Tom Higdon's Lab Notebook
Week 01
January 10, 2006 (1 hour):
We got our group together today in class. We threw around a
few potential ideas for projects directly after class.
January 11, 2006 (3 hours):
I spent some time individually on the web researching possible
project ideas. I spent a lot of time looking at archived projects from the past on the 477 website. I also checked out
websites from senior design classes at other universities for possible ideas. I came up with a few ideas that would be
interesting to me personally, including an object-following robot and an eye-tracking controller.
January 13, 2006 (1.5 hours):
Our group met today to discuss project ideas and to complete the
preliminary proposal homework. We came up with four ideas: an e-book reader (iReader), a 3D mouse using a gyroscope, a
hand-held video game, and sound synthesizer/effects box. We wrote all of these up and turned them in.
WEEK 01 SUMMARY
Accomplishments: Formed group, came up with preliminary project ideas.
Weekly Work Total: 5.5 hours
Project Work Total: 5.5 hours
Week 02
January 19, 2006 (0.5 hours):
The group has agreed that the iReader is the best idea. However,
we have determined that the bi-stable display that we originally proposed is going to be too expensive and difficult to get
ahold of to be able to use for our project. Jeremy then proposed that we modularize the display component of the reader,
allowing us to plug in different displays, such as a high-resolution LCD or even a module that interfaces with a CRT. This
would allow us to build a system that could eventually accept a desirable bi-stable display, while still allowing us to
pursue the
e-book reader idea that each of us is very enthusiastic about. We agreed to meet on Sunday to flesh out the idea more. In
the meantime, each of us is going to independently research what is going to be required to implement the idea, including
the proposed interface to a USB thumb drive, the interface between the base module and the display, the interface to the
LCD display unit that we'll probably use as our primary display, and determining which LCD display we will eventually use.
January 22, 2006 (5 hours):
I spent some time today researching the USB interface. A possible IC
we could use to interface with the USB drive is the Cypress Semiconductor
CY7C68013. It's mentioned on the datasheet for this device that one of its intended applications is interfacing with
memory cards, so I think this might be an appropriate IC for our project. I am worried, though, that interfacing with a
USB drive like this is going to be difficult because our device is going to have to act as a "host" according to the USB
protocol, and I'm not sure if the above-mentioned IC can do that.
I looked at some LCD displays also. It seems that the offerings of Optrex might fit well
with our needs. They offer monochrome LCD displays that fit our size requirements, have backlights, and are under $100,
available from Digi-key. Other manufacturers whose parts seem appropriate include Varitronix and Lumex, whose parts are
available on Mouser, at least.
I also found a few LCD interfacing sites:
Site 1 |
Site 2 | Site 3
Met with the rest of the group today in the evening. We talked about more of the specifics of the project, going into more
detail about which components are to be used. For the USB interface, Mat found a USB host IC manufactured by Cypress
Semiconductor that also includes a microcontroller on chip. We are hoping that this microcontroller with a USB interface
will be enough computing power for the base unit of our reader. We also got more specific on the LCD module that will
be used. We figured out that one of the Optrex modules as mentioned above will be suitable for our project. During the
course of this meeting, we decided who will be doing which homeworks and also what each person's role will be in the
project. I am doing the PCB Layout and Safety and Reliability homeworks, and my role in the project is the hardware design.
We completed Homework #2 during the course of the meeting. For class next Tuesday, we decided that we would each continue
to do research on our respective parts of the project to nail down more specifically which components would be used and
how they would interface together.
Links:
Cypress Semiconductor CY7C68013
Optrex
WEEK 02 SUMMARY
Accomplishments: Completed Homework #2. Decided more specifically which microcontrollers and display
units would be used. Caught up with journal entries.
Weekly Work Total: 5.5 hours
Project Work Total: 11 hours
Week 03
January 24, 2006 (1 hour):
I consulted with Brian Moerdyk during lab office hours today to ask
him about our proposed LCD module unit from Optrex. He seemed to think that this display would be suitable and would
include all the hardware that we would need. He was also able to provide me with a sample LCD display that the lab had
that apparently has the same interface as the one that we will be buying, and is approximately the same size as the one we
will be using. We will be able to use this one for prototyping. I left this in our group box in the 477 lab.
January 26, 2006 (4.5 hours):
Our group met briefly after class to discuss some of the features
of the text-rendering engine (word-wrapping, bitmapped vs. vector fonts etc.). Mat also brought up the possibility of using a
ready-made USB thumbdrive reader that he had found, but the group decided that in the interests of power and space-saving, that
we would implement our own using the Cypress IC. We also discussed the division of work between the two microcontrollers in the
display unit and external interface unit.
