Plate Load Impedance / resistance

[W4RFM]

I am so happy to be able to get on this forum and ask questions, and therefore - hopefully - learn.
Recently the question of plate load impedance raised a question.  I always thought you divided the plate voltage by the plate current, and therein was your figure in ohms. However a member of this forum challanged me to re think that, and I learned somethin'.
I have discovered that their are two schools of thought, maybe the difference in thinking from the west coast to the east coast - (Talking about handbooks).
The 1956 Radio Handbook by Orr states Plate Load is calculated by dividing the plate voltage by two times the plate current.  The ARRL book says to take the plate voltage and multiply times 500, and divide that number by the plate current and get your answer. Funny thing is, it's the same answer!
I wonder if their is additional input on this output thing from the forum.

[k7yoo]

Modulating impedance is as you first described: E/I
RF impedance (for designing tank circuits) can be: plate voltage /1.5. or 2X plate current depending on the class of service. (C, AB1, etc).
Many people get these mixed up.

For example: a class C amp running 3600V and .6A has a modulating imp. of 6K.
The RF Imp. is 3K. In linear (AB) mode the amp has an RF impedance of around 4K. If you use old handbook data to design your linear amp tank you will be scratching your head as to why your amp is not loading right. Another example of this is the B&W 850 tank that never seems to be right for any of the rigs I have built. It is OK as a starter, but is not "plug & play". Ask me how I know.
They finally got it right on the 80's handbooks.

I use ELSIE to design all of my tank circuits.

If I am wrong feel free to correct me.
Skip

[AB2EZ]

I agree.

For those who wish to know where these rules of thumb come from... here is a simplified explanation.

The plate current flowing in the tube(s) is a series of pulses. The attachment illustrates this for the special case of Class B operation.

The series of plate current pulses can be represented by: its average value (the average plate current) + a sine wave at the fundamental frequency of operation (shown on the attachment) + a sine wave at twice the fundamental frequency of operation (not shown on the attachment) + sine waves at 3x, 4x, etc. the fundamental frequency of operation.

The pi network has a low impedance, except at the fundamental frequency of operation (e.g. 3.885 MHz if that is the frequency you are operating on).

Therefore, the only component of the plate current that will have to flow through a high impedance load is the fundamental frequency component of the plate current.

You want the impedance, Z, looking into the pi network (at the fundamental frequency of operation), to be set so that:

Z x the amplitude of the fundamental frequency component of the plate current = the value of the plate voltage (B+).

That way, the plate voltage will swing, sinusoidally, up to 2x (B+) and down to zero volts.

Therefore, you want to pick Z, the impedance looking into the pi network at the fundamental frequency, to be:

Z = B+ / F, where F= the amplitude of the component of the plate current at the fundamental frequency  


The above formula for Z is always true, for any class of operation.

Next, one has to use a little fancier math (calculus) to calculate what F is.

It turns out the for the case of Class B operation (shown in the attachment), the amplitude of the component of the plate current at the fundamental frequency [what I am calling F] is pi/2 x the average value of the plate current. Pi/2 is approximately 3.14/2 = 1.57. That is, for Class B operation: F= 1.57 x A, where A is the average value of the plate current


Therefore, for Class B operation,  the value of Z you want to use is:

Z= B+ / F = B+ / (1.57 x A)

... which is the formula that Skip provided in the post above.

For Class C operation, the amplitude of the fundamenal frequency component of the plate current (what I'm calling F) is closer to 2x the average value of the plate current (what I'm calling A).

Therefore, for Class C operation, you want the load impedance at the fundamental frequency to be: Z= B+ / F = B+ / (2 x A).

For any class of operation, you can find the correct formula for Z, in terms of the value of B+ and the average plate current (what I am calling A) by either: calculating F/A, or looking up what F/A is in a table, or using an on-line application to calculate it.

Stu

[W2PFY]

Nice discussion. please direct me to ELSIE. All I come up with is the cow.

[W4RFM]

I have been using this one:

http://www.qsl.net/wa2whv/pidownload.shtml

BM

[W8ACR]

Bob,

Here is the how I have it memorized in my simplistic RF challenged brain. As per the discussion above, I guess this would only apply to a class C amplifier.

1. plate impedance = plate voltage/plate current - used when determining modulator operating parameters.

2. plate load = plate voltage/2x plate current - used when determining component values of pi network tank circuits.

There is a very helpful explanation and a great chart of experimentally determined component values in the 1959 Orr handbook.

Ron

[KM1H]

Go to:

http://www.ifwtech.co.uk/g3sek/  and scroll down to Pi/Pi-L network design/analysis spreadsheet

It allows you to select any Class for A to C, and change whatever you want. IMO it is the best and easiest design package out there and even the manufacturers use it.

That site is loaded with other good info.

