gfxgfx
 
Please login or register.

Login with username, password and session length
logo
 
gfx gfx
gfx
19681 Posts in 1259 Topics by 5012 Members - Latest Member: theodorosgrigoriou September 23, 2017, 08:49:15 AM
*
gfx* Home | Help | Login | Register | gfx
gfx
Breaktru Forum  |  eCigarette Forum  |  Modding  |  Topic: You asked mamu it's coming.
gfx
gfxgfx
 

Author Topic: You asked mamu it's coming.  (Read 77924 times)

0 Members and 11 Guests are viewing this topic.

Offline CraigHB

  • PV Master
  • *******
  • Joined: Nov 2011
  • Location: Reno, Nevada
  • Posts: 2023
  • Karma: +243/-1
  • Gender: Male
Re: You asked mamu it's coming.
« Reply #100 on: October 16, 2014, 01:09:04 AM »
You know what, I'm not fully understanding your circuit configuration, I think I misread the first drawing.  Regardless, if you want to measure voltage at the atomizer then that's where you should put the tap for your ADC's voltage divider.  The above drawing shows that to be the case. 

If I'm understanding the previous drawing correctly, it shows an additional resistance R1 inline with the atomizer.  In that drawing you are not measuring voltage at the atomizer.  You're measuring atomizer plus R1.  There would be an error with respect to atomizer voltage introduced by the voltage drop at R1.

Putting a cap on the ADC's divider serves a two fold purpose (across R3 in your diagram).  Firstly, it filters the ADC input to keep it stable.  It also provides a low impedance source for the ADC.  The way ADC's work is they charge a small capacitor when they take a sample, usually around 5pF.  That being the case, they are sensitive to hold times defined by ADC software configuration and the impedance of the source under measurement.  A filter capacitor provides the energy to charge the ADC's capacitor quickly reducing error and sensitivity to hold times.  To get the most benefit a low ESR cap should be used, ideally an MLCC cap, those have the lowest ESR.

It can be helpful for accuracy to use a precision voltage reference some amount lower than supply voltage.  Linear regulators are typically not that accurate, they can vary a few percent over changes in load.  If using supply voltage as a reference, readings can also vary over load.  Voltage references are designed specifically for use with an ADC and will eliminate that issue.  Here's an example;  http://www.ti.com/product/lm4040-n


« Last Edit: October 16, 2014, 01:14:23 AM by CraigHB »

Offline CraigHB

  • PV Master
  • *******
  • Joined: Nov 2011
  • Location: Reno, Nevada
  • Posts: 2023
  • Karma: +243/-1
  • Gender: Male
Re: You asked mamu it's coming.
« Reply #101 on: October 16, 2014, 01:59:36 AM »
I'm going with a separate post on the converter stuff.  This is going to be long and I want avoid making one giant reply that obscures the main topic.

You can use a stand-alone MOSFET driver in designing a converter and it's not particularly old school.  There's some benefit in using a separate driver.  Converter controllers with built-in drivers are limited in how much gate charge they can drive.  In some cases, specifically for high power stuff, you need to drive low resistance MOSFETs that have a relatively large gate charge.  This means you need a more powerful driver. 

Most converter controllers with built-in drivers are limited to MOSFETs with a gate charge under 20nC @4.5V (assuming a typical PWM frequency around 500kHz).  That's fine for lower power stuff, but when you get into higher power outputs, you need MOSFETs with lower on-state resistance.  That translates to higher gate charge.  There's a dependence between those.  Lower on-state resistance calls for larger gate junctions which have more capacitance. 

You can find powerful MOSFET drivers specifically designed for switching converters.  They can drive MOSFETs at high speed with gate charges as high as 100nC @4.5V.  However, there is a trade-off there.  Power consumption in driving a switch is gate charge times frequency times voltage so you can see where that goes.  With high gate charges and low power outputs, you can consume as much power driving the switches as the load.

Once you have your MOSFET driver you still need to control it.  You can do it with an analog controller or digitally with a DSP (high speed MCU).  Some MOSFET drivers have built in shoot-through protection which is a critical aspect of design.  I think this may be where you were having issues. 

Shoot through occurs with a timing issue that results in both MOSFETs conducting at the same time.  Adaptive shoot-through protection is best since it minimizes the resulting dead time.  This is the amount of time both switches are off and there's a big efficiency hit that occurs when there is too much.

If your driver does not have shoot-through protection, then the controller must have it.  Either too much dead-time or not enough can cause issues.  Not enough dead time causes excessive shoot-through which can overheat MOSFETs to the point of failure.  Too much dead time can kill efficiency.

You can use a DSP like the Texas Instruments C2000 series to provide the PWM signal and control for your converter, but you must take care to handle dead time properly in code if you go that route using a driver without shoot-through protection.  You should also program cycle by cycle current limiting.  That can be done by sensing voltage drop accross the low side MOSFET when in full conduction.  Again this is required to protect the MOSFETs and keep them from overheating.


« Last Edit: October 16, 2014, 02:07:10 AM by CraigHB »

Offline miskol

  • Super Member
  • *****
  • Joined: Mar 2014
  • Location: Malaysia
  • Posts: 106
  • Karma: +8/-1
  • Gender: Male
Re: You asked mamu it's coming.
« Reply #102 on: October 21, 2014, 02:51:39 PM »
thank you very much for your explanation, very insightful with lots of new information for me.

will take time for me to understand all of it.

but some did makes sense hehe, which is the part regarding the driver's shoot-through protection because there is a design problem on the PCB that makes the driver unstable. will further research on your comments. the driver and mosfets worked on my previous design that uses Buck topology, but with my latest design i used a different and smaller mosfets and also changed the circuit for Boost topology. will design a new PCB to repair the current driver problems and revert back to using the previous working mosfet, this should better clarify where the problems are.

again tq very much Craig and very sorry especially to Claviger and karadorde since some discussions has swayed from the main topic haha, sorry guys.


will continue the development regarding the ADC's voltage divider. i've repaired the connection and followed the latest circuit configuration, unfortunately the problem is still the same but at least by checking with DMM, Vfb relates closely with theory (where the voltage Vfb will always be 1/2 the voltage applied across R2 an R3). so most probably the problem is on software. i'm sure there's no problem with the ADC reading part, supposedly, since there's no problem on reading the battery level. maybe some ports aren't configured properly or some functions of the pin are not disabled.


