Converting signal level from 1.2V to 5V

[mys_721tx]

Hello all,

Our lab has a Picospritzer for microinjection. We want to use it to trigger our LED controller. The problem is that the Picospritzer's output is only 1.2V and the LED controller requires 3.6V-12V at at least 5mA.

From what I find on DigiKey, MOSFETs with 1.2V Vgs(th) at best have 1mA Id. (Please ignore the 3.5mA Id GaN for the moment) Would it be a good idea to use two MOSFETs so that the drain of the 1.2V Vgs(th) is connected to the 5mA Id MOSFET like the following?

[harerod]

Have you considered using a BJT?
https://en.wikipedia.org/wiki/Bipolar_junction_transistor

[pigrew]

First, realize that you will need to have a voltage supply at your output voltage (i.e. 3.3V or 5V).

If you want a simple, few component solution, harerod's solution of a NPN BJT is good (input through a resistor to the base, with a pull-up resistor on the output), but will invert your signal (unless you make it two-stage). Speed might be a bit slow, and it will have higher power consumption when the BJT is "on" due to the pull-up resistor on the output.

If you wouldn't mind using a few extra components (like voltage regulators), I'd suggest a dual-supply logic converter IC like TXU0104-Q1. ESD, input overvoltage, and output short protections could also be implemented. I'd likely power it from a "5V USB" source.

(BTW, your circuit probably won't work properly since there is nothing to pull the middle gate node down to 0V, so the 2nd MOSFET will never be turned off)

[Vtile]

First, realize that you will need to have a voltage supply at your output voltage (i.e. 3.3V or 5V).

If you want a simple, few component solution, harerod's solution of a NPN BJT is good (input through a resistor to the base, with a pull-up resistor on the output), but will invert your signal (unless you make it two-stage). Speed might be a bit slow, and it will have higher power consumption when the BJT is "on" due to the pull-up resistor on the output.

If you wouldn't mind using a few extra components (like voltage regulators), I'd suggest a dual-supply logic converter IC like TXU0104-Q1. ESD, input overvoltage, and output short protections could also be implemented. I'd likely power it from a "5V USB" source.

(BTW, your circuit probably won't work properly since there is nothing to pull the middle gate node down to 0V, so the 2nd MOSFET will never be turned off)

I think you mix with NPN signal (AC) amplifier now, as last time I did ckeck the typical NPN switch, where load is at collector and emitter is ground will not invert signal, since it "pulls the voltage divider" to ground and therefore increase the voltage seen by load
. However PNP switch will invert, but it is not as easy to bias in this application.

Many small signal BJT NPNs do have hfe 200...500 so with 1mA base ( or gate or input) current you would get 200 .. 500mA switching current, plenty for single 2..3mA led. 2N3904, BC547, BC550, BD139, 2N2222 etc.

[Zero999]

Just use a BJT.

This circuit is designed to shift a 3V signal, to 12V, but it will work with 1.2V & 5V.

[raftronik]

Maybe this simple level conwerter will help.
https://www.hobbytronics.co.uk/mosfet-voltage-level-converter

[Zero999]

Maybe this simple level conwerter will help.
https://www.hobbytronics.co.uk/mosfet-voltage-level-converter

Unfortunately that won't work. The 2N7000 is marginal at 3.3V, and stands little chance of working at 1.2V.

[raftronik]

It was more like general idea to use this type of level converter.

[Zero999]

It was more like general idea to use this type of level converter.

The trouble is, finding MOSFETs which will work with a gate voltage of under 1.2V.  If you, or anyone else, can find one, please post a link to the datasheet.

It will work with a BJT. C1 helps speed it up and is optional and if the input is push-pull, then R1 can be removed, but then the circuit is identical, to the one I posted previously.


Level shift bi-direc BJT.asc

[Wallace Gasiewicz]

This Works, I just tried it.
Use a CD 4011 NOR CMOS chip,
https://www.apogeeweb.net/circuitry/CD4011-cmos-nand-gate.html
Use +5 volts on the Vdd pin 14, pin 7 is GND
Tie the inputs 1 and 2 together and put your 1.2 volt input there.
Use a 51 ohm (47) resistor as a pullup from your Vdd (pin 14) to your input gates (1 and 2)
You will get a signal out of pin 3 that is very slightly lower than Vdd.
This setup works maybe up to 7.5 volts Vdd, then the baseline (bottom of square wave signal out) starts to come up
This signal is inverted (NOR) however so use another set of gates on the 4011 to re invert it. There are three more NOR gates on the chip.
Probably a good idea to use two sets of gates anyway for stability.

Wally

You could use a CMOS OR gate or AND gate but I just picked the 4011

[Terry Bites]

Odd, the manual says it outputs 5V trigger pulses. So it the output driver kaput, overloaded or needing a pullup?
How are you measuring it? Is it steady or a steam of pulses?
What is it connected to when you measure 1.2V. Put a 1k load resistor on the BNC socket and have a look.

If that just how it is, then level translator isn't what you need. Don't be mislead by the fudgers and uC nuts.

Use a cheap voltage detector as a comparator. eg BU4809 $1.Slow but sure.
Or use a ref comparator combo like TLV4041R5 $1.5 . Too easy, one bypass cap and you're done.

[bingo600]

I'm no electronics Guru (DS Guru)

But would one of these do it ?

IRLML2502PBF
https://www.farnell.com/datasheets/1912230.pdf

CJ2302
https://pdf1.alldatasheet.com/datasheet-pdf/view/878506/JIANGSU/CJ2302.html

[Siwastaja]

Indeed, 1.2V logic levels exist on modern PCBs (think about high-end CPUs/SoCs/FPGAs), but very rarely in connectors to the outside world. Unless this device really is such special snowflake, adding a level translator is not a good idea; instead, whatever is wrong needs to get fixed, to get the expected output voltage.

[mys_721tx]

Odd, the manual says it outputs 5V trigger pulses. So it the output driver kaput, overloaded or needing a pullup?
How are you measuring it? Is it steady or a steam of pulses?
What is it connected to when you measure 1.2V. Put a 1k load resistor on the BNC socket and have a look.

If that just how it is, then level translator isn't what you need. Don't be mislead by the fudgers and uC nuts.

Use a cheap voltage detector as a comparator. eg BU4809 $1.Slow but sure.
Or use a ref comparator combo like TLV4041R5 $1.5 . Too easy, one bypass cap and you're done.

Indeed, 1.2V logic levels exist on modern PCBs (think about high-end CPUs/SoCs/FPGAs), but very rarely in connectors to the outside world. Unless this device really is such special snowflake, adding a level translator is not a good idea; instead, whatever is wrong needs to get fixed, to get the expected output voltage.

Indeed, the BNC terminal is badly corroded. (It is green now) Copper carbonate probably does not help with the resistance. I will see if I can replace it. I am not very confident given it also has pneumatic parts inside. If all failed I am going to try to dissolve the corrosion with diluted HCl.

[harerod]

mys_721tx, would you mind giving us a link to the manual / technical spec of your actual device and telling us which interface we are discussing?

[mys_721tx]

Here is the Picospritzer III manual from the manufacturer. I cannot remember we have ether Picospritzer II or III on the top of my head. The output is 5V TTL through a BNC connector per page 5.

[Terry Bites]

Replace the dodgy BNC. Can we see a photo of it from inside?

[mys_721tx]

Replace the dodgy BNC. Can we see a photo of it from inside?

It turned out that we have Picospritzer II and II is very different from III.

Here is the front panel.



Here is the wiring on the back. Fortunately, all BNCs are panel mounted and should be easily replaceable.

[mys_721tx]

I have been probing the out terminal, the center pin on the other side, and the pin 2 of the potentiometer. It turned out that Picospritzer II does output ~800mV!

Then I noticed that AA battery and realized that an opto-isolator on the main board acts as the switch for the battery the output circuit. The oscilloscope screenshot shows the output after replacing the battery with a new one.

Also it seems that the corrosion does not affect the voltage too much.

[mys_721tx]

Here is the main board in case anyone interested.

[mys_721tx]

And the back side. Apology in advance for splitting the gallery up but they easily exceed the maximum total size restriction.

[mys_721tx]

I swapped the AA battery to a 9V. The following is my the oscilloscope trace. Channel 1 is on the BNC terminal. Channel 2 is the orange/white wire on the DPDT switch, which is on the LED side of the optocoupler. Channel 3 is the input of the optocoupler.

I modeled the circuit in circuitjs. I used two 1M resistors to represent the input impedance of the oscilloscope. Everything in the simulation works as expected. The potentiometer adjusts the voltage of one terminal and the DPDT switch changes the polarity of the input.

However, the switch does absolute nothing when I am probing it. Since I check with a meter that the the potentiometer resistance is in its nominal range, I suspect something is wrong with the switch.

[mys_721tx]

When I was probing with the 9V battery, the Picospritzer magically sent out the correct voltage. However, when I removed the scope, it did not work again. This is when it hit me. The outer conductor of the BNC terminal is grounded and the battery is floating so to speak. The ground pin on the oscilloscope made it work!

Soon I removed the AA battery holder, fashioned a ground wire from the 9V connector, and spiced the wires. Now the Picospritzer can reliably trigger the Mightex controller!