LSF0101: Is there minimum current value for LSF0101's I/O?

Benjamin Nordman

Part Number: LSF0101
Other Parts Discussed in Thread: SN74LVC2T45, SN75LVDS387, SN75LVDS386, SN74LVC8T245

Hello TI team,

When I read the datasheet of the LSF0101, it seems that the current flowing to the I/O should be of the mA level (there are example for choosing pull-up resistor for 15mA, 10mA or 3mA and the test conditions are presented with current from 10mA to 64mA).

I need to use this translator for SPI communication. However, in the MCU I'm using, the SPI communication port have pull-up/down resistor of about 13kOhms. That means that for 5V there would be a current of approximately 385µA flowing in the switch of the LSF0101.

Would it still be alright or is there a minimum current value that should be flowing in the LSF0101 to make it work properly?

Thanks for your help.

over 5 years ago

0 Dylan Hubbard over 5 years ago

TI__Mastermind 45855 points

Hey Ben,

There is no minimum current needed, just a maximum to avoid damage to the device. This isn't usually the recommended device for SPI so there may be a better solution for this particular application if you give me more details on the signal you are trying to translate. That isn't to say this device won't work, and if you wish to forward with using it, I recommend checking out these training videos. They are very helpful when using the LSF family.

0 Clemens Ladisch over 5 years ago

Guru 317060 points

The LSF outputs open-drain signals, so the size of the pull-up resistors directly affects the speed of the rising edges. You need a high current only if you want to achieve high frequencies.

SPI, by itself, uses push/pull signals, so the pull-up/down resistors are needed only to avoid floating signals when the devices are idle. In other words, these 13 kΩ resistors do not affect the speed.

Do you need to translate up or down? Do you have a MISO signal?

0 Benjamin Nordman over 5 years ago

Intellectual 410 points

Hello Dylan and Clemens,

Thanks for your respective answers :-)

I actually need to translate several kind of signals to and from the MCU: the standard SPI ones (MISO, MOSI, SCLK and /CS), but also two other signals : one sent and one received by the MCU.

The MCU is 3vdc powered and the data shall be transferred to another IC via a cable. The point to translate the voltage level is too keep the integrity of the signals during their path in the cable (which could be at most 2 meter long).

To resume :

MCU (3V) <==> Translator (3V->5V) <== CABLE (at most 2m) ==> Translator (5V ->3V0) <==> IC (3V)

Furthermore, the MCU shall control up to 8 IC the same way, that is to say 8 x 6 = 48 signal to send this way.

@Dylan : For practical reason, I'd prefer use translators which can manage as much bit as possible.
I also found the part SN74LVCH8T45 and/or SN74LVC2T45 (configured for one way or the other depending of the signal, as there are no two direction signals). What would you think of these?

@Clemens : I may have been misled by the following presentation www.ti.com/.../LSF0101
I think I have answered your question in my presentation above :-)

0 Clemens Ladisch over 5 years ago in reply to Benjamin Nordman

Guru 317060 points

A 2m cable has quite a high capacitance, so you will need even more current to get fast rising edges. Are the MCU and ICs capable of driving that current?

There is not really much difference between 3 V and 5 V, but using translators might make sense if you want to use the stronger output drivers of the LVC family.

If you are really concerned about signal integrity, you should use a much higher voltage (as in RS-232) or differential signals (as in LVDS; check out the SN75LVDS387‌/‌SN75LVDS386‌).

0 Benjamin Nordman over 5 years ago in reply to Clemens Ladisch

Intellectual 410 points

Hello Clemens,

Thanks again for your new post.
Indeed, I didn't taken into account the capacitance of the cable. Do you know a simulator (e.g. in LTspice) that could help me check this matter?

In any case, I'm sorry but even if your idea is a good one, I could not use the SN75LVDS387‌/‌SN75LVDS386.
Indeed, I'm limited in number of wires to send the data and the connector I have to use is already full.

If ever there is no other idea, I think I'll take my chance with the SN74LVCH8T45.

0 Dylan Hubbard over 5 years ago in reply to Benjamin Nordman

TI__Mastermind 45855 points

Hey Ben,

Yes this is a very long cable for a translation device in our portfolio to drive, but your best bet will indeed be to use the SN74LVC8T245 and the SN74LVC2T45 devices. They will provide the strongest drive strength and may be able to drive the capacitance of the cable while maintaining good enough signal integrity at the other end. Can't really say for sure, but you can always use the IBIS model to simulate this. My biggest concern would be slower transition time causing some issues on the other end.

0 Benjamin Nordman over 5 years ago in reply to Dylan Hubbard

Intellectual 410 points

Hello Dylan,

Thank for your reply. I'll try to perform some simulations with the IBIS model.

Just for my understanding, could you please tell me the difference between the SN74LVCH8T45 and the SN74LVC8T45?
The electrical characteristics seems the same to me (same pinout, voltage/current maximum rating, operating temperature limit, consumption almost the same switching time characteristics).