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 20th Mar 2019, 10:53 pm   #21
MrBungle
Dekatron
 
MrBungle's Avatar
 
Join Date: Jun 2016
Location: London, UK.
Posts: 3,687
Default Re: Homebrew SSB Transceiver

Yes this was a 50W IRF510 push pull! I have upgraded them to T50-2 and they run around 40 degrees on CW which is better.
MrBungle is offline  
Old 21st Mar 2019, 2:41 pm   #22
Wendymott
Octode
 
Wendymott's Avatar
 
Join Date: Jun 2013
Location: Huddersfield, West Yorkshire, UK.
Posts: 1,782
Default Re: Homebrew SSB Transceiver

Hi Both...... I opted for T50-6 as Minikits supplied me with a module for 80/40metres that used T50-6 's The blurb sheet Quoted 100W on SSB. So thats why I chose those types. Ok Jeremy I could go to T80-6 but I am not thinking of higher powers..Im not of Italian extraction LOL.
See my comments in Homebrew...Things are more clear now.
__________________
Should get out more.

Regards
Wendy G8BZY
Wendymott is offline  
Old 21st Mar 2019, 5:07 pm   #23
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: Homebrew SSB Transceiver

The other reason for going for the larger T68-6 or T80-6 toroid is that they will be easier to wind when using thick wire about 1mm diameter. It might be a struggle to get 15 neat looking turns of 1mm diameter wire on a T50-6 toroid. I think the thick wire is needed to get optimal Q.

One obvious health and safety tip I'll pass on to everyone is that enamelled copper wire becomes hazardous to work with once you get above about 0.8mm diameter because it resembles a wayward skewer as it gets wound around a former. It takes a lot of user force to form it tightly around a core so it can store a lot of energy and can suddenly ping up into your face as you rotate the core in your hands. I would always recommend the use of eye protection when winding transformers or toroids with (say) 1mm diameter wire. Otherwise the free end can spring up and it might poke into an eye for example.
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Old 21st Mar 2019, 7:28 pm   #24
MrBungle
Dekatron
 
MrBungle's Avatar
 
Join Date: Jun 2016
Location: London, UK.
Posts: 3,687
Default Re: Homebrew SSB Transceiver

Definitely agree with eye protection! I tend to wear a long sleeve top and fleece as well. Managed to slash my arm open on some thick wire a while back
MrBungle is offline  
Old 22nd Mar 2019, 5:07 pm   #25
Wendymott
Octode
 
Wendymott's Avatar
 
Join Date: Jun 2013
Location: Huddersfield, West Yorkshire, UK.
Posts: 1,782
Default Re: Homebrew SSB Transceiver

Hi Jeremy and Mr Bungle... Well noted..I have ordered 50-6's for the time being..which will be adequate for my power levels, re "Q"..if not adequate I will get some T68-6's
Thanks
__________________
Should get out more.

Regards
Wendy G8BZY
Wendymott is offline  
Old 22nd Mar 2019, 9:59 pm   #26
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: Homebrew SSB Transceiver

To give you an idea of what Q to expect I had a go at winding a 780nH coil on a T50-6 toroid today and I ended up using 0.71mm wire. 1mm was just too thick to work with because it wouldn't be easy to adjust the inductance when tuning up the filter.

However, I had a bit of a surprise when I measured the Q of the inductor. I was expecting maybe 260 at 14.2MHz but I actually got a Q of just over 300.
This is higher than my ancient set of (Micrometals) Q curves suggested I would get. So I tried measuring the Q three different ways to see if something was wrong. I did the classic notch method on the VNA and got a null of 40.6dB. This suggested the Q was about 316. I then tried the 3dB BW method using my H field probe as the source. This showed the Q was 310. This method is really fast because the VNA measures the 3dB BW in real time when controlled from my little EeePC netbook. It finds the peak, measures the 3dB BW and computes and displays the Q on the VNA screen. I then tried a direct measurement of Q using S11 data from my VNA. This is the least reliable method but it indicated a Q of about 315. The S11 method with the VNA will be bathing in uncertainty so even though it agrees with the other results it doesn't really add much value.