I did a lot of research today on what the requirements are going
to be for the LCD and LCD backlight, especially with respect to power requirements. It turns out that we need an additional
part for the LCD module in order to get the CCFL backlight to work. Because the CCFL uses a high-voltage, alternating source,
than a low-voltage direct current source that we'll provide, we require a part called an inverter that can convert our power
source into one that the backlight can use.
The LCD module that we'll use is probably going to be this
one. The inverter required for the backlight on this device is the CXA-M10A. This part is $12, while the LCD module itself is around $85, making the total
for the display itself around $100. The backlight draws 230 mA at 5V, the LCD module itself draws 20 mA at 5V. Assuming that
each of the two microcontrollers draws around 20 mA apiece (a conservative estimate according to Atmel datasheets), this brings the current draw total to around 300 mA. I found 2000 mAh
AA 4-cell NiMh battery packs on Digi-key for $17. This means that our reader would be able to stay powered on with the backlight
for 6 hours, which isn't too bad. Without the backlight, the total is closer to 30 hours, which is even better. I'm not sure how
much we can assume the backlight will be on. This LCD is supposed to be transflective, meaning the backlight is only used when
needed, so hopefully under good lighting conditions, the backlight will not be necessary. I also looked into the possibility of
using an LED backlight, which might consume less power, but I wasn't able to find a readily available LCD module that was large
enough that had one. If we are using NiMh batteries, a
recharging circuit is going to be necessary, the preliminary design of which I'll try to get into next.
Links:
Proposed display
CXA-M10A
WEEK 03 SUMMARY
Accomplishments: Determined LCD module to be used, determined rough current requirements.
Weekly Work Total: 5.5 hours
Project Work Total: 16.5 hours
Week 04
January 31, 2006 (2 hours):
I did some research on power circuits and batteries that we can use. It
looks like the best thing to use will be NiMh batteries, which will probably be 4 1.2V cells, providing 4.8V.
February 1, 2006 (2 hours):
The group met today and discussed part selection, especially how the
software design will affect part selection.
February 3, 2006 (9 hours):
I spent a lot of time today researching possible power supply circuits.
I've come up with a workable solution that will hopefully allow us to recharge some Lithium Ion batteries in-circuit. Linear
Technologies' LT1980 provides
power regulation and battery charging in the same IC, although a decent amount of discrete components are required for the
circuit. Nevertheless, the part looks relatively easy to use and I think we will be able to get some samples, which I've already
requested.
I've spent a lot of time looking at LCD module datasheets, and I've come to the conclusion that the Optrex model I mentioned
above is not going to be sufficient for our purposes. Contrary to what I thought before, this LCD screen does not include a
controller to make it easy to interface with. It appears that our group has two choices at this point. We can either use the
display that we obtained in the lab (which appears to be missing parts if I'm reading the data sheet correctly), or we can
purchase an LCD screen and a controller separately. A 320x200 LCD screen sans controller is approximately
$50. The controller
board,
which can be purchased on thiswebsite, is about $50.
We could however, build this part by itself on our own board, which means we'd have to include the
$15 part with 32k of SRAM and
a 10 MHz crystal, which wouldn't be all that bad, but would still add to the complexity, which is getting bad enough as it is. If we still wish to use the CCFL backlight,
that part is another $5 or so. The controller, though, is of the same type as the one used in the eval
kit we have in lab, so at least we would be able to prototype on that.
Links:
LT1980
$50 display
Controller Board
SED1335
WEEK 04 SUMMARY
Accomplishments: Had revelation that display didn't include controller. Preliminary power supply design research was conducted. Fleshed out software a little bit more.
Weekly Work Total: 13 hours
Project Work Total: 29.5 hours
Week 05
February 5, 2006 (1.5 hours):
The team met today to discuss the upcoming homeworks and more
component selection. We pretty much decided which microcontroller we're going to use for the display unit. We also figured
out who is doing what for the presentation on Wednesday.
February 5, 2006 (5 hours):
Some notes on the SED1335 (the LCD controller):
- It can take a crystal or an external oscillator as a clock.
- The controller is intended to be mapped to the memory space of the controlling microcontroller.
- Max clock is 10 MHz
- Operating current is typically 11 mA under conditions that are less severe than what we'll subject it to. I'd budget 20 mA.
- Datasheet has two electrical characteristics tables. It's possible that the chip may operate on 3.3V logic voltage(!)
- If Vdd is less than 4.5V (it probably should be 3.3) the max clock rate of the chip is 8 MHz.
- The controller provides a facility to scroll through the memory using a single command.
- Mat: the SRAM needed is 8k x 8.
I also did a very significant amount of research into sources for LCD displays I managed to dig up a site that had a display
for sale for only $10 that will be usable for our design. I went ahead and ordered it.