Offline CraigHB

  • PV Master
  • *******
  • Joined: Nov 2011
  • Location: Reno, Nevada
  • Posts: 2023
  • Karma: +243/-1
  • Gender: Male
Re: You asked mamu it's coming.
« Reply #103 on: October 21, 2014, 07:17:09 PM »
Configuring ADCs can be a bit of a headache sometimes.  You can get error if the timings are too tight.  I usually start them liberally, then tighten them up to where they start having problems to get an idea what the available range is.  There's external factors like source impedance that can have an effect on the limits of ADC timing.  However, using two 3.3K resistors in your divider is plenty low enough.  That shouldn't be a problem.

To make sure it's not an ADC timing issue, decrease the sample rate and increase the hold time liberally just to check it.

Another thing that can give you trouble is excessive noise in the power supply.  You should be isolating the ground planes for your digital and power circuits.  You should also run a 1uF MLCC cap as close as possible to each of the power and ground pin pairs for your MCU, especially the analog power pins if they are separate for the MCU you are using.

Offline norcalreballer

  • Sr. Member
  • ****
  • Joined: Jul 2014
  • Location: USA
  • Posts: 63
  • Karma: +4/-0
  • Gender: Male
Re: You asked mamu it's coming.
« Reply #104 on: October 26, 2014, 02:28:41 PM »
I had my first go at creating a schematic.  Please let me know if you see anything that isn't correct.  There has to be something, as I killed a teensy when I powered it via the LM317 configured like this...  I tried to create a pic, but you can't see anything.. Used designspark to create the schematic.

lt3092 - constant currect source.  Decided to use kelvin measurement for resistance of atty
MIC4420 - Mosfet driver
LM4040  Voltage reference



https://www.mediafire.com/?xur31ozm2rizqi8
« Last Edit: October 26, 2014, 03:30:15 PM by norcalreballer »

Offline miskol

  • Super Member
  • *****
  • Joined: Mar 2014
  • Location: Malaysia
  • Posts: 106
  • Karma: +8/-1
  • Gender: Male
Re: You asked mamu it's coming.
« Reply #105 on: November 04, 2014, 01:33:34 PM »
thanks for the suggestions Craig, i've tested both, its not timing issue.

Another thing that can give you trouble is excessive noise in the power supply.  You should be isolating the ground planes for your digital and power circuits.  You should also run a 1uF MLCC cap as close as possible to each of the power and ground pin pairs for your MCU, especially the analog power pins if they are separate for the MCU you are using.

it might be the isolation problem, although i'm not absolutely sure about it because it could be that maybe my Booster's PIC ADC pin is damaged somehow, maybe lol. but still, i believe isolation is really the problem. i used a different external PIC to read the ADC from the OKR output and also to read ADC from the current sensor, my setup is as below image.



so by using external PIC, the ground planes are isolated. only a single cable is connecting the external PIC's ground to the Booster's ground (joined on the proto board).

the results:
OKR setup, 220 ohm fixed resistor and a 200 ohm trimmer potentiometer. The approximate output range is 3.4v to 6v

0.5ohm(5% tolerance) load  test
top pic: pot set to min, Vout = 2.83V under load; Iout = 4.86Amp; Rload = 0.58ohm
bot pic: pot set so that approx. Vout = 4V under load; Iout = 5.97Amp; Rload = 0.67ohm


1ohm(X% tolerance) load  test
top pic: pot set to min, Vout = 2.99V under load; Iout = 3.33Amp; Rload = 0.89ohm
bot pic: pot set so max Vout = 5.63V under load; Iout = 5.45Amp; Rload = 1.03ohm


so as can be seen, i have accuracy issues with the detected load resistance. might be better to compare with load readout with DNA, to be continued hehe.

tested with DNA 30's ohm reading, for the 1ohm load test it reads 1.0ohm and for the 0.5ohm load test it reads 0.6ohm.
« Last Edit: November 06, 2014, 01:10:09 PM by miskol »

Offline Bruno_B

  • Skydiver
  • Jr. Member
  • **
  • Joined: Nov 2014
  • Location: Atlanta, GA
  • Posts: 8
  • Karma: +0/-0
  • Gender: Male
Re: You asked mamu it's coming.
« Reply #106 on: November 11, 2014, 02:51:43 AM »
I am also using an I2C display so that I can free up the SPI pins on the board to use with a digital potentiometer.

You don't have to free up SPI pins. All your SPI devices share the same MOSI, MISO, and SCK pins. Each device gets a discrete SS pin which you pull low to let that device know that it is the one you are speaking to. https://learn.sparkfun.com/tutorials/serial-peripheral-interface-spi/slave-select-ss You can have as many SPI devices on your Arduino as you have IO pins minus 3. (That 3 is for the MOSI, MISO, and SCK pins.) This is really important when you start using smaller AVR chips like I do here:

Offline CraigHB

  • PV Master
  • *******
  • Joined: Nov 2011
  • Location: Reno, Nevada
  • Posts: 2023
  • Karma: +243/-1
  • Gender: Male
Re: You asked mamu it's coming.
« Reply #107 on: November 11, 2014, 06:26:05 PM »
I2C is a better interface for running multiple peripheral devices.  First the bus is two way requiring only 2 pins and devices are addressed in software rather than hardware.  For example, to run 4 SPI devices you need 3 bus pins plus a CS pin for each device requiring a total of 7 pins, though you can multiplex the CS pins to get that down to 5 pins.  I2C can run four devices with only 2 pins. 