However, see the plots below. I also used an old piece of Micrometals software to work out the turns required and how much wire I would need. See the screenshot below. This software is an ancient MSDOS program the company purchased at work from Micrometals at work nearly 30 years ago. I think they give a demo version away for free nowadays but I think it is still MSDOS based.

There's also a screenshot showing the post processed results of the S11 data from the VNA. This shows a Q curve that peaks at just over 300 and this does seem to be a bit higher than expected. However, the other methods I used agreed with this. For the null and 3dB tests I used 3x47pF ATC 800B caps in parallel and this should give an overall ESR of <0.02R so maybe the Micrometals Q curves are measured with a Q meter that has a relatively lossy air variable capacitor. None of my classic air variable caps can be used in place of the ATC caps up at 14MHz because they degrade the Q from 315 to about 270. This suggests the variable caps only have a Q of a couple of thousand at 14MHz and this isn't really good enough for a Q meter.

With a coil Q of just over 300 I think the 7th order lowpass filter could have an insertion loss of <0.15dB at 14.3MHz if also built with low loss caps.
Attached Thumbnails
Click image for larger version

Name:	T50_6_13T.gif
Views:	70
Size:	23.1 KB
ID:	180345   Click image for larger version

Name:	Q300.gif
Views:	63
Size:	18.2 KB
ID:	180346   Click image for larger version

Name:	Q300dip.gif
Views:	57
Size:	17.4 KB
ID:	180347   Click image for larger version

Name:	Toroid_T506_13T.jpg
Views:	63
Size:	95.6 KB
ID:	180348   Click image for larger version

Name:	Qnotch14.gif
Views:	52
Size:	8.6 KB
ID:	180349  

__________________
Regards, Jeremy G0HZU

Last edited by G0HZU_JMR; 22nd Mar 2019 at 10:04 pm.
G0HZU_JMR is offline  
Old 23rd Mar 2019, 12:09 pm   #27
Wendymott
Octode
 
Wendymott's Avatar
 
Join Date: Jun 2013
Location: Huddersfield, West Yorkshire, UK.
Posts: 1,782
Default Re: Homebrew SSB Transceiver

Hi Jeremy. Many thanks for the additional information. I have posted the new data in the Components section this am... it is obvious the "snide" parts are useless.
I will now build the 20M filter and post the results later.
__________________
Should get out more.

Regards
Wendy G8BZY
Wendymott is offline  
Old 26th Mar 2019, 9:47 pm   #28
Wendymott
Octode
 
Wendymott's Avatar
 
Join Date: Jun 2013
Location: Huddersfield, West Yorkshire, UK.
Posts: 1,782
Default Re: Homebrew SSB Transceiver

Hi Peeps... Finally I can put this filter module to bed..for those that helped.. a BIG thank you.
Info. All the capacitors except for 5 are now SMD 630V NPO. The toroids are all T50-6 from Micrometals.
A spreadsheet explains the markers on each graph.
I am happy with it..
Unless you see any glaring errors.
Attached Thumbnails
Click image for larger version

Name:	Filter.jpg
Views:	83
Size:	79.5 KB
ID:	180486   Click image for larger version

Name:	80 Metre filter.jpg
Views:	63
Size:	101.9 KB
ID:	180488   Click image for larger version

Name:	40 Metre filter.jpg
Views:	64
Size:	100.9 KB
ID:	180489   Click image for larger version

Name:	20 Metre filter.jpg
Views:	61
Size:	100.9 KB
ID:	180490  
Attached Files
File Type: xlsx Filter spreadsheet.xlsx (12.9 KB, 55 views)
__________________
Should get out more.

Regards
Wendy G8BZY
Wendymott is offline  
Old 26th Mar 2019, 9:52 pm   #29
MrBungle
Dekatron
 
MrBungle's Avatar
 
Join Date: Jun 2016
Location: London, UK.
Posts: 3,687
Default Re: Homebrew SSB Transceiver

Looking good. Thanks for reporting back
MrBungle is offline  
Old 27th Mar 2019, 2:48 am   #30
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: Homebrew SSB Transceiver

That looks good now you have the genuine Micrometals toroids

It looks like you have gone for a filter with a zero/notch because you have quite a sharp roll off rate on the 20m filter. The loss also looks quite low considering you are measuring the whole PCB with the relays etc.

