Next step then I think is to sort out the biasing on the driver and PA, so they dont get so hot, and to measure the output power, see if it really is producing what it should be! But these are probably all tasks for another day.
Monday, 30 September 2013
Balancing Act
Several tricks have been tried with the 10m Wispy to try and get the carrier balance right, so far none have been entirely successful. A balance control potentiometer had no effect at all, and changing the diodes seemed to have no effect either. So, I reverted the circuit back to its original form, put a pot in place of R11 to make it adjustable again, and totally rewound T2, physically symmetrically. An RF bypass capacitor has also been put on the AF modulation input
Things do seem a little improved. The carrier is still present, but I can get it at least to look more like DSB should do. Adjusting R11 for best 2nd harmonic resulted in what looks to me to be a sensible spectra for a doubler. The 28MHz 2nd harmonic is dead center, the first two peaks from the left are an image of the 14MHz fundamental, and the machines zero marker
The harmonics are in the form of a 'staircase' going down in relative strength. Adjusting now for best carrier balance at the mixer output, I managed to get the signal looking like this
Yes, its much weaker than the doubler, but I suppose taking into account the differences from changing the analyzer settings, and the insertion loss of the mixer, and the likely mismatch between the analyzer and the input to the driver where this signal was taken, it at least shows that the carrier is down on the sidebands. This is with a 10kHz 50mV AF input. 10kHz is needed to be able to see the sidebands apart from the carrier on this machine. Looking again at the doubler, shows that when set for best carrier null, it isnt quite as good an output on the 2nd harmonic, and the 3rd isnt down as far as it was
Putting the probe now to the output of the PA, just after the Low Pass Filter, we see an amplified DSB signal, still with significant carrier though. Whether this is good enough for on-air use I dont yet know. Im not certain by how much I should expect the carrier to be suppressed, Roger claimed 30dBm, mine I think is about 14dB down, assuming im reading the scale properly!
I shall have to make a point of listening to the output on a general coverage receiver in SSB mode to see how noticeable the carrier is. The 3rd harmonic isnt very well suppressed either, being barely below the wanted second, but interestingly its carrier is almost none existent! There are also more noticeable extra mixer products. The picture isnt so good as the camera captured the interval of the sweep time! but the upper sideband can be made out
The 3rd harmonic then needs looking at, perhaps making the driver and PA inductors as tuned circuits will help here. I could not observe any greater harmonics though.
Next step then I think is to sort out the biasing on the driver and PA, so they dont get so hot, and to measure the output power, see if it really is producing what it should be! But these are probably all tasks for another day.
Next step then I think is to sort out the biasing on the driver and PA, so they dont get so hot, and to measure the output power, see if it really is producing what it should be! But these are probably all tasks for another day.
Sunday, 29 September 2013
Muppet Show
Got the Larkspur mast guying kit out for inspection today. The valise needs a new sheet of plywood gluing in over the original, which is split, but other than that its not in bag nick really. Now have eight guy lines, but only one base plate pin. The ground spike and base plate that came with it are not much use, so have been replaced with those I already had. I might think about putting the mast up for a test tomorrow.
Building the RF diode probe proved rather simple this morning, and the pen selected was spot on for it. Cutting off the tip a little earlier than the main body left a gap the same diameter as the body of the cut down test probe, which fit in snug, held in place with a dab of superglue.
A small hole was bored through the side wall for the ground clip, and the end cap was slit and a small hole made for the coax. It was then slipped over the coax and pushed back into place, sealing the end and securing the cable. A pair of 4mm plugs and a couple of quick continuity tests later, and it was ready for testing with RF.
RF testing was accomplished by connecting it to my DVM, and probing the BNC output connector of my Marconi 2955B communications test set. With the BNC selected, it will give an RF output up to +5dBm. A quick run through at 28MHz allows comparison with the signals on the 10m WSPR unit -
0dBm = 0.38v 3dBm = 0.56v
1dBm = 0.44v 4dBm = 0.63v
2dBm = 0.49v 5dBm = 0.72v
Of course, these are at 50ohms, so the readings in-circuit wont be exactly the same.