SPI does have some advantages, it's more robust and can be a lot faster.  Also, SPI has one way modes that require only two pins for the data bus.  One way SPI is often used by peripheral devices where they only receive and not send.  Genrally, SPI is better for one or two devices running high speeds where I2C is better for multiple devices running at lower speeds.

Offline Bruno_B

  • Skydiver
  • Jr. Member
  • **
  • Joined: Nov 2014
  • Location: Atlanta, GA
  • Posts: 8
  • Karma: +0/-0
  • Gender: Male
Re: You asked mamu it's coming.
« Reply #108 on: November 11, 2014, 10:25:24 PM »
@CraigHB, that is all true. I prefer I2C devices when I can find them. One nice thing about SPI is that it's pretty easy to analyze with a scope. If you have a peripheral that is a little wonkey in its implementation of SPI, It's easy to bit-bang the protocol with only the wikipedia article and an hour of not so focused time. (I did it while watching an episode of The Walking Dead.) In order to bit-bang I2C, you pretty much have to hate yourself or be working on a Thesis.

I love the way they state it in this article. "SPI is quite straightforward – it defines features any digital electronic engineer would think of if it were to quickly define a way to communicate between 2 digital devices." Yep, if I needed to quickly whip up a protocol for 2 devices, that's about how I'd do it. The rest of that article is worth reading too.

Offline CraigHB

  • PV Master
  • *******
  • Joined: Nov 2011
  • Location: Reno, Nevada
  • Posts: 2023
  • Karma: +243/-1
  • Gender: Male
Re: You asked mamu it's coming.
« Reply #109 on: November 12, 2014, 11:18:20 AM »
Yes SPI is very easy to work with and you can quite easily verify the data stream with a couple channels on a scope.  As I2C is not synchronous, it can be a bit tricky to analyse without a DSO having ample memory.  I like SPI a lot myself because it's so easy to work with, but it does have a lot more overhead.  I usually use SPI unless I have a need to run a number of devices and don't have the spare pins for it.

Offline miskol

  • Super Member
  • *****
  • Joined: Mar 2014
  • Location: Malaysia
  • Posts: 106
  • Karma: +8/-1
  • Gender: Male
Re: You asked mamu it's coming.
« Reply #110 on: November 12, 2014, 11:48:35 AM »
Another thing that can give you trouble is excessive noise in the power supply.  You should be isolating the ground planes for your digital and power circuits.  You should also run a 1uF MLCC cap as close as possible to each of the power and ground pin pairs for your MCU, especially the analog power pins if they are separate for the MCU you are using.

Finalizing the PCB design to solve the isolation of ground planes. Two different ground planes are isolated (Analog & Digital) which only inter-connects by a single 0 ohm resistor. This should solve the ADC issues right?


Offline CraigHB

  • PV Master
  • *******
  • Joined: Nov 2011
  • Location: Reno, Nevada
  • Posts: 2023
  • Karma: +243/-1
  • Gender: Male
Re: You asked mamu it's coming.
« Reply #111 on: November 12, 2014, 12:35:08 PM »
The biggest source of noise comes from induction and voltage spikes due to changing currents acting on impedance.  Most generally, it's the parasitic impedance encountered by current flow that gives you trouble.  You minimize those effects by routing return currents from power circuits away from sensitive high impedance connections.

For the board you show, I can't tell for sure where the pickup is located for battery negative.  It should be in the area of the orange arrow you show as common ground.  You want to keep return currents for the power circuits from running through the digital plane.

Offline DOUG36854

  • Jr. Member
  • **
  • Joined: Nov 2014
  • Location: Valley, AL
  • Posts: 2
  • Karma: +0/-0
  • Gender: Male
Re: You asked mamu it's coming.
« Reply #112 on: November 14, 2014, 02:49:54 AM »
Hopefully someone can help me on this. I plan to build a regulated box using a delta electronics chipset. Undecided on 200w/300w/400w as of yet. Regardless, it is a neg logic chip like the OKR so I should be able to swap it into this schematic and use the programming as written. However, I would like for the MCU to be able to read the resistance of the coil. It looks like the n-fet, pin A0, 5v from the MCU and the 100 ohm pot are being used as a voltage drop calculator to read the ohms of the coil (using the signal off of A8 to switch the nfet) but I don't see anything in the code to read this voltage on A0 and calculate the resistance. Did that portion of the code get removed for simplicity sake and the coil resistance is just being manually entered via code? If I am wrong regarding that portion of the schematic being a voltage drop for that purpose, is it possible to switch a mosfet from a signal off the MCU and have it read voltage on a pin to calculate resistance of the coil? Maybe I'm over looking it but I would really rather have the MCU calculate everything rather than have resistance set as a variable entered in the code.

Offline CraigHB

  • PV Master
  • *******
  • Joined: Nov 2011
  • Location: Reno, Nevada
  • Posts: 2023
  • Karma: +243/-1
  • Gender: Male
Re: You asked mamu it's coming.
« Reply #113 on: November 14, 2014, 08:00:13 PM »
I haven't looked at the code and hardware for this project in detail, but my understanding is it uses a voltage divider to determine atomizer resistance.

There's a couple ways to find atomizer resistance.  You can use a current sensor or you can use a voltage divider.  Either way is reasonable, but I prefer to use a current sensor since it provides real time current monitoring and the best accuracy.  However, using a divider can have advantages too. 

There should be a FET in the circuit that momentarily switches atomizer output through a voltage divider, though like I said I'm not intimately familiar with this project.  The code measures voltage at the output and voltage at a known resistance then calculates atomizer resistance based on the math of a regular voltage divider.  Once resistance and voltage are known, current and power are calculated from that.