February 6, 2006 (2 hours):
More research today into what is going to be required in terms of
power by our LCD display. I sent out an email to the site I bought the display from to see if they had any more detail about
it. I also sent an email to the Chinese manufacturer of the display in hopes of getting a more detailed datasheet, but I'm
afraid I won't get any response. Regardless, I'm fairly confident we'll be able to interface with this display.
February 8, 2006 (2 hours):
I spent quite a bit of time today poking around Linear Technology's
website today trying to find power management IC's that we can use in our design. The features that we require are
battery-recharging via AC wall adapter, regulation to 3.3V, 5V, and -22V, and biasing for the LED backlight. I have been
attempting to find a part that encompasses as much of this functionality as possible, but I have been unsuccessful so far.
February 9, 2006 (7 hours):
After class, Danny, Tryba, and I went to the 477 lab to get a demo from
Tryba concerning the software Mat and I will be using to the schematic capture and PCB layout. After that, I spent some more
time researching power ICs that we can use in our device. After perusing the Dallas Semiconductor/Maxim website for awhile, I
came up with a few possible ICs for our project. Later in the day, I solidified the possibilities into a mostly-finished
block diagram that nails down the entire power supply subsystem except for the component that will control the input to the
LED backlight, which is an unknown until the display arrives in the mail.
Links:
S1D13700 (SED1335 replacement)
LTC4150
LTC1129-3.3
MAX1677
Power Supply Block Diagram
February 10, 2006 (3 hours):
I pored over the datasheets for those power supply components in order
to get an idea of how their application circuits worked. I feel like I have a pretty good handle on their operation now, and I
have written up a list of the components that we will need for the power supply circuit as it stands now. We will need a few more
parts I'm sure, once we know what the LED backlight on the display is like, which should arrive in the mail any day now. Tomorrow,
I'm going to try to get the whole power supply circuit down in Orcad.
Links:
Discrete Parts List
WEEK 05 SUMMARY
Accomplishments: Found a new source for cheap LCD screen. Found a new part for the LCD controller that is available from
Digikey. Ordered a lot of parts from Digikey. Requested samples of the majority of the power supply ICs from the
manufacturers. Designed the power supply schematic.
Weekly Work Total: 20.5 hours
Project Work Total: 50 hours
Week 06
February 12, 2006 (8 hours):
Mat and I spent some time in the lab today familiarizing ourselves with the
tools we are going to use to develop the schematic and PCB layout for our project. We managed to figure out how to use the Place
Database Part feature of Orcad Capture, which made getting parts for the schematic much much easier. I got all of the power supply
circuitry into our schematic except for the portion that concerns the LED backlight. The schematic now is almost done, there just
need to be a few things added before it is completely done. Namely, we need to figure out what needs to be done to notate headers
in the schematic, and we have some questions for the TAs about crystal oscillator selection.
February 13, 2006 (4 hours):
The LCD came today. I found the datasheet for it. It's a Hitachi
LMG7521.
I spent some time in the lab figuring out how to work the LED backlight. It's fairly simple, I believe me may just be able to run it
off of the 5V supply we will already have.
Links:
LMG7520 Datasheet
Pertinent LCD/Controller Interface Info
February 14, 2006 (4 hours):
I found a connector for the LCD. I had to manually ohm out the connector from our
LCD to figure out which pins on the cable corresponded with the inputs to the LCD, which took quite a bit of time. I updated the schematic
to show the LCD
connector, which includes pins for the backlight. The backlight, which it turns out is simply 4 white LEDs in parallel, will
be powered by the 5V input from the EIU. I also did some updates to the battery charging and battery gauge circuit. It turns out it
is going to be necessary to buy special 4-wire current sensing resistors for both of these parts, which are ~$3.50 from Mouser.
Capture has a Bill of Materials feature, so once we finalize the schematic, we should print that out, check it over, and order the
remainder of the parts.
Links:
LCD Connector - FH12A-30S-0.5SH
February 15, 2006 (3 hours):
I called Linear Technology today, found out that the sample of the
LTC1980 I ordered 2 weeks ago just shipped yesterday, so it should be coming in the mail within a few days. Mat and I spent
some time in the lab working on the schematic and doing part selection. Mat found a wall wart that we will able to use, so
we spent some time testing that to make sure it would work correctly. We also were able to take our current schematic and
create a netlist that we could import into Layout. When we left the lab, we had a board up in Layout that we should be able
to begin to place the parts on tomorrow.
February 16, 2006 (8.5 hours):
I touched up the schematic today, did some more detailed part selection
for some FETS, finding distributors etc. Tried my hand at Layout, made a few footprints. Making the footprints for these parts is going
to be a pretty big undertaking.
