Level conversion principle:
Between 5V and 3.3V level plus 470 ohm current limiting resistor, to avoid being burned of 3.3V port.
So, when the LOW side no-load, HIGH side plus 5V level,LOW side test is still at 5V level,vice versa.
This is a normal phenomenon, not the level conversion function does not work.
So this is what I am seeing, the low side is at 5v but I need 3.3v and I am confused by the above statement and the actual output of the module.
If I put a 3.3v device or IO to the pin and it is delivering 5v then dead 3.3v device!
I put an led onto the L1 side with a 220 ohm resister and I do see about 3v output on the L1 side now.
So it must be working...
Just need to understand this more.
I am wanting to connect an ESP8266 WIFI module and do not want to blow it up as there is confusing things about the IO pins on the Internet. Some say RX/TX directly to 5v and some say that will blow the IO pins on the WIFI module.
Card seems to rely on input protection/clamp diodes of the receiving device to help out the level conversion.
The resistor limits the current flow, the clamp diodes keep the level within 0.5V of receiving device's Vcc and Gnd pins.
For example, 5V signal into 3.3V device:
(5V - 3.3V)/470 = 3.6mA of current flow
Atmel recommends 1mA of clamp diode current flow, so this would not be a good level converter.
The "471" chip is just four resistors connected side by side -- only for mounting convenience. The 471 is the resistor code for the value "47" followed by "1" zero (470 ohms).
Measuring the floating secondary (3.3V ) side of a 470 ohm resistor with a DMM with a 10 Mohm input resistance and no return path for the current on the 3.3V side of the resistor is going to give you the same
voltage as the 5V side. If you had a 3.3V device connected to that side when you measured the voltage then there might be a problem but trying to measure it with nothing connected is not a valid test for a voltage divider that relies on a flow of current (which, in the case of a DMM is probably nanoamps.)
Level conversion principle:
Between 5V and 3.3V level plus 470 ohm current limiting resistor, to avoid being burned of 3.3V port.
So, when the LOW side no-load, HIGH side plus 5V level,LOW side test is still at 5V level,vice versa.
This is a normal phenomenon, not the level conversion function does not work.
I can't quite grasp what it means except for the " the level conversion function does not work" part.
KenF:
The level conversion works perfectly. It's just that YOUR understanding of THEIR level conversion is wrong
I am not sure we want to be brought down to their level in any way! :o
If it included a Schottky diode array to clip the outputs to 3.3V, I could believe it as a one-way converter, but it never was, nor will be, a useful I2C level converter.
Strangely enough, I can see that it will work in a manner, in some instances, where the pull-ups are at least 10k, but we should not give it the "benefit of the doubt".
The funny thing is that the seller provides an honest schematic that shows a circuit that is clearly not a level converter and still he has sold 44 of them. I am not sure who to blame, there are so many choices.