KSB2 filter, IF signal path impedances, T1/T2 turns ratio

[John Magliacane]

Hi Folks.

I have several questions regarding the impedances involved in the K2's
I.F., and their impact on the design of the KSB2's filter and matching
transformers.  I poked around the K2 mail archives and saw a similar
question posted some time ago
(http://www.ac6rm.net/mailarchive/html/elecraft-list/2003-05/msg01686.html),
but found no definitive answer or information.


Several assumptions that I believe are true:

* The impedance of the K2's I.F. signal path is 150 ohms.  The input
  impedance of the MC1350 I.F. amplifier chip is around 2000 ohms.
  This is matched to 150 ohms through T7, a toroidal transformer
  with a 20:5 turns ratio (16:1 impedance transformation).

* The impedance of the crystal filter in the KSB2 is 1500 ohms.
  T1 on the KSB2 board transforms the 150 ohm I.F. impedance
  to 1500 ohms (22:7 turns ratio, 10:1 impedance ratio).  T1
  is terminated either by R11 (150 ohms) on the KSB2 board,
  or the post mixer I.F. amplifier/noise blanker (150 ohms),
  depending on whether the transceiver is transmitting or
  receiving.

* T2, on the OTHER side of the crystal filter, has a 22:4 turns
  ratio, yielding a 30:1 impedance ratio (a 50 ohm impedance,
  rather than 150).

* The KSB2's crystal filter is operated in "opposite directions"
  in transmit and receive.  On transmit, the input is on the side
  of T2, and on receive, the input is on the side of T1.

* On transmit, the balanced modulator provides a proper 1500 ohm
  source impedance for the crystal filter.  R11 provides the proper
  150 ohm termination after T1.

  --- HOWEVER ---

* On receive, the output of the KSB2's crystal filter is mismatched
  by a factor of 3 to 1 (T2's output is 50 ohms, and it is feeding
  a 150 ohm I.F. amplifier).

* The KI6WX +10db modifications appear to treat the output of T2
  as being 150 ohms (or at least something higher than 82 ohms,
  the value of the emitter follower's emitter resistor), rather
  than 50 ohms, yielding a significant impedance mismatch on
  transmit.


Impact:

* Many of the crystal filter response measurements, such as those
  conducted using programs such as Spectrogram and Baudline, have
  been performed with the K2 in RECEIVE mode only.  Since the filter
  appears to be terminated differently depending on whether the K2
  is in transmit or receive mode, the results may not tell the whole
  story.


Confession:

I built my K2 last year and performed the KI6WX KSB2 filter and
+10dB gain modifications.  The improvement in frequency response
on receive was substantial, but I have never been satisfied with
my on-air audio quality (despite reports to the contrary).

I ran sweeps of the TX chain and found my transmitted frequency
response to be around 2 kHz after performing the 2.5 kHz mod.
Results are posted at http://www.qsl.net/kd2bd/ksb2.html

I put the results of these tests out of my mind for the past
10 months.

Yesterday I checked the batch of 82 ohm resistors I used for the
+10dB gain modification and found they were all measuring between
120 to 140 ohms!  Reflected through T2, this would yield a filter
source impedance of about 6000 ohms, a significant mismatch for
a 1500 ohm filter.

The higher-than-normal source impedance might explain the
narrow filter bandwidth I am experiencing on transmit (but
not on receive).

Before I change the defective resistor, I would like to get a better
understanding of the true terminating impedance of the filter.  T2's
turns ratio might suggest that even an 82 ohm resistor is too high
for proper termination.  If the KSB2 is truly operating in a 150 ohm
I.F. signal path, why aren't T1 and T2 wound with the same turns ratio?


Possible solutions:

* Decrease the resistance of the 82 ohm resistor to provide a
  closer match to the 50 ohm impedance provided by T2.

* Decrease T2's turns ratio to that of T1 to bring the output
  impedance of the filter from 50 ohms to 150 ohms for a better
  match to both the I.F. amplifier on receive, and the KI6WX
  +10dB gain modification on transmit.


Suggestions?  Comments?


Thanks in advance.


73, de John, KD2BD
K2/100 #3563

=====
Visit John on the Web at:

        http://www.qsl.net/kd2bd/

__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around
http://mail.yahoo.com 
_______________________________________________
Elecraft mailing list
Post to: [hidden email]
You must be a subscriber to post to the list.
Subscriber Info (Addr. Change, sub, unsub etc.): http://mailman.qth.net/mailman/listinfo/elecraft   
Help: http://mailman.qth.net/subscribers.htm
Elecraft web page: http://www.elecraft.com

[Jim Wiley-2]

 John -


Just a random thought.  Since you mentioned that several resistors
marked as 82 Ohms came in around 130 Ohms:  Have you verified that your
Ohmmeter is giving you a correct reading?  For example, have you checked
it with other  (new) resistors marked with different values, or compared
the readings by using another instrument?  New resistors will typically
be reasonably close to their marked values, and usually about half of
the allowed tolerance.  While it is not impossible for a whole batch of
resistors to be significantly out of  tolerance,  it is unusual.  It is
also possible for only one range of an Ohmmeter to be wrong, and all the
rest are OK.   With analog instruments, weak meter batteries can lead to
confusing results.


I have fooled myself occasionally by overlooking one of the "basics".


