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 > Specific Vintage Equipment > Vintage Amateur and Military Radio

Notices

Vintage Amateur and Military Radio Amateur/military receivers and transmitters, morse, and any other related vintage comms equipment.

Closed Thread
 
Thread Tools
Old 16th Feb 2021, 9:33 pm   #21
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: PIN diode band switching yaesu ft-767

Quote:
Originally Posted by Steve G4WCS View Post
Ive some hp 5082-2800 and 2810 diodes I can send over if they are suitable for your testing
Thanks for the offer of those classic diodes, however I've got quite a few of them here already. They aren't that well suited to this application but I'll do a few tests on one to show why they shouldn't be used here.

I dug out my old >1995 Technical Topics book and there are several articles in it about diode switches in HF radios. There were lots of people contributing test results but there was not much agreement. Some people recommended 1N4148 diodes and others referred to them as catastrophically poor for example. I suspect that the people who recommended the 1N4148 did so as a result of a listening test (for sensitivity?) with the diodes fitted to a radio. The people who didn't like the 1N4148 based their conclusions on distortion tests taken in a 50 ohm test fixture. I do think the 1N4148 will perform poorly in a 50 ohm test fixture but it might be OK to use these diodes when switching a suite of 450 ohm filters as in the Yaesu FT767. Even so I think I'd rather use something else.

Because I work from home my work bench kind of belongs to the company most of the time so I don't have any free space to do testing until the weekend. However, over the next few days I could post up some general diode info and show why fast switching diodes like the 1N4148 are not an ideal choice for this type of application.
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Old 16th Feb 2021, 10:37 pm   #22
Steve G4WCS
Heptode
 
Steve G4WCS's Avatar
 
Join Date: Sep 2004
Location: Preston, Lancashire, UK.
Posts: 632
Default Re: PIN diode band switching yaesu ft-767

RF board is out. Got to have something interesting to do on a cold dark night. The lamp is intact, but interestingly there is already evidence of resoldering around the RF in socket and the lamp.

Oh and thanks Jeremy, so should I stick with the Iss83, may as well replace the lot while the board is out, its not like Im paying for my time
Attached Thumbnails
Click image for larger version

Name:	AADC06A0-2EA7-44D2-88A4-A38260C50358.jpeg
Views:	105
Size:	195.6 KB
ID:	226928   Click image for larger version

Name:	18671FCD-2934-4BB2-A661-F8C566334B4A.jpg
Views:	78
Size:	112.4 KB
ID:	226929   Click image for larger version

Name:	2E2C3734-D7D0-4DDE-AB5D-4A241590514C.jpg
Views:	66
Size:	117.9 KB
ID:	226930  
Steve G4WCS is offline  
Old 16th Feb 2021, 10:57 pm   #23
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: PIN diode band switching yaesu ft-767

Quote:
so should I stick with the Iss83, may as well replace the lot while the board is out, its not like Im paying for my time
I've never worked on one of these radios or played with an ISS83 diode so I can't really say for sure. However, other people on the thread suggest that these 1SS83 diodes are unreliable and you should replace them all.

I'm not sure what you should replace them with though. There's a lot of diodes so you really only want to have to do this once...

I still don't know why the 1SS83 diodes are unreliable. Maybe they can get toasted if high RF power gets into the antenna/receiver from another HF transmitter/antenna at the same location? 1N4148 diodes are very rugged and very cheap but the manufacturer still doesn't fit them as standard.

Did you do the sig gen test to see how deaf the receiver was on the MW band?
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Old 16th Feb 2021, 11:49 pm   #24
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: PIN diode band switching yaesu ft-767

If the receiver overload protection bulb has been replaced in the past then I guess these band switching diodes may have also been stressed or damaged at the same time.

Looking back at post #12 at the manufacturer's specs for the 0.5-1.5MHz range the sensitivity on SSB drops from 0.25uV to 4uV for 10dB S+N/N. This is a change of about 24dB which seems close to the 21-22dB loss in the T attenuator that is permanently inline for the 0.5-1.5MHz range. So this is a design feature for the radio and you can expect to hear the radio behave as if a 21dB attenuator has been fitted at the front end on this range. This feature obviously won't change with new diodes but maybe the fact the bulb has been replaced means that the diodes might not be as healthy as they once were?