After much testing with parallel capacitors on the doubler stage of the wispy, I finally managed today to get a very high 28MHz output with virtually no 14MHz or 42MHz component. Very good I thought, er, no, actually! When I applied modulation, all I got was a very good AM signal! Whats going on? It turns out i'd forgotten to move the probe from the spectrum analyser from the output of the mixer back to the collector of T2, the doubler transistor!
So, all the work to get the parallel capacitance right is in fact wrong. That all needs to be redone now. I also noticed that the driver stage isnt, it appears to have unity gain, what goes in comes out no bigger. It also gets rather hot for a 2N3904! Something then is definately not right there.
So ive decided to go back to the start and work forward, solving any issues as I come to them. Using my nice new diode probe, I find the output of the oscillator is 0.38v, about 0dBm, or 1mW. The output of the doubler is a touch higher at 0.52v, about 2.5dBm or 1.8mW. This seems low to me, im not sure what the insertion loss of the mixer is, but I bet its more than 2.5dB! Measuring at the crystal itself showed 0.79v, about 5.8dBm or 3.8mW.
So, the questions are - what level output should I be expecting from the oscillator and the doubler? What loss must be overcome in the mixer? What level drive does the mixer need? And, why am I not getting those figures?
Im also concerned about the balance of the mixer. After a lot of searching on the tinterweb, ive finally found a similar design that shows a value for a balance resistor, at 200ohm. Im pretty sure I dont have any 200ohm presets though. Theres also a hint that the insertion loss might be about 8dB.
Building the RF diode probe proved rather simple this morning, and the pen selected was spot on for it. Cutting off the tip a little earlier than the main body left a gap the same diameter as the body of the cut down test probe, which fit in snug, held in place with a dab of superglue.
A small hole was bored through the side wall for the ground clip, and the end cap was slit and a small hole made for the coax. It was then slipped over the coax and pushed back into place, sealing the end and securing the cable. A pair of 4mm plugs and a couple of quick continuity tests later, and it was ready for testing with RF.
RF testing was accomplished by connecting it to my DVM, and probing the BNC output connector of my Marconi 2955B communications test set. With the BNC selected, it will give an RF output up to +5dBm. A quick run through at 28MHz allows comparison with the signals on the 10m WSPR unit -
0dBm = 0.38v 3dBm = 0.56v
1dBm = 0.44v 4dBm = 0.63v
2dBm = 0.49v 5dBm = 0.72v
Of course, these are at 50ohms, so the readings in-circuit wont be exactly the same.
After much testing with parallel capacitors on the doubler stage of the wispy, I finally managed today to get a very high 28MHz output with virtually no 14MHz or 42MHz component. Very good I thought, er, no, actually! When I applied modulation, all I got was a very good AM signal! Whats going on? It turns out i'd forgotten to move the probe from the spectrum analyser from the output of the mixer back to the collector of T2, the doubler transistor!
So, all the work to get the parallel capacitance right is in fact wrong. That all needs to be redone now. I also noticed that the driver stage isnt, it appears to have unity gain, what goes in comes out no bigger. It also gets rather hot for a 2N3904! Something then is definately not right there.
So ive decided to go back to the start and work forward, solving any issues as I come to them. Using my nice new diode probe, I find the output of the oscillator is 0.38v, about 0dBm, or 1mW. The output of the doubler is a touch higher at 0.52v, about 2.5dBm or 1.8mW. This seems low to me, im not sure what the insertion loss of the mixer is, but I bet its more than 2.5dB! Measuring at the crystal itself showed 0.79v, about 5.8dBm or 3.8mW.
So, the questions are - what level output should I be expecting from the oscillator and the doubler? What loss must be overcome in the mixer? What level drive does the mixer need? And, why am I not getting those figures?
Im also concerned about the balance of the mixer. After a lot of searching on the tinterweb, ive finally found a similar design that shows a value for a balance resistor, at 200ohm. Im pretty sure I dont have any 200ohm presets though. Theres also a hint that the insertion loss might be about 8dB.
Saturday, 28 September 2013
Probe Plan
I really should have done this a long time ago, but these past few days working on the 10m beacon have shown just how much I need a diode RF probe. So, construction has started on one, now I have some suitable diodes. To be honest, the idea was in my mind yesterday, which is why I picked up quite a few 1N34s.
