Author:  Mike Boesen  

Last updated: 23 July 2006

1.  Background

I wrote an article about backup strategies which appeared in the October 2005 issue of the Canberra PC Users Group's "Sixteen Bits" magazine.  I have posted an updated version of the article here: ( ).  In the article I referred to the merits of using a hard drive in a removable caddy as a backup drive.  The "caddy" is a removable container in which a hard drive is kept.  The caddy slides into a "cradle" (or "rack") that is mounted permanently in one of the CD or DVD drive bays in the PC.   The caddy containing the hard drive can be completely removed from the cradle.  This setup is also referred to as a "mobile rack" but this is a misnomer - what is mobile is the caddy;  the rack is fixed in place and is not mobile.  

Some caddies are available for hard drives with serial interfaces.  However, most caddies are for hard drives with parallel interfaces. This article assumes that if you get a caddy it will be for a hard drive with a parallel interface.

Power is supplied to the back end of the cradle inside the PC's case, and the cradle is connected to the parallel IDE/ATA interface of the PC.  When the caddy is fully inserted, a Centronics-style socket on the outside rear of the caddy mates with a Centronics-style plug on the inside rear of the cradle.  This mating pair provides a 5v (red) and 12 V (yellow) power line and a pair of ground lines (black) to the hard drive plus 39 IDE/ATA interface lines.  If your PC's motherboard has a parallel UATA (Ultra ATA) interface, the cable from the motherboard to the rear of the cradle will be a flat 80-wire cable going to a 39-pin socket (looks like a 40-pin socket but one pin is absent).  If it is not UATA, it will have either a flat 80-wire cable OR a 40-wire cable going to the 39-pin socket.  (An 80-wire cable will work fine with a non-Ultra ATA interface because all the extra 40 wires in the 80-wire cable are simply ground wires designed to eliminate "cross-talk" between the data wires.) 

I have been using such caddies for about 8 months and have found that  the system is very effective in the context of creating clones or compressed images of my master drive, for these reasons:

If you decide to implement a cradle and caddy arrangement, there are three matters that require some thought: 

2.  Effectiveness of air flow through the caddy

To ensure that there is a cooling airflow over the hard drive inside a caddy, it or the cradle should have at least one fan.  This is particularly important for modern, fast (7,200 rpm) drives because they get very warm in enclosed spaces. 

Such fans have to be small and some caddy setups have three such small fans - two in the cradle and one in the caddy.  I don't like those at all because of the noise and I think that more than one fan is quite adequate providing that:

I have found that "Laser" brand caddies which have only one fan in the cradle are very well designed in terms of maximising the effectiveness of airflow within the caddy.  When fully inserted, the caddy buts up flush against the back of the cradle and the air can only be sucked in through the front of the caddy, across the drive and then out the rear of the cradle.  

I monitor the temperature of my drives.  Such monitoring can be achieved through a number of freeware utilities such as  HDD Thermometer (   ) or HDD Health (  HDD Thermometer is the better of these two utilities in terms of displaying drive temperature because it can be configured to show the temperature of each drive as separate numerals that show in the system tray and the warning and critical level settings can be specified separately for each drive.   If HDD Health and HDD Thermometer don't work for you because "SMART is not enabled", check that your PC's BIOS is configured so that SMART (Self Monitoring and Analysis Technology) is "enabled". That setting will be found probably in the IDE settings section of the BIOS.

Surprisingly, in my system when I have the drive in a Laser caddy operating it is about 3 C to 4 C cooler than an identical 7,200 rpm drive inside the PC's case. The internal drive in the case is well ventillated - it is diectly in front of a 120 mm intake fan and the case has a 120 mm exhaust fan.   I attribute the better temperature of the drive in the caddy to the fact that the speed of airflow over the drive which is created by the fan in the cradle is faster than that for speed of airflow over the drive in the PC's case.

So I suggest that if you buy a caddy/cradle setup you should have a good look at its design from this point of view.

A final point about caddy design - if you leave the sliding cover of a caddy off, the temperature of the drive will be hotter than with the cover on!  Leaving the cover off is detrimental to the objective of maximising a flow of air over the drive from outside the case.

3.  Caddy fan noise and how to reduce it 

Because the fan/s in a caddy have to be small - no greater than 40 mm in  diameter - it/they add to the PC's noise emissions somewhat, especially if the PC is on the desk close to your ear.  So the fewer the fans, the quieter is the PC. 

In all the caddy/cradle types I have seen (including the Laser brand), the fan (or fans) are running all the time, irrespective of whether or not the caddy is inserted.  This is doubly annoying, because (a)  if there is no caddy inserted, there is no need for the fan to be running at all and adding noise, and (b)  there is unnecessary wear on the fan, and that will shorten its service life.   After operating for a good while, the blades on a caddy fan invariably become dirty and this increases the load on the fan's components.  Constant use leads to deterioration of bearings.  I have seen a number of well-used cradles that have seized fans.  Invariably, such caddies were installed quite a long time ago and their fans had been running all the time that the PC is on, until they gave up the ghost.  If a cradle's fan has ceased to operate, the drive it is meant to be ventilating will run abnormally hot, because it is tucked away in what has become a heat retaining plastic coffin.  That will shorten the life of the drive and can lead to data transfer errors.

