LM5156: LM5156 - any alternatives?

Hi Tomasz,

If you are concerning with the input transient of 89V, I would recommend to add some input filters to get ride of this transient (by adding higher input cap or a clamping diode)

But you can use also use LM5020 which can withstand 100V.

https://www.ti.com/lit/ds/symlink/lm5020.pdf?ts=1747298922649&ref_url=https%253A%252F%252Fwww.ti.com%252Fproduct%252FLM5020

Best Regards,

Hassan  

Hi Hassan,

Filters are already added and thanks to them the circuit must survive 80V only.
I need SEPIC since I would rather not to use transformers and reversed polarity (buck-boost) is not an option.

Kind regards,

Thomas

Hi Tomoasz,

Does your circuit working properly with the input caps?

Can you please elaborate your second comment regarding the SEPIC? Thanks

Best Regards,

Hassan

Hi Hassan,

I cannot say whether the circuit with LT3758 works properly since I have not built the prototype yet - but soon I will, test PCB is scheduled to go to production this week.

As far as SEPIC: my understanding is that LM5020 requires a transformer, likely it would have to be custom. Even if I can find something off-the-shelf the selection would be limited which is a concern for production because in case of supply shortage it may be difficult to find a replacement. For this reason I prefer SEPIC topology.

Kind regards,

Thomas

Hi Tomasz,

As I recommended earlier that you can try to add decoupling ceramic input cap to filter out the input transients or add clamping diode at the input. You can try this with LM5156. 

I will wait for your results.

Best Regards,

Hassan 

Hi Hassan,

As I stated earlier, filters, including TVS diodes, are already added and thanks to them the circuit must withstand 80V only.
I just wanted a confirmation that I did not overlook something and currently there is no TI chip which would meet my requirements.

Thank you,

Thomas

Hi Mounika,

Thank you, I guess I did not capture it right the first time. Indeed, LM5020 product page mentions SEPIC topology although it is not even mentioned in product PDF specs.
I cannot use LM5020 since it has Vin(min) = 13V. Note that I need to design and test 12V SMPS working in the Vin range 6-35V. LT3758 works from 5.5V. I guess, it is what it is. You may want to bring this case up to product planning/development team.

Kind regards,
Tomasz Jastrzebski

Hi Mounika,

Somehow, I did not think about this solution. Thank you!
It is 3 additional components, but it is worth it. LM5156 is much more modern chip than LT3758.

Kind regards,

Thomas

Hi Mounika,

I think I can use such an external LDO to power both the BIAS and VCC pins and this got me thinking about it again. I could probably also use 5V from another SMPS for this purpose, but I think in that case the switching MOSFET must have a low Vgs(th). What would be the minimum values ​​of Vgs(th) and Qg required to maintain good performance in such a configuration?

Section 9.3.13 MOSFET Driver (GATE Pin) states that recommended min Vcc voltage is 6-7V but probably 5V with carefully selected MOSFET would work too, as long as Q_G@VCC x f_SW < 35mA but I am not clear about units in this expression.
What is the minimum margin between Vcc and V_GS(th)(max)?

Kind regards,

Thomas

Hi Tomasz, 

May I know why would you want to supply Vcc with external supply? The device also supports the use of an external VCC supply to improve efficiency for higher Vin. When the external voltage in Vcc is lower than Vin, the internal LDO will regulate and will have 6.85V at Vcc anyways. 

Make sure the gate driver volatge is higher than the  Miller plateau of the MOSFET used. Higher the Gate drive volatge lower the Rds. Please check the gate drive volatge of the MOSFET that you would like to use.

Best regards,

Mounika

Hi Munika, power efficiency and proper power dissipation are the reasons. LM5156 can draw as much as 110mA and with Vin(max) 35V const (80V transient) almost 4W need to be dissipated somewhere, partially at the additional LDO transistor and the chip itself but it is a lot of power. The same time I have 5V SMPS on the PCB just right next to 12V SMPS. It would be wise to use it - even if I need to select a better (read more expensive) switching transistor with lower Vgs(th) and Qg. According to section 8.3 min V_BIAS and V_VCC is 2.97V so 5V should be achievable.

If I correctly read Q_G@VCC x f_SW < 35mA expression found in LM5156 datasheet section 9.3.13, at 2.15MHz Qg should be lower than 16.3nC - which is not a problem. The transistor I think I could use (ISZ330N12LM6ATMA1) has Qg(th) just 1.1nC and Vgs(th)(max) 2.2V. IAUZ30N10S5L240ATMA1 might be a good choice too.

Please comment/confirm.

The last thought: I could probably use even newer LM5158 with internal MOSFET, but WEBENCH fails to calculate external component values for Iout > 1A and Vin 6.5-35V.

Thanks,

Tomasz

Hi Moritz,

Thank you for the response.
What if I connect both VCC and BIAS pins to externally provided 5V?

Kind regards,

Thomas

Hi Moritz,

Thank you for the clarification/confirmation.
The last question: What is the minimum margin between VCC and Vgs(th)(max) of a given switching transistor?

Kind regards,

Thomas

Hi Moritz,

Thank you, I got it, V_GS(th) is just not the right value to decide.

Best regards,

Thomas