What has always stopped me was the difficulty I found in building the actual probe part, but I decided to sacrifice one of an old pair of meter probes. The next trouble is a suitable barrel to house it all in. A felt tip pen in my pen pot turns out to be the ideal dimensions. With the probe cut down, and a small twin strip bit of veroboard soldered on, its now ready for the components to be mounted tomorrow
It turned out that the probe rod went the entire length of the probe body, so when I find the rest (somewhere on the shelf where it landed after cutting off!) theres enough to make a few more!
As I was finding this picture out from where the PC had saved it, I spotted again an old mugshot of yours truly in the act of calmly knackering the UKs telly
I think this is a few years old, the two telemetry screens are TMC and ARQOSS, so its prior to the final DSO. Im only actually knacking up Scotland and Northern Ireland there.
What has always stopped me was the difficulty I found in building the actual probe part, but I decided to sacrifice one of an old pair of meter probes. The next trouble is a suitable barrel to house it all in. A felt tip pen in my pen pot turns out to be the ideal dimensions. With the probe cut down, and a small twin strip bit of veroboard soldered on, its now ready for the components to be mounted tomorrow
It turned out that the probe rod went the entire length of the probe body, so when I find the rest (somewhere on the shelf where it landed after cutting off!) theres enough to make a few more!
As I was finding this picture out from where the PC had saved it, I spotted again an old mugshot of yours truly in the act of calmly knackering the UKs telly
Friday, 27 September 2013
National Hamfest
Set off early for the National Hamfest this morning, should have been a nice quick run, but several accidents meant detours, one of which from off the A616, I had no idea where to go, but a local window installers van was in front of me, so I followed him! Arrived only about half an hour after opening, so wasnt too bad.
After paying in, which at a fiver I do think is a bit steep, especially since, despite this being our 'National' rally, its still not a patch on rallies of my early radio days such as Elvington and Drayton, I decided to do my usual 'tour' and take a look at everything before starting to shop. My plan soon went awry, when spotted beside a stall with a few PRC-320s on, was this
an almost complete, to CES, guying kit for the 27ft Mk1 Larkspur mast! Its missing a guy rope, the elusive spanner, and the 4lb hammer, but it does include the guy pegs ive so desperately been searching for, and the base insulator, plus a brand new, unissued, 1976 dated instruction plate! Now, ive seen these go for £40 on ebay without the pegs, so I set myself a £30 limit, and asked the stallholder how much "well, I was going to ask a fiver per item" says he, "or, £15 for the lot", well, I couldnt get a twenty quid note out fast enough!
But, it did give me one small problem - i'd been there under ten minutes and was already carrying a very heavy bag! So I obtained a pass out stamp on my hand and took it all to the car, before heading back in to start over again.
Although reasonably small still, it was none the less interesting, i'd have prefered to see more junk than computer stuff, but there wasnt too much computer kit. Some very nice masts which I can only admire, and even met one of our field engineers, who ive bossed about at daft o'clock, but never met. I managed to avoid the cup cake stall, although I was very tempted! and concentrated on getting what id gone for. Well, several 2nd hand N-types at 50p a pop filled that quota, 1N34 and 1N4148 diodes from Bowood, and the main items were accounted for. One stall had little sealed bags at a pound with a selection of trimmer capacitors in them, so some of those went in the bag. I took a bit of a punt on a 50p valve socket that may or may not have been the right one for the dekatrons, not as it turned out!
I also wanted some single sided PCB material as im running low. Most stalls with this were charging upwards of £2.40 for a 4" x 6" sheet. Whilst traversing the car boot, I spied some big sheets of it right at the back of a stall. These turned out to be double sided, but were something like 6" x 30" at least! I wasnt too keen at first, being double, as its a hassle taking one side of copper off, but when I asked the price and was told "20p a chunk", well you cant pass up a bargain like that! I bought five for a quid, and as you can see, it really was a bargain
My only 'frivolous' purchase was a little antenna, supposedly a dual 2m/70cm, according to the stall holder. Im not convinced, as its under a foot long, but he wanted a pound, and accepted the remains of my change, which was 90p! I'll need to test it somehow and find out what frequencies it works on. I also bought an antenna base spring for 50p, which might not be strong enough for the big HF whips, but im sure will hold a 10 or 12m antenna
Once home, via a well awaited Tower burger meal (large with Pepsi), I was very pleased to find my copy of Sprat awaiting me on the doormat. I could now turn my attention back to the 10m Wispy.