So you might as well change them all. I'm not sure I want to recommend an alternative though. People have been arguing over the merits of alternative band switch diodes for over 25 years with little sign of agreement.
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Old 17th Feb 2021, 1:57 am   #25
majoconz
Heptode
 
majoconz's Avatar
 
Join Date: Jan 2008
Location: Ashhurst, Manawatu, New Zealand
Posts: 570
Default Re: PIN diode band switching yaesu ft-767

Just caught the end (hopefully!) of this post. Here's a couple of schematics of the LF front ends of my Kenwood R2000 and NRD515. Both have much the same passband filters although the NRD515 has a tuneable preselector - I guess to keep the MW band blowtorches out better than a broad filter. Both receivers have a bit of attenuation in the filter, the 2000 says its 5dB and the 515 I believe is 10dB - needless to say in NZ I have disabled both of them with no ill effects. In the R2000 the diodes are all BA282's and the 515 has 1S84's - hope this helps.
Attached Thumbnails
Click image for larger version

Name:	R2000 LF front end.jpg
Views:	105
Size:	38.1 KB
ID:	226933   Click image for larger version

Name:	NRD 515 LF front end.jpg
Views:	102
Size:	66.2 KB
ID:	226934  
__________________
Cheers - Martin ZL2MC
majoconz is offline  
Old 17th Feb 2021, 4:06 am   #26
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,799
Default Re: PIN diode band switching yaesu ft-767

I'd question the use of varactor diodes tuning a preselector. If you try to receive a weak signal near to strong ones you are likely to find intermodulation distortion a problem. The large signals pump the capacitance of the diodes and hence modulate the tuned frequency of the preselector.

David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done
Radio Wrangler is online now  
Old 17th Feb 2021, 10:00 am   #27
Steve G4WCS
Heptode
 
Steve G4WCS's Avatar
 
Join Date: Sep 2004
Location: Preston, Lancashire, UK.
Posts: 632
Default Re: PIN diode band switching yaesu ft-767

In terms of the diodes and at risk of going off topic, in the post war period the japanese were catching up with the american semiconductor industry and would have been fiercely competitive, hence the sourcing of hitachi diodes as opposed to the motorola 1n series. However looking at the packaging, the hitachi diodes are in the older oa90 style glass packaging, and Im wondering whether a factor in their long term reliability would be the glass to wire seals failing. Just a thought, so even new old stock ISS83’s may be hit and miss unless tested before fitting of course. I do have a strip of 1N4148’s that I will use. Im also thinking of changing out the blue rubcon electrolytics whilst the board is out
Steve G4WCS is offline  
Old 17th Feb 2021, 10:09 am   #28
Terry_VK5TM
Nonode
 
Terry_VK5TM's Avatar
 
Join Date: Oct 2010
Location: Tintinara, South Australia, Australia
Posts: 2,324
Default Re: PIN diode band switching yaesu ft-767

If there are any grey coloured low value caps, change those as well.
__________________
Terry VK5TM
https://www.vk5tm.com/
Terry_VK5TM is offline  
Old 17th Feb 2021, 1:31 pm   #29
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: PIN diode band switching yaesu ft-767

Quote:
I do have a strip of 1N4148’s that I will use.
If you opt for the 1N4148, the insertion loss of the 1N4148 diode is fairly easy to predict on the HF bands. It's possible to predict the RF (small signal) ESR vs bias current using the manufacturers curves for the diode. I dug out an old excel spreadsheet that shows my own IV measurement of the 1N4148 diode. The slope of the curve at 10mA bias shows a diode resistance of about 5.5 ohms according to the dynamic resistance chart on the spreadsheet.

Alternatively the ideality factor n for a 1N4148 diode is close to 2. I think it is somewhere between 1.9 and 2.0 and this can be used with a simple equation to predict the ESR vs bias current.

