[SOLVED] I cant find suitable PWM controller IC for 90V to 12V buck converter delivering 5A

[nikee]

I am planning to make a buck converter able to convert from (30V - 90V) to 12V delivering 5A current.
I am fine in calculating the inductor and capacitor and other parts requirement and selection but I am facing problem in choosing the right PWM controller IC.
My converter will be
1. Current mode controlled
2. Non-Isolated
3. Synchronous or asynchronous depending on complexity and part cost

I have been looking for several IC from multiple manufacturers but each IC got some problem.
I started with TI UC3842A but their current sense resistor is monitoring 1V voltage drop which results in 5W of power loss which is unacceptable. (specially since I am planning to increase the current in next iteration)

Then I went for LT3800 but the maximum input voltage is 60V but I need 90 volts.
After looking at many PWM controller IC I can say that my PWM IC should be
1. Either not using Rsense (current detected via mosfet Rdson) or the type of Rsense which is in series to load (Rsense+:Rsense-) which consumes very low power.
2. It should not be directly fed by incoming voltage because I don't think many PWM IC can withstand 90V at Vin, and the one able to take 90V are very costly (in 4-6$ range, what I am looking is like 0.6-1.0$))
3. Should be driving the Mosfet directly
4. Certainly should have external switch (not the one with internal switch and limited to 2-3 Amps of current)
5. Should not be too costly

I think choosing the right IC should not be a big issue and I must be missing something.
Any help choosing the right IC will be really helpful. Also if someone can point me to a similar project
that will be really helpful (I might be missing something but I have been looking for similar project on google for past 3-4 days and believe me I cant find one).

I think I am wrong in interpreting the Rsense of TI UC3842A IC. I assume that you have to choose the Rsense so that it drops 1V across itself which sets the current limit accordingly but that will create 5Ax1V = 5watts of power loss. I must be wrong, can someone help me understand the selection of Rsense, the datasheet seems confusing to me.

Secondly I want to know why almost all IC are always directly fed from main input voltage, instead we can use a small linear regulator to power the IC and feed the mosfet directly from the input voltage.

 

[FvM]

You probably won't find a low-voltage switch mode controller for 90V input. So you should get to like the idea of adding a HV preregulator for the internal supply respectively startup of a standard switcher. Gate drive circuits have to be adapted respectively, depending on the switcher topology.

 

[nikee]

That's what I also want to do, have a linear 5V or 12V converter to power the IC and feed the mosfet the actual incoming voltage of 90V (If i understand you correctly). As for gate driver if I choose a logic level Mosfet (with Vgs(th) about 2V) which completely turns on at below 5V and Vds > 100V do you think the IC will be able to drive the Mosfet? since ground is common for Mosfet (i.e. Source) and PWM IC the output of 5V from PWM IC being fed to Mosfet should turn it on. (i have little bit of theoretical knowledge of Mosfet but zero practical knowledge so pardon me if anything is absurd).

If it's feasible why can't I see any circuit doing this?instead I almost always find the main voltage being fed to PWM IC. Even many manufacturers have separate section for high voltage PWM IC.

 

[FvM]

Of course it's not impossible to make a single chip solution if you manage respective HV processes, at least with external switch, it may already be available from a manufacturer. But the offer of high voltage ICs is small and ICs that interface directly to mains voltage are exclusively made by special manufacturers, e.g. powerint.

If you move to flyback topology, you can use existing off-mains switcher solutions, but I guess you want to avoid a transformer.

 

[nikee]

I think I am getting the idea here.
Not many people will be using buck topology for reducing 100 or 120 volt to 12V, instead they will go for flyback or push-pull etc. But for like 60V to 12V buck will do fine, that's why there are many PWM IC available which can be used with 60V input but not many above that (like 90 or 100 volts).
It actually make sense and falls inline with my line of thinking. I was planning to use flyback for 120V+ and buck for 90V and below but I think I will stick to buck for upto 60V and switch to flyback for 90V and above.
And yes you got me right, I wanted to avoid transformer for this small converter.
My whole hearty thanks for your help.

Regards

 

[erikl]

I think choosing the right IC should not be a big issue ...

You could try dcdcselector.com , a provider-independent search engine with useful selection criteria.

 

[nikee]

thanks erikil, that link is really helpful.

Regards