To avoid unnecessary fan noise and wear, I have mounted a microswitch in all my Laser cradles.  If there is a caddy fully inserted into the cradle, the switch is closed and current is supplied to the fan.  When the caddy is extracted more than about 5 mm the switch turns off and no current goes to the fan.  The only time there is any fan noise is when the caddy is pushed in all the way.  Fan wear is minimal because the only time I have a caddy fully inserted is when I am creating a backup of my master drive onto the drive in the caddy.  After backing up, I extract the caddy, so the fan in the cradle turns off.  

The switch I used is is a sub-miniature microswitch (Jaycar SM1036 costing $3.30.  Go to  Jaycar (  ) and search for microswitch).In an earlier version of this article, I described a way of installing the microswitch that involved cutting one of the power wires between the fan plug and the fan.  Since then I have developed two alternative and better ways of installing the microswitch referred to in that article.  These alternatives mean that there is no need to cut one of the power wires between the fan plug and the fan.  The merit of this approach is that if the fan ever needs to be replaced, you can simply unplug the old fan, unscrew it, throw it away, screw in a new fan, and plug it in.

3.1  First alternative

The photos below show how the first alternative is implemented. 

Switch on power side

Switch track

Track closeup

Track soldered

3.2  Second (and better) alternative

Having done one caddy that way, I took a smart pill and realised that there was an even better way to do things: 

Fan side switch

Unfortunately I am not able to include a picture of the completed soldered wiring done this second way because I have lost it and I no longer own the PC in which the cradle is now residing.  Sorry about that, but I think that you will understand what to do.

4.  Improving the mating of the caddy and the cradle

On one caddy system of mine I noted that on some occasions I was not getting reliable transfer of data and the hard drive operated slowly and erratically.  I have had a report of a similar problem from at least one other user of caddies.  The problem may be caused by inadequate mating of the big 50-land plug at the outside rear of the caddy with the big 50-land socket at the inside rear of the cradle.  With my problem caddy, it seemed to slip into the cradle with very little insertion force being applied and I became concerned that caddy and cradle contacts were unreliable.

After I fine tuned the caddy's socket in the way described below, my problems vanished and have not reoccurred in the subsequent six months of frequent use of caddies.

The Laser system has a 50-land Centronics style socket at the rear of the caddy and that mates with a 50-land Centronics style plug at the rear of the cradle.  My preference is for the caddy's socket to mate with the cradle's plug only after a bit of pressure plus a nice "thunk" indicating that the mating has been effected and all the lands (flat gold-plated springy thingies) in the plug and socket are in firm contact.   If your caddy already has to have such pressure exerted to get the caddy fully inserted, then I'd leave it alone unless you are getting unreliable data transfer.  However, if the caddy slides in with very little resistance, I'd do the following fine tuning.

For the sockets on all my (3) caddies, I have raised the arching of all the lands very slightly, by only about 0.5 mm.  The high-tech instrument I used to do that is a number 9 darning needle carefully inserted under each land in turn.  You can see the increased curvature for the lands to the right of the needle in this picture:

Needle under lands

This may sound complicated, but it only takes about 5 minutes to do all 50 lands in the caddy's socket.

If you do the fine tuning, do not lever a land up - simply slide the needle under the land to increase its curvature a tad.  If you lever up the land too much, the pointy end of the land closest to you will pop up out of its bed.  However, if that happens, it is possible to recover the situation:

If you do that fine tuning, when you insert the caddy into the cradle, you will find it now meets with more resistance at the point when the caddy is almost fully inserted (all but about 5 mm to go).  The ideal resistance is one that:

If the caddy is a bit tight, it will slacken over time.  A little bit of WD40 or other lubricant that is suitable for electrical switches will help a caddy that is a tad tight to fit in properly.

If you have not yet mounted the cradle for the caddy into your PC, an alternative approach (and probably the better one) is to raise the lands on both the caddy socket and the cradle plug slightly - say 0.3 mm for both - rather than raising just the lands on the caddy's plug.  However, if the cradle is already mounted in the PC, raising just the lands on the caddy plug seems to be effective.

5. Conclusions

If you get a caddy/cradle set:


Appendix 1 - One way of fixing the microswitch in place

I attached my microswitch using a very stiff mixture of two-part epoxy resin (West Systems epoxy) and a microfiber-silica thickener.  Technical stuff, but those were the things I had on hand.  Instead of this stuff, you can use any 2-part epoxy (e.g. Araldite, Bote Cote) and any thickener.  A spoonful of very fine sifted sawdust is OK as a thickener.  You can ask you friendly chop saw operator at your local hardware shop for a cup or so of sawdust - take it home and sift it, keeping only the very finest particles.  Mix enough fines with 2-part epoxy to make a VERY stiff consistency that will not slump. 

Place a bed of thickened epoxy into the cradle and sit the switch on top with the spring lever pointing towards the 50-pin plug.

Put a tad of epoxy/thickener mix in the two holes through the body of the switch.  Stick two shortened round-head nails or short lengths of wire through the holes.  Then lay the switch onto the bed of epoxy/thickener, with the nails or wire pushed into the bed.  Slurp a little epoxy/thickener over the nail heads for good measure.  The nails or wire will guarantee that the switch is kept firmly in place when the epoxy is set. 

The position of the switch should be such that when the caddy is fully inserted, the switch will be in the closed position, but the caddy is not prevented from being fully inserted so that it touches the rear of the cradle.