[Solved] They can't all be bricked, can they?

I own 6 Tiny85s and 4 Mega328p. Out of those I can only program 2 Megas and 3 Tinys. The others produce "Invalid device signature" errors.

I have kept everything the same, same programmer, same board, same process on the IDE, just swapping chips. The good chips accept a bootloader and accept "Upload via programmer" blink sketches and work.

So it IS the chips, not the set up.

One Tiny85 I did brick myself by corrupting the bootloader upload due to a conflict on the MOSI pin during flashing crashing the programmer.

The rest however, 2 Tinys and 2 Megas are all brand new chips from RS Components that arrived this way. They are bootloader free of course.

So it's one of a few things.

a) They are DOA bricked.
b) They have suffered ESD damage in my care or in transit / storage.
c) They have some weird funky fuse settings which is preventing them accepting a bootloader.

As I can't really do much about a and b, that leaves c.

For today I am using an UNO and it's ICSP header with a USBAsp to flash these chips. It works for the known working Mega328s and fails for the suspect ones. So the programmer set up is good.

I also made a board with a serial programmer on it and the two working chips accept a program via it. It has a 16Mhz xtal on board. The dead 328s do not work in it either.

My thoughts are pointing to the clocks. Both the UNO and the DIY board have 16Mhz clocks. I have no confirmed that either are working. If the two working chips are running off their internal 8Mhz resonator this might explain things.

How do I check that? Would a 25Mhz/100Msps scope be able to tell the difference between an 8Mhz clock and a 16Mhz?

Basic question....

If a 328p is fused for an internal clock and I connect it to a board with an external clock (of same or different rate) will it appear dead or should it work?

Any other suggestions for ruling out the clock settings?

Also, are there any recommended, commercially available HV programmers, all I can find are really expensive boards or Frankenstein devices on Instructables. I don't want to battle a dodgy HV programmer while diagnosing dodgy chips!


I suppose it skips a lot of testing. If I reset them with an HV and they still don't work they get added to the microwave list.

Is this worth a punt? Looks like it's probably coming from the states though.

I tried probing one of the megas in the UNO board with the ISP programmer attached. The chip has power, but I can't see anything on the ISP SCK or xtal CLK1/CLK2 pins, but I was using a little handheld scope. I could see my 50Hz mains hum noise floor at around 7mV, with a 20Khz modulation, probably from the PC or fluorescent lights, so the scope was working. No determinable digital signals or any life on the pins though.

So the Megas are all sorted. Turns out they were just in blank mode with internal clocks and would probably have worked fine in a breadboard without an xtal. I put them all into Uno mode by default. I'll play more with fuses another time.

I frankensteined this up based on something I found in the link you sent.

Now I need to modify it for the Tinys.

Thanks for the help!