Skip to main content
Topic: Corrections / Suggestions (Read 6759 times) previous topic - next topic

Corrections / Suggestions


C5, C6 should be 15 pF according to page 25 of the data sheet. I realize that the exact capacitance needed varies depending upon the crystal and PCB, but I would expect it to be closer to 15 pF than 27 pF.

Q2-Q8 can only handle 0.1 A each. It would be better to spec a part that could handle at least 0.2 A (25 mA from all 8 segments at once) or even 1.2 A for maximum brightness. A quick search on Mouser shows the Fairchild SS8050CBU for $0.05 each that can take 1.5 A.

The article suggests that R10-R17 could be dropped as low as 22Ω for increased brightness, but that doesn't seem possible if the PIC is limited to 25 mA per I/O pin. Keep in mind that the PIC can only handle 200 mA total on all I/O pins at once, so if all segments were on in a single digit, that might blow the PIC unless you're lucky. (*)

R10-R17 may produce as much as 29 mA if the LED segment operates at its typical forward voltage of 2.1V, depending upon the collector voltage (which I didn't bother to calculate). That much current exceeds the PIC port ratings. The maximum forward voltage is 2.8V, which should be 22 mA. The data sheet doesn't mention a value for the minimum forward voltage, and that unfortunately means the current could exceed 29 mA with only 100Ω.  Maybe you really want 116Ω or greater to be completely protected against PIC burnout.


True Flash chips have a similar 8-pin package as EEPROM chips. If the PIC had PPS, you could use the same socket. Since it doesn't, you might be able to place two sockets on the board and instruct the user to only populate one at a time.

(*) Add bipolar or field effect transistor buffers to the 8 anode port pins so that the brightness can go beyond 25 mA.  With a 1/7 pulse duty cycle or less, you could easily exceed 25 mA, but maybe not the full 150 mA unless you keep the pulse under 10 microseconds.

Others have already suggested SPI instead of I2C for faster destruction!

Re: Corrections / Suggestions

Reply #1
Thanks for looking it over. I'll slip these corrections into the kit:
1. C5, C6 - 15pf
2. Q2-Q8 - higher current (actually already specified in the BOM by current instead of part number)
3. R10-R11 - 116ohm
Got a question? Please ask in the forum for the fastest answers.

Re: Corrections / Suggestions

Reply #2
You're welcome.

By the way, did you calculate the collector voltage for the transistors when they're on? I'm feeling a bit too lazy to do that. The only reason I mention this is that you'd need to subtract the collector voltage from the total as well before calculating resistance.

Now that I think about it, you should also calculate for the worst case supply voltage. The 7805 can deliver as much as 5.25 V, so you might as well use that value. LED current increases with supply voltage, and you want to avoid going over 25 mA, so you should calculate from the maximum possible voltage.

Re: Corrections / Suggestions

Reply #3

How about adding a temperature display to add convincing results

Re: Corrections / Suggestions

Reply #4
picking up on vimark's suggestion: Ian may want to put the chip under test in his oven ... and bake it at about 90-120° while running the test ... or the test may very well reach over 20 million cycles.

Stuff the flash chip on a breakout board together with a temperature sensor and in the stove ... and record the temperature for each cycle. Then even those critics who can't accept that it's rather an educational and fun project than a serious scientific test may calm down ... oh well, they won't but that's part of the fun side of this project ... at least for me.

Once the flash chip is up in smoke and in SiO2 heaven, the Flash_Destroyer could be retired as a thermometer or even as a weather station with a humidity and a preassure sensor attached via the I²C bus ...

Re: Corrections / Suggestions

Reply #5
I've seen temperature sensors (transistors) and chips ... are there humidity and barometric pressure sensor chips or do you have to combine a more physical sensor with an A/D?

Re: Corrections / Suggestions

Reply #6
Yup, there is a number of humidity and MEMS (same technology/principle as used in silicon accelerometers) BAP chips with I²C and other digital interfaces ...

BMP085 by Bosch is an example for a digital pressure sensor with I²C interface.

SHT11 by Sensirion is an example for a digital humidity sensor with I²C interface ... almost ... SHT ... why almost? There is a tweak to get it to work and it can't share an I²C bus with other I²C devices

quoted from Sensirion FAQs
5. Is your bi-directional 2-wire interface compatible with I2C ?

No it is not. The difference is in the start/stop sequence (transition start with SHTxx) and how the SHTxx humidity sensor signals the end of a measurement. SHTxx should not share a I2C bus with I2C components. However sharing of the SCK line and using a dedicated DATA line for the SHTxx humidity sensor will work well.

Re: Corrections / Suggestions

Reply #7
Thanks, iPenguin !!

Re: Corrections / Suggestions

Reply #8
sparkfun caries a lot of sensors ( ) The advantage of them is they also carry the breakout boards.

Not every sensor is I2C but they carry some with serial interface (some have the problem not to be fully compliant with SPI or I2C, like Ipenguin said)