Firing it up, and looking at the spectra on the output of the mixer, I can clearly see that the fundamental, plus the third and fifth harmonics, are present along with the 28MHz signal, albeit at a somewhat reduced level. Setting the spectrum analyser up for observing the modulation, I can now see that the 28MHz signal is showing DSB, and the 3rd and 5th harmonics are much more AM like.
Curious as to whether im getting the maximum 28MHz signal, I looked again at my setup. Rogers original design stated 60pF trimmer, and I have a 35pF trimmer in parallel with a 33pF fixed cap. But, is it enough? The key to that would be if the trimmer was set at maximum. But these little ceramic ones are not easy to work that out with. So, taking my capacitance meter and the rest of the bag, I tested several, setting them to maximum and marking them. Maximum it turns out coincides with the top plates center being aligned with the single pin. Looking at my setup on the Wispy, it is indeed at maximum. I have a feeling that adding a bit more parallel capacitance will give me a bigger 28MHz peak, and allow me to gain that peak at a mid-range position. But thats a task now for another day.
After paying in, which at a fiver I do think is a bit steep, especially since, despite this being our 'National' rally, its still not a patch on rallies of my early radio days such as Elvington and Drayton, I decided to do my usual 'tour' and take a look at everything before starting to shop. My plan soon went awry, when spotted beside a stall with a few PRC-320s on, was this
an almost complete, to CES, guying kit for the 27ft Mk1 Larkspur mast! Its missing a guy rope, the elusive spanner, and the 4lb hammer, but it does include the guy pegs ive so desperately been searching for, and the base insulator, plus a brand new, unissued, 1976 dated instruction plate! Now, ive seen these go for £40 on ebay without the pegs, so I set myself a £30 limit, and asked the stallholder how much "well, I was going to ask a fiver per item" says he, "or, £15 for the lot", well, I couldnt get a twenty quid note out fast enough!
But, it did give me one small problem - i'd been there under ten minutes and was already carrying a very heavy bag! So I obtained a pass out stamp on my hand and took it all to the car, before heading back in to start over again.
Although reasonably small still, it was none the less interesting, i'd have prefered to see more junk than computer stuff, but there wasnt too much computer kit. Some very nice masts which I can only admire, and even met one of our field engineers, who ive bossed about at daft o'clock, but never met. I managed to avoid the cup cake stall, although I was very tempted! and concentrated on getting what id gone for. Well, several 2nd hand N-types at 50p a pop filled that quota, 1N34 and 1N4148 diodes from Bowood, and the main items were accounted for. One stall had little sealed bags at a pound with a selection of trimmer capacitors in them, so some of those went in the bag. I took a bit of a punt on a 50p valve socket that may or may not have been the right one for the dekatrons, not as it turned out!
I also wanted some single sided PCB material as im running low. Most stalls with this were charging upwards of £2.40 for a 4" x 6" sheet. Whilst traversing the car boot, I spied some big sheets of it right at the back of a stall. These turned out to be double sided, but were something like 6" x 30" at least! I wasnt too keen at first, being double, as its a hassle taking one side of copper off, but when I asked the price and was told "20p a chunk", well you cant pass up a bargain like that! I bought five for a quid, and as you can see, it really was a bargain
My only 'frivolous' purchase was a little antenna, supposedly a dual 2m/70cm, according to the stall holder. Im not convinced, as its under a foot long, but he wanted a pound, and accepted the remains of my change, which was 90p! I'll need to test it somehow and find out what frequencies it works on. I also bought an antenna base spring for 50p, which might not be strong enough for the big HF whips, but im sure will hold a 10 or 12m antenna
Once home, via a well awaited Tower burger meal (large with Pepsi), I was very pleased to find my copy of Sprat awaiting me on the doormat. I could now turn my attention back to the 10m Wispy.
Firing it up, and looking at the spectra on the output of the mixer, I can clearly see that the fundamental, plus the third and fifth harmonics, are present along with the 28MHz signal, albeit at a somewhat reduced level. Setting the spectrum analyser up for observing the modulation, I can now see that the 28MHz signal is showing DSB, and the 3rd and 5th harmonics are much more AM like.
