The Amp Hour Electronics Podcast

A weekly show about the trends in the electronic industry.

You are here: Home / Radio Show / #558 – Toasted Marshmallow Connectors

#558 – Toasted Marshmallow Connectors

Play

Podcast: Play in new window | Download

Subscribe: Apple Podcasts | RSS

Thanks to our sponsor Mouser Electronics. This week Chris and Paul discuss Industrial Automation during the ad break.

Comments

  1. A customer of ours is just trying to extend the lifetime of a legacy product with a lifetime buy of an obsolete PowerPC processor they use in a module. Turns out they were not the only ones on the hunt and this chip is no more. They were now offered the security-enabled variant where apparently a few hundred are still somewhere on a shelf in some warehouse. Thankfully (hopefully?) it will be a drop-in replacement with no change in hardware or software. Famous last words, I know. “Security-enabled” always raises some back hair with me.

  2. We bought the T962A years back. First run, it smoked and stunk like hell. Fifth run, the fuse holder got so hot it melted.

    There was no cold junction compensation and there was no ADC front end.

    Although there’s a crystal, the timing was about 20% off to start and would slow down during the profile due to inefficient code. Each run would heat the controller, and would occationally cause the thing to lock up with the heating elements hard ON. Even when it wasn’t locked up, the hot melt would MELT and the LCD would slowly disappear.

    After the thing caught fire, we took it apart and reworked heat shielding, fixed the dangerous wiring and made our own external controller. The damn blower kicked in at the wrong time during one very expensive UVC array build and we stopped using it.

    We took its controller and put it on a $20 toaster oven. Works like a dream.

    If anyone could recommend a REASONABLY priced professional oven with an exhaust vent please tell. We still can’t find a desktop oven for less than $1k. There’s no reason they can’t sell a good one for $500

    • No, LTZ1000 is still king until ADI can figure out how to build high performance analog. The ADR1000 is proving quite a drifter in the lab (worse drift than LTZ in initial tests, only marginal noise improvement)…but at least it uses more power. The ADR1001 is even more ridiculous in an smt package, that way you can get great humidity drift..

  4. (Same post with a few typos corrected:)

    I bought even smaller T962 model and went on a safari to look at modifications. There are quite a few, some are really sophisticated such as modifying and reflashing the controller that the oven came with, adding an temp reference and improving software, without having to add/replace any controller board.
    But the one that blew my mind is the work of Jerry Walker.
    Walker apparently work with precision electronics and its manufacturing at his day job, and recreates boards for vintage electronics in his spare time. He has access to professional equipment to do things like CALIBRATE a thermcoupler…
    Walker bought a T962A and did 6 part series on how he finds out what is worth fixing. He sets a goal at the first video and creates a laundry list of fixes to consider.
    In the 6th part, he concludes that a lot of the planned fixes was unnecessary.
    https://youtu.be/4_7S4vderb4
    His major fix was putting in a S-shaped thread (not a propeller that would create draft, more of an air cocktail stir) spinning at 300 rpm in the ceiling of the oven chamber, creating enough turbulence to get an even temperature distribution, eliminating the need for things like additional heaters.
    Walkers whole video series on T962A system characterization and modifications (for those playing along at home):https://youtube.com/playlist?list=PLzvLbUxGuZ-zK_24L4maC1OOX__b6bArA

  5. I haven’t heard of 2G tip-over before but I have seen this before.
    We once had a product with an Ethernet port that would occasionally fail.
    The SW guy looking at the issue found that the problem never happened if you turned the unit upside down.
    Turns out the Ethernet spec requires a crystal with certain specs. An expensive crystal. Affordable crystals are often a little off frequency and change frequency with temperature and aging. Some units are on the verge of not working and the stress on the crystal due the gravity pushes the frequency just as the electrostatic stress of a DC bias voltage would. Turning the unit upside down would pull the frequency away from the brink of being out of operating range. It turns out that all those cheap NIC cards I have dumped because they stopped working were just using cheap crystals and some drift out of spec. Enterprise NIC cards use high spec crystals and they work reliably. There is always something new to learn.