LM25117: DC-DC converter with constant current source

EL,

The purpose of the DC-DC CC/CV is to supply a pool chlorinator system with power

Since the chlorinator uses electrolysis to separate sodium chloride into a usable form of chlorine, to optimize control and efficiency, I would like to have the programmable option of either providing the chlorinator with a CV in the range of 8-14VDC or a CC in the range of 8-12A. In other words, I need to be able to manually or programmatically select the mode of the DC-DC converter to either CV or CC and also select their operating values. So, in this case, the CC mode isn’t just for protecting again excessive load current but can be used in continuous operation to source a constant current where Vout becomes the variable based on load resistance instead current.

 Can the LM(2)5190(-Q1) provide this function and if so, can you provide some help on making sure I configure the circuit correctly as outline above?

Thx,

Jeff Boyer

TekMagic

Hello Jeff

My recommendation is for you to purchase LM5190 EVM and check your idea on the bench. 

If the CC mode current limit target becomes very high (~ 16A) by pulling down the IMON pin, the converter will work in CV mode in normal condition. Make the variable VOUT circuit works from 8V to ~ 18V, and then use this 18V setting as an OVP. 

If the CV mode regulation target becomes very high ( ~ 18V) by pulling down the FB pin, the converter will work in CC mode in normal condition. Make the variable IOUT circuit works from 8A to ~ 16A, and then use this 16A setting as an OCP. 

There will be a glitch during the transition from CV-only to CC-only, or CC-only to CV-only. You have to check this transition performance on the bench.  

-EL 

EL, after reviewing the latest needs with our product team, a few requirements have changed, and I have a few more questions.

Requirement changes.

The product requirement for CV is now ideally a range of 5 – 24 VDC and for CC a range of .5 – 10 for current. Hopefully this or something close, is still achievable with the right ratio of FB and IMON resistors.

Questions

How did you come up with the 8V to 18V in you statement - “Make the variable VOUT circuit works from 8V to ~ 18V, and then use this 18V setting as an OVP” ?

And how did you come up with 8A to 16A in your statement - “Make the variable IOUT circuit works from 8A to ~ 16A, and then use this 16A setting as an OCP” ?

How/where are you getting the current and voltage number from?

Thx

EL, I haven't tried the eval unit on the bench yet, but I should be able to simulate the LM5190 with Tina correct?

Also above you say "If the CV mode regulation target becomes very high ( ~ 18V) by pulling down the FB pin, the converter will work in CC mode in normal condition". But per spec, if I tie FB to gnd, Vout goes to 5V. So how does pulling down the FB pin put the converter in CC mode? When you say pull down, exactly what are you doing with the FB and IMON pins?

Can you provide enough of a schematic so I clearly see how to connect up FB and IMON?

Thx

El, I don't understand what the LM5190-Q1 spec means in section 6.3.9 by stating "the output regulation target can be adjusted during operation by connecting external feedback divider resistors whose parallel resistance is greater than 5.0kΩ".  The TI xlsm file for the LM5190 quick start design violates this statement by recommending RFBB & RFBT that are consistently < 5K for all Vout value I select. For example, if I select a Vout of 23V, the spread sheet recommends RFBT = 100K and RFBB of 3.6K. These values are correct for a Vout of 23V, but in parallel those values are much less than 5K, not great then 5K as the LM5190 spec seems to requires. 

What am I missing here?

Thx

Jeff

So if I don't connect FB to AGND or VCC and the parallel combination of RFBT and RFBB is less than 5K what happens? In other words if I use 100K for RFBT and 3.6K for RFBB to get Vout of 23V (which less than 5K in parallel), does that mean I can't vary the R ratio after initialization and what would happen if I did?

What I want is for RFBT and RFBB as low as possible to get 23V out so that I can use cheaper digital potentiometers and 100K (values of digital potentiometer) and 3.6K do that for me.

El, what do you mean by "please input 150 at G72"?  

Also IMON is supposed to provide a way to measure Iout. I don't see a scaling ratio anywhere in the LM5190 spec. So what is it ratio of Vimon to Iout? Thx

This is also the case with ISET. ISET purpose is to allow me to set I_CC out. So I need the ratio of V_ISET to I_CC. Can you provide that? Thx

El, the first page of the LM5190 spec says the part can do the following 3 things -

Under the bullet Current monitoring and constant current features

1 – Analog voltage proportional to output current (IMON)

2 – Programmable average output current limit (ILIM)

3 – Dynamic average output current limit (ISET)

I want to do all 3 of things but the spec doesn't seen to provide enough information for me to proceed and accomplish the task. Can you help me? 

El, I what to simulated the LM5190-q1 design with LTspice since Pspice seems so slow. If I sign an NDA, can TI provide a spice model I can use with LTspice? Thx. My email address is Jeff.b@tekmagic.net

Hello Boyer 

Please contact your local FAE if you'd like have a NDA required material from TI. Please contact me via your local FAE by email. 

-EL  

El, more questions. I'm attempting to design AC compensate for the LM5190 and I'm using both the Quick Started calculator for the LM5190 and comparing that with the Power stage designer setup with synchronous CMC buck topology. However, when I set things up almost identically for both design tools, I get different results. Quick Start give me values for Rcomp, Chf and Ccomp that are considerably different then PS designer and so are loop gain and phase. And even when I plug in the recommended values from quick start into PS design and vice versa, I don't to get the same loop responses. If both tools are functional, they should provide similar results... right? So what I'm I doing wrong? And which tool should provide the best results? Thx

Also in Power stage designer, Compensation Network is set for Type II

I think the network should actually be type II transconductance but I tried both type II and transconductance and I don't see much difference

Hello Boyer 

Please also try LM5190 WEBENCH which is on the bottom side of this product folder  https://www.ti.com/product/LM5190-Q1. 

-EL