Question about tube xmitter PI Filter please

[KC9KEP]

Hello all,

I'm building a simple one-tube (circa 1941) QRP rig.

(I've added an isolation xformer, fuse, & 3-wire AC line cord for safety.)

A link of the schematic is here:

http://www.bignick.net/temp/1941_QRP_rig.bmp

My question is about the PI filter in the output section.

For some reason, they have placed the DC blocking capacitor AFTER the PI filter.
This means that both tuning capacitors and the tapped inductor are all at the B+ voltage.
In most schematics that I've seen, the blocking capacitor is wired BEFORE the PI network,
(and an RF choke is usually wired from the output of the PI filter to ground.)

Can I move the DC blocking capacitor to a location BEFORE the PI filter without
affecting performance?

It would seem to me that the RF would pass thru the 0.001 uF DC blocking capacitor
an on to the filter, but I'm not sure.

For that matter, if I'd like to do some circuit analysis of the PI filter, what input
output impedances would I use for calculations?  I would assume that input impedance
would be the plate impedance of the tube, and the output impedance ideally 50 Ohms?

Thanks!

73

--KC9KEP

[AC5UP]

Can I move the DC blocking capacitor to a location BEFORE the PI filter without affecting performance?

If you're as old as dirt you'll recognize the name Lew McCoy.

http://www.af4k.com/imag/ham/6gw8_2.gif

Yup.

.01 blocking cap ahead of the Pi network, no B+ on the antenna line. (for the rest of the pages take the URL above and sub 6gw8_1, _2, _3, _4)

...Ya' gotta' love the #48 or #49 lamp used as a tuning indicator.

[N3QE]

You have to put yourself into the mindset of the folks who built rigs that had "hot chassis". In THAT mindset, the right place to put the antenna isolation capacitor is indeed right before the antenna terminal - because everything else in the rig may in fact be "hot" in the 120VAC sense (which will be far "hotter" than the couple watts you'd get out of the rig RF-wise.)

"Hot chassis" concepts can certainly seem backwards looked at from a more modern perspective. But generally the external terminals was where the isolation had to be done (and in fact consumer hot chassis stuff had components specifically rated for such isolation duties) and extra isolation at interior stages seemed superfluous.

Yes, you can move the capacitor before the pi network if you have an isolation transformer. (Not sure if a 1941 handbook called it a "pi filter" or a "pi network" but ... more modern texts would use the word "network" because it's primary goal is to do impedance matching; the low pass filtering is useful too!)

[W8JI]

There are very serious safety flaws with that design.

First, the switch is in the "chassis" lead. The switch should be in the hot lead.

Second, it has no fuse. It absolutey needs a small fuse (less than 1 amp fast blow 125V) on BOTH leads.

Third, the CW key could have 120 VAC on it. It is tied to one side of the power line. So can the chassis.

IMO unless you are really careful with what outlet you use and use a three wire plug, you better use an isolation transformer.

As for the blocking cap, not having a blocking cap at the tube end of the network doubles the peak voltage the tank input cap sees, and does much worse than that to the loading cap. This is because 170 volts DC is superimposed across the caps.

The input of the tank is approximately E^2 / P where P is power output and E is plate voltage downward swing. So with 150 plate volts and 5 watts output you'd probably have 120 volts plate voltage downward swing, and about 2900 ohms at the tube side of the network. This is a rough approximation.

[KC9KEP]

Thank you all for your replies!

Yes, to reiterate:

1.)  I am using a 3-prong AC line cord, the safety ground lead is connected to the chassis
2.)  I have added an AC line fuse
3.)  I have added a Triad N51-X isolation transformer
4.)  I have moved the DC blocking capacitor to a location prior to the output PI filter

The AC line still gets routed to the 117L7's filament, however, the chassis ground
connection and the plate of the rectifier are both sourced from the secondary of the
isolation transformer.

I made a successful QSO with this transmitter this past weekend on 7049 KHz. 
It was during the afternoon and "only" 104 miles away.
But I'm completely blown away!  Only 2 watts (less than a nightlight) and 104 miles!
(I was using an end-fed wire antenna, home brew antenna tuner and home brew tube receiver.)

Thank again guys,

-- KC9KEP

[W8JI]

Thank you all for your replies!

Yes, to reiterate:

1.)  I am using a 3-prong AC line cord, the safety ground lead is connected to the chassis
2.)  I have added an AC line fuse
3.)  I have added a Triad N51-X isolation transformer
4.)  I have moved the DC blocking capacitor to a location prior to the output PI filter

The AC line still gets routed to the 117L7's filament, however, the chassis ground
connection and the plate of the rectifier are both sourced from the secondary of the
isolation transformer.

I made a successful QSO with this transmitter this past weekend on 7049 KHz. 
It was during the afternoon and "only" 104 miles away.
But I'm completely blown away!  Only 2 watts (less than a nightlight) and 104 miles!
(I was using an end-fed wire antenna, home brew antenna tuner and home brew tube receiver.)

Thank again guys,

-- KC9KEP

You are a smart and wise fellow.

I know people who would not have used an isolation transformer or a fuse.

By the way, if you moved the capacitor you might want to consider a leak resistor of 100k ohms or better yet a safety choke from the loading cap to ground. This will prevent leakage or normal charging currents from charging the variable caps.