Edit-Only need to know this to calculate the voltage on pin 7. Once that is done the diode will be pypassed for the voltage divider.
Thread covers a different type of higher voltage mode for an FET based converted 105C board. See Post 19 for a description and Post 35 for the beginning of the mod.
Has anyone measured the voltage drop on this diode in use ? It’s the one in the shopping cart for many of the Oshpark project boards. The data sheets indicate somewhere between .3-.4v depending on the current to the mcu. The actual value affects the choice of resistors for the voltage divider.
Thank you C_k, I want to parallel RMM’s efforts on the voltage ladder for the Zener mod but it hinges on correct data for this diode and the existing Qlite diode(if different). The Qlite specs quoted from Mtnelectronics are 2.9-3v step down and 2.8v imminent shut off and I assume this is the voltage under load and accounting for the diode drop pin 7 should see less than that. I’ve heard .25v thrown around but the 41 shows to be more like .3~~.4v so pin 7 might trigger when Vcc is as low as 2.4v or as high as 2.55. The actual voltage pin 7 sees is determined by R1&R2 and is somewhere around .47~~.5v depending on the diode. If we keep the 4.7k R2 and just replace R1 then knowing the diode drop allows us to choose the correct value for R1. I’d like to order some 0805’s ASAP.
I compiled a list of different values for R1 keeping R2 the same at 4.7k ohms using different forward voltage drops on D1. This is for a Zener mod of drivers and firmware that use 19.1k and 4.7k for R1 and R2.
Assuming that the battery voltage under load trips lvp at 2.8v and that R/(R1+R2) = .1975 for those two values then when the voltage drop across D1 is:
Case #1, .25V pin 7 sees .1975 x (2.8-.25)=.50V
Case #2, .3V pin 7 sees .1975 x (2.8-3)=.494V
Case #3,.35V pin 7 sees .1975 x (2.8-.35)=.484V
Case#4, .4V pin 7 sees .1975 x (2.8-4)=.474V
I went to the mouser catalog and looked up replacement resistors for R1 that would give these same voltages when the battery voltage was 5.6V or (2 x 2.8V).
For D1= .25V, R1= 47.5k ohm
For D1= .3V, R1= 48.7k ohm
For D1=.35V, R1= 49.9k ohm
For D1= .4V, R1= 51K ohm
Mouser has these values in 0805 1% for ~ $.10 each so as soon as I get some decent values for VD1 I’ll order some to try out.
Even at a full charge of 8.4V pin 7 will still see less than 1.1V so I don’t expect to blow any mcu’s however…
I realize that this is somewhat different than what RMM did so I expect this will be all wrong and I’m not encouraging anyone else to try this but just posting what my plans are.
Why is the voltage divider placed after the protection diode anyway? Seems to me that we’d be just as well off not having the diode in the mix other than providing VCC to the MCU.
Microa - yes, after the diode does work. Results change when the diode is swapped for a different value. If we put the voltage divider before the diode (only use the diode for VCC) then we would have no need to worry about diode value.
The protection diode isn’t in this part of the circuit for the Zener mod but to calculate the correct voltage that pin 7 sees normally it no needs to be accounted for. If you were to remove it from a normal 105c then the protection wouldn’t register until the battery was dangerously low.
My point is that you are setting yourself up for a job that will end up needing to be done again in the future. Why bother when we can eliminate the task you asked about (measuring vDrop on the protection diode and changing formulas in the firmware)?
You didn’t start the conversation about a normal 105c, it was about Oshpark boards. If we modify the design to supply the divider with BAT+ instead of hooking the divider up to the protection diode then we’ll simply need to adjust resistor values to achieve the appropriate ratio for the divider. That’s a one time job, changing brands or models of resistor in the future will not have an effect on the voltage supplied to the ADC (unlike changing brand/model of diode). VREF will be constant, unaffected by the diode or absence thereof. Hopefully you see my point now?
I saw your point before and it is my intention to bypass the diode for Zener mods but to correctly figure out what resistors to use I need to be able to account for any diodes being used now. Doing a Zener mod and keeping low voltage detection with any driver that has the diode in front of the divider requires knowing the voltage drop of the diode. Obviously if the diode doesn’t precede the divider then the correct values can be anticipated.
I have not measured the diodes directly, but, when using the DrJ firmwares that blink out the battery voltage, it gives the same number of blinks as the original Nanjg S4 diode. That's only a resolution of 0.1v but it doesn't have to be that close. It's not like a cutoff of 2.8v/cell vs. 2.7v/cell will make a significant difference in actual use. I think nearly all cells get their factory capacity ratings from being run down to 2.75v anyway (some even cheat and run them all the way down to 2.5v).
I confuse myself sometimes. I should not have mentioned the digicart diode since the S4 is likely the only one I need to worry about but I wanted to solve the general equation before moving to specific examples. CC is quite right about the resolution here but since margin of error is additive I’d like to do what I can to not make it worse. Mainly I’m just stumbling along trying to figure this stuff out and posting my ignorance for all to see. Usually Texaspyro will come and save the day with dry wisdom. I was never shy about asking the stupid question in class.
I think the error factor might be more of an issue using two cells as slight cell imbalance combined with .1v or more error might be an issue. I guess it depends on how far apart the cells are when the combined voltage hits 5.6V.
In other words “get organized”(thank you Wallace&Grommit)
I will be using a 105C board with the 7135’s removed and the mcu replaced with one flashed for me by Dale.
It will have an Lfpak56 FET since I have some(rated for 100A) and fits the board with no overlap of ground ring. The pins don’t match the 70N02 so I won’t be using an Oshpark board.
Existing 4.7k R2 and new value for R1(somewhere between 47k and 51k).
The 7135 position closest too output traces cut, D1 as the jumper for Vin, a jumper wire from Vout to Vcc.
Voltage divider disconnected from Vcc and jumpered to Vin using current L+ pad or somewhere on the pwm trace after D1.
After all the work on the Oshpark boards by others I must be just a glutton for punishment. Here’s a pic of the board I’ll be modifying. Q3 is where I’ll put the voltage regulator and I’m thinking of drilling through the pwm trace for a direct path for L+.
Getting back to the main topic of the thread, if I connect the divider to B+ before the diode I need to know the voltage drop of the diode but if I’m using the same diode and connect the divider after the diode I just use 2.8v per cell to recalculate R1? This doesn’t seem right to me. I think you still need to know the voltage drop across D1 to get the correct value for R1 but wth do I know? Maybe you’re right in that not knowing VD1 won’t give you the exact value of the actual voltage read by pin 7 but will still give you the correct resistance value if you match that offset value. I think I need an Ervin translation.