MRF101 Band 2

[XXL]

What’s a safe maximum power from the mrf101 device ?.

With around 1.2w in I’m getting 150w. But iv turned down the bias because I’m scared it’s going to pop. The datasheet says it’s only meant to be 100w but that could mean 100w approx across 0-250mhz. The temperature doesn’t seem to go up.

[rigmo]

115

[XXL]

I usually run them at 120w. But with more drive it seems to do more. It seems alright but it could pop later on.

[thewisepranker]

It's bad practice to say how much power a device can deliver by just looking at the first page of the datasheet. You need to evaluate it on a per-case basis. The output power is a thermally limited parameter, which is dictated by the power dissipated by the device (a function of drain efficiency and other things) and the thermal resistance of the heatsink arrangement, taking into account all of the thermal resistances. The drain efficiency is also a function of the power being delivered further complicating things.

On page 27 of the MRF101 datasheet is the drain efficiency performance for the 87.5-108 MHz circuit that they provide. Worst case is about 75% at around 104 MHz, with the device delivering something like 113 W to the load, so let's go with 80% efficiency just in case someone claims to be able to do better than NXP.

This results in 28.25 W dissipated into the device from just the drain efficiency. We've also got to factor in Idq (100 mA at 50 V = 5 W) and the drive power (1 W) for a total of 34.25 W. From experience that's well above what I'd be comfortable dissipating into a TO-220 device but let's continue anyway.

The thermal resistance is listed on page 2 as 1.1 °C/W, junction to case.
On the same page is the table of maximum ratings where we can find the maximum junction temperature, which is 175 °C. The device has a linear power dissipation derating above a case temperature of 25 °C, -0.92 W per °C of case temperature rise. The numbers on the datasheet are a bit optimistic and you won't be able to achieve them in practice without expensive machined parts, which we're not going to use. For example, the datasheet claims that you can dissipate 7 Watts at a case temperature of 175 °C, at which point the junction temperature would be 192.5 °C and the transistor would be destroyed.

Going back to the example.
We have to choose what choose as a starting point. You can choose maximum junction temperature (for MTTF prediction), heatsink temperature, case temperature or thermal resistance of an existing heatsink. Let's go with maximum junction temperature of 150 °C.
We know the junction to case thermal resistance from the datasheet as mentioned earlier, but there are more thermal resistances we need to factor in, such as the junction to heatsink thermal resistance and the thermal resistance from the heatsink to ambient. Mounting thermal resistance will in practice be about 1.5 °C, so we've got a total thermal resistance of 2.6 °C/W from the junction to the heatsink.

We now need to know the ambient temperature, which will probably be around 45 °C by the time it's in a box. This gives us a total allowed temperature rise of 150-45 = 105 °C.
We can now work out how big a heatsink we need. Divide the allowed temperature rise by the power dissipation to get the total allowed thermal resistance from junction to ambient: 105/34.25 = 3.066 °C/W.
Take off the mounting and junction to case thermal resistances: 3.066 - 2.6 = 0.466 °C/W. This is easily achievable using a medium sized heatsink without forced cooling so this shouldn't be a problem, but bear in mind that if your mounting thermal resistance is a fraction higher (for instance if you believe that more thermal paste is better, which it isn't), the junction temperature will also be higher.

Also bear in mind that a junction temperature of 150 °C isn't something to aim for, but let's continue anyway.

In practice we can do better than 0.2 °C/W with a forced cooling solution, so let's put the numbers back in and see what the junction temperature is.
Rth (j-a) = 2.8 °C/W
Pdiss = 34.25 W
Tamb = 45 °C
Trise = 95.9 °C
Tjunc = 45 + 95.9 = 140.9 °C

The MRF101 is considered "rugged" partly because it can handle a mismatch of over 65:1 and that means that you don't usually need SWR protection, saving you money. This reflected power is also dissipated in the device and has to be accounted for if you want to be able to consider it as "rugged". Performance will be massively reduced because of the high junction temperature.
Worst-case we can continuously dissipate as much power as takes us to a junction temperature of 175 °C where the device just starts to exceed its maximum rating. The power required to do this is the distance from Tj(max) (175) to the ambient temperature (45) (which is 130 °C) divided by the total thermal resistance (2.8), which is 46.43 W. Minus the 34.25 W we're already dissipating = 12.18 W.

Assuming 113 W forward power, maximum SWR will be 1.978 which is terrible considering most people aim for 1.5 for their aerials. A far cry from the 65 or greater from a rugged transistor.

[teckniqs]

I know someone who has used one on the N Proctor 100w NXP competition amps at 125w and it worked fine, after 10 minutes bench test on dummyload the power only drops to 124w.

[Albert H]

My two prototypes couldn't get rid of the heat really efficiently. I found that one of the devices had a distorted case and didn't fit flush to the heatsink! The gentle application of a fine file and some emery cloth sorted that problem out.

I found that running at 94 Watts out, they'd trundle along happily all day. and survive most kinds of abuse. They didn't get too hot.

Driving them harder, and tweaking the bias a little would get them to >120 Watts, but they temperature derated quite quickly... These are NOT high power devices! I blew one up at 155 Watts out, when I was seeing just how far I could push it.

To be honest, nobody is going to notice a difference in field strength between a 95 Watt rig and a 110 Watt rig. There's no practical difference except that the finals will be more robust at the lower power. I'd just run these things at around the mid 90's to 100 Watt and make sure that there's plenty of heatsink and there's a copper clamp to take the topside heat away too.

[XXL]

Thanks everyone for the info.

[rigmo]

mrf101 PCB

[Albert H]

Nice looking board! Run it at 95 - 100 Watts and it should last until it's removed!

[jvok]

Looks like a direct copy of the enigma board

[Albert H]

Don't think so.... it looks like it's based on the Application Note from Motorola!

[RF-Head]

Show the the Application Note from Motorola
It's the design from Enigma

[Krakatoa]

Show the the Application Note from Motorola
It's the design from Enigma

Don't worry. Soon there will be Chinese knockoffs everywhere. It happens every time a device becomes interesting to market.

BTW, the MRF101 is an NXP design, not Motorola.

[rigmo]

https://www.moutoulos.com/eshop/rf-pcb/ ... dband.html

[XXL]

Someone needs to make an MRF300 version. Iv tried many times with no luck.

[rigmo]

Someone needs to make an MRF300 version. Iv tried many times with no luck.

single 300W AN or AN BN? 600W

[XXL]

Single. Theres no need for 600w. 300 is fine.

[jvok]

Don't think so.... it looks like it's based on the Application Note from Motorola!

Its definitely the Enigma board, I know because I'm building a rig with one

mrf101 enigma.jpg

They've changed the silkscreen design a bit, beefed up the via stitching, removed the solder mask over the striplines (or covered them with silkscreen, hard to tell) and got it done in purple. But its clearly the same layout.

Here's the NXP reference design for comparison. Much smaller, no onboard bias regulator, no harmonic filter (just the matching network) and doesn't use striplines for matching.

mrf101 nxp.jpg

[rigmo]

[rigmo]

Single. Theres no need for 600w. 300 is fine.