[Amps] 3.5 kV 2A REGULATED Power Supply: Schematic ?

[Ian White, G3SEK]

R. Measures wrote:
>
>
>>GGLL wrote:
>>>I've been following this nice discussion, and want to ask something
>>>about Ian's post, ¿what's bad with running exciter at a lower level and
>>>to have the amplifier "seriously under-driven"? (it seems to be not
>>>only a lower power output matter).
>>>My modest guess is something related with operating points due to
>>>distortion considerations, is it?.
>>>
>>By "under-driven" I simply mean that you're not utilising the full power
>>output capability of the amplifier.
>>
>>With a good tetrode amplifier - the right tube, the right power supply
>>and the right operating conditions -  you can obtain low distortion
>>right up to the point of grid current, and even beyond that into class
>>AB2 if the grid dissipation allows it.
>>
>>Rich's line of argument is that you under-drive the amplifier in order
>>to ensure that it can never be driven into grid current. But he is
>>forced into that position because he advocates a high-impedance bias
>>supply that cannot handle grid current without changing voltage.
>
>**  Not quite.  I do not use a regulated bias supply because the tube
>manufacturers specify a maximum bias supply impedance for AB1.
>Typically, the max Z spec is 100k-ohms.

Not for the amplifier tetrodes we're talking about. Check the data 
sheets - it's a few K at most.

>>
>>Darn right I'm being stubborn about rejecting that! Not only will such a
>>bias supply *cause* IMD if there's any trace of grid current, but also
>>it is contrary to tube manufacturers' recommendations. For example,
>>Svetlana's recommendations for the 4CX800A are that "for stability, the
>>source impedance [of the grid bias supply] should not exceed 1K ohms."
>
>**  I read it and they are talking about stability vs, the RF source
>impedance of the exciter, not the impedance of the DC grid bias supply.

No they aren't. Those words about "vs the RF source impedance of the 
exciter" are a complete fabrication - they simply aren't there in the 
4CX800A data sheet.


>>
>>Simply design the bias supply to have a low impedance, so that it *can*
>>handle grid current (in either direction) if necessary. If the bias
>>voltage is held constant, there will be no sudden jump in IMD if the
>>tube is ever driven into grid current.
>
>**  Higher DC grid bias supply impedance means that less grid current is
>possible since the more drive applied the greater the negative DC voltage
>on the grid.  .  Does more grid-current equate to less IMD?

IMD is a function of both grid current and bias supply impedance. 
Compared to a high-impedance bias supply, then yes, a low-impedance 
supply conceivably can give less IMD with more grid current.



>>
>>Almost all "class AB1" amplifiers actually do nudge into a small amount
>>of grid current in order to develop an ALC signal. The tube data sheets
>>allow this.
>
>**  Eimac/CPI's do not.  Svetlana's do not.

Yes they do! For example, although the rated grid dissipation of the 
4CX350A and 4CX1000A is zero watts, the Eimac data sheets specifically 
allow peak grid currents of up to 2mA (4CX350A) or 5mA (4CX1000A) "for 
signal monitoring purposes". Other Eimac and Svetlana data sheets 
contain wording that is equivalent in engineering terms: a small amount 
of grid current is allowed.

How else do you think tetrode ALC circuits operate?

>>
>>
>>There is still a need to protect the grid against excessive power
>>dissipation due to overdriving. A high-impedance bias supply will
>>protect the grid, but it also creates IMD with any trace of grid
>>current.
>
>**  An unregulated grid bias supply causes a brief reduction in ZSAC
>whenever grid-I flows.  It would take a steady flow of substantial
>overdrive to reduce ZSAC enough to put it into the unlinear region
>provided the ZSAC had initally been set correctly with the screen supply.
>

That simply isn't true. The non-critical nature of ZSAC doesn't mean 
that grid bias can wander about as the drive level changes.

>
>>We should *all* reject that "solution" because it's unethical -
>>you're solving your own problem by dumping IMD onto everyone else.
>>
>**  The amplifier shown at:
>http://www.somis.org/pb.top.gif
>used a DC grid bias supply with an impedance of c. 15k-ohms.   I heard
>this amplifier on the air after I sold it and it had minimal IMD when it
>was producing 14 out.

You yourself stated that a 100W transceiver couldn't ever drive that 
monster amplifier into grid current, so your bias supply was never put 
to the test. The same will be true of the next one that you're about to 
build.

But please don't use those examples to tell the rest of us how to design 
smaller, legal, tetrode amplifiers.



-- 
73 from Ian G3SEK         'In Practice' columnist for RadCom (RSGB)
                            Editor, 'The VHF/UHF DX Book'
http://www.ifwtech.co.uk/g3sek