UK Vintage Radio Repair and Restoration Powered By Google Custom Search Vintage Radio and TV Service Data

Go Back   UK Vintage Radio Repair and Restoration Discussion Forum > Other Discussions > Homebrew Equipment

Notices

Homebrew Equipment A place to show, design and discuss the weird and wonderful electronic creations from the hands of individual members.

Closed Thread
 
Thread Tools
Old 13th Aug 2016, 5:12 pm   #21
Wendymott
Octode
 
Wendymott's Avatar
 
Join Date: Jun 2013
Location: Huddersfield, West Yorkshire, UK.
Posts: 1,782
Default Re: SSB Exciter... HELP !!

Hi John. All noted and thanks. As I have the USB and LSB filters it is a pity not to use them... I have two pairs actually. I bought 2 sets of 3 from Ebay a while back.
I will persevere with the two filter version and get the termination right. It may mean two separation buffers to isolate the filters. but with SMD components, not too much real estate will be needed.
__________________
Should get out more.

Regards
Wendy G8BZY
Wendymott is offline  
Old 13th Aug 2016, 6:45 pm   #22
Jon_G4MDC
Nonode
 
Jon_G4MDC's Avatar
 
Join Date: Nov 2015
Location: Welwyn Garden City, Hertfordshire, UK.
Posts: 2,013
Default Re: SSB Exciter... HELP !!

OK Wendy. Most SSB crystal filters are about 500 Ohm termination.
I like to test and drive them in a 50Ohm situation when possible.
You could try

..................100pF-----------9MHz Filter------------------100pF
50R In ------..............|...............|..............|.... ................----------50R Out
................220pF.....4.7uH Var....|..........4.7uH Var.......220pF
....................|______|_______|_______|______ ___|

{Ignore the dots - the forum software hates spaces!! Best I could do intead.
------is a wire while .......is just thin air }

The 220pF is slightly large, ideally 210pF. This has worked for me.

Tuning the Var inductors is enough to let the filter see slight capacitive
or slight inductive match. This can affect passband shape and allows adjustment for optimum.

Pay attention to grounds and don't let the input and output sides couple.
Canned Toko 10K inductors or similar are good!

Last edited by Jon_G4MDC; 13th Aug 2016 at 6:55 pm.
Jon_G4MDC is offline  
Old 14th Aug 2016, 12:50 pm   #23
Wendymott
Octode
 
Wendymott's Avatar
 
Join Date: Jun 2013
Location: Huddersfield, West Yorkshire, UK.
Posts: 1,782
Default Re: SSB Exciter... HELP !!

Hi Jon. I think I get your meaning.. please see attached. Or if not please advise. I have a number of Toko coil formers, so thats no problem.
I did think of separate fet buffers and select the outputs with the Teledyne.
But as you say its matching that is the problem.
Attached Thumbnails
Click image for larger version

Name:	Filter matching cct.jpg
Views:	116
Size:	18.1 KB
ID:	128666   Click image for larger version

Name:	Filter pcb bottom side.jpg
Views:	88
Size:	101.2 KB
ID:	128677  
__________________
Should get out more.

Regards
Wendy G8BZY
Wendymott is offline  
Old 14th Aug 2016, 4:15 pm   #24
Jon_G4MDC
Nonode
 
Jon_G4MDC's Avatar
 
Join Date: Nov 2015
Location: Welwyn Garden City, Hertfordshire, UK.
Posts: 2,013
Default Re: SSB Exciter... HELP !!

Hi Wendy, that had a lot of potential to be misunderstood.
Looks like it was.
Here's what I meant, mirror it to the output side.
Attached Thumbnails
Click image for larger version

Name:	Captap_opt.jpg
Views:	114
Size:	25.9 KB
ID:	128688  
Jon_G4MDC is offline  
Old 14th Aug 2016, 5:14 pm   #25
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: SSB Exciter... HELP !!