When using a current sensor, you measure output current and output voltage then determine resistance and power from that.  In terms of the math and code, it's more simple to use a current sensor, but you can't measure resistance on demand.  The atomizer has to be normally powered.  Plus there's more hardware involved in a current sensor.  Still, the ability to monitor current on the fly is a big advantage in my opinion.

Offline DOUG36854

  • Jr. Member
  • **
  • Joined: Nov 2014
  • Location: Valley, AL
  • Posts: 2
  • Karma: +0/-0
  • Gender: Male
Re: You asked mamu it's coming.
« Reply #114 on: November 14, 2014, 08:08:49 PM »
From a size standpoint,  won't a current sensor capable of up to 80A with accuracy into the .1st range be a little large to fit into a portable box mod?  Current shold be able to be real time concerning voltage changes (resistance changes should be small while firing and not need to be re-read unless the coil is changed). This project appears to already have a voltage divider planned I just can't find anything in the code for the MCU to switch the fet or read the voltage input from that circuit. Maybe the original poster/designer can chime in with some insight.

Offline miskol

  • Super Member
  • *****
  • Joined: Mar 2014
  • Location: Malaysia
  • Posts: 106
  • Karma: +8/-1
  • Gender: Male
Re: You asked mamu it's coming.
« Reply #115 on: November 14, 2014, 08:28:26 PM »
It looks like the n-fet, pin A0, 5v from the MCU and the 100 ohm pot are being used as a voltage drop calculator to read the ohms of the coil (using the signal off of A8 to switch the nfet) but I don't see anything in the code to read this voltage on A0 and calculate the resistance. Did that portion of the code get removed for simplicity sake and the coil resistance is just being manually entered via code?

Based on the code, the resistance of the Atty is first known and then its value is inserted in the R2 variable (float R2 = .90). karadorde have also verified this on post #94 that Claviger's work is not reading resistance.

Offline CraigHB

  • PV Master
  • *******
  • Joined: Nov 2011
  • Location: Reno, Nevada
  • Posts: 2023
  • Karma: +243/-1
  • Gender: Male
Re: You asked mamu it's coming.
« Reply #116 on: November 14, 2014, 08:34:54 PM »
The OP for this project has not signed on to the forum for several months now so you probably won't be able to get any advice from him in this thread.  Others have used this project as a starting point for their own and you might get some advice that way. 

In terms of current limitations, there's one that comes from the FET used to switch the divider in and out.  However, current sense resistors and current sensing chips typically have limits lower than high power FETs.  That's going to be a consideration. 

There is actually more physical size involved with a divider since the high power FET required to handle large currents would be larger than the sense resistor or sensing chip required for current sensing.  Plus, the lower the value of the reference resistor the better the accuracy and the size of resistors goes up quickly with power rating.

The part count can be lower for a divider, but then current sensing chips provide a single chip solution.  Current sense resistors require an amplifier using a number of parts unless you happen to be using an MCU with an available milli-Volt ADC.

Offline miskol

  • Super Member
  • *****
  • Joined: Mar 2014
  • Location: Malaysia
  • Posts: 106
  • Karma: +8/-1
  • Gender: Male
Re: You asked mamu it's coming.
« Reply #117 on: November 14, 2014, 09:20:25 PM »
The biggest source of noise comes from induction and voltage spikes due to changing currents acting on impedance.  Most generally, it's the parasitic impedance encountered by current flow that gives you trouble.  You minimize those effects by routing return currents from power circuits away from sensitive high impedance connections.

For the board you show, I can't tell for sure where the pickup is located for battery negative.  It should be in the area of the orange arrow you show as common ground.  You want to keep return currents for the power circuits from running through the digital plane.



Thanks Craig for the comment.

following up, hope this will help you advise if the PCB should be okay. shown is the battery negative contact (Black rectangle). basically top and bottom PCB plane is the same, there is no Analog grounding that runs through or under the Digital grounding or vice versa.

Offline CraigHB

  • PV Master
  • *******
  • Joined: Nov 2011
  • Location: Reno, Nevada
  • Posts: 2023
  • Karma: +243/-1
  • Gender: Male
Re: You asked mamu it's coming.
« Reply #118 on: November 14, 2014, 10:19:50 PM »
The digital plane is still not fully isolated since return currents for the digital circuits are going to be common with return currents for the power circuits.  Think of it as running two separate ground connections on two separate boards.  One plane should connect next to the the negative pickup and the other plane should connect next to the negative pickup with no other connections between the planes.  This is a two connection version of the "star topology" for ground connections typically recommended for minimizing noise in electrical circuits.

Offline norcalreballer

  • Sr. Member
  • ****
  • Joined: Jul 2014
  • Location: USA
  • Posts: 63
  • Karma: +4/-0
  • Gender: Male
Re: You asked mamu it's coming.
« Reply #119 on: November 15, 2014, 01:08:27 AM »
DOUG36854  I just scored an ACS709 75a version, but haven't tested it yet.  It's on a little board.  Very small.  There's also other current sensors that can handle(supposed to, anyways)75a+.  I'm hoping to test it this weekend.  Unrelated, my idea of using 4 wire measurement worked with a multimeter, but it has problems... The voltage being read by the adc in this setup is way too low.  I read an InAmp is the way to go, but there's another problem.

I was using the differential input feature of the teensy 3.1 and didn't account for that circuit always being on.  Needless to say, the MCU smoked when the analog pin got 5v over the max it could handle.  One option would be a mosfet, which adds even more hardware.  So current source + InAmp + mosfet.  Even if it worked, that's too much stuff for me.  I was hoping to get resistance before firing, but doesn't matter. Seems like it's the best way is the current sensor. :)

I'm not understanding the get resistance with a voltage divider thing.  Could someone link me or explain that?  I'm guessing you'd need some actual high quality resistors?

