GPIO.input voltage levels vs edge detection

[Brandon92]

Ehm, do you know what for kind of brand that is. Because the mark doesn't look like the two we discussed in here:
A3700, take a look at page 3. And the exact location of the print.
HCPL, take a look at page 8.

So, I wonder if they are the real deal. Or the are fake ones.

edit
And the datasheet that you profited, look also not the same as your (deleted) picture.

Edit2
To give you a idea how much those devices are by a well know seller (at my location). They are €3 (euro) without tax.

[tlfong01]

Ehm, do you know what for kind of brand that is. Because the mark doesn't look like the two we discussed in here:
A3700, take a look at page 3. And the exact location of the print.
HCPL, take a look at page 8.
So, I wonder if they are the real deal. Or the are fake ones.
edit
And the datasheet that you profited, look also not the same as your (deleted) picture.

Well, I am a cheap hobbyist, making cheap things using fake things.

The real thing
http://hk.element14.com/search?st=hcpl3700

It's the real thing
https://www.youtube.com/watch?v=dDSnjjdGh5M

[Brandon92]

It's the real thing
https://www.youtube.com/watch?v=dDSnjjdGh5M

Okay, good to hear that. How did you find out that they are real?

[tlfong01]

It's the real thing
https://www.youtube.com/watch?v=dDSnjjdGh5M

How did you find out that they are real?

I don't know. I can make 101 guesses. Below is one.

1. Any die manufacturer can easily make legally 3700 type dies, which may not have the same circuitry, but same functionality of the original HCPL3700. No one copycats the original circuit, which is so simple, so many late comers actually improves the function, or even add more functions, but still backward compatible.

2. One such 3700 manufacturer legally sells the wafers to assembly and testing factories in ShenZhen, under some sort of legal commercial contract.

3. The assembly, testing, and quality control standard of the ShenZhen factory is of course very poor, much below military, industrial, or even consumer standard, but nothing illegal.

And for a very poor hobbyist like me, paying 2 yuan for one A3700, what do I expect?

4. The ShenZhen factory cannot mark their chip Fairchild HPCL3700 for copyright reasons. But I think it is not illegal to mark A3700, B3700 etc.

5. So it is like AMD and Intel, or Pepsi Cola and Coca Cola, ...

The above is just an educated guess. I did once read an Ali-Express seller of A3700 says that their dies are imported, ...

[Brandon92]

With a chip that is not made by the right company there are, I believe, two options.

1. Its possible that it is a complete other chip than you would expect.
2. It will "work" like the real one. But it does not have the same specification and maybe the biggest point is the safety point of view. Will it have the same safety barrier if you will use it at high voltage.

And that's why you need a bit careful if you connect the "fake" chip to the mains voltage. And in that case it's maybe better to buy a real one that is cheaper

[tlfong01]

With a chip that is not made by the right company there are, I believe, two options.

1. Its possible that it is a complete other chip than you would expect.
2. It will "work" like the real one. But it does not have the same specification and maybe the biggest point is the safety point of view. Will it have the same safety barrier if you will use it at high voltage.

And that's why you need a bit careful if you connect the "fake" chip to the mains voltage. And in that case it's maybe better to buy a real one that is cheaper

I agree. It is possible that the fake chips, though legal, does not meet all the specifications.

So the Ali-Express guy may say to me, "Ah this thing is A3700, no spec, no tech support, no nothing. Two yuan each, take your own risk！"

I can now give you one more reason that I don't connect any fake things to my real Rpi GPIO. I only connect my fake A3700 to my fake MCP23017 - fake to fake, what a perfect match!

I appreciate your warning about high voltage risk. So I think I better allow my fake A3700 only touch AC below 24V. When I later test mains 220V, I might consider getting a real Fairchild HCPL3700 from Farnell for 30 yuan.

[Brandon92]

I appreciate your warning about high voltage risk. So I think I better allow my fake A3700 only touch AC below 24V. When I later test mains 220V, I might consider getting a real Fairchild HCPL3700 from Farnell for 30 yuan.

That is a wise decision. I often use "one" rule, if the part is connected to the mains. I only want the official product. And from a well known company. Then you know for sure it is designed right and with the good test report. And in this case I will buy the part with the "v" in the part number. The voltage here is 230-240 Vac.

