Re: [Amps] Transformer shop update

[craxd <craxd1@ezwv.com>]

Ian,

I'll go ahead and answer this here so folks can get a grasp of what were 
discussing. What all should know is were working together to try and 
come up with a one size fits all transformer for several tubes up to 
about a 4CX1500.



 Will accidentally replied to my e-mail on-list. Not a problem, but there 
 is one point that I'd like to clear up before taking this back to 
 e-mail...


> > Ian,
>  
[...]

> > The voltages I shown for the screen was the voltage needed for DC plus 
> > 10%. Does your regulator require more than 10% extra? I want to make 
> > these so they will work with everything, including your boards.
>  

 My Tetrode Boards require exactly the same as any other shunt regulator. 
 Perhaps the difference is that I looked at the screen voltage on the 
 'scope!

 ...and realised that the charts we normally use for 
 transformer/rectifier/smoothing design are not right for this particular 
 application.

 Those charts are based on *average* output voltage; but when you're 
 feeding into a voltage regulator, what matters is not the average but 
 the *instantaneous minimum* voltage. This occurs at the lowest point on 
 the ripple cycle, which is the average voltage minus 0.5 x  the 
 peak-to-peak ripple voltage at full load.

 If you have a lot of 100/120Hz ripple (too much winding resistance 
 and/or not enough smoothing C), the instantaneous minimum voltage can be 
 a long way below the average that you'd read from most charts.

/// Exactly, and the main reason for upping the wire size on the windings. I used 26 gage for the screen windings and for the bias, even though I had the bias rated for 125 mA. 26 AWG is rated for 361 mA at 750 cir. mils per ampere, and I derated that down to 250 mA. Anything smaller, and you can have problems. Most would use 28-1/2 AWG for 250 mA and something like 30 AWG for the bias winding. 30 is too bad to heat up and or break. I wont use anything under 28 gage on a power transformer. 26 AWG has 41.02 ohms per 1000 feet, 28-1/2 has 73.2 ohms per 1000 feet. The resistance almost doubles between the two. To top it off, the cost of the wire is almost the same between the two sizes per pound.

/// I want to see this transformer run cool as a cucumber even though someone is hammering an amp. This has to be rated as CCS and not ICAS because the supply runs the heater and bias continuously. ICAS only works for plate voltage and or screen voltage derived from it.

 If this point is ignored, the voltage regulator may 'drop out' on the 
 negative peaks of input ripple, so you won't have a constant regulated 
 DC output any more - it will have negative blips at 100 or 120Hz. This 
 is a feature of *all* shunt regulators, and series regulators too.

 To stay clear of the problem, the transformer AC voltage has to be 
 higher than you thought it oughter be. There aren't any ready-made 
 design charts for this, and I had to draw my own using the 'PSU 
 Designer' software.

/// Yes it has to be higher. The rectifier circuit used has a lot to do with this also. Most are using either a full wave bridge, or a full wave center tapped circuit. Again, not having enough capacitance is a big no-no in screen and bias circuits. In either, in my opinion, the stiffer, the better.

 -- 
73 from Ian G3SEK 'In Practice' columnist for RadCom (RSGB) 
http://www.ifwtech.co.uk/g3sek Best & 73's Will Matney


_______________________________________________
Amps mailing list
Amps@contesting.com
http://lists.contesting.com/mailman/listinfo/amps