ad5933

[KeepItSimpleStupid]

You realize this is an extremely complex project and even the evaluation board needs support. It will likely need a bipolar supply, be battery powered and have the USB isolated. I2C may or may not be necessary, but I haven't even looked at the oscillator issue. Some of this stuff is high speed. The Eval kit does have the software sources available. I think it is VB, so you could use it and modify. Labview would just be another layer to learn. LabView is relatively easy to learn when the instrument drivers are written. They would not be in this case. The evaluation program would be a start. The modifications apparently change the output to the reciprocal and I have no idea what those effects mean. I don't know your ultimate goal, so I can't even suggest a language or system.

You also have to characterize the modifications as well and make sure they work.

 

[grr]

can you suggest a circuit? in the analog site has the evaluation board schematic I think, could you suggest the other schematics it needed? Shall I send my email with pm?
Thank you

 

[KeepItSimpleStupid]

What frequency response are you going to try for? There are two chips in this series. The 5933 or the 5934.

I don't know what I can do.

 

[grr]

whatever maybe 5934. Do you mind me sending my email?
Thank you

 

[KeepItSimpleStupid]

I have no problem with you sending me your email, but if at all possible we should keep the communication open.

Ideally, this design should be simulated and then built. It is a relatively high frequency design. I did grow up in an era where simulation wasn't used in school.

Those two papers provide the basics:
1) What the AFE (Analog front-end) should look like block diagram wise.
2) What the offset free current source should like.
3) Both of the paper have LOTS of missing parts.
4) Both IA's (Instrumentation amplifiers) and regular OP Amps are required.

Some other comments
1) You will need a USB isolator. B&B Electronics does sell one certified for medical use. It's not needed right away.
2) It's very likely that a +-5 volt supply will be needed. Some options are a medically certified one, or a DC-DC converter that operates out of the frequency band of interest and battery power the device. That raises the questions of externally rechargeable batteries?
3) If the signals are low and/or the impedances are high, then guarding techniques may be required. A guard surrounds the shield with a voltage of the same potential. Usually a guard is surrounded by a ground shield too. So,the cables have two shields and one center conductor. This may not be necessary.
4) Offset voltages, bandwidth and low bias currents seem to be an issue.
5) The OP amps will all likely require bias return paths (A given) and Offset Nulling. Both happen to be temperature dependent.

I have worked with sub-pA currents and have designed stuff that worked in the nA range.

I have NO idea what your skill set is.

 

[grr]

Hi, thank you for all the details. Please chek your messages. I sent you a pm.
Do I need a software so I can simulate it? Which do you think it is suitable for this kind of circuits?

 

[KeepItSimpleStupid]

See this article on Circuit Simulations: https://en.wikipedia.org/wiki/Circuit_simulation

LTSPICE: https://www.linear.com/designtools/software/ is free and popular. Naturally, LTSPICE only simulates Linear's parts. Models seem to be hard to come by.

PCB layout tools that are nearly free and popular are: KiCAD (open source), Eagle (limitations) and Designspark

Take a peak at this datasheet for the AD621: https://www.analog.com/en/specialty...tation-amplifiers/ad621/products/product.html

Although, it's not likely suitable, I think it is easy to understand. The voltage to current source application is likely how that design would be based.

You also need to be looking at this page in great detail: https://www.analog.com/en/digital-t...hesis-dds/ad5933/products/EVAL-AD5933/eb.html I haven't even brushed on this stuff.

You will also have to figure out what you need for a clock.

 

[KeepItSimpleStupid]

If the AD5934 is needed, there are most likely other better chips to use because that one is 5 years old and it looks like it uses SPI. I2C is called the Inter Integrated Circuit Communications Bus. You might want to take a brief look here: https://www.electro-tech-online.com/custompdfs/2012/08/an168.pdf

What was talked about before is that you needed a USB to I2C converter to talk to the AD5933. I2C is a bus, so it can support more than one instrument, but evaluation kits make all sorts of assumptions. So, you might require two evaluation kits and their USB interface for the freq synthesizer/divider and the network analyizer and that also means two USB isolators.

Now, you could "break" the evaluation kit(s) so it uses just the I2C bus and use a single USB to I2C converter with an isolator.

This was an interesting I2C USB solution: https://www.i2cchip.com/ and this company, https://www.diolan.com/ probably had the most comprehensive software interface.

You do have to get an idea of the entire hardware/software interface, at least for now, in an overview fashion.

So, I don;t know if your looking for the smallest, nearly wearable gizmo that doesn't require a PC or something that's wearable; a gizmo that returns a number or two or something that records info to be analyzed by a PC or some quick, dirty and safe contraption that needs a PC. For now, I'm assuming the last one.