February 17, 2006 (7.5 hours):
I found distributors for all of the FETs, almost completely finished the schematic.
I had to make a change in the power distribution. Now the power rail going between the two boards will be 3.3V instead of 5V. This was necessary
because I found out that the MAX1677 is not capable of outputting 3.3V from a 5V supply because it has only a step-up converter in it. I made this
change to the schematic.
WEEK 06 SUMMARY
Accomplishments: Did the schematic design. Got the LCD and figured out the interface. Finished selection of most of the parts.
Minor redesign of the power supply subsystem.
Weekly Work Total: 35 hours
Project Work Total: 85 hours
Week 07
February 19, 2006 (12 hours):
Mat and I completely finished the part selection and put in an order to Mouser. It turns
out that the rest of our discrete components are going to run us around $35 + shipping, which is actually a little bit better than I thought it would
be. Mat and I also finished finding and creating the rest of the PCB footprints needed to begin the layout. I spent a few hours doing some laying
out of the board. It looks like we might be a little bit cramped on the EIU side, but we'll see how it goes. I printed out a 1:1 rendering of what the
board looks like at this point and left it in our box.
February 20, 2006 (7.5 hours):
Spent some time in the lab today on part selection and PCB
Layout. I came to the unfortunate realization that many of the electrolytic capacitors that we ordered yesterday are of the
wrong type. We ordered aluminum electrolytics capacitors for some of the filter capacitors, but further research indicates
that tantalum electrolytics, with their lower ESL, are more suitable. We made a list of capacitors to replace the ones that
we had selected before, and added those to the list, along with the FPC connector which we forgot to order yesterday.
Parts list: Discrete parts list
February 21, 2006 (6 hours>:
Mat and I seriously got started on the layout of the DU. At first
we were really frustrated with the tools, but eventually we kind of got the hang of it. Mat logged some serious hours getting
the DU ready for the presentation tomorrow.
February 22, 2006 (11 hours):
I stayed up all night working on the routing for the power supply on
the EIU.
February 23, 2006 (15 hours):
I worked on the report, picked up where Mat left off on the routing on
the EIU. I fixed all of the design rule errors on both of the boards. I worked through the night and finished everything 45
minutes before the deadline.
WEEK 07 SUMMARY
Accomplishments: Completed part selection, schematic. Ordered discrete parts.
Completed PCB layout. Received the majority of the parts.
Weekly Work Total: 51.5 hours
Project Work Total: 136.5 hours
Week 08
February 28, 2006 (10 hours):
The group met today to work on the presentation for the design review tomorrow.
Although we only intended to spend a few hours putting together some slides and rehearsing the presentation, further examination revealed that
some work needed to be done on the PCB layout and the schematic before we could begin presenting tomorrow. Also, the generation of the slides took
a little longer than we anticipated. We managed to make some significant changes that improved the DU PCB by reorienting the microcontroller and LCD
controller to make the traces more direct. I also adjusted the EIU power PCB to put the traces for the MAX1677 into a star ground configuration
as suggested by the datasheet. We got everything together, and then ran through the whole presentation a couple of times. By then, we had just enough
time to get a couple hours of sleep.
March 1, 2006 (1 hour):
Mat and I figured out which parts we had received and ordered the remainder of
the parts that we had yet to order, including the FPC connector, current-sensing resistor, and the battery.
March 3, 2006 (2 hours):
I spent some time and made the changes to the schematic and PCB layout that
were suggested to us during the design review. I changed clear signal on the battery gauge IC to go to the microcontroller
instead of being wired directly to the interrupt line, because on further examination of the data sheet for that part, I decided
the course staff was correct in their suspicion about the behavior of this component. I also placed a copper pour connected to the
ground net over the power lines on the EIU to reduce noise. I fixed the remainder of the design rule errors in the PCB. I also
changed the via drill hole size to 20, as suggested. I made once copper pour for the ground, but I still want to make another one
over the remainder of the -22v power line on the EIU, but for some reason I couldn't figure out how to make a second one. The
issue of route spacing is something that needs to be resolved, but I want to consult with one of the TAs before I make any changes
on that front, so it will be Tuesday most likely before anything happens on that front. The PCB is basically done now though. When
the remainder of the parts come in next week, we'll be ready to put them all on the board to complete the homework assignment.
WEEK 08 SUMMARY
Accomplishments: Design review presentation. Mostly completed the PCB. Ordered the remainder of the parts.
Weekly Work Total: 13 hours
Project Work Total: 149.5 hours
Week 09
March 6, 2006 (4 hours):
The remainder of the discrete components arrived in the mail from Mouser today.
