Power Series Fans
Last updated: 24/6/98

Pictures and approximate specs of those noisy fans from a predator cabinet

Close up of the Nidec Beta V TA600DC, 12V DC, 1.9Amp ball bearing fan used in the Power Series Predator cabinet. The model number is AM32590-10 and the fan looks/feels like it is constructed from aluminium. Looking on Nidec's web site (http://www.nidec.com) I can no longer find this model number. However, the TA600DC series is still being manufactured and the current A34453 fan looks to have the same specs and dimensions as the AM32590-10. These specs (taken from Nidec's web site) are:

240CFM air flow, 6-18V range, 1850mA draw, 22.2W power consumption, 66 decibel sound level.
Dimensions are 6.375 inches between mounting screws, 5.91 inches wide, and 1.5inches deep - a full dimensions diagram can be found on Nidec's web site.

Here's a picture of the entire fan draw from a Power Series predator.

Trying to reduce the noise generated by the fans

Warning: If you follow any of the steps I have outlined here and either cook your cards or burn down your workshop/workarea/environs - then don't blame me! This information is offered without any guarantee and I don't claim to be an expert or a professional in these matters.

As far as we can make out, the noise generated by these fans seems to be due to two factors:

  1. The speed at which they rotate (this is by far the biggest contributor)

  2. Turbulence caused by the air flow from the fans hitting the card cage slots
Unfortunately reducing the speed of the fans has the unpleasant consequence of reducing the airflow. Given that we arent running a full complement of CPU/MC2 cards but that the RealityEngine Raster Manager boards produce a fair bit of heat, then by how much can we reduce the speed of the fans without cooking the boards? I wish I knew! :-)

Looking at the problem from the users side, how much do we need to reduce the fan speed to satisfy noise constraints? According to the sound experts, we need a reduction of 10 decibels to make the beasts quiet enough for peaceful office coexistence.

What I am planning to do is to reduce the voltage supply to the fans in steps of 1V at a time until I can satisfy the noise constraints. To do this I'll use silicon diodes (about a 1V drop across these I believe) that can handle 8amps and see what happens to the noise level and the air flow. Note: another alternative, using a DC rheostat, is not really sensible because of the heat that will be produced in the rheostat.

Another alternative might be to go with AC fans (Nidec make an AC version of this fan (TA600AC model A30322 for 240V AC supply) which has the same dimensions and approximately the same airflow, noise level) and use the wave chopping capabilities present in most household light dimmer switches to reduce the fan speed. As a sidenote, I believe the Power Series Twin Tower machines use AC fans but I'm not sure whether they are the same as these or not - I'll find out soon when I get one of these beasties in for a closer look.

Comments to: simon@dpiwe.tas.gov.au