For example:
Vt = thermal voltage = 0.026V at room temperature. Therefore at 10mA bias current the small signal resistance of the diode will be 2*(Vt/0.01) = 5.2 ohms. This agrees with the spreadsheet and I have also measured a 1N4148 at 10mA on a VNA and it showed 6 ohms series resistance. Close enough I think...

With 10mA bias current and a series resistance of 5.2 ohms the insertion loss of the 1N4148 in a 50R system will be about 0.45dB.

With with 10mA bias current and a series resistance of 5.2 ohms the insertion loss of the 1N4148 in a 450R filter bank system will be about 0.05dB.


It looks like there is one diode (D06) in a 50R section of the FT767 signal path and maybe 4 in the 450R BPF section. So the combined insertion loss of the diodes will be about 0.45 + (4x 0.05) = 0.65dB based on what I can see in your schematic if the bias current in the diodes is about 10mA. I think this is why lots of people are happy with using 1N4148 diodes. I'd expect the insertion loss using 1SS83 diodes to be very similar.

However, the problem with a fast switching diode like the 1N4148 will be that its series resistance is also 'fast' and will be able to change dynamically with the introduction of large broadcast signals or local amateur signals. The diode resistance will be modulated by the RF signals in other words.

Other diodes like the BA282 will have much lower ESR at 10mA bias and diodes like the 1SS83 are 'slower to respond' so the ESR stays more stable with large signals. A decent PIN diode with a long carrier lifetime would be even better as the series resistance will stay largely constant even with large signals present. This means these diodes will generate less distortion.

This issue is probably only relevant in D06 on your schematic as this looks to be the only diode operating in a 50 ohm section of the signal path. The other diodes all seem to be in the 450R BPF switching section. With a 450R system, the impact of a dynamic change in diode ESR with large broadcast signals should be a lot less because the RF signals will only be modulating a 0.05dB insertion loss. Therefore, 1N4148 diodes might be OK here but I can only guess. It would be best to test them in the real radio.

If you definitely want to fit 1N4148 diodes then they will probably be OK in the 450R section and this is most of the diodes. However, it might be better to fit something else at D06. Maybe even consider keeping a healthy 1SS83 diode here.

I can show you how badly a 1N4148 will perform in the D06 location compared to other diodes but a lot depends on your antenna system. If you only use a small antenna system then you probably won't notice any issues.

The distortion problems with a 1N4148 at D06 will maybe be detectable if you can receive signals at S9 +50dB on the lower bands.
Attached Thumbnails
Click image for larger version

Name:	1N4148_curves.jpg
Views:	73
Size:	72.7 KB
ID:	226957  
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Old 17th Feb 2021, 2:32 pm   #30
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,799
Default Re: PIN diode band switching yaesu ft-767

Ordinary diodes, whether plain silicon junctions or Schottky cannot carry reverse current, so the effective current in the reverse direction is caused by reducing the DC bias current, so the effective reverse-going signal current cannot be less than the bias current... thereafter you get rectification and massive distortion.

PIN diodes will do plenty of reverse current at RF. Orders of magnitude more than the bias current. Just so long as the frequency is above the carrier decay time limit. It's their superpower!

David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done
Radio Wrangler is online now  
Old 17th Feb 2021, 10:01 pm   #31
Steve G4WCS
Heptode
 
Steve G4WCS's Avatar
 
Join Date: Sep 2004
Location: Preston, Lancashire, UK.
Posts: 632
Default Re: PIN diode band switching yaesu ft-767

So. This afternoon I lifted and tested all the switching diodes, checked all the electrolytics around the bandpass filter, and resoldered the rf input connection, and lamp pins.

