12-24V DC speed controller queries

I have been studying the motor controller circuit in the March 2008 issue and I notice that all components are on the same side of the board, so it means that the Mosfets are lying on the epoxy rather than the copper. No wonder the FETs run hot! If you don’t mind me saying so, that is a very strange arrangement indeed.

They cannot be turned over and put on the copper side as the leads would not line up. I do intend to build one of these so I have decided I would stand them upright and bolt an angled 16G copper or 3mm aluminium heatsink onto them where they are situated now under the tab. That should enable loads of 50 or 60A at a significantly lower operating temperature.

I happen to want to run two 500W motors in parallel on 24V in a model train so I know I will be running the controller at its limit. If necessary, I would be prepared to use a separate supply for each motor but a much cheaper solution is to just slightly beef up a single controller, especially when I think that a sizable heatsink is going to be necessary anyway.

My other concern is with the 55V rating of the FETs. I once came across an old rule of thumb that stated that the FETs should be rated four times the battery voltage. Is this a valid rule still? The IRF1405 is designed specifically for automotive use and that means 14V. If you multiply 13.8V by four you get almost exactly 55. This cannot be a coincidence.

Now I wonder if you have run this circuit in a real life vehicle on a 24V power supply? I once blew a 60V (60N06) Mosfet in a five-Mosfet parallel array on a large heatsink on a 24V supply, powering a 5-inch gauge model train. I could never be sure of the cause of the failure so I decided to replace them all with 100V types and had no more trouble, even under some periods of "rough" use when the train was dragging its brakes. Even that big heatsink got hot then, so it was a real test!

On that rule, these Mosfets should be rated 100V for a 24V supply. But there is no 24V equivalent to the IRF1405 (that I can find) so it would probably have to be an IRF540N with a rating of 33A. I have not yet worked out the thermal gradients for this one but it would certainly mean a decent-sized heatsink and the use of perhaps two extra Mosfets. I would appreciate your comments on these two issues. (P. D., via email).

A continuous 40A rating also means a very large battery capacity, otherwise the batteries would be exhausted very quickly medium load, and would The reason we decided against heatsinking the Mosfets was simply to save space on the board and to save on costs.