LTC4365 resistor values

[qrper]

Gang....

I guess my old high school math is lacking. I'm unable to come up with the three necessary values to use with the Linear LTC4365.

I have the SPICE model working, and I've been tossing values in it all day long.

The required over voltage set point would be 15V and the low voltage set point 11.5 V.

Quick frankly, I don't know why Linear had use the examples of 18 and 3.5 with a 12V source. But then again.... who knows?

I can set the OV setpoint, but when it comes to LV set point, I'm not even close.

so, does anyone what to help a guy with a few numbers?

Many thanks

mike

[MagicSmoker]

Attached is a snippet of a window comparator circuit I use for OV/UV lockout based on TI TPS3700. I didn't check the LTC4365 datasheet very carefully, but I did notice it uses a 0.5V reference, so the 0.4 in my equations would need to be changed to 0.5. See if they work for you and then let us know the result, as all too often around here we get drive-by problems, so to speak.

edit: bay -> by

[qrper]

That's close.... running the spice, the low voltage kicks in a 12 V

I'll diddle with the values a bit more.

Is it me or does this seem a bit much to calculate a few resistors

Mike

[MagicSmoker]

That's close.... running the spice, the low voltage kicks in a 12 V

I'll diddle with the values a bit more.

Is it me or does this seem a bit much to calculate a few resistors

Mike

Well, now you know why I put the equations in the schematic, because I got tired of re-deriving them every time I need this kind of OV/UV lockout!

[digsys]

....  I use for OV/UV lockout based on TI TPS3700 ...

I also use that comparator, with 2+ in series, with up to 5 windows. Just made up an EXCEL spreadsheet, all the complex formulas plugged in, and now I can easily change resistors / or voltages
and instantly see the values change. In one case, I even linked a chart.  Use the tools Luke, use the tools !! :-)

[Zero999]

Here's a handy tool which can help you select resistor values for a potential divider, given the input and output voltages.
http://sim.okawa-denshi.jp/en/teikokeisan.htm

[qrper]

....  I use for OV/UV lockout based on TI TPS3700 ...

I also use that comparator, with 2+ in series, with up to 5 windows. Just made up an EXCEL spreadsheet, all the complex formulas plugged in, and now I can easily change resistors / or voltages
and instantly see the values change. In one case, I even linked a chart.  Use the tools Luke, use the tools !! :-)

If I buy you a beer and a cheeseburger, would you send me the excel file?

Mike

[digsys]

... If I buy you a beer and a cheeseburger, would you send me the excel file? 

:-) I can't give you my equations, as they are part of our products, but I can push you in the right direction -
If you google search - "BQ24450 calculations" - Which uses a similar resistive divider PITA chain :-) , one guys formula is at - oliver.st/blog/quick-calculator-for-bq24450-lead-acid-charge-controller/
Which is where I started. There are also other formulas for that IC, which we also use with the series battery stack, hence the complicated requirement.
One more hint - SET a FIXED current for your string, eg 10uA / 20uA, and derive the formulas from that. It actually becomes a piece of cake once you have 1 fixed reference. See how you go :-)
Edit: another formula
https://e2e.ti.com/cfs-file/__key/telligent-evolution-components-attachments/00-179-01-00-01-32-75-90/correct-BQ24450EVM-Battery-Charging-Calculations.xlsx

[Dave]

You know, there's a reason why IC manufacturers provide datasheets...

They even go as far as showing an example calculation for the mathematically challenged.

[digsys]

You know, there's a reason why IC manufacturers provide ... They even go as far as showing an example calculation for the mathematically challenged.

Yeah thanks, I must have been too dumb to understand them :-) attitude issue?

[qrper]

... If I buy you a beer and a cheeseburger, would you send me the excel file? 

:-) I can't give you my equations, as they are part of our products, but I can push you in the right direction -
If you google search - "BQ24450 calculations" - Which uses a similar resistive divider PITA chain :-) , one guys formula is at - oliver.st/blog/quick-calculator-for-bq24450-lead-acid-charge-controller/
Which is where I started. There are also other formulas for that IC, which we also use with the series battery stack, hence the complicated requirement.
One more hint - SET a FIXED current for your string, eg 10uA / 20uA, and derive the formulas from that. It actually becomes a piece of cake once you have 1 fixed reference. See how you go :-)
Edit: another formula
https://e2e.ti.com/cfs-file/__key/telligent-evolution-components-attachments/00-179-01-00-01-32-75-90/correct-BQ24450EVM-Battery-Charging-Calculations.xlsx

Ah, no problem. I was worried about how to package a cheeseburger for the mail.