This evening I had a quick go at making a basic 7th order cheby filter. This is a basic filter and it doesn't roll off quite as sharply as your filter but the reason I made it was to add confidence that the 800nH toroids really were giving a Q of >300 at 14.3MHz. This filter is very similar to Mr Bungle's design in that the cap and inductor values turned out to be very similar. I made it using multiple ATC 100pF SMD caps to make up the capacitance total at each null. So my filter has 200pF, 400pF, 400pF, 200pF at each node.

With a coil Q of 310, the simulation predicted just under 0.15dB loss at 14.3MHz and 55dB rejection at 30MHz. I've posted up the plots of both the VNA measurement and the simulation in case anyone wants to see what the loss of 'just' a lowpass filter is when built with the T50-6 toroids and SMD caps.

At these frequencies, the passband response of the simulation and the 'real' filter can be expected to agree very closely and you can see that they do agree even on a scale of 0.2dB/div. The insertion loss at 14.3MHz is about 0.14dB. I've used an exotic Ecal module to calibrate the VNA and this is able to minimise the overall measurement uncertainty such that a low insertion loss like this can be measured with a degree of confidence. I also measured the insertion loss on a lab power meter (with a padded sig gen as the source) and it also measured 0.14dB loss at 14.3MHz. So it does look like the 800nH T50-6 toroids can achieve a Q of about 300 at 14MHz.
Attached Thumbnails
Click image for larger version

Name:	W14lpf.gif
Views:	54
Size:	18.2 KB
ID:	180499   Click image for larger version

Name:	W14lpfm.gif
Views:	54
Size:	18.5 KB
ID:	180500   Click image for larger version

Name:	W14_sim_LPF0.gif
Views:	59
Size:	15.5 KB
ID:	180501   Click image for larger version

Name:	W14_sim_LPF100.gif
Views:	51
Size:	16.0 KB
ID:	180502  
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Old 27th Mar 2019, 11:42 am   #31
Wendymott
Octode
 
Wendymott's Avatar
 
Join Date: Jun 2013
Location: Huddersfield, West Yorkshire, UK.
Posts: 1,782
Default Re: Homebrew SSB Transceiver

Hi Jeremy.... I suppose we could go on trying various configurations .. but the main purpose of mine was to clean up the PA output and that is achieved now. Hopefully I will be "on air" with the transceiver in the next few days...in fact the 20 metre band was only added as a ..... nod to higher HF as the sunspot level is quite poor at the moment. I omitted the 15 and 10 metre bands.... for now..they could be added relatively simply if required as the VFO is a DDS circuit.
Thanks again Jeremy and Mr Bungle
__________________
Should get out more.

Regards
Wendy G8BZY
Wendymott is offline  
Old 7th Apr 2019, 2:47 pm   #32
Wendymott
Octode
 
Wendymott's Avatar
 
Join Date: Jun 2013
Location: Huddersfield, West Yorkshire, UK.
Posts: 1,782
Default Re: Homebrew SSB Transceiver

Hi Jeremy. I have just "re jigged" my Filter pcb.. Output of the Exciter to a Bandpass configuration..I used your circuit that you sent me a while ago, with 3 X 37 -2 toroids with L1 / L330 T and 7T for L2. I found the tuning was slightly low, so I reduced the series caps (150 pf) and fitted 100pf with a 5 - 30pf poly spaced variable in each leg. It tunes ok but the insertion loss was 9 db... this is with using 10 db pads on both input and output of the Rigol. I thought about it... then I replaced the T37.2 cores with T50-6... same turns ratio...heavier duty wire at 0.5mm, to hopefully increase the "Q". THe tuning was now "Peaky" with resonance at 7.1 mhz and 7.0 and 7.2 about 2 db down... the 9 Mhz was wayyyyyyyyyyy down in the noise, however the insertion loss is still 8 db........... I seem to remember your figures were for a flat top and an insertion loss of approx 1 db. As this is the pre PA filter I need as much drive as I can get and - 8db compared with -1 db for the 80 and 20M bands is a bit of a loss I could do without. If I used the 80 and 20M config for 40, I can get the drive ok and the spurious carrier at 9 Mhz is then -30 db.
Incidentally the caps for the 40M filter are all SMD.

