Can't find Roea value for LT3864

[LTnet]

Hi guys I'm new here and this is my first post

I have a homework to simulate and calculate the poles and the zeros of the compensation circuit of a buck converter.

I've set up all the circuit and done all the simulations

The problem is that i'm stuck at calculating the near zero pole because i cant locate on the datatsheet the Roea value for this chip.

The component used is LT3864(datasheet https://www.electro-tech-online.com/custompdfs/2012/12/3864f.pdf) from Linear

As shown in this linear application note (https://www.electro-tech-online.com/custompdfs/2012/12/an76.pdf page 5) the first pole is calculated using the Roea

I cant really find whats the value of Roea for this component.

Can you you please give a look at the datasheet?

Thanks

 

[LTnet]

I made some progress guys

i found this Application Report from TI (https://www.ti.com/sc/docs/psheets/abstract/apps/slva352a.htm)

if you see on Table 1 in page 4 they say that Roea=DCgain(ea) / gm(ea)

i can find gm_ea on the datasheet of LT3864 and its 1.8mS but i cant find the DCgain

Can you please help me?!

 

[crutschow]

I can't find a data sheet for a LT3864.

 

[LTnet]

I can't find a data sheet for a LT3864.

link on first post
https://www.electro-tech-online.com/custompdfs/2012/12/3864f.pdf

 

[LTnet]

So no one can figure this out?!

Do you think they(linear) will ever reply if i write them about this through their tech support form?
(https://www.linear.com/designtools/tech_support.php)

Anyone has any experience in this?

Thanks

 

[crutschow]

You are looking at the data sheet for one type of switching regulator and the application note for a different type. You need look at the ap notes for the regulator you are using.

 

[LTnet]

You are looking at the data sheet for one type of switching regulator and the application note for a different type. You need look at the ap notes for the regulator you are using.

"For a detailed explanation of OPTI-LOOP
compensation, refer to Application Note 76."

This is written on page 20 on the converter datasheet(https://www.electro-tech-online.com/custompdfs/2012/12/3864f.pdf) and it points to the Application Note linked on the first post

 

[MrAl]

Hi,


Yes App Note 76 does seem to apply, but it looks unclear whether or not the other note applies as they are using the two divider resistances not the innate characteristic of the error amp itself.

I took a quick look and it looks like a catch 22. If i give you a single parameter:
A=3.2

and then tell you to look in app note 43 to find out how to calculate the fish population in a given tank at a fish farm in New Jersey and app note 43 says:
Fpop=A*B*7

we can plug in A and get:
Fpop=3.2*B*7

but we still cant figure out the fish population because we were never told in reference 1 what B was.

If they state that you need to know what B is and they dont tell you what it is, then then only way to find out is to ask them. They clearly state the value in their app note for the other amplifier, so we would expect to see it clearly stated for the other amplifiers too. It could be that their chronology was that one app note came before the other, but that still doesnt explain why there is no way to easily calculate the required compensation values for the several examples they provided. They provided specific values for their several examples but they never show the way to achieve those values without referring to the app note, and the app note tells us we must know the error amp output resistance in order to calculate those specific values. So somebody made a mistake somewhere, and it doesnt really matter that much if we can extrapolate the required value from the parts already specified (this may be an option) because that much work should not be required and i seriously doubt they would want their customers to have to go through all that work when a simple spec would clear it all up immediately.
So yes you should contact them even if you can figure this out from the reference designs.

 

[LTnet]

Hi,


Yes App Note 76 does seem to apply, but it looks unclear whether or not the other note applies as they are using the two divider resistances not the innate characteristic of the error amp itself.

I took a quick look and it looks like a catch 22. If i give you a single parameter:
A=3.2

and then tell you to look in app note 43 to find out how to calculate the fish population in a given tank at a fish farm in New Jersey and app note 43 says:
Fpop=A*B*7

we can plug in A and get:
Fpop=3.2*B*7

but we still cant figure out the fish population because we were never told in reference 1 what B was.

If they state that you need to know what B is and they dont tell you what it is, then then only way to find out is to ask them. They clearly state the value in their app note for the other amplifier, so we would expect to see it clearly stated for the other amplifiers too. It could be that their chronology was that one app note came before the other, but that still doesnt explain why there is no way to easily calculate the required compensation values for the several examples they provided. They provided specific values for their several examples but they never show the way to achieve those values without referring to the app note, and the app note tells us we must know the error amp output resistance in order to calculate those specific values. So somebody made a mistake somewhere, and it doesnt really matter that much if we can extrapolate the required value from the parts already specified (this may be an option) because that much work should not be required and i seriously doubt they would want their customers to have to go through all that work when a simple spec would clear it all up immediately.
So yes you should contact them even if you can figure this out from the reference designs.

Thanks for your advice
I wrote to Linear on the 4-th of January but unfortunately no reply has arrived yet(and i doubt it will ever come...).

 

[MrAl]

Hi,

If all else fails we might be able to back track from the reference designs, but im not sure how much detail they provided there either. If they are skimpy on details we wont know what they intended to get from the design anyway so it will be very hard if not impossible to back track.
LATER: Cant find any info that would allow us to backtrack the reference designs either on the data sheet.

If you let me know what you asked them i can mirror that and maybe they will reply to one of us.

 

[LTnet]

i Just asked them that i was unable to find the Roea value for the LT3864 device as described on the Application Note 76 and if they could provide that value.
Not the exact words but the meaning was that.

I wrote them through their website contact form using my university account.

Thanks for your help!

 

[LTnet]

MrAl
any news? Did you write them?

 

[MrAl]

Hello again,

I have to sincerely apologize as i did not get this written and sent until today. I expect a reply within 3 days.

Again im sorry i did not get this out sooner, but at least now it's done and so we just have to wait a little bit. I wanted to take the time to explain the situation clearly so they know what we are talking about here. It may be that they have to look up the data themselves or do more testing to get the original value as it must have gotten swept under the table somehow. It does appear that they must have had it at some point though, so may be they will try to contact the original engineers.

 

[LTnet]

Hello again,

I have to sincerely apologize as i did not get this written and sent until today. I expect a reply within 3 days.

Again im sorry i did not get this out sooner, but at least now it's done and so we just have to wait a little bit. I wanted to take the time to explain the situation clearly so they know what we are talking about here. It may be that they have to look up the data themselves or do more testing to get the original value as it must have gotten swept under the table somehow. It does appear that they must have had it at some point though, so may be they will try to contact the original engineers.

Thank you for you time.
I do really appreciate it.
Hope they do reply this time.

 

[LTnet]

well more than 10 days have gone
I really doubt they will ever reply at this point...

 

[MrAl]

Hi,


Yeah im starting to think even they dont know anymore and obtaining the data again is difficult for them. It's a fairly simple question so they must have a problem getting the information. They expect engineers to bench test the circuit i guess.

Maybe you should consider using a different part.