TL2844: Frequency Calculation Question

Darren (FAE)

Expert 5483 points

Part Number: TL2844

Figure 16 in the DS has Note A) which defines an equation for calculating the oscillator frequency.

Using 3.6kΩ and 1nF, the expected frequency was around 480kHz.

Figure 10 in the DS seems to confirm this.

But the actual measured frequency was ~200kHz.

I noticed Note A) states the equation is valid for Rt > 5kΩ; so it is possible the equation changes for Rt < 5kΩ?

But then we expect this to be expressed in Figure 10 for Rt < 5kΩ.

Either the equation, or the Figure 10, need to be updated to reflect actual behavior...I think we may need a Datasheet update...

Can you confirm?

Darren

over 3 years ago

0 Hong Huang over 3 years ago

TI__Guru 60195 points

Hi,

I think you should always use Figure 10 for estimation. Page 11 formula provides shorthand estimation but only when RT > 5k. RT = 5k is a rough cutting condition, so the formula gets closer to Figure 10, when RT becomes much larger than 5k. You can take a few points, such as when RT = 30k, or 40k, to check.

0 Darren (FAE) over 3 years ago in reply to Hong Huang

TI__Expert 5483 points

Hi Huang,

The problem is Fig.10.

Look at Y-axis and draw a horizontal line at 3.6kOhms.

Then, look at the trend line for 1nF.

You will see they cross at around 500kHz.

Bit customer is observing closer to 200kHz.

The problem here is the trend line under 5kOhm is not accurate, and causing confusion / wasted time in design/test process.

1) Can we update figure 10 to show accurate frequency vs Rt and Ct under 5kOhm?

0 Hong Huang over 3 years ago in reply to Darren (FAE)

TI__Guru 60195 points

Hi,

I will request to re-test and check.

Note Fig 10 was based on the tested numbers while for reference only as typical. There are tolerance factors both on CT (what is your 1nF capacitance tolerance?) and on IC (roughly 10%).

0 Darren (FAE) over 3 years ago in reply to Hong Huang

TI__Expert 5483 points

Hi Huang,

Hong Huang said:
I will request to re-test and check.

Thank you.

I understand tolerances will cause sub-ideal results, but in this particular case the trend-line in Figure 10 going below 5kΩ and continuing linearly, doesn't align with the equation stating that linear trend only works above "Rt > 5kΩ"

When the customer tried to pick values using Table 10, they saw this issue and are hoping the trend line can more accurately reflect how the linearity changes with lower Rt (sub 5kΩ) values.

Darren

0 Hong Huang over 3 years ago in reply to Darren (FAE)

TI__Guru 60195 points

Hi,

Ok, noticed.

0 Darren (FAE) over 3 years ago in reply to Hong Huang

TI__Expert 5483 points

Hi Huang,

Any update? Customer was hoping for a reply by 7/14 - today.

0 Hong Huang over 3 years ago in reply to Darren (FAE)

TI__Guru 60195 points

Hi,

I have made request. A people has been assigned to review Fig 10. Upon his review result, he will decide how to proceed. No report has been at moment yet.

+1 Hong Huang over 3 years ago in reply to Hong Huang

TI__Guru 60195 points

Hi,

Any further inquiry, please send it to the person is in charge of this request. Please copy to me as well. In case some help is needed from me, such as help explaining what it has been on this.

+1 Darren (FAE) over 3 years ago in reply to Hong Huang

TI__Expert 5483 points

Hi Huang,

Thanks for giving me a contact. I have copied your email and will send a message offline.

Darren

Power management

Power management forum

TL2844: Frequency Calculation Question