[Elecraft] IMD and supply voltage

[Don Wilhelm]

Frank,

I think the "issue" is one of some hams wanting to make what is good 
already into something better.
If you look closely you should discover that this is a situation common 
to *all* 100 watt class amplifiers that are supposed to run on a nominal 
'12 volt' supply.  That voltage level is chosen because it is common for 
vehicle batteries, so most transceivers are designed around that voltage 
- the K3S is just one of those.

Using an increased voltage power supply will drastically reduce the 
transmit IMD, and that is true for all those 100 watt class amplifiers.
Bottom line is that I don't think it will be "fixed" in the K3S - I 
don't think it can be fixed in any of the 100 watt class solid state 
amplifiers.

So the thing to do is to run the power supply voltage up to near 15 
volts (but not above), and do not operate with "all knobs to the right".
Many hams seen to want to squeeze all the power output that they can 
from a transceiver, so you will put a cleaner signal on the air if you 
set the power to 100 watts rather than trying to squeeze that last 20 
extra watts out of an amplifier.

73,
Don W3FPR

On 10/29/2015 11:09 AM, frank wrote:
> This subject keeps cropping up.  Is this a serious issue ?
>
> I have heard that IMD improves by 10db with a one volt increase in supply voltage.  Is this really the case or just a measurement error?  10db is a large difference.  One volt is a small variation.
>
> What is the minimum acceptable IMD?  How linear does the amp need to be?  If 36db is better than 26db is 100db even better or just overkill?
>
> Are other solid state amplifier products similarly effected?
>
> What is the IMD performance of the K3S at 10 watts over a supply voltage variation from 11 volts to 14.7 volts?  I ask because I believe those ten watts are generated by Dmos devices as well.  Do these devices only misbehave with varying supply voltage at higher (100w) power levels and lower supply voltages?
>
> Should I wait to purchase a K3S until this issue is resolved?  (Assuming this is a valid issue)
>
>