EDIT:

I'm working on a little side project where I'm using a 3.3v pro mini(328p) with a 128x32 oled display.  I'm switching an NFET via digital pin that's switching a PFET to the atomizer.  PWM frequency is 480hz.  Reason for the PFET is because I can't read the average atomizer voltage at the 510 any other way.  If you know of one, please let me know.  I know an NPN transistor will work instead of the NFET, but I just used what I had.  That part is working great.

I mapped the analog input from a pot(going between a high digital pin and ground)from 3.3 to 5.5.  This is what I'm using for target voltage.  With no filter(software or hardware)I'm getting mostly reads of battery voltage.  I averaged the reads in software from 10-250 times, and it just jumps like crazy. 

I tried some different combos with resistors(in the voltage divider to read the atty avg voltage) and capacitors along with different amounts of averaging, but they seem to either jump too much or the cap takes too long to charge.  Any ideas here?  I'm using a 10k and 2.2k for the voltage divider now. 
« Last Edit: November 15, 2014, 04:44:30 AM by norcalreballer »

Offline CraigHB

  • PV Master
  • *******
  • Joined: Nov 2011
  • Location: Reno, Nevada
  • Posts: 2023
  • Karma: +243/-1
  • Gender: Male
Re: You asked mamu it's coming.
« Reply #120 on: November 16, 2014, 05:41:00 PM »
I'm not understanding the get resistance with a voltage divider thing.  Could someone link me or explain that?  I'm guessing you'd need some actual high quality resistors?

With two resistors in series one known and one unknown, you can find the unknown resistance if you know the applied voltage and and voltage at the common point of the two resistors.  It's just a matter of some algebra with the the equation for a voltage divider.

By putting an NMOS on the return for the atomizer jumped with a resistor, you can switch the divider in and out by turning the MOSFET on and off.  You would run an ADC connection to both the high and low side of the atomizer.  You can alternately do it on the supply side using a PMOS, but an NMOS would be preferable.

You don't need a particularly high precision resistor.  A typical 1% resistor would be fine.  The one consideration is that the lower the value of the series resistor the better the resolution of the measurement, however, that also increases power demand and resistor size up goes up exponentially with power rating.  Though if you minimize the "on" time for the divider, you can use a typical 1/4 watt resistor without issue even for larger currents.

Quote
I know an NPN transistor will work instead of the NFET, but I just used what I had.  That part is working great.

Yes a bipolar junction transistor will "work" but one of the issues in using a BJT over a FET is they have a collector-emitter drop similar to a diode.  It's a lot less than a diode, typically around 200mV.  That's not much, but when you get into big currents, you can start getting into appreciable losses and excessive component heating. 

Also the CE drop varies depending on the BJT so that's something you have to check in the part's data sheet.  The reason MOSFETs were invented was to alleviate those issues encountered in using BJTs for switch applications, especially for high power stuff.

Offline norcalreballer

  • Sr. Member
  • ****
  • Joined: Jul 2014
  • Location: USA
  • Posts: 63
  • Karma: +4/-0
  • Gender: Male
Re: You asked mamu it's coming.
« Reply #121 on: November 16, 2014, 11:22:27 PM »
Thanks for the info!  Last night I tried 10x lower value resistors. Ended up with 1k and 220ohm. My batteries don't last anyways. :)  I got it to read smoothly and quickly with a 220uf cap and only reading 3 times then averaging.  It's a very tiny bit jumpy, but not bad for only 3 lines of code for the MCU to try to get it to regulate the average voltage.   

Offline miskol

  • Super Member
  • *****
  • Joined: Mar 2014
  • Location: Malaysia
  • Posts: 106
  • Karma: +8/-1
  • Gender: Male
Re: You asked mamu it's coming.
« Reply #122 on: November 17, 2014, 12:24:08 AM »
The digital plane is still not fully isolated since return currents for the digital circuits are going to be common with return currents for the power circuits.  Think of it as running two separate ground connections on two separate boards.  One plane should connect next to the the negative pickup and the other plane should connect next to the negative pickup with no other connections between the planes.  This is a two connection version of the "star topology" for ground connections typically recommended for minimizing noise in electrical circuits.



okay Craig, i think i've understand it and got it now hehe. the digital ground now goes directly alone to the negative pickup.

hope you could advice on this revision, thanks :)

Offline CraigHB

  • PV Master
  • *******
  • Joined: Nov 2011
  • Location: Reno, Nevada
  • Posts: 2023
  • Karma: +243/-1
  • Gender: Male
Re: You asked mamu it's coming.
« Reply #123 on: November 17, 2014, 10:56:36 AM »
That's the idea, but you should do a copper fill on the whole board, unless you're trying to conserve copper.  Otherwise use a liberal width on that trace feeding the digital plane.  The fill would look something like this;

Offline zotek

  • Jr. Member
  • **
  • Joined: Apr 2014
  • Location: san diego
  • Posts: 16
  • Karma: +0/-0
  • Gender: Male
Re: You asked mamu it's coming.
« Reply #124 on: November 24, 2014, 01:03:19 AM »
hi what kinda of Potentiometers are u using? thannks

Ya craig, I expected in a failure of the Atty it would all be safe inside the box, but its one thing for a data sheet to say so, and entirely another for it actually work as designed in the real world, a point I am very sure is not lost on you!  As a result of your response, I am now considering adding a simple 1 wire thermistor to measure the temp of the bottom of the 510 connector and lock out firing if it reaches a certain temp, no idea what I'll set it at yet, but thinking 170f should be the right ballpark.  That kayfun was H O T, definitely over 300f, quickly tapping it with a finger felt hot similar to when you accidentally touch a soldering iron tip.

Final layout of the screen for this version. A short video showing the multi button function and sleep timer.