- Jim, KL7CC


John Magliacane wrote:

>Hi Folks.
>
>I have several questions regarding the impedances involved in the K2's
>I.F., and their impact on the design of the KSB2's filter and matching
>transformers.  I poked around the K2 mail archives and saw a similar
>question posted some time ago
>(http://www.ac6rm.net/mailarchive/html/elecraft-list/2003-05/msg01686.html),
>but found no definitive answer or information.
>
>
>Several assumptions that I believe are true:
>
>* The impedance of the K2's I.F. signal path is 150 ohms.  The input
>  impedance of the MC1350 I.F. amplifier chip is around 2000 ohms.
>  This is matched to 150 ohms through T7, a toroidal transformer
>  with a 20:5 turns ratio (16:1 impedance transformation).
>
>* The impedance of the crystal filter in the KSB2 is 1500 ohms.
>  T1 on the KSB2 board transforms the 150 ohm I.F. impedance
>  to 1500 ohms (22:7 turns ratio, 10:1 impedance ratio).  T1
>  is terminated either by R11 (150 ohms) on the KSB2 board,
>  or the post mixer I.F. amplifier/noise blanker (150 ohms),
>  depending on whether the transceiver is transmitting or
>  receiving.
>
>* T2, on the OTHER side of the crystal filter, has a 22:4 turns
>  ratio, yielding a 30:1 impedance ratio (a 50 ohm impedance,
>  rather than 150).
>
>* The KSB2's crystal filter is operated in "opposite directions"
>  in transmit and receive.  On transmit, the input is on the side
>  of T2, and on receive, the input is on the side of T1.
>
>* On transmit, the balanced modulator provides a proper 1500 ohm
>  source impedance for the crystal filter.  R11 provides the proper
>  150 ohm termination after T1.
>
>  --- HOWEVER ---
>
>* On receive, the output of the KSB2's crystal filter is mismatched
>  by a factor of 3 to 1 (T2's output is 50 ohms, and it is feeding
>  a 150 ohm I.F. amplifier).
>
>* The KI6WX +10db modifications appear to treat the output of T2
>  as being 150 ohms (or at least something higher than 82 ohms,
>  the value of the emitter follower's emitter resistor), rather
>  than 50 ohms, yielding a significant impedance mismatch on
>  transmit.
>
>
>Impact:
>
>* Many of the crystal filter response measurements, such as those
>  conducted using programs such as Spectrogram and Baudline, have
>  been performed with the K2 in RECEIVE mode only.  Since the filter
>  appears to be terminated differently depending on whether the K2
>  is in transmit or receive mode, the results may not tell the whole
>  story.
>
>
>Confession:
>
>I built my K2 last year and performed the KI6WX KSB2 filter and
>+10dB gain modifications.  The improvement in frequency response
>on receive was substantial, but I have never been satisfied with
>my on-air audio quality (despite reports to the contrary).
>
>I ran sweeps of the TX chain and found my transmitted frequency
>response to be around 2 kHz after performing the 2.5 kHz mod.
>Results are posted at http://www.qsl.net/kd2bd/ksb2.html
>
>I put the results of these tests out of my mind for the past
>10 months.
>
>Yesterday I checked the batch of 82 ohm resistors I used for the
>+10dB gain modification and found they were all measuring between
>120 to 140 ohms!  Reflected through T2, this would yield a filter
>source impedance of about 6000 ohms, a significant mismatch for
>a 1500 ohm filter.
>
>The higher-than-normal source impedance might explain the
>narrow filter bandwidth I am experiencing on transmit (but
>not on receive).
>
>Before I change the defective resistor, I would like to get a better
>understanding of the true terminating impedance of the filter.  T2's
>turns ratio might suggest that even an 82 ohm resistor is too high
>for proper termination.  If the KSB2 is truly operating in a 150 ohm
>I.F. signal path, why aren't T1 and T2 wound with the same turns ratio?
>
>
>Possible solutions:
>
>* Decrease the resistance of the 82 ohm resistor to provide a
>  closer match to the 50 ohm impedance provided by T2.
>
>* Decrease T2's turns ratio to that of T1 to bring the output
>  impedance of the filter from 50 ohms to 150 ohms for a better
>  match to both the I.F. amplifier on receive, and the KI6WX
>  +10dB gain modification on transmit.
>
>
>Suggestions?  Comments?
>
>
>Thanks in advance.
>
>
>73, de John, KD2BD
>K2/100 #3563
>
>=====
>Visit John on the Web at:
>
> http://www.qsl.net/kd2bd/
>
>__________________________________________________
>Do You Yahoo!?
>Tired of spam?  Yahoo! Mail has the best spam protection around
>http://mail.yahoo.com 
>_______________________________________________
>Elecraft mailing list
>Post to: [hidden email]
>You must be a subscriber to post to the list.
>Subscriber Info (Addr. Change, sub, unsub etc.): http://mailman.qth.net/mailman/listinfo/elecraft   
>Help: http://mailman.qth.net/subscribers.htm
>Elecraft web page: http://www.elecraft.com
>
>  
>

_______________________________________________
Elecraft mailing list
Post to: [hidden email]
You must be a subscriber to post to the list.
Subscriber Info (Addr. Change, sub, unsub etc.): http://mailman.qth.net/mailman/listinfo/elecraft   
Help: http://mailman.qth.net/subscribers.htm
Elecraft web page: http://www.elecraft.com