By the way, I like to do sometime to over design my hobby projects :p

[tlfong01]

The following works with 5vdc power. Press the button and the led goes on. Release and it goes off.
I used a 1k breadboard potentiometer ... to roughly find the threshold where the led would come on. It happened at about 310 ohms and continued to work at lower resistances. I chose 220 because it was one I had and it was not far from the 280 suggested. I measured r1 amperage at 2.1 ma making the power dissipation 10.5 mW well under the 230 in the Abs Max Ratings and I-in well under the 50 average max. I'm happy to report that I did NOT fry my chip, at least pins 2,3,5,6 and 8. Next, I will try this same type of thing with my 24vac on pins 1 and 5. We're getting closer!

tlfong01 wrote: ↑ 2018-Aug-09 Thu 4:59 pm
Yes, HCPL3700 is good. I will very likely use it later (My home automation project is about 2 months lagging behind yours.)

I have almost zero experience in playing with power electronics, and of course don't know anything about HCPL3700.
However, I just order some. I will follow the expert's steps to get some experience.

Following the expert's steps

update 2018aug26hkt1156

• 1. I think I will take 2 big steps: (1) Testing DC signal, (2) Testing AC signal.

• 2. I will try to confirm the hysteresis thing as described in Note 5 in the picture below.

• 3. petermeigs (a) uses a fixed DC input voltage of 5V, and (b) by trial and error, found the best resistance Rx.

• 4. I will do the reverse, (a) use a fixed Rx = 480R ==510R, and (b) by varying input voltage Vin, increasing from 0V to 7V, then decreasing from 7V to 0V, to find the hysteresis voltages Vth+ and Vth-.

Sister act - I will follow him - 10,818,008 views
https://www.youtube.com/watch?v=ghHE_kVWXxM

[tlfong01]

• 2. I will try to confirm the hysteresis thing as described in Note 5 in the picture.

• 3. petermeigs (a) uses a fixed DC input voltage of 5V, and (b) by trial and error, found the best resistance Rx.

• 4. I will do the reverse, (a) use a fixed Rx = 480R ==510R, and (b) by varying input voltage Vin, increasing from 0V to 7V, then decreasing from 7V to 0V, to find the hysteresis voltages Vth+ and Vth-.

A3700 Hysteresis Analysis

/ to continue, ...

[Brandon92]

3. petermeigs (a) uses a fixed DC input voltage of 5V, and (b) by trial and error, found the best resistance Rx.[/list]

• 4. I will do the reverse, (a) use a fixed Rx = 480R ==510R, and (b) by varying input voltage Vin, increasing from 0V to 7V, then decreasing from 7V to 0V, to find the hysteresis voltages Vth+ and Vth-.

To make your measurement more precise. I won't use a led as part of the pull up circuit. And you don't provide a extra load to that pin. But the best way is to measure the voltage or use your oscilloscope

[tlfong01]

3. petermeigs (a) uses a fixed DC input voltage of 5V, and (b) by trial and error, found the best resistance Rx.[/list]

• 4. I will do the reverse, (a) use a fixed Rx = 480R ==510R, and (b) by varying input voltage Vin, increasing from 0V to 7V, then decreasing from 7V to 0V, to find the hysteresis voltages Vth+ and Vth-.

To make your measurement more precise. I won't use a led as part of the pull up circuit. And you don't provide a extra load to that pin. But the best way is to measure the voltage or use your oscilloscope

A3700 Hysteresis Analysis Results

I very much agree with you. Using LED is just testing the water. My clever strategy is to let the brave guy try his luck first. If he does not fry his chip, then I will follow him, using LED and voltage at the same time. The LED is necessary, because Vo is open drain, and Iout max is 30mA, so a little LED (I am using green LED with R = 4k7, so only 1mA, the strong guy won't notice anything! )

I just finished the I will follow him test, with results attached below.

• 1. I tested 4 samples, using Rx = 180R and 220R

• 2. at Vin == 5V I emitter 180R == 4.2mA, 220R == 3.9mA

• 3. The hysteresis loop seems shifting up when Rx increases from 180R to 220R. Does the scary equation agrees with my observation?

• 4. My objective hysteresis loop is centred at 5V, because my goal (well, boasting again! ) is to set up a dual Rpi system, each monitoring the other's 5V power. If one goes down, the other will take over immediately (well, within milliseconds, I hope).

