|
Vintage Amateur and Military Radio Amateur/military receivers and transmitters, morse, and any other related vintage comms equipment. |
|
Thread Tools |
15th Sep 2017, 1:41 pm | #1 |
Dekatron
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
|
KVN/KVG? 9MHz Crystal Filter Board
This info is mainly of interest to anyone interested in old school ham radio homebrew.
It appears that a large batch of old (TEL-RAD?) radios has been dismantled and the 9MHz IF boards contain some nice 9MHz KVN/KVG crystal filters. These are available from various sellers either as the complete IF PCB or just the filters. I paid £23 + £3p&p for the IF board from a seller in Israel a couple of weeks ago and the board has just arrived today. See below for an image of the board that I received today. I think there is a 9MHz USB, LSB and AM crystal filter on the board. XF-9 S42 XF-9 S43 XF-9 S44 My board was fully populated so there are some other things on there of interest. eg my board also has MC1496P modulator chip qty 3 (plastic DIL 14 package) Plessey SL1612C IF chip qty 1 (plastic DIL8 package) MC1350P IF chip qty 1 (plastic DIL8 package) 2N2222A qty 3 (metal can package) 2N4416 low noise JFET qty1 (metal can package) Microtran T2104 audio transformer 600R 1:1 The little transformer may be of use as a 1:1 isolation transformer when fed to a PC soundcard if ground loop connections are a problem? But that's just a guess. I'll post up plots of the filters later. I'll look at them on a VNA and export a 2 port model and work out the correct impedance match.
__________________
Regards, Jeremy G0HZU |
15th Sep 2017, 2:21 pm | #2 |
Dekatron
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
|
Re: KVN/KVG? 9MHz Crystal Filter Board
There seems to be some useful info about this board here:
https://gw4sae.wordpress.com/2017/03...-ssb-receiver/ it looks like someone in that link has reverse engineered the schematic. It also appears that some people are using it as the heart of a homebrew HF receiver because the first mixer, the IF strip and AGC and product detector are all there. Obviously the LO and 9MHz BFO signals are missing but this would be easy to add. I'm not sure how much AGC range it offers but probably enough for casual HF listening on a ham band for example.
__________________
Regards, Jeremy G0HZU |
15th Sep 2017, 5:11 pm | #3 |
Dekatron
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
|
Re: KVN/KVG? 9MHz Crystal Filter Board
I removed the crystal filters from my board and sadly only one of them works properly. It may be possible that I damaged them taking them out but I used a decent solder sucker and only minimal heat. The retaining screws are a tight fit in the PCB so it may be possible that removing the filter stresses the filter in some way. I did try to be careful but either way, only my LSB filter works properly.
I did also buy a set of these filters 'loose' and I found that two of the second set seemed OK but the USB filter was showing signs of problems with some asymmetry in the passband. But nothing too major. So it looks like these filters may have reliability issues. Both of the LSB filters and one of the AM filters work fine so I'm happy enough in this respect. I'm not sure I can face the hassle of sending the others back so I may just take them apart at some point. The USB filter from the 'complete' IF board is very poorly as it shows loads of loss and a narrow passband. The faulty AM filter shows a big bump in the ripple and return loss at the top end of the passband. I guess it's a case of buyer beware. I did notice that two of the filters in my board had loose screws before I even attempted removal. But I don't know if these were the two that were actually faulty.
__________________
Regards, Jeremy G0HZU |
15th Sep 2017, 5:49 pm | #4 |
Dekatron
Join Date: Apr 2012
Location: Wiltshire, UK.
Posts: 14,010
|
Re: KVN/KVG? 9MHz Crystal Filter Board
These are "KVG" filters, the details of which have been discussed in several other historic threads on this board.
http://www.vintage-radio.net/forum/s...d.php?t=119808 KVG pulled-out of the HF xtal-filter business some time back but the details are available out there. |
15th Sep 2017, 6:00 pm | #5 |
Dekatron
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
|
Re: KVN/KVG? 9MHz Crystal Filter Board
Ok thanks, I wasn't sure if they were the same type. Mine are marked KVN but I think it's the same people. Note that I don't think these are 560 ohm filters.
Based on my tests I think they are best matched with a source (and load) that looks like 1000 ohm in parallel with about 23-24pF.
__________________
Regards, Jeremy G0HZU Last edited by G0HZU_JMR; 15th Sep 2017 at 6:09 pm. |
15th Sep 2017, 7:03 pm | #6 |
Dekatron
Join Date: Jun 2016
Location: London, UK.