A preview of things to come, just waiting on my Femtoduino and the new screen (33x12x2mm screen with 128x32 I2C for 9 bucks from china :P) to arrive and all of it will fit in the silver case 32x32x95:



https://www.youtube.com/watch?v=THTb6blhewM&feature=youtu.be

Offline miskol

  • Super Member
  • *****
  • Joined: Mar 2014
  • Location: Malaysia
  • Posts: 106
  • Karma: +8/-1
  • Gender: Male
Re: You asked mamu it's coming.
« Reply #125 on: December 04, 2014, 12:03:58 AM »
Update :)

That's the idea, but you should do a copper fill on the whole board, unless you're trying to conserve copper.  Otherwise use a liberal width on that trace feeding the digital plane.  The fill would look something like this;

yes thanks Craig, my board's final version is copper-filled.



it is working nicely :) again, thank you very much Craig!

will be playing with the programming now :)

Online Breaktru

  • Administrator
  • PV Master
  • *****
  • Joined: Nov 2008
  • Location: NY
  • Posts: 3297
  • Karma: +784/-4
  • Gender: Male
  • PLEASE participate in this forum
Re: You asked mamu it's coming.
« Reply #126 on: December 04, 2014, 07:29:02 AM »
Looking really great. Very professional   :thumbsup:

Offline Visus

  • PV Master
  • *******
  • Joined: Oct 2013
  • Location: Nexxus
  • Posts: 929
  • Karma: +60/-0
  • Gender: Male
Re: You asked mamu it's coming.
« Reply #127 on: December 04, 2014, 08:36:34 PM »
Badarse Miskol


That is awesome, so many modders want that board. 

Offline sKyZ

  • Newbie
  • *
  • Joined: Dec 2014
  • Location: Charlotte, NC
  • Posts: 1
  • Karma: +0/-0
  • Gender: Male
Re: You asked mamu it's coming.
« Reply #128 on: December 31, 2014, 11:38:53 PM »
Hello :)

I joined this forum just for this thread. I'm a huge nerd and love vaping. With the new chips coming up, new mods, and other such technology, I have been really interested in programming my own chip to do so and stumbled upon this thread.

I am looking to order the parts to get started and was wondering, after the 3 pages of discussion, what parts you guys suggest ordering and any other (updated) information you may want be able to share before I start this adventure.

Thank you!

Offline cadcoke5

  • Newbie
  • *
  • Joined: Jan 2015
  • Location: Lancaster, PA
  • Posts: 1
  • Karma: +0/-0
  • Gender: Male
Re: You asked mamu it's coming.
« Reply #129 on: January 03, 2015, 11:45:13 PM »
I am interested in doing something like this, however, rather than collect the pieces myself, I would prefer to purchase either a full kit, or even an assembled version.

Do you plan on eventually offering a kit, or if not, is there any E-cig VW mod kit, that includes a programmable processor like the Arduino?

Thanks for any leads.

Offline hill115side

  • Jr. Member
  • **
  • Joined: Aug 2014
  • Location: HIllside nj
  • Posts: 9
  • Karma: +1/-0
  • Gender: Male
Re: You asked mamu it's coming.
« Reply #130 on: January 06, 2015, 03:09:09 PM »
Just want to leave this here...

Offline hill115side

  • Jr. Member
  • **
  • Joined: Aug 2014
  • Location: HIllside nj
  • Posts: 9
  • Karma: +1/-0
  • Gender: Male
Re: You asked mamu it's coming.
« Reply #131 on: January 06, 2015, 03:12:10 PM »
OKL2

Offline Bruno_B

  • Skydiver
  • Jr. Member
  • **
  • Joined: Nov 2014
  • Location: Atlanta, GA
  • Posts: 8
  • Karma: +0/-0
  • Gender: Male
Re: You asked mamu it's coming.
« Reply #132 on: January 06, 2015, 05:28:52 PM »
Looks promising! I'm just not a fan of stacking batteries. I don't need 100W. (I don't know why anyone does.) I have looked at the OKL2 http://www.digikey.com/product-search/en?vendor=0&keywords=okl2 but for the 2.4v input version which can run off of a single battery, it only has up to a 3.63v output. So in order to get the full 44w it is capable of, you have to use a 0.3ohm build. For the 73w version that's a 0.18ohm build. I know some RDA vapers get that low, but I think .5ohm is the basement for a lot of us.

I'd like to be able to drive an Aspire Atlantis, but at .5ohm that's just 26w when 35w is recommended.
« Last Edit: January 06, 2015, 05:38:31 PM by Bruno_B »

Offline atu

  • Jr. Member
  • **
  • Joined: Jan 2015
  • Location: Belgium
  • Posts: 2
  • Karma: +0/-0
  • Gender: Male
Re: You asked mamu it's coming.
« Reply #133 on: January 28, 2015, 12:07:07 PM »
Hi guys, congrats for your work!!

I'm really interested in working on an Arduino controlled Raptor but my electronic skills are not as good as my development skills are.
I've already written a good part of the code but I'm still trying to figure out how to get an accurate current measurement (or an accurate ohm reading but from what I understand based on your posts, it seems that calculating it based on current measurement is more accurate) and how to control the raptor output (digital pot or other solutions ??).

If you have some clear explanations to give me including part numbers that would be awesome !

Thank you guys for sharing all this knowledge !!!

Offline doobedoobedo

  • Super Member
  • *****
  • Joined: Dec 2013
  • Location: UK
  • Posts: 165
  • Karma: +35/-0
  • Gender: Male
Re: You asked mamu it's coming.
« Reply #134 on: January 28, 2015, 01:20:45 PM »
Hi guys, congrats for your work!!

I'm really interested in working on an Arduino controlled Raptor but my electronic skills are not as good as my development skills are.
I've already written a good part of the code but I'm still trying to figure out how to get an accurate current measurement (or an accurate ohm reading but from what I understand based on your posts, it seems that calculating it based on current measurement is more accurate) and how to control the raptor output (digital pot or other solutions ??).

If you have some clear explanations to give me including part numbers that would be awesome !

