Are the inputs to the TNT4882 3.3 volt tolerant?

[SteveW]

I plan to interface the TNT4882 directly to a uC that uses 3.3 volt I/O.

[DiegoF]

Hi Steve,

The TNT4882 doesn't. However, the TNT5002 does have 3.3 v signaling. You can check the user manual at the GPIB register-level programming site.

Hope this helps.

DiegoF
National Instruments.

[Collin M]

Steve,

I think your concern is whether your processor is 5V tolerant not whether the TNT4882 is 3.3V tolerant, right?

The CPU side IO on the TNT4882 is all 5V and Vin-high for the TNT4882 is 2.0V. This means that if your uC drives to 3.3V the TNT4882 should detect a high. However, this may result in higher sink current on the TNT IO pins.

The TNT4882 will drive 5V signals to your uC so if your uC isn't 5V tolerant damage could occur.

The suggestion to use the TNT5002 is very good because in Generic mode it appears almost identical to a TNT4882 and it has a 3.3V core and 3.3V IO.

Collin

[SteveW]

Collin,

Thanks for the super response. You interpretted my question correctly. My uC, a Rabbit 3000 (enhanced Z180), is 5V tollerant.

I hadn't previously considered the TNT5002 for a couple of reasons. First, there is the "16-bit" comment on the first page of the data sheet (I'm using an 8-bit device). Secondly, and probably more importantly, there is a fair amount of sample code for using the TNT4882 in an embedded mode (e.g., AN077, Using the TNT4882 with an 8051). How applicable is this information to the TNT5002?

Steve

[SteveW]

Diego,

After I reviewed the data sheet for the TNT4882 that you kindly pointed me to, I concluded that the TNT4882 could accept 3.3V inputs (Vinh > 2.0). My uC can handle 5V inputs, so it would appear that all is well. Am I missing something?

As far as the TNT5002 is concerned, I hadn't seriously considered it due to a lack of application notes, etc. Are the application notes (i.e., AN077) also applicable to the TNT5002?

Steve

[Collin M]

Steve,

In 16-bit generic mode, it is possible to only implement an 8-bit bus to the host. The only drawback of this is that DMA is not supported in 8-bit mode. If you don't need DMA you can simply place pullups on the IOD[8-15] and connect the uC to IOD[0-7].

For One-Chip mode the GPIB register set of the TNT4882 and TNT5002 are identical so any code written for one should be able to run unmodified on the other.

App Note AN077 specifically applies to the TNT4882. There are some differences for the TNT5002. For example, ABUSN and BBUSN are not implemented in the TNT5002 because all 8-bit accesses are on IOD[0-7]. I believe the assembly code in AN077 should also run on the TNT5002 but I don't think anyone has tried it before.

From your
description of your application you should be able to use either the TNT4882 or TNT5002, unless you want to implement DMA. In that case you must use the TNT4882.

Collin

[SteveW]

Collin,

I plan to communicate directly with the FIFO, so I don't need DMA.

In the one-chip, 8-bit mode, why are pull-ups needed on the upper 8 bits?

Regarding AN077, I will code in C, but the assembly listing seems to provide a simple outline of the basic tasks. From what you said, I can still do this with the TNT5002, but must look out for differences like you mentioned.

Your assistance is greatly appreciated.

Steve

[Collin M]

Steve,

The TNT5002 does not have pullups or pulldowns on any of the 16 data lines. If you use the TNT5002 in 8-bit mode you should not allow the upper 8 bits to float.

Collin

[SteveW]

Collin,

Thanks again for all of your help.

Steve