[BC] Xmitters at aol.com

Wed Apr 11 11:39:35 CDT 2007

I think that most people do realize the importance of having the correct 
source impedance for transformers. This would be one reason why a 6 db 
pad was standard practice when driving telco lines.

R

Dana Puopolo wrote:

>What many don't realize about transformers is that they have to be sourced
>properly. For example, if you connect the average Distribution amp directly to
>a transformer wired 1:1, you will be sourcing the unit with about 200 ohms
>(the average DA uses 100 ohm build out reisitors on each output). This will
>reflect through into the secondary, which will now also look like a 200 ohm
>source (give or take, because the wire used to wind the transmformer has a bit
>of resistance). If telco has set up an equalized circuit using an oscillator
>with a 600 ohm output    
>impedance, then sourcing this transformer with this DA (or many other
>compressors which have low output impedances) will cause frequency response
>errors in the program circuit. This is why I put a pair of 270 ohm resistors
>in each output leg of a Compellor directly feeding a program line for example.
>The compellor has a very low output impedance (about 75 ohms-read the book)
>and you need to "build up" the source impedance to 600 ohms to properly match
>the line.  Also, it's essential to put some kind of pad between two
>transformers connected back-to-back lest they screw up each other's response.
>
>-D
>
>
>------ Original Message ------
>Received: Wed, 11 Apr 2007 08:06:23 AM EDT
>From: RichardBJohnson at comcast.net
>To: Broadcasters' Mailing List <broadcast at radiolists.net>, "Broadcasters'
>Mailing List" <broadcast at radiolists.net>Cc: "Dana  Puopolo"
><dpuopolo at usa.net>
>Subject: Re: [BC] Xmitters at aol.com
>
>
>No! No! No!
>
>The transformer does __NOT__ represent an impedance in itself. The transformer
>is specified to work in a circuit with certain impedances. To do this, it was
>designed so that its own reactances and resistance (complex impedances) are
>inconsequential for the range of frequencies and range of impedances in which
>it was to operate. The transformer does __NOT__ have a bunch of "300 ohm"
>windings that can be connected in series or parallel. It has four identical
>windings that can be connected as desired. This gives one the following turns
>ratios: 1:1, 2:1, and 3:1. The impedance ratios vary as the square of the
>turns ratio. Therefore, if a transformer was fed from a 600 ohm source, the
>possible output impedances are 600, 150, and 66.6 ohms. To best distribute the
>internal reactances, when the 1:1 configuration is used, pairs of windings are
>series-connected. When 2:1 is used, the input pair is series-connected, and
>the output pair is connected in parallel. These are the connections that 
>would be used in program channels. For mid frequencies, there would be no
>difference in impedance transformation for the 1:1 configuration if the
>windings were paralleled. However, the transformer would not have enough
>inductance to properly pass frequencies below about 100 Hz without distortion
>(the volts-per-turn would be too high). That's why the higher impedance side
>has the windings in series. The lower impedance side ends up with the same
>volts-per-turn because of the impedance transformation. To correctly
>distribute the leakage reactance (for high frequency response) the low
>impedance windings are paralleled.
>
>To correct other misconceptions about wire transmission, one only  needs to
>look at the limits. Since an audio transmission line is a distribution of
>series inductance, series resistance, shunt capacitance, and some shunt
>(dielectric loss) resistance, we have relatively low Q delay line. If a delay
>line has identical elements, or is physically long enough so that its
>distribution "seems like" a bunch of identical elements, then the following
>occurs when you muck with its source impedance: If the source impedance is
>equal to its load impedance, and the delay line is matched, the line acts like
>a lowpass filter with a smooth (single pole) roll off at its high frequency