[qrper]

You know, there's a reason why IC manufacturers provide datasheets...

They even go as far as showing an example calculation for the mathematically challenged.

Yeah, I know why they publish datasheets. My original post states quite simply, that the math is beyond my abilities. It's been 60 years since I've been in high school, and the numbers I come up with aren't correct when I run the LTSPICE simulation.

I guess I'm only a stupid old man that asked from some help.

[Zero999]

You know, there's a reason why IC manufacturers provide datasheets...

They even go as far as showing an example calculation for the mathematically challenged.

Yeah, I know why they publish datasheets. My original post states quite simply, that the math is beyond my abilities. It's been 60 years since I've been in high school, and the numbers I come up with aren't correct when I run the LTSPICE simulation.

I guess I'm only a stupid old man that asked from some help.

See the attached spreadsheet.

The V1, V2 & V3 outputs are for use with the optimum resistor finder tool I linked to previously.
http://sim.okawa-denshi.jp/en/teikokeisan.htm

V1 = 10
V2 = 0.434782608695652
V3 = 0.333333333333333

Enter them into the tool and it spits out the following values:
R1 = 4k3
R2 = 47R
R3 = 150R

Where R1 to R3 is R_LOWER to R_UPPER. Those values are a bit high and it's only the ratio which matters, so they can be multiplied by 100.

R_UPPER = 430k
R_MIDDLE = 4k7
R_LOWER = 15k

[digsys]

  ... I guess I'm only a stupid old man that asked from some help.

Don't worry about Dave, he (and some others) have a habit of belittling people who aren't as smart as they is :-). Happens pretty much in all forums.
And welcome to EEVBlog, hope this didn't put you off. If you really get stuck, I'll make you up a generic spreadsheet. Just PM me.

[qrper]

  ... I guess I'm only a stupid old man that asked from some help.

Don't worry about Dave, he (and some others) have a habit of belittling people who aren't as smart as they is :-). Happens pretty much in all forums.
And welcome to EEVBlog, hope this didn't put you off. If you really get stuck, I'll make you up a generic spreadsheet. Just PM me.

gang, after all these years on the planet, my skin is kinda thick. but sometimes my wagon does come out of the barn.
Thanks everyone for the help. For some reason the spice simulation doesn't seem to like the values used for the low end turn off. It became quite apparent that the way the designers made the IC, changing one resistor  value ever so slightly upsets the whole cart.

So what I did was pull the voltage dividers from the OV and LV pins. I put separate voltage dividers on each input line. I will order some of the ICs and find a carrier so I can put power to the things in real time. It also looks like a 10 turn trimmer might be needed as tolerance in even 1% resistors can really swing the outputs.

Since I don't have one here to look at, the datasheets say the footprint is a TSOT-23. Tiny little sucker!

Made up a pcb device pattern using a SOT-23 footprint to start with. According the dimensions for the pcb cad software, the one I made should be okay. We will see.

thanks everyone

mike

[Zero999]

  ... I guess I'm only a stupid old man that asked from some help.

Don't worry about Dave, he (and some others) have a habit of belittling people who aren't as smart as they is :-). Happens pretty much in all forums.
And welcome to EEVBlog, hope this didn't put you off. If you really get stuck, I'll make you up a generic spreadsheet. Just PM me.

gang, after all these years on the planet, my skin is kinda thick. but sometimes my wagon does come out of the barn.
Thanks everyone for the help. For some reason the spice simulation doesn't seem to like the values used for the low end turn off. It became quite apparent that the way the designers made the IC, changing one resistor  value ever so slightly upsets the whole cart.

So what I did was pull the voltage dividers from the OV and LV pins. I put separate voltage dividers on each input line. I will order some of the ICs and find a carrier so I can put power to the things in real time. It also looks like a 10 turn trimmer might be needed as tolerance in even 1% resistors can really swing the outputs.

Since I don't have one here to look at, the datasheets say the footprint is a TSOT-23. Tiny little sucker!

Made up a pcb device pattern using a SOT-23 footprint to start with. According the dimensions for the pcb cad software, the one I made should be okay. We will see.

thanks everyone

mike

It doesn't work because the formula on the data sheet, my spread sheet and that posted above are wrong! I don't have time to correct it at the moment.

I agree, it's easier to use two potential dividers! The disadvantage is higher power consumption.

Using the tool I linked to above, with V2 set to 0.5V and V1 set to 11.5V and 15V for under and over voltage, gives out the correct values. Use a trimmer pot. if you like but I wouldn't recommend it because it's a bodge. If it's possible, it's better to select the correct resistor values in the first place.