I knew it was deaf, its like a different radio. Over the moon with the results

https://youtu.be/VTn6Y6Epu5A

(Id also bypassed the 0.5-1.5mhz attenuator, but its not that thats made the difference)
Attached Thumbnails
Click image for larger version

Name:	FC04639B-C74B-449D-B9B1-3A3AA5B31F63.jpg
Views:	66
Size:	140.0 KB
ID:	226989   Click image for larger version

Name:	2F42EB12-BF61-45F2-B0A9-F23487C264E6.jpg
Views:	78
Size:	129.2 KB
ID:	226990   Click image for larger version

Name:	F0384A64-65BA-4C0A-B0CD-A19456FAF82C.jpeg
Views:	66
Size:	175.2 KB
ID:	226991  
Steve G4WCS is offline  
Old 17th Feb 2021, 10:49 pm   #32
Steve G4WCS
Heptode
 
Steve G4WCS's Avatar
 
Join Date: Sep 2004
Location: Preston, Lancashire, UK.
Posts: 632
Default Re: PIN diode band switching yaesu ft-767

Plugged the extension speaker in. The audio quality is superb. Love these old radios, they sound so much better than the modern dsp jobbies

https://youtu.be/FCczG28PxV0

https://youtu.be/GH69IkoCyRA
Steve G4WCS is offline  
Old 17th Feb 2021, 11:16 pm   #33
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: PIN diode band switching yaesu ft-767

A good result! From what you say, it sounds like the radio was deaf on all ranges and not just 0.5-1.5MHz?

Did you spot any obvious duds amongst the old 1SS83 diodes?

On your youtube demo video I did spot a few signals near S9 +60dB. I'm not sure how accurate the s meter is but this implies signals up near -13dBm at the input. For a bit of fun I dug out my diode test rig this evening and tested a 1N4148 diode at 15mA bias with a -10dBm test signal at 950kHz. The idea was to look for distortion in the diode in a 50 ohm system at 1900kHz.

The spectrum analyser showed a second harmonic at -76dBm at 1900kHz. The analyser is protected by a diplexer at the front end so this distortion is caused by the 1N4148 diode.

This means that the 1N4148 diode would generate a second harmonic at about an s8 signal on an equivalent s-meter in this harsh test. I quickly swapped the 1N4148 diode for a decent PIN diode (also biased at 15mA) and the second harmonic distortion term fell about 40dB. It is probably better than this but this is the limit of the test setup at the moment. This shows how relatively poor the 1N4148 diode can be in terms of the even order distortion it can generate in a 50 ohm system. When fitted into a radio it might be different because the diode will be terminated in a BPF. This might makes things worse or better.

I think it is this kind of difference that prompted Ulrich Rohde to experiment with PIN diodes about 25+ years ago. I think the slower 1SS diodes will perform somewhere inbetween. They will definitely be better than the 1N4148 but they won't match a decent PIN diode here.
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Old 17th Feb 2021, 11:42 pm   #34
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,799
Default Re: PIN diode band switching yaesu ft-767

Just remember that PINs only work their magic at frequencies above a limit set by their carrier lifetime spec. For lower frequencies, they steadily turn into ordinary diodes.

For the best off-state isolation, some reverse voltage bias id good to widen the depletion region, and even more reverse bias is even better if you have it.

David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done
Radio Wrangler is online now  
Old 17th Feb 2021, 11:56 pm   #35
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: PIN diode band switching yaesu ft-767

Agreed. The PIN diode I used was a bit special with a carrier lifetime of about 11us when biased at 15mA. This is why it was still performing very well as a PIN diode down at 950kHz.
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Old 17th Feb 2021, 11:58 pm   #36
Steve G4WCS
Heptode
 
Steve G4WCS's Avatar
 
Join Date: Sep 2004
Location: Preston, Lancashire, UK.
Posts: 632
Default Re: PIN diode band switching yaesu ft-767

Quote:
Originally Posted by G0HZU_JMR View Post
A good result! From what you say, it sounds like the radio was deaf on all ranges and not just 0.5-1.5MHz?

Did you spot any obvious duds amongst the old 1SS83 diodes?

On your youtube demo video I did spot a few signals near S9 +60dB. I'm not sure how accurate the s meter is but this implies signals up near -13dBm at the input. For a bit of fun I dug out my diode test rig this evening and tested a 1N4148 diode at 15mA bias with a -10dBm test signal at 950kHz. The idea was to look for distortion in the diode in a 50 ohm system at 1900kHz.

