Has anyone here used the BUZ341 N-Channel SIPMOS for a PA design ?

I am considering using a bunch for a variation on Mr Holden's interesting quasi-complementary N-Channel FET amp.

the data sheet looks encouraging - Ptot 170W , Ciss 2600pF, Gfs 23 S, and derates to zero at 150C ....

I've mailed the man himself to see what he thinks, but I wondered if anyone else has any experiences with these devices??

I'm looking for 450-500 WRMS from a set of 4 pairs at +/-70volts/rail into a 4ohm load.

any thoughts?

I'm sorry I can't give you any information on the BUZ341 because it's kind of hard to get around here. I have seen BUZ used in PA amps (BUZ900/905), the IRFP240/9240 and 2SK1058/2SJ162. I'm sure the designer has his reasons for using this part. Maybe due to the 'ON' resistance.

If you are going for a 500WRMS @4 ohms output, 4 pairs seems a bit thin. It is not that it will not be able to deliver. With 3 pairs of BJT's MJ15003/4, I can get 500W RMS @4 ohms. But this is on my test bench and it is into resistive load.

In practical use, I would suggest at least 6 pairs. Because when it's out in the field, they are going to parallel 2 bass bins and there is a good chance it will dip below 4 ohms at certain frequencies. Furthermore, a loudspeaker is not a pure resistive load.

For 500W output, you will be working at about +-85V on BJTs. With IRFP's, about +-5V more due to the gate 'ON' threshold.

Hope I've been of some help.

Mike, thanks for your thoughts.

I ought to point out that I'm unlikely to actually get much more than about 400W due to psu limits, the 450-500W figure was more in line with commercial amp claims .....

why do you think 4 pairs is too light? even with a heavily loaded o/p I would have thought that 4 pairs could dissipate plenty even at 100C , or do you think the margin is too small?

I could add another 2 pairs quite easily, but the o/p stage constitutes the major cost of the project and pushes the cost of the design up by almost a third!

the sheets claims an Rds(on) of 0.06R typ. not sure quite how relevent it is 'cos ( I hope ) the o/p stage won't be driven that hard into conduction :bigeyes:


Hi Ray

You can quite easily use 4 pairs. More outputs will basically give you a better safety margin. But bear in mind that how much power an output device is capable of dissipating is very much dependant on temperature and the voltage it is working at. In a nutshell, the higher the rail voltage, the lower the current. The higher the temperature, the lower the current also.

There are actually 2 schools of thought here. You can design it under FTC specs (which means RMS into load) or do like what NAD amps are doing, go for Dynamic Headroom instead. It lowers the cost as the transformer can be made smaller.

As for operating the outputs at 100C, I don't think that's a good idea. It is common in Pro Amps to install a thermal cutout on the heatsink or an output trans (basically the hottest part they can find) so that it cuts off the mains supply before the outputs fry. The cutout temperature is generally between 75-85C.

If you expect your amp to be heavily used, find a suitable heat sink and force cool it with a 5" axial fan. Basically, keep the temperature down. The lower it is, the more power the outputs can deliver.

Have fun

I quite agree about the temperature, no I was thinking worst-case overheat.
The design I'm working towards calls for chunky heatsinks and dual fan forced cooling. I'll probably use a 85deg thermal cutout too 'cos I've got a few left over from another project ;-)

.. and soft-start on the psu too, with another cutout of the dropper resistors in case the relay circuit fails....

many thanks for your thoughts.