eevBLAB #6 - Don't Assume It's Faulty

[EEVblog]

[george graves]

 

I've had worse.  Not afraid to admit it.  Reason #203.  For me... it's why breadboards have a limited life span, and you should just thrown them out after a few years.

[bbjk7]

Hi dave, Love the videos

Ran into a similar problem at work last week except I'm pretty darn sure this one is in fact faulty. The exact power supply in question is this one here http://us.tdk-lambda.com/ftp/specs/hws1500.pdf

I was trying to run 3 of them in parallel, when I powered the them on (with a multimeter across the output because I was about to set it up for 28V), straight away I smelt smoke and turned it off within 2s 

Turns out one the little connectors (see under positive bar) used for setting the mode, sensing load etc... had the sense wires connected backwards!! (V+ to -Sense, V- to +Sense) I'm not even sure what this would do??

So I tested each independently and turns out only the middle one might have blown something.. PHEW!!   
When it's turned on, the voltage shoots up to where it's supposed to be, but then slowly drains away till 0V 

Took it apart and there are three main boards roughly equal in area.. with nothing obvious blown.. all tracks are intact, no blown cap or black marks... hmmm no idea where to go from here.. boss said to just get a new one, even with these things going for near $2k 


 

[mcinque]

So why the "fault" didn't show up in lower ranges?

[IanJ]

Nice one Dave.......it happens to us all.

Actually, on my own EDC Voltage Standard when I modified it I actually got rid of the sense terminals completely when I moved the terminals from the rear of the unit to the front...............I figured simplicity wins (in my case).

Ian.

[EEVblog]

Actually, on my own EDC Voltage Standard when I modified it

Woah, you got details of this mod?

[george graves]

Woah, you got details of this mod?

Duh!  He's got a great youtube channel.  Worth subscribing too.

[N2IXK]

Faulty sense connections were the first thing that came to mind, but you showed the unit working on lower ranges, and specifically stated that the problem wasn't a bad connection.

Is the remote sense disabled on this unit on lower ranges, and only enabled on the 10V range? Why would the 1V and 100 mV ranges not be affected by this?

[c4757p]

and specifically stated that the problem wasn't a bad connection.

yup, hence why I had absolutely no idea what the problem was!

[EEVblog]

Ah, already a digital unit, just added remote controls.
I thought it was a hacking an existing analog unit in some way.

[EEVblog]

and specifically stated that the problem wasn't a bad connection.

as I wiggled the banana plugs which is what I was talking about.

[nitro2k01]

and specifically stated that the problem wasn't a bad connection.

as I wiggled the banana plugs which is what I was talking about.

And also stated, "we've got out sense terminals here, everything is fine". On the other hand, not much point in complaining about being fooled for 5 seconds in an off the cuff video.

[hobbs]

It doesn't seem too likely that the output really rails at 10.47 volts.  There are probably inverse parallel diodes between the sense and force terminals to protect the sense circuitry.  They never turn on on the lower ranges, and wouldn't conduct under reasonable circumstances.  There could be a voltage divider loading the sense terminals, which might not be present on the lower ranges.

 A highish value resistor in parallel with the diodes would explain why the lower ranges work fine.

Good warning, though.  I've made a few mistakes like that!

Cheers

Phil Hobbs

[SeanB]

Had the same today. Replaced a 40 year old Siemens relay with a new one, as the switch driving it has been changed from a reed switch and magnet to an inductive sensor ( either the reed or the magnet is getting weak, it will not turn off intermittently) and a more sensitive relay. Changing out the block relay to an 11 pin socketed relay meant I has to move the terminal wiring, and in placing the common relay leads into the socket I got one on top of the clamp and not under.  Did not find this till the power came back on ( Load shedding, ending just as I type this on generator power) and I went to start the machine again. Powered up and no operation, on touching the relay I saw the flash ( 220VAC from a local transformer) on the contact. Turn off, pull the relay and get the wire into the clamp properly. Was hard to see the brown wire inside the machine under the relay socket, with the only light being supplied by a 7W LED lamp in the lead light, and a temporary 1.2M fluorescent lamp on the floor nearby.

[scopeman]

Usually when I get one of these I unscrew the caps as far as I can, pull the links and clean all of the contact areas and the links with Caig DeOxit. Then a take a nut driver and drive them by hand until snug, taking care not to apply much torque.

Sam
W3OHM

P.S. Just picked up a EDC 521 that was nearly pristine but had only a bad bridge rectifier in the 5V supply. Fixed that, powered it up and in 20 minutes later it was within 2ppm.

[miguelvp]

Just last week while upgrading my laptop lcd screen converted to a monitor the 12 V 4 Amps brick supply that the display card came with (one of those ebay VGA/DVI  to LVDS adapters with no HDCP of course) it wasn't powering up.

I did use another one I had from my Tegra Jetson TK1 (12 V 5 Amps) and positive center as well and that worked fine.

So I decided to take the cheap power brick appart which was glued. Everything looked fine, plug it into the mains again and no power out, no power anywhere in the board....

So I measured the power cable, one like the ones you use to power a PC and checked for continuity.

Ground, was good, Neutral just fine, Live ... Nothing!!

I didn't cut the cable yet to look at the gauge they used, but it's the first time one of this type of cords failed on me, but all of the ones I have are from a while back from older PCs so I do have plenty of good ones.

The cable from the outside doesn't feel flimsy at all but what it's funny is that I immediately thought "power brick" when it was the always considered reliable type of cord that failed. It goes to show you that we assume where the faults are at because what we think the common failing point is.

At least It gave me peace of mind to open up the glued up power brick to make sure it was safish, I still won't leave it plugged in unattended, and only plugged in when I'm using that monitor for my dev boards.

I have not thrown away the cable yet because I'm curious where did it fail and how thick is the copper or whatever used to make it.

[SeanB]

Likely failed at the laptop plug end, unless you had previous ly bent the plug end to get it in a socket. They tend to break at the most flexed point, which is almost always the socket end.

I had some cables where the strands were supposed to be 1.5mm, but the six or so strands of what possibly could have been copper at some time came nowhere close. Black brittle stuff that was enclosed in a massive plastic sheath which in turn had an even thicker outer sleeve. Plastic was double insulated though with a white inner and a coloured outer and the overall outer was black. Genuine junk, i put it in the pile of cable to be sold as scrap that was not worth stripping the copper out.

[ornea]

at 4:18 I assume you meant to say 4 wire sense and not 4 ohm sense.

[owiecc]

I have a general comment about terminals. I can't remember the last time I saw terminals like these. All the stuff we have is shrouded banana. All the cables we have are stackable. I don't think we can even buy equipment with banana posts.

There is a bit more cabling involved (short sense wires) but all connections are very reliable. What do you use in your labs? Do you get new equipment with posts?

[kfitch42]

At least you found your loose connection faster than those faster-than-light-neutrino scientists.