Your thoughts are requested please.
PS Hopefully you can read the schematics ok
Attached Thumbnails
Click image for larger version

Name:	Bandpass filter 1.jpg
Views:	54
Size:	73.3 KB
ID:	180994   Click image for larger version

Name:	Bandpass filter 2.jpg
Views:	58
Size:	90.4 KB
ID:	180995  
__________________
Should get out more.

Regards
Wendy G8BZY

Last edited by Wendymott; 7th Apr 2019 at 2:52 pm. Reason: More info and typo
Wendymott is offline  
Old 7th Apr 2019, 9:52 pm   #33
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: Homebrew SSB Transceiver

I've had a quick look at your circuit and C5 and C6 don't look to be the right values. You have 1500pF in parallel with 680pF here and I'd expect to see a total of just 1000pF here. Then adjust the inductor L3 in series with them to give a null at 9MHz. I'd expect L3 to be just over 300nH once tuned correctly for a null at 9MHz.

I'll have another look tomorrow, I recall making this filter and I probably still have it somewhere. With an upgrade to T50-6 toroids I'd expect the insertion loss to be about 1dB at 7MHz but I'll double check this tomorrow.
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Old 8th Apr 2019, 10:48 pm   #34
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: Homebrew SSB Transceiver

I managed to find the original BPF but it had lost a few parts. However, I rebuilt it quickly this evening using T50-6 toroids for the 3900nH coils. I managed to get just under 1dB loss at 7MHz and I used the same schematic as the one in post #31 in this thread below.

https://www.vintage-radio.net/forum/...=138534&page=2

I'll try and tidy up the filter and post up a plot tomorrow. I'll also try and measure the Q of the 3900nH coils at 7MHz. I used 31 turns of 0.4mm En Cu wire on a T50-6 toroid for the series 3900nH inductors and I had to spread the turns over most of the toroid to get the response to peak up at 7MHz. It might have been easier to use 30 turns but I didn't try this. I used a T37-2 toroid for the 310nH coil and all the caps were SMD apart from the 150pF series caps. These were cheapo NP0 ceramic disc caps.
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Old 9th Apr 2019, 10:17 am   #35
Wendymott
Octode
 
Wendymott's Avatar
 
Join Date: Jun 2013
Location: Huddersfield, West Yorkshire, UK.
Posts: 1,782
Default Re: Homebrew SSB Transceiver

Hi Jeremy. I re looked at the previous posting and apart from the 9 mhz trap being way out..." I dont know what happened there" I think my version was "near enough" I used the trimmers to get the best possible tuning....maybe I should use "turn" adjustment rather than trimmers... I have a test jig so I will go back to it this PM.
__________________
Should get out more.

Regards
Wendy G8BZY
Wendymott is offline  
Old 9th Apr 2019, 5:05 pm   #36
Wendymott
Octode
 
Wendymott's Avatar
 
Join Date: Jun 2013
Location: Huddersfield, West Yorkshire, UK.
Posts: 1,782
Default Re: Homebrew SSB Transceiver

Hi Jeremy. I re visited the component values as suggested and replaced the 1n5/680pf caps with a 1n0 cap. I also replaced the L1/L3 windings with 29 T of 0.4mm wire. Keeping the trimmers and the values shown.. I now get a very acceptable insertion loss. see photo's. Thanks for your valuable help..
Attached Thumbnails
Click image for larger version

Name:	Bandpass filter ver 3.jpg
Views:	60
Size:	57.3 KB
ID:	181102   Click image for larger version

Name:	40 Metre filter.jpg
Views:	56
Size:	89.8 KB
ID:	181103   Click image for larger version

Name:	Bandpass filter Ver 3 pcb.jpg
Views:	67
Size:	81.9 KB
ID:	181104  
__________________
Should get out more.

