Quote from: T3sl4co1l on Today at 07:37:16 amI should be more specific; what do the waveforms show, at the input, current sense and switching nodes?
Is this with just the 100R load, or anything else attached -- i.e. do the load(s) ever have
[sizable] bypass caps of their own and are they connected?
And which signals are being switched when? EN_12V I gather, with 10_12V as respective?
MC34063 has appalling efficiency at low input, making it particularly ill suited here; even if you're within its current capability, you're easily going to exceed the nominal 500mA capacity of a USB port.
Yes, drive resistor must be smaller, that's what, about 4V and 1k so 4mA, and min hFE is 50 so expect it to desat around 200mA? The switch node waveform is diagnostic, it won't be pulling down near GND under load if this is the case. More like 3-5 times more current or 330-220 ohms would be called for, it would seem.
If you have the option to change it, NCP3063 is a roughly upgraded equivalent (but actually has worse efficiency from low input voltages), CS5173 even better (a proper "fast" regulator, reasonable performance, still a bit quirkly), AP3012 maybe (a bit low on current, but probably adequate still), etc. And zillions of TPSxxx from TI, etc. The latter being so much faster, significant cost and space savings can be had on capacitors and inductor.
Tim
I was curious to see what would the output waveform be with the original Rsc (13 Ohm in parallel with the 1 Ohm) as it was when the device arrived. In that case the output is always the same regardless of the state of MOSFET for selecting 10V or 12V. I replaced Rsc with the original one more time to capture the output waveform, then I once again replaced Rsc with 0.33 Ohm and since you said: "Yes, drive resistor must be smaller", I just removed 390 Ohm drive collector resistor and replaced it with 180 Ohm as shown in the MC34063's datasheet and the output is now good :-))
The first picture shows the output when running the test script with the factory components, in the 2nd picture Rsc is 0.33 Ohm and in the last picture drive collector resistor is 180 Ohm.
By looking at the differences it seems to me the 12V would be even better if Rs was 0.22 Ohm. I don't have resistors smaller than 1 Ohm but tomorrow I'll try to solder one more 1 Ohm resistor in parallel to check the waveform with Rs = 0.25 Ohm
But the output is already good enough and the device is passing the test. Although 10V is at the upper margin of tolerance measured as 10.4V and 12V is at lower margin measured as 11.5V
Thank you very much for your help!!To answer your questions:
The load is just a resistor, here is the schematics of the adapter which has to be used with the test script.
The signals could be swithed by using the user interface or by running the test script. For example, the user could press
12V --> 10V or
12V --> 12V button to select the voltage at the 12V output and then the
Enable / Disable 12V buttons could be used to enable the output. The fulltest script does all kinds of tests by itself. When testing 10V and 12V the script first sets the signal at the gate of the MOSFET to configure the 10V at the output of the MC34063, then it enables the 10V/12V output, then it measures the voltage at the voltage divider connected to the 10V/12V output which goes to the AD converter of the µC and after 0.2 seconds it sets the signal at the gate of the MOSFET to configure 12V at the MC34063 output and measures the voltage one more time.
There are two types of the scripts - .blr scripts where you can see end edit the source code and .ipr compiled and encrypted scripts where you can't examine how the script works (at least not in an easy way). Each script consist of 3 parts - the first part defines the user interface (similar to winforms), the second part defines the event handlers and the third part is compiled and uploaded to the AT91SAM7S microcontroller on the main board. Unfortunately I don't have fulltest script in .blr format, I have only encrypted version but I can upload some .blr script if you are interested to see how the device works.
iProg is an excellent device, I bought it because I must read and reprogram the locked MC68HC908AZ60 processor from the Column Integrated Module of Opel Vectra but now that I have it I am using it with everything else because it can program "everything". It can replace every Atmel, PIC and EEPROM programmer I was using before I got it + it can do many more things. For example, it can be used as a vehicle emulator for testing car instrument tables - the device emulates the traffic on the CAN BUS and instrument table then works like it is connected to the car.
I tested it with some old car computers and it works great. I tried to find the PIN code from some car ECUs, tried to clear the crash data from some car airbag modules etc. and everything worked well.
But to read the MC68HC908AZ60 the 10V voltage must work well besides 3.3 and 5V because "high voltage" signal is used to enable monitor mode on that processor.
There are many adapters ready to be used with the device and the scripts to be used with those adapters but the device could be used as the µC development environment too and by making your own adapter and the script you can make anything - even multichannel sound generator :-)
I am not sure if I could replace the MC34063 with some other controller. Maybe if there is one with the same pinout.
BTW, seems as the word "sizable" I marked with red is breaking the quote parser.
ps
I couldn't find the way to type the message outside of the quotation when using full editor. The only way I could type outside of the quotation was by using Toggle View button to disable the WYSIWYG editor.