Not sure how you arrived at those numbers, but a casual glance that network won't give very good results with the values posted. Assuming that the frequency is 9MHz and you want to match 500R to 50R then the reactance of the shunt inductor will need to be 167 ohms or less with that capacitive tap network. Obviously, the cap values will need to change as well

So the 4.7uH inductor looks to be too big in value as it has a reactance of about 267 ohms at 9MHz.

For a capacitive tap it will typically have a much lower reactance than 167 ohms. So I think the shunt inductor needs to be less than about 3uH.

With a capacitive tap you can go much lower than 3uH and gain some extra stopband selectivity a few MHz away. 4.7uH looks to be way too large IMO.

Try changing the 100pF to 150pF and change the 220pF to 180pF and use an inductor around 2.65uH ? That should give a good match to 500R at 9MHz?
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Old 14th Aug 2016, 6:43 pm   #26
Jon_G4MDC
Nonode
 
Jon_G4MDC's Avatar
 
Join Date: Nov 2015
Location: Welwyn Garden City, Hertfordshire, UK.
Posts: 2,013
Default Re: SSB Exciter... HELP !!

and when at parallel resonance what part does XL play?
Jon_G4MDC is offline  
Old 14th Aug 2016, 7:00 pm   #27
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: SSB Exciter... HELP !!

I'm not sure what you mean by your question but this looks to me to be a simple matching task from 50R to 500R at 9MHz using a capacitive tap network. If so, the network you posted up will deliver a VSWR somewhere in the region of 3:1 at 9MHz if the filter looked like a 500R resistive load. Without your matching network the VSWR would be 10:1 but I don't see 3:1 as an acceptable design target for the VSWR.

I'd expect to see a slightly higher insertion loss and more passband ripple because of this and also a poor group delay response in the filter.

Is there more to it than this? Otherwise, I think you need to recheck your design calculations?

The network changes I suggested would improve the VSWR (into a flat 500R load) to less than 1.1:1.
__________________
Regards, Jeremy G0HZU

Last edited by G0HZU_JMR; 14th Aug 2016 at 7:08 pm.
G0HZU_JMR is offline  
Old 14th Aug 2016, 7:09 pm   #28
Jon_G4MDC
Nonode
 
Jon_G4MDC's Avatar
 
Join Date: Nov 2015
Location: Welwyn Garden City, Hertfordshire, UK.
Posts: 2,013
Default Re: SSB Exciter... HELP !!

Not starting an argument. 73 HZU
Jon_G4MDC is offline  
Old 14th Aug 2016, 7:58 pm   #29
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: SSB Exciter... HELP !!

I didn't intend to appear like I was 'arguing'. I was merely advising you that there may be something wrong in your calculations... a calculator typo perhaps?

One thing I will warn everyone about is that there are plenty of 'approximation' formulas out there for capacitive tap networks and these don't actually deliver an accurate impedance match at the design frequency.
They typically work best if the impedance transformation ratio is very large.


So this is risky with a 50:500 ratio and can give big errors if your goal is to match to a specific (resistive) part of the smith chart for something as critical as a crystal filter.

I wrote my own software for tasks like this back in the days before I had a PC. I did it all in QuickBasic. It's been proven time and time again to be accurate over the last 25++ years or so.

Try drawing where 500R, 0j is on a smith chart. Then add a shunt inductor. You will quickly see (without any need for calculations) that the reactance of this inductor needs to be low enough to get you to travel anticlockwise and cross the circle of constant 50R resistance.

Until the reactance gets low enough to cross this circle/line (shown by the green dots below) you will never be able to go on to get to the centre (matched) part of the smith chart using a series and shunt cap that makes up the rest of the capacitive tap network. If you stop at this line with an inductor of about 3uH you can then use a series 118pF cap to quickly navigate to the centre of the chart to get a perfect match. This is the basic L match.

But you can choose to go past this circle of constant 50R using an inductance below 3uH and use a capacitive tap network like the one you posted. This will give greater selectivity in the matching section but this comes at the expense of circuit complexity and a bit more insertion loss.