Regards
Wendy G8BZY
Wendymott is offline  
Old 9th Apr 2019, 9:42 pm   #37
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: Homebrew SSB Transceiver

Looks good! It should be much nicer to tune up using those trimmer caps. By contrast, I don't find it rewarding to tune up a filter by squishing or stretching the turns on a toroid.

I've not had a chance to check out the Q of the T50-6 3900nH toroid but I expect it will be somewhere around 250 or so.

You must be fairly close to having a complete signal path now!
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Old 9th Apr 2019, 10:38 pm   #38
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: Homebrew SSB Transceiver

I had a quick go at measuring the Q of the T50-6 3900nH coil and the results are as below. It looks like the Q is up around 270 at 6.5MHz. I measured it in my two Q jigs using 3 x 47pF ATC caps as the resonating capacitance.

This resonated at about 6.5MHz and this is close enough to 7MHz. The 3dB method gave a Q of 268 at 6.5MHz as in the image below and the notch method showed a Q of 271 at a similar frequency.

I also measured it on my VNA. Despite the high cost of the instrument, this is the least reliable method but my VNA generally does surprisingly well in tests like this. The post processed VNA data produces a Q curve and the Q was about 268 at 6.5MHz as in the Genesys screenshot below.

The agreement between the three methods is very good and this is partly because the self capacitance of the coil isn't having much effect on the (measured) inductance at around 7MHz. Otherwise, the -3dB BW method would begin to show different readings for Q compared to the other methods.
Attached Thumbnails
Click image for larger version

Name:	Q3900.gif
Views:	44
Size:	17.8 KB
ID:	181128   Click image for larger version

Name:	Q3900shunt.gif
Views:	46
Size:	17.3 KB
ID:	181129   Click image for larger version

Name:	3900Q_excel.gif
Views:	45
Size:	8.6 KB
ID:	181130   Click image for larger version

Name:	3900Q_Genesys.jpg
Views:	44
Size:	91.1 KB
ID:	181135  
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Old 9th Apr 2019, 11:08 pm   #39
Wendymott
Octode
 
Wendymott's Avatar
 
Join Date: Jun 2013
Location: Huddersfield, West Yorkshire, UK.
Posts: 1,782
Default Re: Homebrew SSB Transceiver

As you say it is getting close to complete the signal path now..I am very pleased with the results..I am getting quite happy using toroids in these circumstances now..as I have good user tables, to enable turns/uH to be calculated easily. I didnt realise how close the T37-6 was to the T50-6 in permeability. OK about your "Q" measurements.. as you say all 3 methods seemed to agree. Thank you again for your input, I hope this info has been useful to others in the group...Just got the PA and antenna relay switching to sort out now.
__________________
Should get out more.

Regards
Wendy G8BZY
Wendymott is offline  
Old 19th Apr 2019, 3:29 pm   #40
Wendymott
Octode
 
Wendymott's Avatar
 
Join Date: Jun 2013
Location: Huddersfield, West Yorkshire, UK.
Posts: 1,782
Default Re: Homebrew SSB Transceiver

Hi peeps. Another issue... "not problem" is causing me a teensy bit of concern.

The RF output drops from >10W pep to < 10W with a continuous dual sinewave as per the scope display WF1. It seems to be heat related to the drive transistor Q2 which is a 2SC1971 as supplied in the kit. If I use freezer the PA current rises as does the RF output. Q2 is well mounted and the heatsink is fan assisted.
If you see the lower waveform it seems to be operating in a "non linear" mode..is the biasing wrong ?? Should it be working in class "A" or class "C". I have not had much to do with this area before..... I never investigated my 40M pa which is dismantled at present.
The eventual output looks good, monitored from the RF sniffer to the dummy load.
Your comments would be appreciated
Attached Thumbnails
Click image for larger version

Name:	SCH 1.jpg
Views:	97
Size:	40.4 KB
ID:	181521   Click image for larger version

Name:	WF 1.jpg
Views:	84
Size:	60.0 KB
ID:	181522  
__________________
Should get out more.

Regards
Wendy G8BZY
Wendymott is offline  
Closed Thread

Thread Tools



All times are GMT +1. The time now is 5:05 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.