Hackaday Links: March 13, 2016

Way back in 2014, Heathkit was a mystery. We knew someone was trying to revive the brand, but that was about it. Adafruit pulled out all the stops to solve this mystery and came up with nothing. The only clue to the existence of Heathkit was a random person who found a geocache in Brooklyn Bridge Park. Since then, Heathkit has released an odd AM radio kit and $150 antenna. These offerings only present more mysteries, but at least [Paul] was finally rewarded for finding the Heathkit geocache. Heathkit sent [Paul] the AM radio kit. He says it’s neat and well documented.

[David] is doing his masters thesis on, “The motivation of the maker community”. That means empirical data, and that (usually) means surveymonkey. You can take his survey on the motivations of the maker community here.

America’s best loved companies, Verizon and Makerbot, together at last.

The BeagleBone Black was launched in 2013. The BeagleBone Green – a Seeed joint – showed up last August. The BeagleBone Blue, released just a few months ago, is a collaboration between the UCSD engineering department and TI. Now there’s the BeagleBone Enhanced. Yes, they should have picked another color. Perhaps ecru. The BB Enhanced sports one Gigabyte of RAM, Gigabit Ethernet, two USB ports and two USBs via an expansion header, optional serial NOR Flash for a bootloader, optional six-axis gyro, and optional barometer.

Atmel is changing a few AVRs. There is a new die for the ATMega 44, 88, 168, and the ‘Arduino chip’, the ATMega328. Most of the changes are relatively inconsequential – slightly higher current consumption in power save mode – but one of these changes is going to trip up a lot of people. The Device ID, also known as the source of the avrdude: initialization failed, rc=-1 error, has changed on a lot of chips.

Makeit Labs in Nashua, New Hampshire has a problem. They were awarded $250,000 in tax credits to help them move and renovate. Sounds like a very good problem, right? Not so: they need to sell these tax credits before the end of the month, or they lose them. They’re looking for a few businesses in New Hampshire to buy these tax credits. From [Peter Walsh]: “Under the credit program, a typical business donating $10,000 would save $9,000 on their state and federal taxes! That $10,000 donation would cost them only $1006!” Does that make sense? No, it’s taxes, of course not. If you’re a business in New Hampshire and are looking to reduce your tax burden, this is the solution.

So I mentioned MRRF, right? You should go to MRRF. It’s next weekend.

26 thoughts on “Hackaday Links: March 13, 2016

  1. > The Device ID, also known as the source of the avrdude: initialization failed, rc=-1 error, has changed on a lot of chips

    Incorrect.

    The ID returned by debugWIRE has changed to be the same for the 328 and 328P. But the *signature* of these chips has not changed. The signature is what you get fed back over the programming interface with avrdude.

    Since Arduino-land doesn’t use debugWire, this is not a problem, and anyway, the serial bootloader used simply fakes the signature anyway (hard coded values).

  2. Whoa who whoa now, the biggest change for the ATmega chips is actually that the support for full-swing crystal clock has been removed. This means that you HAVE to use the low-power oscillator, and means that they no longer support using any crystal faster than 20MHz. So if you’re currently using a 20MHz crystal in an ATmega application, you have a design change coming down the pipeline. Either slow it down to 16MHz and use the slightly-more-finicky low power crystal osc, or move to an external 20MHz osc.

    1. There are no benefits – all the changes are downgrades. Looks like they are moving the chip to a larger nano-meter manufacturing process.

      Microchip … why compete when you can buy the competition!

      1. How is Microchip related in Atmel design change from 2015?
        The change is perhaps related to cheaper manufacturing process to save up pennies per chip. No benefit to customer, perhaps except for decreased price in future.

      2. Are you really perfectly on that edge where the old one was ok in terms of power, but the new one is not?
        Full swing oscillators are something that will go to the past…
        on top of these, the new one seems to be cheaper and has extra stuff. You did check the datasheet, did you?

        1. Extra stuff? Are you referring to the new mega328PB? Because this is not about the mega328PB (gets more peripherals) but about changing the existing mega48/88/168/328(P), and it does NOT get new features.
          Imagine the confusion of having two chips with the same part number but one works up to 20Mhz and the other only 16Mhz…

        2. They have destroyed the serviceability of whole range of existing products. Changing a specification without changing the product part number is unheard of in this industry and this is the reason why. Some product manufacturers will now refuse to use *any* Atmel products because of this.

      3. Actually they have been moving to smaller geometries over the year. The older fabs are closing and that’s inevitable.
        http://www.avrfreaks.net/forum/atmega328p-and-atxmega-64a3-transistors-technology
        >Early AVRs were on 0.35um, later they used 0.18um then since they closed the Atmel fabs and outsourced I think they’re now using 0.12um
        >Where necessary Atmel suffixed an A onto model numbers (such as mega328P becoming mega328PA) to show that the device was slightly different with new electrical characteristics introduced by the new fab geometry. (in one case (and I think it might have been the 328P to 328PA switch) they took the opportunity to also fix a silicon bug so the signature bytes changed too).

        The smaller voltage swing tells you that. There are more leakages unless you are paying for the more expensive process.

Leave a Reply to RenCancel reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.