Author Topic: Dead bug is not dead - yet  (Read 1051 times)

Offline PekkaNF

  • Hero Member
  • *****
  • Posts: 2142
  • Country: fi
Dead bug is not dead - yet
« on: March 12, 2018, 04:57:39 PM »
Dead bug construction that is.

Needed something to check short 50 ohm coax cables/terminations.

Found this TDR article and build it semi dead bug style. Stripped out all the extras I did not need: One capacitor, no extra provisions for other output impedance than 50 ohm.

 I had a bit too small capacitor, it oscilated funny ---think 65 MHz and sort of sine, but not that sure - scope is 50 MHz and pretty sure signal folded. Soldered next size up and then is works.

I think it needs a metal box.

Also I'm considering a different circuit, the pulse could be shorter and sharper.
http://support.fccps.cz/download/adv/frr/Reflectoprobe/reflectoprobe.html

Last picture with one metre of coax, T-connector and 50 ohm terminator. Needs a bit more work.

Pekka

Offline PekkaNF

  • Hero Member
  • *****
  • Posts: 2142
  • Country: fi
Re: Dead bug is not dead - yet
« Reply #1 on: March 13, 2018, 05:00:12 AM »
I know that my scope is a bit slow for this one, but have to try and see if I can produce a pulse that will yield to response I can use.

Offline Rubes

  • Jr. Member
  • **
  • Posts: 26
  • Country: us
Re: Dead bug is not dead - yet
« Reply #2 on: March 14, 2018, 05:35:03 PM »
back in the day as an engineering tech...the engineer gave me a schematic and said build this...make the leads as short as possible.
There wern't many parts, so I gave up  the perfboard and just started soldering component leads together. It ended up being a pile of parts a little bigger than a golfball. he  :bugeye: :wack: and said how am I supposed to troubleshoot that??? I said you wont have too...worked like a champ...he dubbed them 3d circuits.

Offline PekkaNF

  • Hero Member
  • *****
  • Posts: 2142
  • Country: fi
Re: Dead bug is not dead - yet
« Reply #3 on: March 15, 2018, 11:58:23 AM »
Once Pretty badly documented Siemens PLC would not start. Power went on and measured the correct voltages on rack backplane. PLC just would not start. In France. In a pretty hard spot. And really bad timing, mill would go down fast. The whole works.

Yanked the PSU out and found a dead bug in the SMPSU "power good" circuit. Took my toothbrush and some IPA to clean the fried bugger out.

There was still asort of cocroach print on the PCB when I gingerelly put the PSU back and poewered the thing up. All the right leds started blinking and the PLC sprung back to life and my bacon was saved. I really should have paid better to save the day, but I was young and it was a compliment if I didn't got biched up couple times a day.


eskoilola

  • Guest
Re: Dead bug is not dead - yet
« Reply #4 on: March 16, 2018, 01:16:21 AM »
The usual reward of a job well done - is to be able to continue doing that good job.

Offline PekkaNF

  • Hero Member
  • *****
  • Posts: 2142
  • Country: fi
Re: Dead bug is not dead - yet
« Reply #5 on: March 19, 2018, 03:39:27 AM »
Little tweaks, had problem with ringing. Turned out to be insufficient decoupling and the IC GND-pin had weak ground.

I asked help from one forum and got a link to this excelent paper
https://www.murata.com/~/media/webrenewal/support/library/catalog/products/emc/emifil/c39e.ashx
Unfortunately it gets pretty advanced after a easy start.

This is where I start: Power Supply Noise Reduction - An introduction into the basics of power supply noise reduction using supply bypassing and decoupling.
http://designers-guide.org/Design/bypassing.pdf

I had some serious ringing on the output, I could see the effect on the board IC supply pins. Tried adding more bypass capacitors and the result would change, but not improve. Changed the PSU from 5V SMPSU to linear supply and it did nothing, but I could estimate how pow hungry the circuit was.

After some more measurement I finally figured that the decoupling was a problem. Needed thigter routing for power wires (specially the IC GND pin had a large swing) and some more decoupling. Maybe if it were build with SMD:s and had a proper decoupling capacitor near IC, there were not problem at all.

It is funny how much stuff I studied (formally and hobby) is coming back, some of that looks familiar and much is new to me.

Pekka