Actually, the MOS upgrade is almost a necessity.
The 2078 is not like it used to be and even as the years pass, it's not being used as often anymore.
Instead the Big 3 (Galaxy, Cobra and Uniden) they used the atypical 520 - only they "sand it off" so you don't know what is under the hood.
Below is a pic of the 687 those 6-pin Unidens that they say are their replacements to the PC-6X series...
Pay attention to the "chip resistor" markings...they provide you a clue...
Hint: 8.2K to as low as 6.8K and firmly paired with and use 4.7K in a voltage divider configuration.
Hope this helps
(EDIT - To Add)
The 2078 is a Bipolar. To still use it shows the "Backward" compatibility of some of these MOS upgrades. The problem lies in the support parts you'll need to help the swap work.
So if you're looking for 40-watts, DONT! I mean that, DO NOT try for that - better to use the MOS upgrade to develop a better punch and signal to audio envelope ratio of what many call 4:1 that being a fully modulated RF signal is 4 times the original carrier power - but that's if you trust your meter to be a PEP-reading-worthy meter.
If you get 4 watts carrier and it swings to 12~16 watts on a typical SWR/Power Combo Meter - you're fine.
If you go too extreme you can wind up doing more damage than it's worth.
Here's some examples...
They (read CB'ers wanting to make the conversion from Bipolar to MOSFET using the AN2030 methods) used to make their own self-powering Bias - which in light of the audio power presence - the self-powered ones make better sense than to use a simple voltage divider.
Here's what I guide myself with in MOSFET...
Why, well a voltage divider circuit is power flowing ACROSS your RF signal to your Gate (on MOS) or Base (on the 2078) ... Note the schematical references above. Bipolars can "self bias" (See R56 22 ohm Ferrite Bead and R53 1 ohm...)
The "self bias" method works due to the Bipolar and it's structure - it has a DIODE (well kinda - work with me here its a semiconductor) with DIRECT access across the BASE to Emitter/Collector - hence the Diode - and that means it can rectify RF. Yes, rectify - even if it's power level is TOO LOW to pass the built-in "voltage Drop" the Bi-polar exhibits - so RF can rectify to DC and once the Base conducts - power flows across the Collector/Base/Emitter can then provide the power needed to "Stay on" - the resistors provide the mechanism to stop this flow of power once the RF power stops flowing into the circuit (when you unkey to RX).
But MOS needs a little help, so you can "self bias" using a few extra parts - Zener or not, a diode is needed - ZENER would provide a "crowbar" to help keep the 520/13N10/2030 working because it would keep Gate bias low but cannot prevent excessive power from blowing up the gate and perforating your work into a dead short or popped part...
Hope this graphic helps show what I'm talking about here...
The Voltage Divider method is used in these cases.
In the old days Bipolar (2078) uses the Base as part of a Diode Junction NPN=the P=the Base and as trickle voltage can keep the transistor in the on state (Class C onto AB2) - The MOS isn't set up the same way - you need a P-N junction to form the rectification and a means to keep power flowing because the "Gate" is insulated (read very high resistor value) to the Drain and Source.
@loosecannon - a Cobra 19 Graphic that may help...
So hope the above hints helps - just don't overdo it and destroy your work. The IRF520 works - ok - as a Drop-in. But thre 13N10 have a power curve to offset it's limitations and couple nicely with the output networks used in todays radios.
