Resurrecting a VWR Galaxy 14D Benchtop Centrifuge

VWR Galaxy 14D, back alive!

It had been advertised as “does not power on” and I got it pretty cheap on eBay, below 100USD including shipping from the U.S. to this beautiful country.  Assuming the usual culprits for a “complete failure”, when the cosmetics are fine: a broken switch or a blown fuse, it should have been an easy fix and a good bargain for a used item sold new for 1400USD.

But nope. Not really: After powering it on, the fan started, so it was neither the fuse nor a switch problem. Just the display kept staying dark and, of course, not any reaction when pressing any key.

Naturally, I decided to disassemble the dingy. First, I had to clean it: The plastic parts of the housing (including the a little bit dusty fan) went into the shower, the rotor went into dishwasher soap for 3 days.

First, I noticed a corroded micro switch which checks if the lid is closed, asking for replacement. An intermediate test to my surprise resulted in a working display and the centrifuge greeting me with a beep! A bad contact probably, I concluded. Thus, I proceeded with checking the connectors and the circuit board (PCB).

It took me some time to find out that there was a real problem with the power supply (although actually obvious: the controller module in the lid didn’t get any power, thus no display :) ). First I assumed it was in the mains, as there are some connectors which might have been faulty (inconsistent results when checking voltages with the DVM, but probably due to oxidized leads) and during measurements (i.e. touching solder points with the probes of my DVM), the display suddenly came alive again. I resoldered them, but it didn’t help.

Abrasions on the heat sink

During the further checks I noticed some abrasions at the heat sink which probably originate from a serious mishandling of the centrifuge (either a completely imbalanced run, a failure of many tubes during a run or the poor machine had been dropped). Actually, this must have been a pretty heavy impact, as on closer examination, I noticed that the whole heatsink was shifted. An impact, caused by the motor, had exerted a really high force on the semiconductors mounted on the sink.

Detail of the shifted heatsink

Then I began to check the semiconductors on the heat sink: Two voltage regulators (a MC7815C and a MC7805C) and several power FETs (IRF4905 and 532). The MC7815C  (35V (!!!) input voltage, 1A max) was measuring 0V at it’s output, so maybe it was faulty? No! I severed the lead to the PCB and the 15V were back. As the 7815 was feeding (not only) the MC7805C, I severed its input lead, too (after resoldering the 7815). Result: The 7815 was outputting 15V now. Now I resoldered the 7805’s input and severed the output. Result: Again 15V at the 7815’s output. The problem must be after the 7805!

Or in it: I resoldered the 7805’s output lead. Now the 7815 still was delivering 15V. At the 7805 there were 2.5V (with the controller as load 1.8V) . Strangely, I had almost broken off  the output lead of the 7805 during the test (probably it was already damaged by mechanical overload), so I’ll have to replace it anyway. Is this the clue to the problem? Result after replacing the 7805: Nothing. Again 0V at the outputs. I decided to replace also the 7815 with a L7815C which also accepts 35V input voltage (uA types will not survive more than 30V).

After the repair, the controller finally had power: The display lit up and it made “beep” again. The motor still was not working yet. Some more checks and finally I found that I had reconnected one of the micro switches in the wrong way.

Total cost: (8+8+20) NT$, somewhere between 1 USD and 1 Euro, for 2 pcs. 78xx regulators and a micro switch. BaDer-Rd in Taipei City (link) is my place for this (and one of the hubs this world is spinning around). Plus a few hours of measuring and reasoning.
Total reward: a practically new benchtop centrifuge for less than 100 USD.

During the following tests I noticed that the machine is very sensitive to imbalance (which might have caused the previous failure). The rotor is sitting directly on the motor axle. The motor is fixed to the base with 3 rubber dampers on its remote end, so the whole system is somewhat prone to oscillation. The rotor is hollow (in contrast to e.g. Thermo/Heraeus Picofuge machines), thus leaky or breaking tubes may cause a (sudden!) imbalance of the system which might result in disaster, esp. at a higher speed. If being able to acquire a new machine, I would not choose this VWR model, as a massive rotor, preferably made from re-inforced plastics, definitely is the better (and safer) option. One ruined experiment or even injuries and damage caused by a failing centrifuge may incur more cost and hassle than the money you need to spend for a really good machine (like the Picofuge), when you need to use a benchtop centrifuge for your daily labwork.