Mat and I spent some time in the lab today. I verified that all of the footprints we had on the PCB matched the parts that we had
ordered. A few minor changes had to be made to the footprints, mostly just lengthening pins. We had to reorient the FPC connector,
which required some rerouting on Mat's part. Tomorrow during office hours we'll talk to the TAs about the trace/pad spacing on the
layout and hopefully get the component demonstration out of the way, allowing us to take the rest of the week off prior to Spring
Break.
March 8, 2006 (2.5 hours):
I made a few minor cleanups to the PCB. I got the PCB prepared for submission,
but there still a lot of non-show-stopping problems with the board that the production check found. The majority of these are due to
the fact that all of the pads on the layout have no spacing between the soldermask and the copper layer, which is apparently the default
in the Layout libraries, yet 4pcb.com claims that there should be 6 mils between them. Their software can apparently automatically correct
this problem, but I am going to consult with the TAs before I submit what we have.
March 9, 2006 (0.5 hours):
I came into lab, talked with Brian about the PCB submission. The errors I
mentioned in the last entry are apparently fine. I submitted the PCB to Chuck.
WEEK 09 SUMMARY
Accomplishments: Received all of the parts, did the component demonsration. Prepared the PCB for submission. Submitted the PCB.
Weekly Work Total: 7.0 hours
Project Work Total: 156.5 hours
Week 10
March 24, 2006 (5 hours):
After the rest of the group populated all of the parts on the power supply, I was in the lab
today working with Tryba to get the power supply circuit working. We spent most of our time on the battery-charging circuit thta is centered around
the LTC1980. Unfortunately, due to a misunderstanding about the power supply outputs, we blew up one of our bypass capacitors, and simultaneously
destroyed most of the semiconductor parts on the board. After a frustrating several hours, we adjourned without having gotten any of the functionality
of the power supply circuit working.
March 25, 2006 (8 hours):
Tryba and I spent a lot of time in the lab today working on getting the LTC1980 and MAX1677 circuits
working. After spending a lot of time ohming out traces and looking for damaged parts, we ended up replacing all of the semiconductor parts on the board
that were presumably damaged by the accidental application of a drastic overvoltage yesterday. We made a minor change to the MAX1677 circuit that required
fly-wiring on the PCB. Despite this change, and replacing the MAX1677 part, we were unable to get any functionality of the circuit working, and we have
concluded that our remaining MAX parts are damaged. We have ordered new ones, which will hopefully arrive soon. As for the 1980 circuit, after replacing
a few parts, we were able to get basic functionality of the built-in LDO voltage converter operation from the wall-wart power supply to work. We are
seeing 3.1V on the output, which is exactly what is expected. Unfortunately, this is but a small part of the circuit's operation.
March 26, 2006 (7 hours):
Tryba and I spent more time today working on getting the 1980 circuit to work. After making a change to
the PCB to allow operation from two series lithium-ion batteries, we were able to successfully hook up the batteries to the circuit and witness the IC exhibit
switching at 300 kHz. We were unable to determine if battery-charging behavior was taking place, but we think we are fairly close to getting complete
functionality out of the 1980 circuit. We also ordered some extra 1980 IC's from a distributor, in case it turns out the one we have is damaged or that we damage
it in the future. We have also decided to order another EIU PCB at least, in order to incorporate changes we have had to make to the power supply circuit,
in addition the changes that we have made to the Cypress chip circuit. We intend to put the order out by Tuesday, hopefully to get back by the middle of next week.
Here's how the EIU PCB looks now:
WEEK 10 SUMMARY
Accomplishments: Debugged power supply circuitry. Made fixes/changes to schematic/PCB. Ordered more parts.
Weekly Work Total: 20.0 hours
Project Work Total: 176.5 hours
Week 11
March 27, 2006 (6 hours):
Today, Tryba and I continued to try to get the power circuit working. We believe that we were able to get some sort of battery-charging action
out of the circuit, but we are still unsure whether the chip we have on the board right now is damaged or not. We attempted to charge the battery, and got a slowly
increasing voltage from about 6.5V to about 6.9V, which isn't really right, but I think it's close.
March 28, 2006 (8 hours):
I figured out today that the inductors I used on the MAX1677 circuit were the incorrect type (ferrite beads instead of ferrite cores). I selected
a new part for it that will be suitable, and put it on the list of parts to be ordered for the incoming new board. Because the footprints for the new parts are significantly larger than the old
ones, I had to do some significant reorganization of the PCB layout for the MAX1677 circuit. I think the layout this time is significantly better, because of the insight gained in debugging the
circuit before. I know more about the criticality of placing certain components like bypass capacitors and switching inductors as close to their respective pins as possible. As always, the
layout took a significant amount of time. I also tried to place some more descriptive markings on the silkscreen, such as negative/positive markings on some of the connector jumpers. I got the PCB
to a state that's ready to be submitted, but Mat still may add another part to the schematic, and he'll also have to actually do the act of submission, because I don't have time as my flight to Seattle
leaves tomorrow. I will be gone until late Sunday on a visit to the University of Washington.
