BQ4802Y: BQ4802 Does not cycle time

Would a crystal being at 12.5pF instead of 6pF cause this issue?

Hello Chris,

This could be the issue, could you probe the oscillator on a working board and the non functioning board to see if there is a difference on the oscillators?

Sincerely,

Wyatt Keller

no difference on the oscillators on working and non-working boards.

Also both the working and non-working boards are on the same production run boards.

I have ensured the 32.768kHz crystal are the required 6pF.  Both are SMD's.  I have even swapped the SMD crystal out for the standard through hole, but it yields the same results.

I'm at a complete loss of what is wrong, especially since the circuits are identical on the same production PCB.

I did notice on the boards that are working the CEOUT, Pin # 25 stays at .05v or less, but on the boards that are not working the voltage is anywhere from .5v to 1v. That strikes me as odd behavior. According to the datasheet:

 CEOUT goes low only when CEIN is low and VCC is above the power fail threshold. If CEIN is low, and power fail occurs, CEOUT stays low for 100 µs or until CEIN goes high, whichever occurs first.

I have verified there are no solder bridges, etc.

Hi Chris,

I'm not an expert on this device. I recommend BQ32000 or BQ32002, these are our most popular RTCs.

I'll still provide some debug support to help you resolve your issue. Due to the age of this device, design support is limited.

What other debug steps have you taken so far? It seems as though you already tried investigating this through the TI quality channel, correct?

Kind regards,
Lane

the reason I chose the BQ4802 line is because the pinout I need is identical to the obsolete BQ4847 with the exception that the crystal and battery backup is internal to the BQ4847 where as the BQ4802 is all external.

this clock chip goes into a vintage computer (believe it or not) a Texas Instruments TI-99/4A.  It is required to be on a IDE Hard disk controller for the TI-99/4A.

It must be able to control an external SRAM, which the BQ4802 does.

chris

Hi Chris,

Can you use the state of the RST or WDO pins to detect whether the IC detects it is in a power failure condition? Per the datasheet:

RST goes low whenever VCC falls below the power fail threshold. RST remains low for 200 ms (typical) after VCC crosses the threshold on power-up.

I assume you've checked the state of the CEIN pin as well and found it to be correct (Low) for operation, including while it should be counting?

Also, have you tried swapping the units on two boards, where one board will not count, and one board will?

Regards,

I have already swapped the units from one board to another.  When swapping from the working board to the non-working board, that board starts working.  and vice versa.

I will double check the CEIN pin tomorrow.  I believe I did, but that was days ago.

Any chance this is some kind of ESD failure shorting CEIN pin high internally? This is a 20-year-old device with no ESD ratings listed in the electrical characteristics, and a lot of us are operating outside of our typical lab environments where ESD mitigation devices and tactics may not be as well-enforced.

Where did you order the devices? TI.com? A distributor?

I ordered all devices from TI.com

The computer itself is actually almost 40 years old!

But, this is a fairly new design, the TI-99/4A did not have an IDE disk controller until 2005.  This production run was done this year and they are new boards from the 2005 design.  Since the original BQ4847 was obsolete and I have seen counterfeit clock modules I went with the BQ4802 as a replacement.

As I said, I got 3 out of the original purchased IC's from TI.com to work on the production boards with no issues.  I only purchased 3 to prove it out.  but with the big lot I purchased (well I call it big) about 30, after 5 i gave up.

Hi Chris,

The only other thing I can think to try is swapping the crystals instead of the ICs, to rule out some crystal impedance-related phenomenon. Maybe as a long-shot you could try using the BQ4802LY with a 5V/3.3V LDO and some level shifters (or just resistor+zener clamps on the inputs, if the output at 3.3V would still be compliant).

TI doesn't guarantee failure analysis for devices. If TI won't take your return despite filling out the form and meeting the criteria (A-B-A swap, cleaned part, etc), there isn't really anything the applications team can do about it. That said, I would be surprised if the issues you're experiencing are due to a bad lot, as we would expect to see other customers raising similar issues simultaneously.

Regards,

I got around to finally messing with this some more.  I have built a PCB that has the circuit requirements for the BQ4802LYDW, 5k resistor between pin #6 and #28 and 32.768 crystal.  Instead of using a SMD crystal as I did before i used a barrel type this time as I noticed with the original PCB's, if I heated up t he crystal pins it would start working.  My only assumption is a connection issue witih the SMD and since the SMD's, so I decided to build the test board.

I also use a SOP28 to DIP28 IC Socket Programmer Adapter Socket so I can quickly change the BQ4802YDW for testing without soldering.

out of the 25 I ordered, 5 will not pass but the other 20 work flawlessly.

I would like to be able to get a replacement on the 5 that will not pass.

Perhaps you could request the five free samples from samples.ti.com. More information about our sample program here: 

Kind regards,
Lane

i would like to be able to do that, but I was told i could not since I am an individual and not a business or educational.  I am just supporting the old TI-99/4A home computer for the 1980's era.

Hi Wyatt,

Please send us an email at cts_dc@list.ti.com and we can help you with the sample request.

Kind regards,
Lane

Lane Boyd said:

Hi Wyatt,

Please send us an email at cts_dc@list.ti.com and we can help you with the sample request.

Kind regards,
Lane

Lane,

There is no one on the thread named Wyatt.  Did you mean me to send you an email about the samples?  If so I don't have permission to that distribution list.  I have already tried.