With your network values in the interactive simulation below you never reach the constant 50R circle with the 4.7uH shunt inductor so you can never get to the centre of the chart using the two remaining caps in the network. This is because 4.7uH never reaches the constant 50R circle. The smith chart easily proves this and the smith chart agrees with my old software.
Attached Thumbnails
Click image for larger version

Name:	ctap_bad.jpg
Views:	97
Size:	60.6 KB
ID:	128697  
__________________
Regards, Jeremy G0HZU

Last edited by G0HZU_JMR; 14th Aug 2016 at 8:23 pm.
G0HZU_JMR is offline  
Old 14th Aug 2016, 8:25 pm   #30
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: SSB Exciter... HELP !!

When designed correctly I think it should look like the interactive simulation below? This uses my suggested values and you can see that it's now possible to navigate to the centre of the smith chart with this network as long as you cross the circle of constant 50R (green dots) with the initial shunt inductor. I used approx 2.6uH in this simulation and the values for the caps also agree with my old software
In case anyone is confused by the chart, the location marker 1 is 500R resistive on the chart. the shunt inductor takes you anticlockwise to position 2 where you can stop at 2.6uH if you choose. Then the series 150pF cap takes you to position 3 on the chart and then the shunt 180pF cap takes you to the cente of the chart and this indicates a good match has bee achieved at 9MHz

Anyone can do this without the need for a calculator. You just plan a route from 500R to the cente of the chart using the toolbox of series or shunt components it contains.

This isn't my smith chart software by the way. My capacitive tap design SW is just an old clunky command line design interface that looks very crude. That snazzy smith chart program dates back 15-20 years and the version I use is dated 2000. It's been a useful tool many times over the years.
Attached Thumbnails
Click image for larger version

Name:	ctapJMR.jpg
Views:	89
Size:	59.8 KB
ID:	128698  
__________________
Regards, Jeremy G0HZU

Last edited by G0HZU_JMR; 14th Aug 2016 at 8:38 pm.
G0HZU_JMR is offline  
Old 14th Aug 2016, 9:05 pm   #31
Jon_G4MDC
Nonode
 
Jon_G4MDC's Avatar
 
Join Date: Nov 2015
Location: Welwyn Garden City, Hertfordshire, UK.
Posts: 2,013
Default Re: SSB Exciter... HELP !!

Good. Wendy - you better talk to Jeremy .
Jon_G4MDC is offline  
Old 14th Aug 2016, 9:31 pm   #32
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: SSB Exciter... HELP !!

Hey, I'm not trying to take over... I'm just trying to offer advice

Stuff like this is close to my heart because one of my passions is RF simulation and network design so apologies if I appear to be going a bit OTT on this stuff. Note that there are various RF publications/books that list design equations for captap circuits and I think they are often just approximations for impedance in terms of magZ. Often there are some weasel words in the text to alert the reader to this fact. So they are very risky to use and I abandoned these crude equations many years ago and wrote my own SW to nail this network really accurately.

See below for a screenshot of the output data from my little program running in a command window. In the background is a simulation showing how accurate it is. It probably still does have bugs for 'crazy' resistance ratios but for anything remotely sensible it usually nails the match dead on. You can see the network intersects 50.00R with no reactance at 9MHz when it is simulated. It also shows 500.00R resistive when looking back into the network from port 2.

However, I think the smith chart program is much more versatile. I can show Wendy how easy it is to use the smith chart program to do impedance matching if that helps? It takes about 20 seconds to do a design and there are no equations required. Just enter the design frequency and load impedance and then 'navigate' to the centre of the smith chart using the components in the toolbox. You basically plan a route and each type of component can move in a circle/arc/straight line and you just have to come up with a combo of parts that hits the centre.