Posts: 3,687
|
Re: KVN/KVG? 9MHz Crystal Filter Board
Just as a note, GQRP club sell new 9MHz filters. 2.2KHz bandwidth, 500R impedance for £12. Also a pair of SSB crystals +/-1.5KHz for USB/LSB for £4.
http://www.gqrp.com/sales.pdf Spectrum comms still sell CA3028's as well. You could build a new IF strip for not much |
18th Sep 2017, 5:57 pm | #7 |
Octode
Join Date: Jun 2013
Location: Huddersfield, West Yorkshire, UK.
Posts: 1,795
|
Re: KVN/KVG? 9MHz Crystal Filter Board
Jeremy.... Just seen this after my return from holiday.... I have bought 3 of these modules over the past year and the filters and the MC1496 are what my 40M ssb transceiver is based on....... I did wonder about a schematic and if it was available... but the vendor did not reply.... I guess because he is a scrap dealer not a technical person.
I will try the link above to obtain the schematic
__________________
Should get out more. Regards Wendy G8BZY |
19th Sep 2017, 10:26 pm | #8 |
Octode
Join Date: Jun 2013
Location: Huddersfield, West Yorkshire, UK.
Posts: 1,795
|
Re: KVN/KVG? 9MHz Crystal Filter Board
I downloaded the schematic of the module.......... even if it is not totally accurate .. it gives me a good idea where the design was coming from....I thought the wider band filter was for FM... but not so.. it is as suggested a "roofing" filter.....Thanks Jeremy for the link.
__________________
Should get out more. Regards Wendy G8BZY |
21st Sep 2017, 10:20 pm | #9 |
Dekatron
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
|
Re: KVN/KVG? 9MHz Crystal Filter Board
I think there may be a few versions of the schematic depending on where you click once into the web page link I gave. But it does look like someone has spent a fair bit of time working out the circuit.
I'll try and find time to measure the LSB filter on the VNA this weekend. I'll aim to produce a decent two port model of the filter and then design a suitable matching network for a 50 ohm test system on an RF/linear simulator. I'll then post up my plots and findings on here.
__________________
Regards, Jeremy G0HZU |
22nd Sep 2017, 10:33 pm | #10 |
Dekatron
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
|
Re: KVN/KVG? 9MHz Crystal Filter Board
Here's my initial attempt at measuring a two port model of the LSB filter. I generated this quick 2 port model with the VNA when the filters arrived last week. Note that it only models the filter over a 20kHz span centred on 9MHz.
To get the forum upload tool to accept the file I've had to rename it from KVGXF9S43.s2p to KVGXF9S43.txt. So once it is downloaded it can be renamed to KVGXF9S43.s2p to make it load into a simulator easily. Hope it uploads OK. The source and load that the filter sees needs to be the equivalent of about 1000 ohm in parallel with about 24pF to get a good match. See below for a first attempt at a matching network based on this VNA s2p file. Also a computer simulation of the filter once fed with this matching network. It looks like the insertion loss will be about 2.5dB. You can load this KVGXF9S43.s2p two port model into RFSIM99 and create a matching network for it or try the one posted below in the RFSIM99 image. But the model is only valid over a 20kHz span so RFSIM99 would have to be set up for a 20kHz span I'll try and find time to create a better two port model over the weekend as I think the stopband performance could be better than that shown in the quick and dirty simulation below. But I'll also build the matching network for real and test it. Note that the inductor used in the matching network will need to be adjustable in order to tune for a good passband response.
__________________
Regards, Jeremy G0HZU Last edited by G0HZU_JMR; 22nd Sep 2017 at 10:57 pm. |
22nd Sep 2017, 11:42 pm | #11 |
Dekatron
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
|
Re: KVN/KVG? 9MHz Crystal Filter Board
I loaded my two port model KVGXF9S43.s2p into RFSIM99 and this particular LSB filter seems to be optimal at about 850 ohm in parallel with 24pF. But the other filters seem to be closer to 1000 ohm in parallel with 24pF. So there will be some spread from filter to filter in terms of the optimal matching network.
See below how the port impedance in RFSIM99 can be changed to 850R and the 24pF can then be placed across the ports 1 and 2. It shows low loss and low ripple. The other way to do it would be to use the series equivalent at the ports and the series equivalent of 850R in parallel with 24pF at 9MHz is 365R in series with 42pF. I've posted up two RFSIM99 simulations one with the ports set up as 850R in parallel with 24pF and one with 365R in series with 42pF. They both give a very similar response in RFSIM99. I think that most of the filters I have here are either faulty or maybe they have aged slightly because they are not all optimal with 1000 ohm in parallel with 24pF. The spread seems to cover about 850R to 1050R from filter to filter. Maybe this is down to component ageing? The schematic provided in the earlier link seems to suggest it was designed to provide 1000 ohm in parallel with some capacitance and this seems to agree fairly well with my initial experiments. The last image below looks at the parallel R and parallel C equivalent looking backwards into the capacitive tap matching network that I posted up earlier. The network was optimised on the computer and it has presented an Rp of 820R in parallel with about 23.5pF to the filter. So this agrees quite well with the RFSIM99 ports when they were optimised to 850R in parallel with 24pF I'll build this captap matching network over the weekend and test it on a real LSB filter when tested with 50R ports.
