micro AC current generator and voltage controlled switch

Thread Starter

Michele Losito

Joined Nov 23, 2017
13
Hi all!

I'm working on a device that produces an AC current with an RMS value between 50uA and 1mA and a frequency between 50kHz and 150kHz on an unknown load.
I'm looking for some kind of safety device to place after the V / I converter, that enable / disable the flow of current under certain conditions or redirect the current on a known load... I have some doubt about what kind of device could be good for my application.

Since the signals are weak (and sure noisy), I need a device with low noisy and also with a low Ron (I'm thinking about mosfets...) , but the device must work with an AC current, then I need a device that will works bidirectional...

I was looking an analog switch like this ADG419 but this device have an high Ron and this could be a problem for me ....

I also need to measure continously the value of this microcurrent:
- first on the known load;
- after on the unknown load.

If, for some reason, the current if more than a certain value, I stop the flow of the current on the unknown load and redirect the current on the known load.

Do you have any suggestion, idea or advice to give me?
Thank you in advance for your help or even for a brainstorming! ;)

Cheers,
Mik
 

Thread Starter

Michele Losito

Joined Nov 23, 2017
13
Without knowing the load value, or at least the upper and lower limits, it's difficult to give any advice.
Is this a medical treatment project?
Hello Alec_t,
the load is unknown could be from 3 to 50 Ohm.
This is not for medical treatment but yes, is a device for medical use. :)

Thank you for your attention ;)
 

Alec_t

Joined Sep 17, 2013
14,280
If the load resistance is 50 Ohms max, then the maximum signl voltage needed is 50mV. Perhaps analogue gates (e.g. CD4066) could be used to pass/block/re-route that signal.
 

Thread Starter

Michele Losito

Joined Nov 23, 2017
13
If the load resistance is 50 Ohms max, then the maximum signl voltage needed is 50mV. Perhaps analogue gates (e.g. CD4066) could be used to pass/block/re-route that signal.
Thank you again Alec!
I see the CD4066 datasheet , interesting solution, but still with a high Ron between 125 and 1050 Ohm... this could be a problem.
What do you think about this: TS3A5223 0.45 Ω 2-Channel SPDT Bidirectional Analog Switch

I think that is not very clear if this device could work with an AC signal, because in the page 6 of the datasheet, typical caracteristics, is shown only the graph of Ron VS Vin with positive values of Vin...

What do you think about this?
 

DickCappels

Joined Aug 21, 2008
10,152
You can use an opamp to close the loop around the transmission gate/mux. With 1K you only need to loose 1 volt across the device to get your 1 ma.

For example, use a Howland current pump with a transmission gate between the output of the opamp and the rest of the circuitry.

Or do you think you need galvanic isolation for safety?
 

Thread Starter

Michele Losito

Joined Nov 23, 2017
13
About TS3A5223 , I see this discussion on the TI E2E Community ...
This in particular :
The TS5A5223 is passive FET switch. The Gate of the FET is biased by the Vcc pin and the drain and source is the COM and NO pins. This part is not capable of sourcing or sinking current but will pass what ever signal placed on the COM pin to the NO pin ... [CUT]
 

Thread Starter

Michele Losito

Joined Nov 23, 2017
13
You can use an opamp to close the loop around the transmission gate/mux. With 1K you only need to loose 1 volt across the device to get your 1 ma.

For example, use a Howland current pump with a transmission gate between the output of the opamp and the rest of the circuitry.

Or do you think you need galvanic isolation for safety?
Hi Dick and first of all, thank you!
I'm using a LT1990 as a current source generator (Howland Current Pump) because I think this is a good solution since it have precision resistors inside.
In output of this device I was thinking to place the trasmission gate in order to close the circuit on the output (patient or other) and/or on the known load.

I don't think that we need an isolation in output since the device is supplied by a certified medical power supply and we are working on very low currents and voltages (10uA to 1mA and max 5V).
I see that in other devices is commonly used a couple of diodes (like FDH300 ) with a series resistor in order to protect the input of the devices connected to the patients (e.g. by electrodes).
If you have any observations to make, I would be very happy to read your opinion and discuss the possible and multiple solutions.
Confronting is the secret to growing professionally ;)

For sure I will place the galvanic isolation in the communication part for safety reason.

Thank you again,
Mik
 

Thread Starter

Michele Losito

Joined Nov 23, 2017
13
That TS chip looks a good choice. It will be happy with AC signals as it is stated to be suitable for audio.
Thank you Alec !
Yes I was looking exactly for something like this but I never use this kind of switch !
I will continue the design with this chip and I will give you some news about it ;)

Now I need to choose a very good and low noisy solution in order to measure the voltage across the return line (e.g. electrodes).
In the prototype made previously by a colleague, the signal was really very noisy also because of the very poor layout :rolleyes:
 

Thread Starter

Michele Losito

Joined Nov 23, 2017
13
Easy: Just put your transmission gate in series between pin 6 and RSENSE.
Hi Dick,
why do you want to put the transmission gate in series between pin 6 and Rsense?
For my purposes I only need to redirect the same current on two different loads depending on many conditions.

Please, could you explain me?
Thank you again,
cheers
Mik
 

DickCappels

Joined Aug 21, 2008
10,152
If you want to do it with a single LT1990 -there may be good reasons for that, try this. Still easy but a few more parts.

upload_2017-12-21_17-45-38.png

All switches are ganged. The voltage follower that drives the REF input might not be be needed -I was unable to determine the characteristics of that input.
 
Last edited:

Thread Starter

Michele Losito

Joined Nov 23, 2017
13
If you want to do it with a single LT1990 -there may be good reasons for that, try this. Still easy but a few more parts.

View attachment 142155

All switches are ganged. The voltage follower that drives the REF input might not be be needed -I was unable to determine the characteristics of that input.
Thank you again Dick !
Interesting solution!
I remember a discussion about the voltage follower for the REF input but it'n not clear for me the usefulness of the upper OPAMP.
Please, could you explain? As I see, V+ and V- and Vout of that opamp are in shortcut (out of the transitory conditions).

Anyway it can be a good solution but I think it's more expensive and more noisy for my purpose, unless I'm using low-noise OPAMP.

Thank you again,
Mik
 
Top