Thank you guys for sharing all this knowledge !!!
I'm using this current sensor http://uk.farnell.com/webapp/wcs/stores/servlet/ProductDisplay?catalogId=15001&langId=44&urlRequestType=Base&partNumber=1791413&storeId=10151 it requires an opamp to boost the output to give the 0-5V scale used by the arduino as it only outputs 28mV/A.

I'm using a dual 5k digipot wired in parallel to set the voltage http://uk.farnell.com/webapp/wcs/stores/servlet/ProductDisplay?catalogId=15001&langId=44&urlRequestType=Base&partNumber=1840762&storeId=10151 I use a lookup table to set each pot to give the overall resistance for the required voltage. I wrote a little program which generates the table declaration to give me 0.01V steps from 1V to 8V although I'm not sure yet how high I'll take the voltage when it's done.

Using a scale of 0-25A I have 0.024A resolution and a scale of 0-8.5V gives me a 0.008V resolution for the output which should translate to pretty accurate resistance measurements.

Offline atu

  • Jr. Member
  • **
  • Joined: Jan 2015
  • Location: Belgium
  • Posts: 2
  • Karma: +0/-0
  • Gender: Male
Re: You asked mamu it's coming.
« Reply #135 on: January 29, 2015, 04:49:30 AM »
Thank you very much for your feedback doobedoobedo !
I'll look into it tonight  :)

Good luck for you project !

Offline mamu

  • PV Master
  • *******
  • Joined: Dec 2013
  • Location: IN
  • Posts: 476
  • Karma: +70/-3
  • Gender: Female
Re: You asked mamu it's coming.
« Reply #136 on: February 25, 2015, 04:50:30 PM »
Claviger - where are you?  Hope you're ok.

Here's something as a kind of drop in module for those of us illiterate Arduino programmers lol.  Shows batt charge level, volts, amps, ohms, and you can set high/low batt voltages.  Would be perfect if it showed watts too, but then that's easily calculated.

Got some coming in to tinker with to see how compatible it is with the Raptor 20A.

http://www.aliexpress.com/item/Lithium-Li-ion-battery-tester-LCD-meter-Voltage-Current-Capacity-18650-26650-Free-shipping/32266677192.html




Offline Visus

  • PV Master
  • *******
  • Joined: Oct 2013
  • Location: Nexxus
  • Posts: 929
  • Karma: +60/-0
  • Gender: Male
Re: You asked mamu it's coming.
« Reply #137 on: February 25, 2015, 06:18:01 PM »
Excellent find Mamu

Wonder what happen over 10A tho
waiting for your testing with a huge bag of  :popcorn:

Clavinger disappeared off every forum,  ya,  hope is he is ok..

If only the guy who did the raptor programming with temp control was as giving as you guys lol..
30A, ohms,  tc, on/off, etc etc  someone will post it eventually lol..

I haven't played around with it but it'll probably be me by 2020 lol...

Offline mamu

  • PV Master
  • *******
  • Joined: Dec 2013
  • Location: IN
  • Posts: 476
  • Karma: +70/-3
  • Gender: Female
Re: You asked mamu it's coming.
« Reply #138 on: February 25, 2015, 09:28:50 PM »
Thanks Visus! 

But it's not my find - I saw a few posts about it on a FB modding group a while back.  ebay has it, but I bought mine directly from aliexpress.

From what I've read, there's a few modders that are testing it with an unreg MOSFET mod and seems to hold up ok with the higher amps - 30A I believe he said and so far aok.  Time will tell though.


Offline Visus

  • PV Master
  • *******
  • Joined: Oct 2013
  • Location: Nexxus
  • Posts: 929
  • Karma: +60/-0
  • Gender: Male
Re: You asked mamu it's coming.
« Reply #139 on: February 26, 2015, 01:26:31 AM »
Ya, it would be nice to know the data sheets on the chips on the board so it could be jumpered a bit so when you fire off the reg the bat v-meter doesn't get all discombobulated lol its gonna freak the frack out if tryin to measure the reg vout and compare against battery lol...  :laughing2:


Definitely a tell what the reg is requesting which is cool 2.. 


On a mech mod it would be fine..  Hope ya figure it out.   

Offline mugenman1111

  • Jr. Member
  • **
  • Joined: Mar 2015
  • Location: iowa
  • Posts: 10
  • Karma: +0/-0
  • Gender: Male
Re: You asked mamu it's coming.
« Reply #140 on: May 17, 2015, 06:36:17 PM »
I am reviving what seems to be a dead thread but I plan on making something either based off claviger's work or independantly sourced.

I was planning on putting everything on Github or Bitbucket some type of git based repo in order to allow people to fork the work and make their own SW mods
as well as hosting PCB plans there.

The only things I have not figured out is how to calculate resistance of atty. I am sure that is simple resistance math with a divider as I have briefly researched.

Hopefully if anyone is still around they would be on board to make a open source mod :-p

Offline PoppaVic60

  • Full Member
  • ***
  • Joined: Aug 2012
  • Location: United States
  • Posts: 31
  • Karma: +5/-0
Re: You asked mamu it's coming.
« Reply #141 on: May 17, 2015, 09:51:55 PM »
I'm still poking around arduino code, (arms still not arrived). I've been toying with the build-system, makefiles - the software.

Much as I like the OKR, I can't seem to find a reasonable alternative to the silly trimpot: I don't want to be installing and replacing pots (btdthts).

A single cell, (automatically covers parallel), and series build would cover the bases - satisfying anyone but the cloudchasers.

Let me know when you have some stuff up, I'll keep tinkering/doodling.