__________________
Regards, Jeremy G0HZU Last edited by G0HZU_JMR; 23rd Sep 2017 at 12:12 am. |
23rd Sep 2017, 9:54 am | #12 |
Dekatron
Join Date: Apr 2012
Location: Wiltshire, UK.
Posts: 14,010
|
Re: KVN/KVG? 9MHz Crystal Filter Board
The "traditional" designs for these generally drove them with a common-collector bipolar transistor using a 910-Ohm resistor as the collector-load, shunted by a small capacitor.
The filter then fed into another bipolar - usually an emitter-follower - whose emitter-bias resistors were a pair of 1.8K or 2K resistors. So the input and outputs 'saw' around 1KOhm. When testing real-world filters it can be worth remembering that though in theory they're symmetrical, in practice they're not! Try them both ways round and see which gives the best [lowest passband-ripple, best near-in rejection] performance. The biggest issue I had with these sorts of filters was 'group delay' - but explaining the details involves maths just too gnarly for a Saturday morning. |
23rd Sep 2017, 12:42 pm | #13 |
Dekatron
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
|
Re: KVN/KVG? 9MHz Crystal Filter Board
Yes, it's partly because of group delay issues that I'm going to the trouble of trying to match the filter as optimally as possible. Ultimately, I want to box up a couple of these filters as 50R in/out 'bench test' filters with SMA connectors. So I want to try and get the return loss and the group delay to look fairly good across the centre of the passband.
I had a play with RFSIM99 to see if it could plot group delay when my s2p model was matched with the captap network. It seems to be able to do this based on the S21 data from RFSIM99. However, I had to adjust the number of display points to be a sub multiple of the number of points in my s2p files. My s2p file spans 20kHz with 801 data points and this has an aperture of 25Hz per step. I had to adjust the display/analysis points to get a sensible group delay plot in RFSIM99. See the plots of group delay below in RFSIM99 and Genesys. Genesys is set to 801 analysis points Note that this is taken with the captap matching network on either side of the filter in both cases. This should agree quite well with the group delay measured with the VNA once I build the matching network for real. However, all of this depends on how well I extracted that first S2p model from the VNA. I did it in a hurry but it should be OK
__________________
Regards, Jeremy G0HZU Last edited by G0HZU_JMR; 23rd Sep 2017 at 1:06 pm. |
23rd Sep 2017, 11:09 pm | #14 |
Dekatron
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
|
Re: KVN/KVG? 9MHz Crystal Filter Board
This evening I made up a basic jig to test the XF-9 S43 LSB filter in and added the captap matching network. Then I tested it on the VNA.
I used smaller size toroids that have slightly lower Q for the inductors and I think this explains the 2.8dB insertion loss instead of the expected 2.5dB but it still looks good. The metalwork in the jig needs optimising wrt screening to get low stopbands but the plot below is already showing promise. At present the jig has no screening lid and only a basic partition across the filter. So there is still some leakage. I'll post up some more detailed VNA plots of the passband and return loss and group delay tomorrow.
__________________
Regards, Jeremy G0HZU Last edited by G0HZU_JMR; 23rd Sep 2017 at 11:27 pm. |
24th Sep 2017, 3:17 pm | #15 |
Dekatron
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
|
Re: KVN/KVG? 9MHz Crystal Filter Board
I'm using the VNA for another task today but I did find time to remove the cover from the faulty XF9S44 filter.
I'm still studying it under a microscope but see below for a quick image of the insides. It has 11 HC49 cans inside but I don't think there are crystals inside all of them. This thing is quite a strange mix of SMD and leaded caps, inductors, gimmick caps and HC49 crystal packages with access/feed holes in the side of the HC49 can. I suspect that there are 8 or 9 crystals and the other two or three HC49 devices contain matching parts. But that is just a guess. The soldering of the SMD caps on the top isn't great when looked at under a microscope and I suspect that these were built by an operator with limited magnification. Also, there is a lot of flux residue on the top PCB. I suspect that this filter was damaged before I got it, possibly a crystal has been damaged by signal overload or maybe it has been dropped and a crystal has failed. Having seen the insides I don't think I damaged it when unsoldering it. I expect that one or more of the crystals will be duff. The SMD caps look OK and I doubt any are cracked. Hopefully, I'll be able to salvage some healthy 9MHz crystals from it if I can't repair it easily.
__________________
Regards, Jeremy G0HZU Last edited by G0HZU_JMR; 24th Sep 2017 at 3:31 pm. |