[qrper]

BASICALLY, that's what I did

I used the resistor calculator and solved for 11 in and .5 at the input of the LV divider.

Did the same for the OV input, too.

The trimmers would make it a bit easier to trim the settings. Current draw won't be an issue, considering if this works, the supply will be a high capacity (200amp/hrs) battery.

[Zero999]

BASICALLY, that's what I did

I used the resistor calculator and solved for 11 in and .5 at the input of the LV divider.

Did the same for the OV input, too.

The trimmers would make it a bit easier to trim the settings. Current draw won't be an issue, considering if this works, the supply will be a high capacity (200amp/hrs) battery.

I don't know what I was thinking last night. The formula above, on the data sheet and my spreadsheet are correct.

One thing you'll notice and possibly what might have confused me, is the turn on voltages for the rising and falling side of the input ramp, are slightly different. This is because the LTC4365 has hysteresis, which is a good idea, to prevent oscillation.

[digsys]

I forgot to mention, how I dealt with the resistor tolerance issue -
I use 2x resistors for each Ref point. The 1st resistor is a value below calculation, the 2nd resistor isn't installed, but has a 2 pin cal connector, in which I plug 2 - 4 trimpots.
I have 2 DMMs also plugged into test points. Plug it all in, wiggle the pots, find the closest R1b / R2b / R3b value and add it. I keep a huge set of E192+ resistors.
It's perfect for a production environment, and you don't end up with a trimpot of dubious tempco / tracking.

[Zero999]

I forgot to mention, how I dealt with the resistor tolerance issue -
I use 2x resistors for each Ref point. The 1st resistor is a value below calculation, the 2nd resistor isn't installed, but has a 2 pin cal connector, in which I plug 2 - 4 trimpots.
I have 2 DMMs also plugged into test points. Plug it all in, wiggle the pots, find the closest R1b / R2b / R3b value and add it. I keep a huge set of E192+ resistors.
It's perfect for a production environment, and you don't end up with a trimpot of dubious tempco / tracking.

If the optimum resistor values are chosen, the tolerance is fairly small.

With the values shown in my previous post.
R_UPPER = 430k
R_MIDDLE = 4.7k
R_LOWER = 15k
Overvoltage = 14.99V
Undervoltage = 11.414V

Using 1% resistor values:
R_UPPER = 165k
R_MIDDLE = 1.74k
R_LOWER = 5.76k
Overvoltage = 14.974V
Undervoltage = 11.5V

[MagicSmoker]

Generally 1% resistors are sufficiently accurate when using a window comparator for OV/UV lockout. For example, the circuit excerpt I posted earlier controls the operation of a much larger dc/dc converter with an operational input range of 100-400V, while the window comparator trip points are closer to 95V and 425V so as to not run into any nuisance tripping of UV or OV lockout (and believe me, the market this product is aimed at has customers that will light up my inbox if I don't deliver better than promised specs - it's like the exact opposite of how engineers should interpret component datasheets...  ).

[Zero999]

Generally 1% resistors are sufficiently accurate when using a window comparator for OV/UV lockout. For example, the circuit excerpt I posted earlier controls the operation of a much larger dc/dc converter with an operational input range of 100-400V, while the window comparator trip points are closer to 95V and 425V so as to not run into any nuisance tripping of UV or OV lockout (and believe me, the market this product is aimed at has customers that will light up my inbox if I don't deliver better than promised specs - it's like the exact opposite of how engineers should interpret component datasheets...  ).

I concur. There's no need for precision components or trimmers. I'd even say 5% is good enough for a voltage cut-off circuit, where you ordinarily want to keep the voltage a good 10% away from either extreme, to avoid nuisance tripping.

[digsys]

I concur. There's no need for precision components or trimmers. I'd even say 5% is good enough for a voltage cut-off circuit, where you ordinarily want to keep the voltage a good 10% away from either extreme, to avoid nuisance tripping. 

And for most situations, that may be quite true, but since the OP asked about tolerances, I offered a solution. He may well have been content with 1-5%
This is my setup - During cal / setup, insert the service PCB, and run all the relevant tests required, including sweeping the Input voltage ranges. Quite nifty :-)

[MagicSmoker]

And for most situations, that may be quite true, but since the OP asked about tolerances, I offered a solution. He may well have been content with 1-5%
...

The OP specified a voltage window of 11.5V to 15V, so 1% resistors at each position in the divider should suffice.

[qrper]

gang,

I received my supply of LTC4365 ICs.

I know what size the footprints are, but when you see the part in real time.

Damn! That thing is tiny.

Mike