. Our Data Converter Applications group has Tech Day presentations about mixed-signal circuit layout techniques that help maximize performance. Is this what you are looking for?

Thomas

Hi Thomas.

Yes, I am looking for a presentation on mitigating earth potential rise problems.

Could you please give me the full name of the presentation you are referring too and maybe the city it took place in. I am not sure I can find it in the 09/10 Ti Tech Day presentations.

Regards

Alex

Hello Alex,

I think it is your lucky day. I've located a TI presentation on analog PCB layout considerations. It was developed by Rick Downs, our Applications Manager, and is one of the most complete treatments on the subject that I believe you will find. I've attached the PowerPoint file. Be sure to view the presenter notes where slide content is explained and additional information is provided.

Regards, Thomas

Thanks for the presentation Thomas.

The presentation is very interesting, but I am looking for something more relevant to circuit isolation.

I am trying to understand how I can isolate all the components on the analog side (active filter and ADC) of a data acquisition PCB, from harmful surges.

A number of interesting TI Tech Day presentations deal with the issue of isolation.

However, I was hoping to find a presentation that discusses the issues of isolation, when the isolation device is placed after the ADC as you suggested I could do, a couple of slides above.

The presentation you sent and similar ones (e.g. Protecting Against EOS of Op Amps) explain how to protect an op-amp from EOS.

However, as I mentioned on the previous slide, the ADC GND pin might be subject to surges in the range 2kV. So will any other component that has its pin grounded, like a LDO regulator.

My question is, if I isolate after the ADC, as far as I understand, I have no way of protecting any device that has its pin grounded.

Simply providing a better ground impedance will not be adequate, as most IC will get damaged if the GND pin rises (or falls) a few hundred milli-volts above (or below) 0V.

…or at least that is my understanding.

Is there an advice you can give me on a circuit that has its isolation after the ADC or should I go back into thinking of using the ISO124.

As far as the ISO124 circuit is concerned, I still have to find a way of protecting the LDO regulators from a ground potential rise.

Regards

Alex

Hello Alexander,

I'll look some more, but I haven't found any TI presentation that addresses your questions.

Regards, Thomas

HI, I NEED A HELP ABOUT ISO 124! Please, here says: 

http://e2e.ti.com/support/amplifiers/precision_amplifiers/f/14/t/156139.aspx#566041

Hi all!

I don't speak 100% english. I apologize if you find mistake in this. I bought iso 124 cause i need isolate 2 stage, 

and i am confused because i wanna use only one battery (9V=VCC) for my circuit, iso 124 requires 2, 

but i have questions, can i use voltage regulators for the other VCC and -VCC? or what supposed to going to do? and the battery has one ground,

 and i dont know if can i use that ground for stage 1 and stage 2? (Figure 1) My teacher told me with DC/DC Converser, 

i found MC34063A but this does not have its own ground, somebody can help me how can i use iso 124 with only one battery?

Thank you, i hope your answer please! I need it soon.