Offline iusedtoanalog

  • Full Member
  • ***
  • Joined: Jan 2012
  • Location: Pennsylvania
  • Posts: 28
  • Karma: +4/-0
Re: You asked mamu it's coming.
« Reply #142 on: May 18, 2015, 12:33:57 PM »
Mugen and Poppa,
I have had much interest in building an open source type of mod based around an arduino controller. This is what a couple of modders have ultimately decided to run with their arduino based mods...... http://www.ericsson.com/ourportfolio/products/bmr450-series?nav=productcategory003%7Cfgb_101_378%7Cfgb_101_379 It is a POL module with 20A capability, but the reason this got the nod over the Raptor, it is controllable via PMBus. Which from what I understand can be done with an arduino......There is a thread here http://ukvapers.org/Thread-VapeShield-VV-VW-VA-Mod-Board where a few guys have been using these for (as best I can tell) at least a few months....The thread is ~138 pages long so it chronicles all the different incarnations from PWM to Raptor with digipot to the use of the erricsson module. They pass around code within the forum and do a generall nice job of keeping everything documented. Give it a look and see if it stirs your imagination......

Offline PoppaVic60

  • Full Member
  • ***
  • Joined: Aug 2012
  • Location: United States
  • Posts: 31
  • Karma: +5/-0
Re: You asked mamu it's coming.
« Reply #143 on: May 18, 2015, 03:25:37 PM »
well, I've been thru 30 of the 130+ posts and seen very little code.

They are also building "hookah" stations - not mods. Although, I definitely approve of the nokia display (cheap and easy). I was interested in the nano once, but the pro mini and pro micro are my smallest at hand.

I did pick up a pair of those specialized displays Mamu had posted about, as well as a 1.8 and 2.4" tft. (The latter is probably good for building someone a desktop picture-frame). Eventually, I'll try one of the small graphic oleds (smaller meaning "than the nokia"), but the price/value of the 5110 nokia display is just too good to miss.

Offline iusedtoanalog

  • Full Member
  • ***
  • Joined: Jan 2012
  • Location: Pennsylvania
  • Posts: 28
  • Karma: +4/-0
Re: You asked mamu it's coming.
« Reply #144 on: May 18, 2015, 03:39:09 PM »
If memory serves me correctly they branch to handheld devices around page 70...... there eventually are two folks who build different versions side by side.  And yes it was a long read but we'll worth it IMO....

Offline PoppaVic60

  • Full Member
  • ***
  • Joined: Aug 2012
  • Location: United States
  • Posts: 31
  • Karma: +5/-0
Re: You asked mamu it's coming.
« Reply #145 on: May 18, 2015, 04:27:58 PM »
still reading... ;-)

BTW, is it just me or are a ton of videos defunct? (could be I lack membership, but I doubt it)
« Last Edit: May 18, 2015, 11:17:15 PM by PoppaVic60 »

Offline mugenman1111

  • Jr. Member
  • **
  • Joined: Mar 2015
  • Location: iowa
  • Posts: 10
  • Karma: +0/-0
  • Gender: Male
Re: You asked mamu it's coming.
« Reply #146 on: May 18, 2015, 10:56:18 PM »
So realistically.

What is more feasable. Using a PMBus compatible buck chip or using a few digipots?

I honestly have just started dabbling in this and from what I can gather Claviger did not have control of the chipset only a readout of what was going on with it and to fire it?

Things I am wanting ..... seeking to accomplish

Live Ohm readout of atty
Wattage display with an easily adjustable method (hopefully with choice of either rotary or buttons both should be simple to setup with some ifdefs)
Voltage display and obviously intelligent voltage based off ohm reading and desired wattage
Current battery status. (not super important but would be nice to read the current value of battery vs expected and set a icon for it)
other things to me seem meaningless.... i vape..alot so i really dont care about how many times i push the button or how long i hold it.


I am going to start piecing together a repo I will make a github page for it and can assign a couple admins to the page to allow others to join and manage projects.

https://github.com/Open-Source-Vape

I will be happy to add members to the project



Offline PoppaVic60

  • Full Member
  • ***
  • Joined: Aug 2012
  • Location: United States
  • Posts: 31
  • Karma: +5/-0
Re: You asked mamu it's coming.
« Reply #147 on: May 18, 2015, 11:10:35 PM »
I'm interested. Although, my Git-Fu and admin-fu are near zero.

I'm up to page 110 over there.. So far, almost nothing they expected from the PMbus seems to do what they wanted.

Further, we have NO digipots in the under 1K zone.. An EE buddy mentioned in a drive-by that a DAC might work, but I never heard more.

Part of the fun for me is the MCU programming, plus tinkering/soldering again. I've little hope of designing a 2-sided PCB myself. Further, I can happily live w/o ever replacing a pot in a mod again..

Back to reading..

Offline mugenman1111

  • Jr. Member
  • **
  • Joined: Mar 2015
  • Location: iowa
  • Posts: 10
  • Karma: +0/-0
  • Gender: Male
Re: You asked mamu it's coming.
« Reply #148 on: May 18, 2015, 11:29:33 PM »
I have extensive PCB design experience...

Git is more to let people who want to mod the code see what's been done and why and allow their own versions/forks/branches on a public space


git is SUPER easy to learn I taught a 12 year old how to do it

Offline PoppaVic60

  • Full Member
  • ***
  • Joined: Aug 2012
  • Location: United States
  • Posts: 31
  • Karma: +5/-0
Re: You asked mamu it's coming.
« Reply #149 on: May 18, 2015, 11:42:48 PM »
The fun part is: I ran a photoplot dept. and did all sorts of layup, laminating, developing and press work.. Fixed inner layers and surface.. Just never understood the DESIGN part of it ;-)

I'd certainly stay with avr's and this thing, at least myself.. I don't feel like suffering mplapX again and I never did bother to order a PICkit 2/3.. (I've still got a half-dozen pic24's here I was unable to program).

In any event, sure. Sign me up. Can't help but learn something. Did you want to start a wiki as well, or create a thread herein, (with Breaktru's permission)?



Breaktru Forum  |  eCigarette Forum  |  Modding  |  Topic: You asked mamu it's coming.
 

gfxgfx
gfx gfx
Powered by MySQL Powered by PHP Valid XHTML 1.0! Valid CSS!