Equivalent Level Shifters ICs

[Korawy]

Hi ALL

Im working on a project that uses communications between atmega32 and an atheros based module, So there must be Level shifting between the two sides (from 5 to 3.3 or 2.5), The schematics I have is using NTB102, NTB104 and MAX3375EEK as Level Shifters.

Im asking about equivalent ICs as Im still in testing phase and not intending to connect such small ICs.

Thanks..

 

[d123]

Hi,

I don't know if any of these pdf's can help, I don't know specific ICs...

 

[Korawy]

Thank you.. I will check them

 

[KlausST]

Hi,

Level shifters transform signal voltages from one device to the other, not power supply voltages...

You just give the power supply voltages, but more interesting are V_OH, V_OL, V_IH, V_IL of the devices.
And V_input_max.

Some devices simply don't need a level shifter, even if they are powered from different supply voltages.
Fir other solutions you may use a simple current limiting resistor or a two resistor voltage divider..

In your description it is not clear
* how many data lines you have,
* wheter they are unidirectional or bidirectional
* and in case of bidirectional: what device controls the direction
* and speed (data rate) is of interest.

Klaus

 

[Korawy]

Hi KlausST

Yes I know that Level shifters are for transforming signals and you have to supply it with the two power supplies from the two devices, sorry if I type that in a wrong manner.

The Scheamtics I have (Arduino Yun)are using those ICs to do the Level Shifting as follows:

-Max3375EEKA+T for UART Connection ( 2 pins TX-RX)
-NTB0104 for ICSP Programming (4 pins SS-MOSI-MISO-SCK)
-NTB0102 For 2 GPIO PINS

There are not any pins for the direction which seems it has a self-controlled mechanism for determining the direction.
The Data rate for Max3375EEK+T is 16Mbps, and for NTB is 80Mbps

Thanks

 

[KlausST]

Hi,

There are not any pins for the direction which seems it has a self-controlled mechanism for determining the direction.
The Data rate for Max3375EEK+T is 16Mbps, and for NTB is 80Mbps

I doubt the atmega32 can handle such high datarates.

With ISP I don´t use level shifters at ll. My programmer can handle 3.3V devices (I´m not sure if 5V devices)
For one direction logic signals I either use an analog switch or resistive dividers. For higher speed and/or higher reliability I use AHC devices.
For a parallel bus I used LVXC devices (74LVXC4245 for example)

Hope this helps

Klaus

 

[Korawy]

Those Data rates are maximum rates the ICs can handle and thus it can operate the both sides whatever their data rates.

Again I forgot to tell that the programming not through the Level shifter.. In fact the ICSP Pins are directly connected to the atmega and also to the level shifter but not to program the Linux Side but it's used to load sketches from the Linux Side (atheros 3.3v) to the atmega.

So For a parallel bus those LVXC ICs use Direction Pins ??If yes Are there ICs that have self-controlled mechanisms to determine the direction like the ICs I post ??!!

Thanks

Thanks

 

[KlausST]

Hi,

I don't know if every LVXC IC uses direction pins, but the given example does..

But does your application need alternate directions?
All Uart signals and all SPI have unique directions...

Klaus

 

[Korawy]

Yes it's a special condition where data can be in both ways, but I think I should do some digging to find out if there are ICs that uses no direction pins !!

Thanks

 

[ads-ee]

TI makes some auto sensing level shifters (I've never used them) https://www.ti.com/lit/ds/symlink/txb0108.pdf

 

[Korawy]

Hi ads-ee

That's an excellent equivalent .. the problem is it has a very small pitch pin (around 0.7mm for txb0104) but I have no experience in soldering such small ICs in addition that I will use a basic soldering unit (not hot air).

Thanks

 

[ads-ee]

Take a look at the following selection guide and see if you can find a package more suitable for you to hand solder.

https://www.ti.com/lsds/ti/interface/bidirectional-voltage-translation-products.page#

 

[Korawy]

Thank you ads-ee

After some digging I found the link you post, In Fact I had to know the packages and its pitch pins so as to decide if this IC will be ok for me as soon as I see the first page in the datasheet.

However In youtube I found it's very easy to solder the TSSOP Package (0.65 pitch) or greater packages.

Also Im little confused in something.. In the datasheet of the Unidirection Level Translator it says that the data direction must be in one direction either from low to high or from high to low, So why there is a bidirection level translator
considering that you cant pass a couple of data in a single line at the same time ??!!

I think the reason is that a bidirection level translator can reverse the direction of the data anytime during operation, while a unidirection level translator can only accept one direction (ANY DIRECTION) during the IC operation, That's, the Unidirection level translator waits for the signal that comes from either direction and fix that direction till the power is OFF.
Am I right in this??!!

Thanks

 

[Korawy]

Hi again

I have another question...the level shifter max3375eeka+t is a very small package (0.5 pitch), I know it can be replaced by txb0104, but instead Can I use a voltage divider circuit?, The max3375eeka+t is a unidirection level shifter that is used for UART Connections between atmega (5V) and atheros (3.3V).

So if i know the direction .. is it possible to use a voltage divider circuit instead ??

Regards

 

[d123]

Hi,
I assume so, if it will always be a 5V output signal that needs to be reduced to a 3.3V input signal. Maybe 3K3 and 1K7 divider?

 

[Korawy]

Hi d123

I saw in a different topic that a voltage divider circuit may act like an RC Filter that will affect the rise and fall time of the signal, Im searching now for level shifter that have dir pins .

Thanks

 

[d123]

I saw in a different topic that a voltage divider circuit may act like an RC Filter that will affect the rise and fall time of the signal...

Hi,
I didn't know that, that's interesting to learn, thanks.

 

[KlausST]

Hi,

Hi,
I didn't know that, that's interesting to learn, thanks.

Every trace acts like a capacitor and every IC input, too.
Therefore every (source) impedance has influence on the rise time.

If you use the recommended 3k3/1k7 then the equivalent source resistance is about 1k1.
Maybe the IO capacitanc is 10pF, let´s add 5pF of trace capacitance, to get 15pF in total.

This creates a tau (delay) of about 16ns. The risetime form 0.7V (TTL_low) to 2.0V (TTL_high) is about 11ns.
I think this is not critical....but you have to decide...

Klaus

 

[Korawy]

d123.. You are welcome

Klaus

For Voltage divider, What's the max rate of data it can handle ?!

Also the Max has three-state pin which protect against thermal short-cut, I dont know why he uses that pin, It's just UART Connection!!

Thank you

 

[KlausST]

Hi,

For Voltage divider, What's the max rate of data it can handle ?!

I´ve calculated the timings for you. But I don´t know where you set the limits for delay and/or risetime.
With the calcualted values I see no problem with 1MHz and more..

The values you gave (16Mbps and 80Mbps) will be difficult. On th other hand you don´t need to use 1k7 and 3k3. if you use 170 and 330 you are ten times faster.

Consider: some developers use RC filters to avoid EMC problems and to reduce EMI ....

Klaus