Carl

[G3UUR]

Carl,

Do G3SEK's calculations allow for tube non-linearity?

Dave.

[KM1H]

I didnt ask him

[G3UUR]

Well,

Plate load = Plate Voltage/2xDC Plate Current

does work well for Class C using real tubes according to Terman's book "Radio Engineering" so why would anyone want to use G3SEK's spreadsheet?

[KM1H]

Because Ians works perfectly well if you bother reading it

[G3UUR]

Why would I bother to waste my time reading Ian's information when the calculation above is trivial and works fine?

Duhhh....

[KM1H]

Well, if you were to actually build an amp you would find it most helpful as you need a working RF deck before applying modulation. The spreadsheet makes it easy for a Pi net or Pi-L.

For just the Rp part I do that in my head.

[G3UUR]

If you'd built as many amplifiers over the years as I have, Carl, calculating either a Pi or Pi-L tank network would be second nature to you by now. They become trivial calculations after a while and you ought to be able to do them quicker than you can bring up Ian's website, let alone scroll down and find the right plate impedance!

Dave.

[k7yoo]

Go to PIEL which is on the Tonne WB6BLD website. He has some really cool free and for sale stuff. ELSIE is a network design package on the same site.
http://www.tonnesoftware.com/

It is easier to use the available software than try to dig through handbooks--that may or may not have the correct info. Between the formulas and a cheap MFJ analyzer even I can put together tank circuits that play when I turn them on. This really came home to me about 15 years ago when I pre-tuned a PA, applied voltage, and barely had to tweak the TUNE cap to get full output. It wasn't luck. Before that it was all clip leads and aggravation.

[W2PFY]

Go to PIEL which is on the Tonne WB6BLD website. He has some really cool free and for sale stuff. ELSIE is a network design package on the same site.

  He changed his call sign to: W4ENE

My computers warns me that this site may be loaded with malware? Has anyone experienced any problems after a visit?

This is not an attack on the poster

I just sent him this email.

Hello James, When ever I try to look at anything published by you, my computer sends up a red flag suggesting that there is malware or worst on your site. This is not an attack on you but I thought you would like to know this.
 
 
73
 
Terry
W2PFY

[G3UUR]

Terry,

I'm sure Jim Tonne would want to know if there is a problem with his site. I certainly would.

I haven't looked at his website for several years. He was offering a free download of a limited version of ELSIE at the time, but it didn't really offer me anything new compared with my own crude DOS analysis program and was limited to quite a small number of components. The main thing missing from my program is a group delay facility and I have to use AADE's software for that these days. The AADE software is great for filters because of the library of sub-circuits available, but lacks the flexibility for other RF applications. I'll have to look at Jim's website again and see what's new.

Dave.

[K1JJ]

Just to confirm, I also got the message below and it was blocked by my software. I'm sure the owner will appreciate your heads up..

T

---------------------------


Website blocked by Trend Micro Worry-Free Business Security Services
Malicious website blocked
http://www.tonnesoftware.com/

Rating:
Dangerous

What You Can Do:
Contact your administrator about security settings on your network
Provide feedback to Trend Micro
Copyright © 2006-2011. Trend Micro™ Incorporated. All rights reserved.

[k7yoo]

I just clicked the link on your message and went right to his website with no issues. Maybe MY link was corrupted. Oh well--I have no explanation for that.
Skip

[W2PFY]

I just clicked the link on your message and went right to his website with no issues. Maybe MY link was corrupted. Oh well--I have no explanation for that.
Skip

It's not you Skip. No mater how I try to get to his site, I get the same response. Not to get into a software issue, but perhaps you don't have any filters on your computer to weed this out or if you do, it doesn't see the web site the way mine does. I use WOT which is an ad on for Firefox and it has saved my butt many times. I use WOT on all my search engines.

[KM1H]

If you'd built as many amplifiers over the years as I have, Carl, calculating either a Pi or Pi-L tank network would be second nature to you by now

Well, I'll let you enjoy your mine is bigger than yours nonsense. Im long beyond that stage.

I still recommend Ians site for anyone who wants to see the results of making on the fly changes to maximize what he has to work with. While not that many on here get on more than 75 it helps the others to move Q's and component values around for 160 and up on 10.

Its a popular program on forums where there are many all band builders for SSB/CW/digi.

[G3UUR]

Carl,

I'm not sure what the first line of your last reply means, but that's of no importance.

You were right to mention Ian's website on this thread, and I'm sure many will find it useful. But, having observed you baiting other people on this forum I just couldn't resist 'doing a carl' on you to see how long it would go on for. It wasn't as long as I thought it'd be.

Carl, you've been carled!

73,

Dave.

[KM1H]

I dont think so Dave, you have to be able to understand that earlier first sentence to know why

Carl