>limit. If its input is mismatched by feeding if from a lower impedance, the
>frequency response with contain a peak in amplitude just before its roll-off.
>Also, the roll-off with be faster, containing an additional pole that the
>mismatch creates. This characteristic can be used to "bring up" the high end
>when attempting
> to equalize the line. If the line is fed from
>an impedance that is higher than its designed impedance, there will be a
>high-frequency roll-off that occurs  at a lower frequency than the natural
>lowpass characteristics of the line. Playing with the load impedance does
>nearly the opposite, a lower impedance starts an early roll-off, and a higher
>impedance causes peaking.
>
>In the limit of a zero source impedance, and an infinite terminating
>impedance, we have a resonant
>circuit as f = 1 /  2pi * sqrt(LC)  where the L is the series inductance and
>the C is the shunt capacitance.
>The wire-line has become a pi-network.
>
>
>--
>Cheers,
>Richard B. Johnson
>Read about my book
>http://www.AbominableFirebug.com
>
>
> -------------- Original message ----------------------
>From: "Dana  Puopolo" <dpuopolo at usa.net>
>  
>
>>You can use ONE winding to source 300 ohms. Put a 300 ohm resistor across
>>    
>>
>the
>  
>
>>ubused winding.
>>
>>-D
>>
>>
>>------ Original Message ------
>>Received: Tue, 10 Apr 2007 09:05:25 AM EDT
>>From: RichardBJohnson at comcast.net
>>To: "Broadcasters' Mailing List" <broadcast at radiolists.net>,
>>broadcast at radiolists.netCc: K7qa at aol.com
>>Subject: Re: [BC] Xmitters at aol.com
>>
>>
>>The impedance of various telephone cables is well documented. 16-gauge
>>spiral-four toll entrance cable is 540 ohms -j460, 22 AWG emergency cable
>>    
>>
>is
>  
>
>>464 ohms -j449, paper insulated (inter-city toll cable) is 910 ohms for 26
>>AWG, 22 AWG is 756 ohms, 19 AWG is 453 ohms, 16 AWG  is 320. All from Page
>>823, Wire Transmission, Reference Data for Radio Engineers, fourth edition.
>>Nothing is close to 150 ohms. It is well known that the series 500
>>    
>>
>"standard
>  
>
>>black telephone" was called the "500" because its hybrid (anti sidetone
>>network) was designed
>>for 500 ohms. Historically, the telephone lines have been called "500 ohm"
>>circuits.
>>
>>Also, there are no 300-ohm windings on a 111-C. The transformer consists of
>>four identical windings that are electrostatically shielded so they can be
>>connected in any manner. The usual (read correct) connection is for two
>>windings to be series-connected for the output, and two windings to be
>>series-connected for the output.
>>
>>See http://www.oldradio.com/current/111c.html , It shows the two usual
>>    
>>
>methods
>  
>
>>of connection. Note that both the line-side and the drop side can be
>>    
>>
>connected
>  
>
>>either in series or parallel. Don't be fooled by the "typical use" note.
>>    
>>
>It's
>  
>
>>for "short" i.e., unequalized circuits.
>>
>>I would guess that your response was either a test to see if I was still
>>awake, or not very well thought out!
>>
>>--
>>Cheers,
>>Richard B. Johnson
>>Read about my book
>>http://www.AbominableFirebug.com
>>
>>
>> -------------- Original message ----------------------
>>From: K7qa at aol.com
>>    
>>
>>> 
>>>In a message dated 4/9/2007 6:41:47 AM Pacific Daylight Time,  
>>>RichardBJohnson at comcast.net writes:
>>>
>>>The  circuit feeding the telco pair should not be connected as 150 ohms  
>>>unless
>>>there has been a problem obtaining the required frequency response.  
>>>
>>>
>>>Most standard balanced telco multiconductor trunk lines used for many
>>>      
>>>
>years 
>  
>
>>>had an internal surge impedance of 135 to 150 ohms at audio frequencies.
>>>      
>>>
>The
>  
>
>> 
>>    
>>
>>>111C had individual 300 ohm windings for a good reason.  Connect the  111C
>>>      
>>>
>
>  
>
>>>transformer split windings in parallel to achieve 150 ohms on the line
>>>      
>>>
>side 
>  
>
>>and 
>>    
>>
>>>in series for the 600 ohm equipment side. 
>>> 
>>>tm
>>>
>>>
>>>
>>>************************************** See what's free at
>>>      
>>>
>>http://www.aol.com.
>>    
>>