Saturday, July 18, 2009

It's all VSE's fault

Remember this photo I put in my last blog post?

Something's wrong with this picture. And it was about to cause me a bunch of extra work. We'll revisit this...

I received the MTA-156 connectors from Mouser and made the harnesses to connect the +5V supply to the distro board, like so:

The DC output:

The AC input:

Here's the completed wiring to the distro board. The jumper wire is tying the returns together for the two supplies:

Next, I soldered the AC-side connections. For this module, I used 18-gauge zip cord as the power cord. On past modules, I've use 16-gauge, but that's really overkill; these things don't draw much power. Below you see the fuse holder, which is connected to the AC hot; the neutral is connected directly to the transformer, and there are some jumpers per Condor's installation instructions. There are also smaller jumpers that go over to the +5 supply. In the photo, the connections are complete, and I'm gooping all of the exposed terminals with insulating RTV.

The connections are complete, and I'm about to install the fuse. I'll start with a 0.25A fuse. Note the cable staple retaining the end of the power cord on the base.

Putting a plug on the end of the cord, being careful to get the hot side of the cord on the correct prong. (The hot side is the brass screw; the neutral side is the silver screw.) This is an "old fashioned" one that uses screw terminals; I don't trust the quick-connect plugs. In the second photo, note the built-in grommet for strain relief.

Now, it's time to test. But first, let's go back and look at something again. Remember from the last post, when I said that I always wire the +15 with a red wire, and the -15 with a green wire? Well, note carefully the silkscreen designation next to the red wire below.

Yes, I connected them backwards. And no, I didn't catch it before I powered it on. And yes, the smoke test succeeded -- in producing smoke. When I applied power the first time, I got nothing. Checking, I found that the fuse was open. It didn't appear to be actually blown, so I figured it was a defective fuse. I didn't have another 0.25A one, so I put in the next highest amperage I had, which was a 0.75A. I powered it up and I was initially encouraged to see voltage. But then I was puzzled as to why my voltmeter was indicating a negative voltage when I had it connected to the +15 side. Then the voltage started to drop. Then I heard a fizzing noise and saw smoke coming from the distro board's filter caps...

The result. Note the capacitor in the center, and compare to the one on the right. (The one on the right is for the +5V, which wasn't effected.)

So I had to replace them. 330 uF, which of course I don't have. Another order to Mouser. While I was at it, I ordered 35V, 105C types (the originals were 25V). Here, the two failed ones have been replaced; note that they are a bit taller than the remaining original:

Screw the distro board back down, power it up, and everything works. I didn't even have to adjust the voltages; they were all dead on.

Here's the (nearly) completed Block 4, with an MOTM-820 VC lag processor installed. The slot to the left of it is awaiting my MOTM-650. I still need to put on the top, but that's easy; it's just another piece of wood, held on with Velcro at the corners.

So why did I forget to verify the power supply connections? Probably because, just before I worked on it, I was reading this VSE thread on stupid accidents with gear. They jinxed me.

No comments: