DS89C21: Check the circuit diagram

Hi Ishiwata,

This device is from the 1990s and was created by national semi (TI acquired them in the 2010's) so we don't have the internals for the device (they are probably sitting in a cabinet on paper somewhere in California if they still exist). We also don't provide any internal drawings for our devices anyways as this is considered maximum restricted information. 

What specifically are you looking for from a circuit diagram? (Why do you need it?)

-Bobby

Hi Bobby,

I'm sorry for the lack of explanation.

I would like to ask you to check the customer's circuit diagram.

I sent you the circuit diagram in a private message.

I checked it and I think it needs a termination resistor, is that correct?

Best Regards,
Ishiwata

Hi Ishiwata,

You're correct. The customer is not using a terminating resistor. They need one if this is the end point of the cable. Usually customers don't use 100 ohms in series with the RS422 inputs (R15 and R16). I would normally expect 10 ohms if they wanted to have pulse proof resistors. 

Di should also not be floating, I recommend you tie to this Vcc or GND with a resistor. 

Please verify if this device is at an end point of the cable or if this is one of multiple boards that may not be an end point. 

-Bobby

Hi Bobby,

Thank you for your reply.
And thank you for checking the circuit diagram.

I understand your answer well enough, but are there any circuit diagrams or EVMs that can be used as reference?

If there are any, I would like to contact the customer with those as well.

Best Regards,
Ishiwata

Hi Bobby,

Thank you for your reply.

I have confirmed how the customer is using it. The customer will use it on a one-to-one basis.

I also have an additional question. Is it okay to leave the DO and DO* pins open?

Best Regards,
Ishiwata

Hi Bobby,

Thank you for your reply.

I have an additional question.

What is the maximum transmission rate that can be input?

The data sheet states that it operates at 20Mbps and 10MHz.

On the other hand, can it also input slower signals such as 1MHz?

Best Regards,
Ishiwata

So 10MHz is equal to 20Mbps so they are basically the same thing.

I did some rough math on the differential rise/fall time max values. If we assume 9ns for rising, falling and also for the bit hold time and another rise or fall time (basically my math was =1/36ns) then the device in theory could support 27MHz (54Mbps) but this is kind of at an extreme case with very low loading conditions (50pF). I believe the 10MHz (20Mbps) to be a value with more margin. 

Yes, you can use a much slower input like 1 MHz. The main disadvantage for using a 10MHz driver with a 1MHz input is you will have faster edges which has more EMI. 

-Bobby

Hi Bobby,

Thank you for your reply.

Are there any restrictions on the rise and fall times of the signal input to RI/RI*?

Please let me know if there is a maximum time.

Also, I understand that there is no problem in inputting a signal with a 5V amplitude to RI/RI*.

Is my understanding correct?

Best Regards,
Ishiwata

Shuji Ishiwata said:

Also, I understand that there is no problem in inputting a signal with a 5V amplitude to RI/RI*.

Yes, that should be fine.

Shuji Ishiwata said:

Are there any restrictions on the rise and fall times of the signal input to RI/RI*?

Please let me know if there is a maximum time.

We don't really have anything like this defined. Ideally the rise/fall time on both lines should be the same rate to get a smoother differential transition. We don't have a max time for this but going very slow could affect prop delay values a little.

-Bobby

Hi Bobby,

I always appreciate your support.

I have an additional question.

A customer reported that the IC's top temperature is 77°C. (Ta=25°C)
Does the DS89C21 really heat up that much? What are the possible causes of the heat?

Also, please tell me the θj-a of the DS89C21 package so that I can calculate Tj.

Best Regards,
Ishiwata

Hi Ishiwata,

Are you able to measure the current going into the device through the Vcc pin? May be helpful to verify its not drawing 100s of mA. Since the device is one driver, I would expect maybe 50mA current draw. If you have a lot of fast switching/datarate it could be a bit more. 

Maybe also confirm that they have good GND/PCB layout for heat dissipation. I do think 77C is a bit high but the device's abs max is 150C so it shouldn't damage the device.

Shuji Ishiwata said:

Also, please tell me the θj-a of the DS89C21 package so that I can calculate Tj.

I don't think we have this data available right now. I would need to submit a ticket for this which generally takes 2 weeks for us to get the data. Are you okay with that wait time? If so, I can submit the ticket. 

-Bobby

Hi Bobby,

Thank you for your reply.

I am asking the customer about the VCC current value. I will contact you when I know.

Please make a request about θj-a. I don't know if the customer can wait two weeks, but...

The customer's original question is to confirm that Tj=150℃ will not be exceeded when Ta=50℃.

Can it be used at Ta=50℃ when there is a Δ52℃ rise? Please let me know your thoughts.

Best Regards,
Ishiwata

Hi Bobby,

I confirmed the VCC current with the customer. The current is 18mA.
I don't think the heat generation is caused by the device's own consumption.

If I get a layout from the customer, can I check it?

Best Regards,
Ishiwata

Hi Ishiwata,

It depends on what the customer's PCB source files are. At TI, we mainly use Altium for PCB building and reading. If they have a file extension I can try to review it. We also have PCBviewer though that interface is harder for me to use. Other possibility is you could try to take pictures of the PCB layers around the DS89 device.

-Bobby

Hi Bobby,

Thank you for your support.

A customer asked me to check if there were any problems with the RI input signal.
The rise and fall times input to the RI are about 120 nsec. Is this ok?

Also, regarding heat generation, it appears that the DO and DO* terminals are grounded to GND.
Will this cause heat generation?

Best Regards,
Ishiwata