• 5. Yes, next step is using a scope. I am thinking of using a fake 555 pulse gen to feed a 5Vpp square wave to the fake A3700 Vin and see if the fake guy explodes or not.

Time for Sunday evening supper. See you later.

[tlfong01]

My first test is summarized below.

• 1. I tested 4 samples, using Rx = 180R and 220R

• 2. at Vin == 5V I emitter 180R == 4.2mA, 220R == 3.9mA

• 3. The hysteresis loop seems shifting up when Rx increases from 180R to 220R.

• 4. My objective hysteresis loop is centred at 5V, because my goal

• 5. next step is using a 555 pulse gen to feed a 5Vpp square wave to A3700 Vin and see how it goes.

Shifting A3700 Hysteresis Up

I first found Rx = 1K2 for detecting 5V power low (lower than 4V).

/ to continue, ...

[petermeigs]

My clever strategy is to let the brave guy try his luck first. If he does not fry his chip, then I will follow him,

You guys crack me up! Just to let you know I was very careful when trying different resistances. For example, I verified potentiometer was at max resistance before adding power. Also, I had a fixed resistance in series just to make sure the potentiometer that the risk of damage was minimized. In general, I find myself adding extra resistors here and there just to make sure the current can't go crazy. I learned to do that the hard way by melting the insulation off a wire on my ADC chip. Fortunately that is a very hardy chip and could handle the current with no damage.

[tlfong01]

• 5. next step is using a 555 pulse gen to feed a 5Vpp square wave to A3700 Vin and see how it goes.

I first found Rx = 1K2 for detecting 5V power low (lower than 4V, 20%).

A3700 Vth-, Vth+ Hysteresis Threshold Values for 5V, 12VDC, 12VDC/AC, 24VDC/AC

My little adjustable psu can go to about 10V. So I need to find another 30V switching power supply, and perhaps another step up module to get over 50VDC for testing.

[Brandon92]

My clever strategy is to let the brave guy try his luck first. If he does not fry his chip, then I will follow him,

You guys crack me up! Just to let you know I was very careful when trying different resistances. For example, I verified potentiometer was at max resistance before adding power. Also, I had a fixed resistance in series just to make sure the potentiometer that the risk of damage was minimized. In general, I find myself adding extra resistors here and there just to make sure the current can't go crazy. I learned to do that the hard way by melting the insulation off a wire on my ADC chip. Fortunately that is a very hardy chip and could handle the current with no damage.

You should think about buying a (prober) power supply. Witch such power supply you could limit the current to a couple mA. And they will not burn your wires down if you make a mistake. And the basic ones are not that expensive I believe.

[tlfong01]

A3700 Vth-, Vth+ Hysteresis Threshold Values for 5V, 12VDC, 12VDC/AC, 24VDC/AC
[/quote]

I ran out of input voltage because my highest bench switching ps is 30VDC. So I need to find a voltage step up module to get voltage up to 50VDC (with input 12
VDC)

Now I have 3 Rx values to detect 5VDC, 12V DC/AC, 24VDC/AC power loss.

Ah, I forgot to mention that I saw in the datasheet and app notes that the Vth-, Vth+ characteristics graphs for DC and AC voltages are almost overlapping each other. So I am too lazy to repeat my experiments for AC. I will later do AC 12V/24V relay switching and displaying scope waveforms anyway.

[Brandon92]

A3700 Vth-, Vth+ Hysteresis Threshold Values for 5V, 12VDC, 12VDC/AC, 24VDC/AC

I ran out of input voltage because my highest bench switching ps is 30VDC. So I need to find a voltage step up module to get voltage up to 50VDC (with input 12
VDC)

Now I have 3 Rx values to detect 5VDC, 12V DC/AC, 24VDC/AC power loss.

Ah, I forgot to mention that I saw in the datasheet and app notes that the Vth-, Vth+ characteristics graphs for DC and AC voltages are almost overlapping each other. So I am too lazy to repeat my experiments for AC. I will later do AC 12V/24V relay switching and displaying scope waveforms anyway.
[/quote]
I didn't read it al. Becuase of time.
But if you have a couple of power supply, you can connect them in series to get a higher voltage.

[tlfong01]

1. Now I have 3 Rx values to detect 5VDC, 12V DC/AC, 24VDC/AC power loss.
2. I will later do AC 12V/24V relay switching and displaying scope waveforms anyway.