If you get it wrong just delete a component to go back a step. I can go through this in detail with a few screenshots if that helps?
Attached Thumbnails
Click image for larger version

Name:	ctaptest.jpg
Views:	79
Size:	76.4 KB
ID:	128699  
__________________
Regards, Jeremy G0HZU

Last edited by G0HZU_JMR; 14th Aug 2016 at 9:39 pm.
G0HZU_JMR is offline  
Old 14th Aug 2016, 11:44 pm   #33
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: SSB Exciter... HELP !!

Hi Wendy,
The Smith chart program is by Fritz Dellsperger and you can get this off the internet here:

http://fritz.dellsperger.net/smith.html

Note that this links to the latest (all singing all dancing?)version so it may require some initial setting up to get it to look like my old retro version dated from 2000.

I did a quick and dirty demo of my older version on youtube here:

https://youtu.be/jJELSKsqq3I

There is no sound because I have no microphone here but you can see how easy it is to use. I designed two different networks, the one with 2.6uH as in my previous posts and also another with 1.2uH. TAs you can see, the second matching network only took a few seconds to design
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Old 15th Aug 2016, 12:07 am   #34
Wendymott
Octode
 
Wendymott's Avatar
 
Join Date: Jun 2013
Location: Huddersfield, West Yorkshire, UK.
Posts: 1,782
Default Re: SSB Exciter... HELP !!

Hi Peeps.... I really did not intend to start a disagreement.. it was a plea for help of much more intelligent persons than I am... I am not a theorist by any means.. I "hack" it.
I am much indebted to BOTH of you for your time and help... and will follow the suggestions posted. I will report back.. when I have tested the circuit suggested... "Monday after Asda"...
This SSB exciter should be worth the time and effort in the end ... I Hope.
__________________
Should get out more.

Regards
Wendy G8BZY
Wendymott is offline  
Old 15th Aug 2016, 12:26 am   #35
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: SSB Exciter... HELP !!

Ideally, you need to work out what the design impedance/termination is for your filter. I think I looked up some 9MHz filters for you a while back and somebody on another (foreign) forum stated 560R and 25pF after web translation.

https://www.vintage-radio.net/forum/...d.php?p=789309

If you have a tracking gen with your Rigol analyser I can show you a cheap and cheerful method to determine (or at least check) that the 560R figure is accurate? It's a method I used for many years with fairly good results and it is cheap and quick and simple if you have an analyser and Tgen.

The other (modern and snazzy) way to do it is to measure the filter directly (unmatched) with a 2 port 50R VNA to generate a model you can use with a simulator to design the matching and lots of people seem to be getting test gear like this nowadays. The VNA method is really neat because you can design and verify your matching network on the simulator using the VNA derived s2p model of the crystal filter. So you get to see the predicted response of the matched filter on the simulator before you build the matching network for real.

But the old school method with the spectrum analyser and tracking gen (followed by a few seconds generating the matching network on the smith chart program) is typically just as effective.
__________________
Regards, Jeremy G0HZU

Last edited by G0HZU_JMR; 15th Aug 2016 at 12:31 am.
G0HZU_JMR is offline  
Old 15th Aug 2016, 9:27 am   #36
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,801
Default Re: SSB Exciter... HELP !!

With crystal filters, the termination impedance they are designed for can be anywhere in a wide spread.

If you have a network analyser or a spectrum analyser with a tracker, sweep the filter with a resistive load, then pad out the resistance. Play around with the resistance until you get minimum ripple in the passband. Then try adding a bit of C to ground to see if you can get it a bit lower. If so go back and muck about with R a bit more. You should see a filter of the described bandwidth and with reasonable ripple now.

I bought boxes and boxes of filters back in the eighties and had a bit of fun finding out what each did and what Z it was designed around.

Once you've got this sorted, you can design a matching network.

David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done
Radio Wrangler is offline  
Old 15th Aug 2016, 12:40 pm   #37
Wendymott
Octode
 
Wendymott's Avatar
 
Join Date: Jun 2013
Location: Huddersfield, West Yorkshire, UK.
Posts: 1,782
Default Re: SSB Exciter... HELP !!