 

[grr]

Hi, I did have a look all the above links you recomended!I dowloaded and installed LTspice and I tried. Can I simulate a circuit with ad5933?

 

[KeepItSimpleStupid]

My guess is no. Goto www.analog.com and search for ad5933 model. Hopefully youll come to the same conclusion. Use the free multisim, but the ad5933 is not likely model material. You can't muck with the registers because the sims only model electrical parameters.

The low Z amp would be.

The IFA would be and so far AD parts are a top contender. TI doesn't look too good.

 

[KeepItSimpleStupid]

Try something like the AD620 and the other OP amp suggested with a +-5 V supply in a voltage to current conversion. Current would be 350uA max RMS, I think with the max voltage of the ad5933. So you could look at the freq response of the circuit. You may be able to look at output offset voltages and input bias currents - don't know.

You could look at the effect of changing the input Z as well. Remember that the AD5933 should be able to resolve 10 M ohms.

You will find that tthe circuit doesn't have enough bandwidth.

 

[KeepItSimpleStupid]

Here is another tutorial you should look at relating to Input Offset Voltage: https://www.electro-tech-online.com/custompdfs/2012/08/MT-037.pdf

 

[KeepItSimpleStupid]

Here is an interesting site: https://www.rdm-apps.com/modules.html The reason why it's particularly interesting is that it nearly contains the building blocks you need at the right bandwidths, but not at the right price.

 

[grr]

I agree with you, it is expensive! So lets summarize up to here, at least the diagram: an evaluation board with ad5933, an I2C to usb to connect to pc, what about a circuit for connecting the 4 electrodes? what else do you think?
Any other circuit based in a different chip to measure biompedance and it could be possible to simulate?
Thanks

 

[KeepItSimpleStupid]

an evaluation board with ad5933, ad5934?
an I2C to usb to connect to pc (I2C is a maybe)
what about a circuit for connecting the 4 electrodes (That's the problem)

5933 vs 5934?
Bandwidth required?
Remember that GBW or the product of gain and BW is a constant. so a 10 MHz amp at A=1 is a 1 MHz amp at A=10
Isolated USB. Medically certified, preferred unless your working with rats.
What's required for the clock and/or divider? I2c? Isolated?
Power supply? Probably +-5
The AFE interface

Assume PC based?
What language? Modify AD's BASIC code?
There was a mentioned effect where the V/I functions are inverted. Probably because with the constant (V+v)/(I+i )never blows up, but V/I does and I/V doesn't when I=0.

OP amp selection process. Bandwidth, Offset and Ib; Which matter the most and where.

USB controls the evaluation board. I2C controls some of the suggested frequency synthesizer. So, the issue of control has to be worked out.

Not sure if having two evaluation boards makes any sense?

 

[grr]

codeforge has some code for this chip: **broken link removed**
The more I am looking at it the more confused I am......At least the diagram for the beginning but even that I think it is impossible....

 

[KeepItSimpleStupid]

This isn't going to be easy.

There was a blurb that the source of the basic code is available from analog devices, but you have to ask. PDF page 15 from here: https://www.electro-tech-online.com/custompdfs/2012/08/537700023EVAL_AD5933EB.pdf

'This section outlines the evaluation board code structure required to set up the AD5933 frequency sweep. The sweep flow outline is shown in Figure 26. Each section of the flowchart will be explained with the help of Visual Basic code extracts. The evaluation board source code (Visual Basic) is available upon request from the Analog Devices Technical Support Center. The firmware code (C code), which is downloaded to the USB microcontroller connected to the AD5933, implements the low level I2C signal control (that is, read and write vendor requests).

 

[KeepItSimpleStupid]

This https://www.analog.com/en/all-opera...plifiers-op-amps/ad8599/products/product.html with the 4627 doesn't look like a bad starting point.

You can add offset correction, again making things complicated, but with a 4 pA bias current and a 15 uV offset voltage and high enough BW, the combination looks like a very good building block with +-5V supplies.

I'm not sure at this point, but it MIGHT be possible to build a circuit based on these parts that works in 1 of two ways:
a) Adds Vdd/2 as an offset - which means the part doesn't change the operation of the chip
b) Adds an offset/null which means it's the part you need.

If it's the case, then this part (voltage to current converter) of the AFE could be tested independently. Not sure though.

 

[KeepItSimpleStupid]

Input offset voltage design article

This is a good read: **broken link removed**

 

[grr]

Is it possible to send me a diagram of the circuit? what it needs?
Thank you