SN74FB1653: Thermal Management

Hi Brandon,

Please see FAQ: https://e2e.ti.com/support/logic-group/logic/f/logic-forum/723481/faq-where-do-i-connect-the-thermal-pad-of-the-logic-qfn-devices

Let us know if you have any further questions.

Regards,

Jack

Hi Jack, 

Thanks for the response. The issue is that there is no exposed thermal pad on the actual devices. It looks to me like these just have a heat slug requiring a heat sink to be mounted on top rather than any thermal management through the bottom of the package. There seems to be a direct conflict between the documentation and the parts I have in hand. 

Hi Jack, 

Thanks for the response. The issue is that there is no exposed thermal pad on the actual devices. It looks to me like these just have a heat slug requiring a heat sink to be mounted on top rather than any thermal management through the bottom of the package. There seems to be a direct conflict between the documentation and the parts I have in hand. 

Hello Brandon,

Could you please share the top marking? I don't see any PCNs where we changed the packaging. 

The CAD models I see on Digikey and Mouser have the thermal pad.

Regards,

Josh

Josh, 

Here is the top marking. 

-Brandon

Hello Brandon,

I'm still working to understand this issue. Where did you buy these parts? From TI or a 3rd party?

Regards,

Josh 

Josh, 

This set of 10 was purchased from Rochester Electronics. Here is a picture of the info on the packaging if that's of any help. However, I would note that we previously purchased Qty-2 direct from TI about a year ago that are identical. I am seeing on the website that there's another P/N ending in G4, but it's not clear to me what the actual difference is. 

-Brandon 

Hello Brandon,

Could you also share the packaging and top marking of the units purchased through TI.com

Regards,

Josh

Those are both actually already in use on a prototype board, but I was able to get a picture. 

-Brandon

Part of my concern is coming from the fact that the device is actually getting pretty hot when not in use, which makes me think there may actually be something wrong in my implementation. I seem to be pulling about 250mA on the Vcc(5V) input. Any ideas as to what might be causing that high current draw? Am I correct that BG Vcc and BIAS Vcc are supposed to be connected to Vcc(5V)? 

Hello Brandon,

Looking at the recommended conditions, connecting BG VCC and BIAS VCC to 5V is okay.

Could you share your schematic? When not in use, are any of your inputs floating?

Regards,

Josh

Hello, 

Here is a snippet from my schematic. It may be a little hard to follow since it's using references, but the signals ending in _O are digital outputs (3.3V) from an sbRIO-9603 and signals ending in _I are corresponding digital inputs. The signals labeled SPARE did wind up being left floating, so I can try to pull those high or low and see if it changes anything. On the B side, those signals are pulled up to 2V through 50 Ohm resistors (except the SPARE lines, which are floating). Otherwise all enable, clock, select lines are tied to 5V or GND. Please let me know if you see anything that stands out aside from the floating inputs. 

-Brandon

I believe I solved the packaging mystery. I found this document from 1997 that seems to show that these chips were originally produced with a heat slug on the top of the device rather than an exposed thermal pad underneath. I think that's what became the G4 variant, which I believe is now obsolete. We were sent G4 chips by TI even though that's not what we ordered, and the parts from Rochester Electronics are likely new old stock of the original obsolete part, before it was even designated as the G4 package. 

www.ti.com/.../scba003c.pdf

Hi Brandon,

We are still looking into this and will provide an update soon.

Regards,

Jack

Hi Guys, 

Following up to see if you had any other thoughts about the schematic? I was able to tie the floating pins to GND today, but it didn't change anything. The chip still gets very hot, seemingly drawing far more current on the 5V supply than it should be. I feel like I have to be connecting 5V somewhere it shouldn't be. Vref is floating, but that's just an output from the IC, correct? 

-Brandon