Monday, October 21, 2013

An Incorrect Guess About The Voltage Regulator

In a previous post of mine about the Zalman fan controller, I guessed that the voltage regulator was an LM317T. I came across a blog post about the same controller and the author said that the voltage regulator was a 7805. So I decided I had to verify that I was wrong. And I was. I desoldered the regulator from the board so I could identify it.
Fairchild KA7805. I’m not sure what E33 means. “Although designed
primarily as fixed-voltage regulators, these devices can be used with external components for adjustable voltages and currents.”
The resistor measures 708 ohms. I was wrong about it and the potentiometer as well. In-circuit measurment is not always ideal. Likewise reading resistor color codes in bad light through a small aperture is difficult.
Violet, brown, black, black, brown, a 710 ohm, 1% resistor.
The potentiometer measures between just over 0 ohms and 1K ohms.
Hey, I have some 7805 regulators in my parts library as well, so I can still make some fan controls if I want. Not that I will. And yes, I soldered it back together.

Sunday, October 13, 2013

Keithley 160B Digital Multimeters and Failure.

I bought 3 Keithley 160B Digital Multimeters at one of the OSU sales last month. They were $5.00 each. Why did I buy all three? Because who knows if they would all work. And because, that’s why. Also they all worked except that none would display amperage…and one other thing.

When I said they might not work, notices like this on the top of all three are why. They came out of LBCC and were likely abused by students.

The back has Analog Output. Except it doesn’t.

The underside. The cord wraps around the outside of the legs.


I opened it up upside down. Don’t do that.

The screws are held with o-rings underneath the 4 brown pillars/risers. I didn’t take any pictures because it’s obvious. The red arrow points to the location of the 3rd, non obvious, fuse. The two obvious fuses are fine.

The non-obvious fuse.

The fuse socket.

The analog output isn’t hooked up to anything.

This is the one with the most notes on top. The “LSD BAD” is my own note. As the Least Significant Digit in the display does not work (that one other thing) on this unit.

The battery pack was present, only without batteries. It will not be used as it requires a slew of NiCad cells.

Maybe I’ll put it up on Ebay.

Testing a non-obvious fuse. Open is bad.I bought this multimeter at an OSU sale as well.

Worlds of circuits…

On the left, the replacement fuse. I don’t know why it’s a LFMX rather than a LF. The Keithley manual (you can download one from their site and there are other copies eiaser to access as well on the web) states that it’s a Fuse, 3A, Microfuse, part 273003 made by Littelfuse. I bought 3 from Newark. With shipping, they cost more than the multimeters did.

With fuses installed in all units, I tested the amperage readings. They all work now…except for the one that doesn’t.

While the 160B has a 3-1/2 digit readout, meaning that it displays 1XXX or XXX as appropriate, the bad Keithley doesn’t display the last digit. Curiously it displays the dot.

This is what I get for probing carelessly, I blew up all the digits.

Those are some wrinkled traces (or wrinkled solder mask). The wrinkling is worse on this unit than the others.

I removed the board…DSCF1454
I had a bunch of identical LED displays, so I desoldered the bad ones and put new ones in.

Testing…this time with better current limiting.

I still wasn’t able to get the last digit to light up. So I plugged the display from the bad unit into the board of a good unit. It lit up, sort of. I possibly damaged the new LED digit when soldering, or probing, or it was already bad. But it lit up more than when hooked up to the bad unit’s board. which means that it isn’t the display that’s bad. I didn’t want to hook a good display up to the bad one’s board as there was a possibility it was burning out the digits, and I didn’t want to make things worse.

I think it’s either the transistor that switches the LED (it’s a multiplexed display) or the input to that transistor. Or a faulty trace. Or black magic. Anyway I’m probably not going to invest more time tracking down the problem. The unit isn’t worth it. I’ll either sell it, trade it, scrap it (it has some nice .1% resistors) or just let it sit in my junk pile. Not sure, I have two others that work fine.