A3700 Testing Hysteresis Under DC 5V

I first used Rx = 1k5, thinking that 555 200Hz 4Vpp input to A3700 should get same frequency output. But nothing happened.

I guessed 4V was not high enough. So I lowered Rx = 1k5 to 220R, but still nothing. I lowered to 110R, then got something to see.

[tlfong01]

1. Now I have 3 Rx values to detect 5VDC, 12V DC/AC, 24VDC/AC power loss.
2. I will later do AC 12V/24V relay switching and displaying scope waveforms anyway.

555 200Hz 4Vpp input to A3700, 110R, then got something to see.

HCPL3700 Detecting 10Hz, 12Vpp Square WavevSignal

I have done the following:

1. Set 555 timer to generate 10Hz, 3.8Vpp square signal.

2. Connect 555 signal to IN of high level trigger relay module.

3. Connect 12VDC power to COM terminal of high level trigger relay module.

4. Connect NO terminal power to HCPL3700 input DC- DC+.

5. Scope to display Vo of HCPL3700.

[tlfong01]

HCPL3700 Detecting 10Hz, 12Vpp Square WavevSignal
I have done the following:
1. Set 555 timer to generate 10Hz, 3.8Vpp square signal.
2. Connect 555 signal to IN of high level trigger relay module.
3. Connect 12VDC power to COM terminal of high level trigger relay module.
4. Connect NO terminal power to HCPL3700 input DC- DC+.
5. Scope to display Vo of HCPL3700.

HCPL3700 Detecting 10Hz on/off, 12VAC/24VC Signal

HCPL3700 seems OK detecting 10Hz 12VDC signal. So I think detecting 12V/24V AC signal should be as easy.

I did the following.

1. By hand use 5V power wire to touch relay module inputs 2, 3, 4 and hear clicking sound OK. Inputs 2, 3, 4 switch relays connected to 12VAC, 24VAC, and 12VDC. I am not testing Input 1 which control relay connected to 220VAC.

2. I used a multimeter to measure AC12V, AC24V, and DC12V when switched on at NO2, NO3, NO4. No problem at all.

3. I used scope to display switched on DC12V at NO4. No problem.

4. 555 10Hz switched AC12V power at NO2 is display below. Looks OK.

/ to continue, ...

[tlfong01]

HCPL3700 Detecting 10Hz on/off, 12VAC/24VC Signal
HCPL3700 seems OK detecting 10Hz 12VDC signal. So I think detecting 12V/24V AC signal should be as easy.
I did the following.
4. 555 10Hz switched AC12V power at NO2 is display below. Looks OK.

Mistake!

I connected the 10Hz switched AC12V to DC-, DC+，

Ah, I made a mistake here. No wonder I got the following Vo signal (blue trace below).

[tlfong01]

Mistake!
I connected the 10Hz switched AC12V to DC-, DC+，
Ah, I made a mistake here. No wonder I got the following Vo signal (blue trace below).

Mistake corrected (signal to AC1, AC2), but Vo still not a square wave, as DC12V case.

I used the same Rx 220R for both DC12V and this time AC12V. I am not sure if I need to use another Rx, ...

[Brandon92]

How did you measure the sine wave and the output of the device. I hope, in this case, that you didn't connect the ground pin of the scoop to the "ground" of the ac. And also connect the ground of the device to the ground of pin scoop. By the looks of the picture you did this. And you are probably one A3700 down

[petermeigs]

You should think about buying a (prober) power supply. Witch such power supply you could limit the current to a couple mA. And they will not burn your wires down if you make a mistake. And the basic ones are not that expensive I believe.

A great idea! Do you have a make/model to suggest?
Along the same line of interesting items tlfong01 is showing: What are these?:

meters.PNG (221.98 KiB) Viewed 3950 times

They look like ammeter / voltmeter. Did you build these yourself or can I just buy one? Where can I get one? They look useful for instrumenting a circuit.

[Brandon92]

Hello,
I don't know how much you have to spant. But EEVBlog, review have made some movies about lap bench power supply. And might give you a idea what you want and can effort. I don't know where you could buy one in you country, but maybe this link will give you somehow a idea how they look (a store in my country).

The importend thing is that it have constant voltage (CV) and constant current (CC). And it should be capable of a short circuit.