New Inductor
WEEK 11 SUMMARY
Accomplishments: Debugged LT1980 circuit. Made more fixes/changes to schematic/PCB. Ordered the remainder of the parts we'll
need in order to repopulate the board when we get it back.
Weekly Work Total: 14.0 hours
Project Work Total: 190.5 hours
Week 12
April 3, 2006 (3 hours):
We still haven't gotten the board back. Today I spent some time working on the LCD backlight circuit. I made a mistake on the schematic that required
me to cut a couple of traces and
do some fly-wiring. I also managed to locate a p-channel MOSFET at the 207 lab shop that is suitable for use in our circuit. I also had to fly-wire an apparently broken trace on the flex-PCB on the
LCD. Once I diagnosed and fixed these problems, the LED backlight circuit seems to work correctly. I have also posted the PDFs of what the new top and bottom copper of the
EIU look like now. I will be gone until from Thursday until Wednesday of next week on visits to UT Austin and Georgia Tech.
EIU Top Copper
EIU Bottom Copper
WEEK 12 SUMMARY
Accomplishments: Selected part for the LED backlight MOSFET. Got the LED backlight working.
Weekly Work Total: 3.0 hours
Project Work Total: 193.5 hours
Week 13
April 12, 2006 (8 hours):
I got back from my grad school visits today. In my absence, we received the new EIU PCB
and Mat populated the power supply portion of the board. The MAX1677 portion of the circuit appears to be working perfectly. Unfortunately, the
LTC1980 portion of the circuit is not working correctly. It is giving the correct regulated 3.3V output, but the battery-charging portion of the
circuit does not appear to be working correctly. About 10 mA of current will go into the battery, but the circuit is set up to put about
1A of current into the battery. As I described in a previous entry, the battery voltage increases slowly to about 6.9V, but then stops. The
output from the timer pin does not seem to be correct, either. It only shows the correct output when the power supply is plugged into the
circuit, and not when the battery is plugged in as it should. I had to spend most of my day today, however, working on the FMECA worksheet for
our group's presentation in class tomorrow. I read the reference article
that Dr. Johnson used for his
presentation in class. I was able to finish the worksheet and prepare a Powerpoint presentation for our group to present in class tomorrow.
I also spent some time assisting Mat in debugging the power circuit.
April 13, 2006 (6 hours):
Today I wrote the safety and reliability report which is due tomorrow. I was able to
successfully use the Military Handbook 217F as a reference to complete the report and estimate the reliability of various components that
are used in our design. Here's a link to the report on our website:
Homework 11
April 14, 2006 (8 hours):
Today was spent attempting to debug the 1980 circuit. I haven't made much progress.
I verified the integrity of the traces on the PCB and verified the resistor component values. I also verified that my circuit is the same as the one
on the datasheet. I verified that the resistors that have been populated are the same ones that are on the schematic. Unfortunately, it still doesn't
work correctly.
April 15, 2006 (8 hours):
Another day was spent debugging the LTC1980 circuit. I spent most of today poring over the
datasheet, trying to get some insight as to how the circuit works so I may have some more insight debugging it. I figured out some points on the
circuit that I can test, but they all seem to check out ok. I'm still at a loss as to what the problem could be.
WEEK 13 SUMMARY
Accomplishments: Wrote safety and reliability report, including FMECA worksheet. LTC1980 circuit debugging.
Weekly Work Total: 30 hours
Project Work Total: 223.5 hours
Week 14
April 16, 2006 (9 hours):
More 1980 circuit debugging. It appears that the circuit is working more correctly when
it is in 1-battery mode rather than 2-battery mode, but unfortunately the battery we have is 2 cells. The timer pin output looks ok only when
the circuit is in 1-battery mode. Mat and I went to Interstate Batteries in West Lafayette to buy a single cell lithium ion battery. I spent some
time trying to solder lead wires to it, but the solder wouldn't stick to the battery contacts. I eventually ended up simply wrapping the battery
with electrical tape to hold wires against the contacts, and this seems to work pretty solidly. Unfortunately, it seems that the battery charging
circuit still does not work correctly, even in one-battery mode. The battery will only draw about 70 mA of current, still incorrect.