The spectrum analyser showed a second harmonic at -76dBm at 1900kHz. The analyser is protected by a diplexer at the front end so this distortion is caused by the 1N4148 diode.

This means that the 1N4148 diode would generate a second harmonic at about an s8 signal on an equivalent s-meter in this harsh test. I quickly swapped the 1N4148 diode for a decent PIN diode (also biased at 15mA) and the second harmonic distortion term fell about 40dB. It is probably better than this but this is the limit of the test setup at the moment. This shows how relatively poor the 1N4148 diode can be in terms of the even order distortion it can generate in a 50 ohm system. When fitted into a radio it might be different because the diode will be terminated in a BPF. This might makes things worse or better.

I think it is this kind of difference that prompted Ulrich Rohde to experiment with PIN diodes about 25+ years ago. I think the slower 1SS diodes will perform somewhere inbetween. They will definitely be better than the 1N4148 but they won't match a decent PIN diode here.

I think something was deading it on all frequencies, and also I think it was ahead of where the marker signal was inserted. It could have been the lamp connections, rf input socket connections, or a dry joint on one of the diodes I lifted ahead of the bandpass filters. All the bandpass filter switching diodes were reading consistently with no leakage, so Ive left them in place . May recommission the attenuator now tonsee what happens

Looking forward to see how it behaves on 60 and 20m tomorrow.
Steve G4WCS is offline  
Old 18th Feb 2021, 9:38 pm   #37
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: PIN diode band switching yaesu ft-767

Hope it all performed well today. If it helps at all I quickly loaded the circuit components into a simulator for the 0.5-1.5MHz filter range. The simulator is set for 450 ohm ports and this does seem to match well with the filter design. The plot shows the impact of the T attenuator when it is fitted (red trace) and when it is removed (green trace). It does show the expected response across 0.5-1.5MHz and the attenuator does introduce the expected 21dB attenuation.

It's good that all the 1SS83 diodes appear healthy. The 1SS83 has a max reverse recovery time of 100ns compared to the 4ns of the faster 1N4148. I think the slowness of the 1SS83 will help a bit with distortion across a fair bit of the HF band.
Attached Thumbnails
Click image for larger version

Name:	Preselector_Att.jpg
Views:	79
Size:	79.1 KB
ID:	227038  
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Old 19th Feb 2021, 11:57 am   #38
Steve G4WCS
Heptode
 
Steve G4WCS's Avatar
 
Join Date: Sep 2004
Location: Preston, Lancashire, UK.
Posts: 632
Default Re: PIN diode band switching yaesu ft-767

Quote:
Originally Posted by G0HZU_JMR View Post
Hope it all performed well today. If it helps at all I quickly loaded the circuit components into a simulator for the 0.5-1.5MHz filter range. The simulator is set for 450 ohm ports and this does seem to match well with the filter design. The plot shows the impact of the T attenuator when it is fitted (red trace) and when it is removed (green trace). It does show the expected response across 0.5-1.5MHz and the attenuator does introduce the expected 21dB attenuation.

It's good that all the 1SS83 diodes appear healthy. The 1SS83 has a max reverse recovery time of 100ns compared to the 4ns of the faster 1N4148. I think the slowness of the 1SS83 will help a bit with distortion across a fair bit of the HF band.
thank you for that , all very interesting. yes I lifted one end of each diode and checked them with a multimeter, all were sound. its a revelation just how lively the radio is now, it was literally receiving nothing down on the broadcast bands and was obviouslu ( now ) quite deaf all across its frequency range.

The only thing of note when I had a look at the solder connections on the board was one of the terminal pins on the protection lamp wasn't fully inserted into its hole in the pcb, I heated it from above and it pressed in a couple of mm. the only thing I can really think is that it wasn't actually making a solid connection, and was relying on capactive coupling between the bottom of the terminal pin and the pad, that's the only thing I can think of that would make such a dramatic difference.
Steve G4WCS is offline  
Closed Thread

Thread Tools



All times are GMT +1. The time now is 2:49 pm.


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.