RF LDMOS failures

[pwalker]

I am testing a new 25W PA design which uses the MRF1535N in the output stage, and I have experienced 2 failures whereby the gate goes low impedance. The second failure occured when I removed/applied gate bias when 1W of RF drive was present. Gate drive was set to ~2.5V for an idle drain current of ~200mA. Prior to the failure the device was outputting ~30W into 50R and was stable. The switch-on transient seemed to kill it. I'm not sure if the failure occurred at the moment I removed the bias (gate was pulled down to 0V via a 47R resistor) or when the bias was re-applied.

I'm assuming that the failure is due to the lack of gate protection diodes. The datasheet suggests adding zener diode protection but I believe a single zener is not sufficient. In fact, most of the old Motorola RF Line datasheets going back to the 1990's say exactly the same thing but no other info is provided. Is there a recommended protection circuit for a 35W or 50W device? Would someting like the Semtech LCDA12C-1 low capacitance transient suppressor be suitable here? My frequency of operation is 180-190MHz.

Any feedback much appreciated.

Thanks.

[milanbaca]

The part could have been killed if the switching times were very fast.  Anything shorter than 100 nSec is at risk of killing the older, MRF1535N parts.

That entire 12V land mobile portfolio is being updated with newer, better, faster switching, more rugged devices.

It would be advisable that if you are working on a new design, you should be working with the latest and greatest
parts to ensure the greatest longevity for your design.

[pwalker]

Anyone able to offer an input on this?? I've since changed to a 50W device (MRF1550) and this device has just failed with the same fault - gate going low impedance. I had fitted a LCDA12C-1 suppressor and the gate was not being switched on and off when the failure occurred. This is getting frustrating (and expensive!).

[donnavigneri]

By now you must be far into, or past your design. However, for the benefit of the community, I would recommend the following:

1. The 135-175 MHz broadband circuit on the datasheet shows a simple bias network, which has been very successful in the past. Your gate bias network should work off of this bill of materials and network suggestion. If you must use the MRF1535N with your own gate bias design, then a voltage regulator with slow switching speed (in case of fast transient) and a large value shunt capacitor may help where your Zener diode has failed. Also, try using a series resistor to connect to the gate of the transistor, as shown in the diagram. If possible, measure the input return loss of your system. It looks like a good design should use only a half Watt of input power to achieve 30W, where you have had to use 3dB more (1W). That lowered gain may indicate bad input return loss.

2. The updated Freescale part, AFT05MS031N, may be a better choice for your application. http://cache.freescale.com/files/rf_if/doc/data_sheet/AFT05MS031N.pdf?pspll=1

There is also a 75W device in the same portfolio, AFT05MP075N. http://cache.freescale.com/files/rf_if/doc/data_sheet/AFT05MP075N.pdf?pspll=1

It has more than the 50W you need, but the 55W device, AFT09MP055N, is likely unsuitable for your operating frequency.

Best of luck with your future designs!