April 17, 2006 (9 hours):
More time was spent today trying to get the 1980 circuit to work correctly. I feel like
I have a pretty good grip on how the circuit is supposed to work, and it appears that the voltage regulation when the battery only is hooked up
is working correctly, so it mystifies me why the battery-charging portion of the circuit continues to elude me. I examined almost every relevant
signal in the circuit with the oscilloscope, but I was unable to find a thing wrong. I talked to Nick the TA today and he said that if I send him an email describing the problem,
he will try to help me out. I additionally decided to call Linear Technology's technical support line to try to get help. When I called the line,
however, I was told to send an email describing the problem and to give as much information as possible. So I drafted an email to send to the
gentleman at LT, and also sent this along to Nick, with the hope of getting outside help on the problems I'm having with this circuit.
Links:
Email that I sent
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6
Figure 7
Figure 8
Figure 9
April 18, 2006 (6 hours):
I spent much of today trying out some things that Brian suggested might be the problem
with the 1980 circuit. I tried replacing the timer capacitor, shorting out the battery-gauge sense resistor, and cleaning off the flux from
the board. None of these worked. I sort of made a breakthrough today, though. It turns out that the one-cell battery that we have been using
is a non-rechargeable lithium battery, not a rechargeable lithium ion battery. This at least partly explains why we have not been able to get
the battery to accept much charge. I still want to find a one-cell lithium ion battery because I still believe that the circuit might work
correctly in one-cell mode. Mat has said that when he is Indy tomorrow for an interview he will go by Fry's and a hobby shop to see if he can
find a one cell lithium ion battery that will fit our needs.
April 19, 2006 (2 hours):
We had our TCSP presentation today. In the meantime while I have been slaving away
getting our power circuit to work correctly, the rest of the group has been able to get the Cypress chip up and running, including USB support.
This is a major breakthrough, but it also means that the last remaining major hurdle for the project is getting the battery-charging circuit to
work correctly. I spent a couple of hours in the lab working with the 1980 circuit, but was unable to make any major headway.
April 20, 2006 (8 hours):
Mat brought back the new battery today. The battery he found is made for digital cameras
and has two different sets of terminal for charging and for use. When we got it, the only terminals that showed voltage were the charging
terminals, so these are the ones that I soldered wires to. I spent a fairly significant amount of time on the internet attempting to verify that
the particular battery we had was suitable and had the correct voltage. I have theorized that the battery is intended for use in cameras that need
3V, and that the non-charging terminals are connected to a built-in switching regulator that converts the 3.6V output from the battery cell and
converts it to the 3V within the battery pack. So if we simply use the voltage output from the charging terminals, it should be fine.
However, I connected this battery up to our circuit and it still did not work correctly. Oddly enough, though, varying the voltage coming in from
the power supply causes the current consumption to spike considerably at certain voltage points, although the behavior is very erratic. Danny and I
spent a considerable amount of time debugging the circuit using multiple oscilloscope probes and we were able to figure out that during the battery
charging, the switching regulator will attempt to start, but it will stop immediately as the voltage at the output reaches the undervoltage
threshold. We were able to trace this problem to the sense resistor, where we found that the sensing leads were transposed, meaning the sense lead
was grounded and the ground lead was connected to the sense pin on the IC. Danny and I are fairly confident that we have found our problem with this
circuit but we decided to save the fixing for tomorrow, because I have some other things I need to do tonight.
April 21, 2006 (4 hours):
I got it working! By cutting the traces going to the sense resistor and fly-wiring the leads
to the correct pins on the 1980, I was able to correct the problem with the PCB we had identified the previous night. The circuit now draws 1.1A when
battery-charging, which is exactly what I would expect out of that circuit. This is a major breakthrough with this circuit, and the last remaining major
hurdle for our project, because battery-charging is one of our PSSCs. I am confident that the power circuit will work correctly now.
April 22, 2006 (7 hours):
I came in today to test the power circuit and verify that it is working correctly. I found
that the resistors that are used to control the current going to the battery apparently did not control the current. I found that the reason for
this is that power supply we were using apparently is not able to supply the amounts of current that we need to power the battery. When I aggresively
chose resistors to limit the current below what the power supply could supply, I got correct behavior. I then decided to test the functionality of the
LTC4150, which is our fuel gauge IC. In the process of testing this circuit, I shorted out the leads from the voltage regulator and burned up one of
the IC's on the board. I had to salvage the IC from our previous EIU board. One of the pads on the part's footprint was lost, so I had to fly-wire
one of the pins on the part. When I did all of that, I powered it up and it still worked. I was able to verify the functionality of the fuel gauge
and it seems to be working correctly. I still need to modify the board to put pull-up resistors on the two input pins to this chip.
WEEK 14 SUMMARY
Accomplishments: Debugged LTC1980 circuit. Sought outside help to get circuit working. Modified PCB to get it working correctly. Completed
population of working power circuit.