Hi Jeremy and David...... Thanks for the comments.... so much to do... Yes Jeremy those were / are the filters discussed earlier.... I have the data sheets.. somewhere...
Yes the Rogol has a tracking generator... I would not have bought it without..
The modulator circuit is attached....The output transformer was wound to ensure resonance at 9 Mhz, and the secondary was an arbitary 3 turns...this may or may not be 50 R... but I can check that by shunting with a resistor to find the 50% level. and adjust the turns to suit.
Attached Thumbnails
Click image for larger version

Name:	Modulator Mixer V3.jpg
Views:	102
Size:	65.4 KB
ID:	128721  
__________________
Should get out more.

Regards
Wendy G8BZY
Wendymott is offline  
Old 15th Aug 2016, 3:19 pm   #38
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,801
Default Re: SSB Exciter... HELP !!

MC1496 have open-collector outputs. The output impednce they present to the outside world is pretty high. The gain they produce is dependent on the internal emitter degeneration resistors and the load impedance the outside world applies to the open collector output.

It's in the transmit chain, so you aren't fighting for gain and noise floor. You can afford to throw away some signal in order to present a nice source impedance to your filter, and a nice load to the mixer/modulator. This may be where some of that visible distortion is coming from. A load resistor in the output circuit would be a great help.

David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done
Radio Wrangler is offline  
Old 15th Aug 2016, 5:02 pm   #39
Wendymott
Octode
 
Wendymott's Avatar
 
Join Date: Jun 2013
Location: Huddersfield, West Yorkshire, UK.
Posts: 1,782
Default Re: SSB Exciter... HELP !!

Hi David... I have conducted tests this Pm to ascertain the output impedance of the Transformer..the method was to measure the output with a X10 scope probe and a small length of high grade coax...to a resistive load....measure the non loaded voltage and the loaded voltage to 50% of the un loaded, then measure the resistance (variable pot). I found the best output Z was 610 ohms (4 turns), which is quite near 500 R the Z of the filter. I realize the load resistor was not a perfect load as there would be capacity and inductance, but I think its.. near enough..I do not have any distortion of the sinewaves now, that was due to the homebuilt dual AF osc. I have also realized the 9 mhz filter is an FM filter with a +/- 6 Khz bandwidth instead of +/ 1.5 Khz. Thus... I must use both the 8.998 and 9.0015 Filters............... or... just make it LSB for 40 metres and stick with one filter. I will ADD a load resistor anyways.. which ever avenue I go down.
Thanks for your comments
__________________
Should get out more.

Regards
Wendy G8BZY

Last edited by Wendymott; 15th Aug 2016 at 5:05 pm. Reason: Typo error
Wendymott is offline  
Old 15th Aug 2016, 10:51 pm   #40
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: SSB Exciter... HELP !!

Glancing over your circuit for the modulator in post #37 I think you may have missed out a (47uF?) decoupling capacitor across R5?

Without this cap I think there will be an issue with AF common mode/cancellation and you won't get very good conversion efficiency because of this? I think you need to drive it with a single ended input and maintain this single ended operation and I don't think your circuit achieves this? You may find it gives more conversion gain and a flatter response at low AF frequencies if you correct this issue by adding the decoupling that will prevent the CM drive issue.

I notice that the gain setting resistor is just a wire link and you may find that you can insert a resistor here to turn down the gain once you add the cap above?

I haven't used this chip in about 25 years so I am rusty on its operation but I think there is this decoupling issue in your schematic in post #37.
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Closed Thread

Thread Tools



All times are GMT +1. The time now is 2:52 am.


All information and advice on this forum is subject to the WARNING AND DISCLAIMER located at https://www.vintage-radio.net/rules.html.
Failure to heed this warning may result in death or serious injury to yourself and/or others.


Powered by vBulletin®
Copyright ©2000 - 2024, vBulletin Solutions, Inc.
Copyright ©2002 - 2023, Paul Stenning.