Verified functionality of the LTC4150.
Weekly Work Total: 45 hours
Project Work Total: 265.5 hours
Week 15
April 23, 2006 (5 hours):
Today I replaced the SI3443DV discrete MOSFET on the power supply board. This chip has not been
populated while we had been debugging the power supply because we did not want strange voltages going to the other part of the power supply circuit,
the MAX1677. Now that the LTC1980 circuit appears to be functioning correctly, we populated this MOSFET and verified that the part works in the
intended fashion. Mat and I worked together and selected a resistor value for a resistor divider that determines the output voltage for this MOSFET.
In the datasheet for the LTC1980, this low dropout regulator voltage is set to be 3.1V, but because we want a lower voltage drop across this LDO
MOSFET in order to minimize the heat and power loss across this component, we decided to set this voltage to 3.2V. Because the swtiching regulator voltage output
is at 3.3V, this means the voltage drop across this part will only be 0.1V instead of 0.2V, effectively halving the heat dissipation and power loss
across this chip at a given current level.
Mat and I also spent some time in the lab today populating the rest of the new EIU board. After
re-verifying the correct operation of all parts of the power supply, we decided to go ahead and put on the rest of the parts on the board,
including our other Cypress EZ-host chip. After doing this, we spent some time verifying that the functionality of this circuit is the same as
what it was on the other EIU, and we successfully able to get full functionality, including communication with the DU.
April 24, 2006 (3 hours):
Jeremy and I spent some time in the lab today getting the signals to come out of the
LTC4150. Even though I had verified correct operation prior to populating the Cypress chip, we were not immediately able to get the correct
interrupts out of the LTC4150. After trying some things and looking at the layout, we determined that the VDD and GND pins needed to have
bypass capacitor directly between them, rather than having it go from the VDD to another ground trace that was much further away electrically.
This approach worked, and we were successfully able to get the correct interrupts out of the chip. Jeremy did some programming to get a nice-looking
fuel gauge and fuel meter on the LCD display
April 25, 2006 (7 hours):
Even with the change in the voltage drop across the LDO MOSFET that I was talking about yesterday,
I am still worried about heat dissipation across this component. I decided to run some tests on the part to see what its limits were. I also tried
some heat sinking techniques to increase the heat dissipation capacity of the component. Reading the datasheet
for this part, it says that if there is a 1.0 square inch pour beneath the part that is connected to the drain pins, that the part is able to sink
1.6W. We need the part to supply about 400 mA of current with as large of a voltage drop across it as possible in order to accomodate AC adapters with
different voltage ratings. Because we do not have any sort of copper pour beneath the part on the PCB due to me not anticipating this problem, I made
some attempts today to fashion a makeshift heatsink. I soldered some copper ribbon to the drain pins of the MOSFET and felt how hot the part got
with this measure in place compared to how hot it got with the same voltage across it without the heat sinking. I was not able to detect much if any
difference between the two. I decided to temporarily give up on this approach and insteaded explored the possibility of using a larger MOSFET for this
application. I acquired an IRF9530 (datasheet) from
the EE402 instrument room and attempted to use this part to satisfy our needs. Unfortunately, because of the difference in the turn-on voltage of this
part compared to the SI3443DV, the attempt did not work.
Later in the day, the rest of the group came in as I was attempting to replace the original SI3443DV. Once this part was populated, the LTC1980
circuit appeared to be working incorrectly. We were unable to determine a cause for the problem, but we eventually narrowed it down to it being a
problem with the LTC1980 itself, because I possibly had connected the power rails backward momentarily at one point. Once we replaced the chip, the circuit
still did not appears to be working correctly. We spent a couple of hours going through debugging procedures with a DMM, but the voltages read
still appeared to be incorrect. Someone suggested that the DMM itself might be the problem because it was showing the low-battery signal, so we
tried another DMM. This turned out to be the problem. The LTC1980 circuit appears to work correctly now, and we adjourned for the day having
accomplished basically nothing.
April 27, 2006 (1.5 hours):
Mat, Danny, and I ran into each other at the lab today and finished up the packaging for the
project for the presentation tomorrow. This mainly consisted of coming up with a workable solution for the hinges and implementing it. We hot-glued
some hinges together and glued them onto the packaging.
April 29, 2006 (1.5 hours):
The group met today and came up with a division of labor for the rest of the paperwork involved
in finishing the class.
April 30, 2006 (5 hours):
I worked individually to clean up the homeworks I have done and write my individual contributions
for the final report.
WEEK 15 SUMMARY
Accomplishments: Testing possibilities for heat-sinking the LDO regulator. Power supply debugging. Finished up packaging.
Weekly Work Total: 18 hours
Project Work Total: 283.5 hours