Curious as to whether im getting the maximum 28MHz signal, I looked again at my setup. Rogers original design stated 60pF trimmer, and I have a 35pF trimmer in parallel with a 33pF fixed cap. But, is it enough? The key to that would be if the trimmer was set at maximum. But these little ceramic ones are not easy to work that out with. So, taking my capacitance meter and the rest of the bag, I tested several, setting them to maximum and marking them. Maximum it turns out coincides with the top plates center being aligned with the single pin. Looking at my setup on the Wispy, it is indeed at maximum. I have a feeling that adding a bit more parallel capacitance will give me a bigger 28MHz peak, and allow me to gain that peak at a mid-range position. But thats a task now for another day.
Thursday, 26 September 2013
DSB finally
Working today on the 10m Wispy transceiver, which is the design here
After a lot of headscratching, discussion with others etc, it looks like im finally getting somewhere with it. The odd phenomenon of the 42MHz AM output started me investigating the doubler stage further. I then noticed that C16, 35pF in this design, was marked as 60pF in the original Tx only version. 42MHz is the 3rd harmonic of 14MHz, was the stage working as a tripler instead? Patching 15pF across C16 brought the 2nd harmonic peak up more than the 3rd, whereas before they were equal. Hmmm, lets try 33pF in parallel then, lo and behold, an even bigger 2nd harmonic peak! Switching back then to probe the output of the mixer, I see this
a 28Mhz double sideband signal! At last! OK, so the carrier suppression isnt very good yet, but its DSB, and at this stage thats whats important. I can now tidy this signal up, and start getting the amplifiers up and running properly.
Spreading the spectrum
The 10m WSPR beacon still causes me headaches. After a lot of twiddling, and much technical advice from Ken G8BEQ (Thanks Ken!) I realised a mistake i'd made with the bias setting of the doubler, and now once again have some sort of signal from it. The modulator however, continues to give trouble. There seems to be no carrier balance at all, all the 28MHz LO in all comes out, AF input or not. So, I have added a preset resistor to try and null the carrier.
Before I can sensibly try it though, I need to get around the problem of observing the spectrum properly on the spectrum analyser. Thanks to Bob G3OOU, I now know how to set the analyser up for this, although it shows that the machine is rather drifty! (either that or the 2955 is!). It seems that since the machines minimum span is 100kHz, I need to really give it some welly with a modulation frequency if at least 10kHz to be able to make the spectrum spread out over a few divisions. Now I know this, I can give the WSPR transceiver 10 or 15kHz of AF drive and see what the output of the modulator really is doing.
One further thing that Ken has advised me with, is the standing current of the PA section. In the last post I mentioned it was getting too hot even with no drive, well, it seems its dissipating nearly a Watt! This is supposed to be a 200mW output PA! Ken has given me some new bias values to try which bring the standing current down to a mere 20mA. But the PA has to wait, the modulation needs sorting first!
I still have my doubts about either the transformer or the diodes. Its the National Hamfest tomorrow, so if im able to go, i'll get some proper OA91 or 1N34s!
Rang Geoff at Castle Electronics today, my FT-857D is repaired and ready. RF preamp transistor, set of IF filters and a few PIN diodes needed changing. Learnt an interesting trick from Geoff on using both antenna ports to see if the front end streering diodes are leaky. Should have the rig back mid week (being sent to work) so finally back on air HF mobile!
Before I can sensibly try it though, I need to get around the problem of observing the spectrum properly on the spectrum analyser. Thanks to Bob G3OOU, I now know how to set the analyser up for this, although it shows that the machine is rather drifty! (either that or the 2955 is!). It seems that since the machines minimum span is 100kHz, I need to really give it some welly with a modulation frequency if at least 10kHz to be able to make the spectrum spread out over a few divisions. Now I know this, I can give the WSPR transceiver 10 or 15kHz of AF drive and see what the output of the modulator really is doing.
One further thing that Ken has advised me with, is the standing current of the PA section. In the last post I mentioned it was getting too hot even with no drive, well, it seems its dissipating nearly a Watt! This is supposed to be a 200mW output PA! Ken has given me some new bias values to try which bring the standing current down to a mere 20mA. But the PA has to wait, the modulation needs sorting first!
I still have my doubts about either the transformer or the diodes. Its the National Hamfest tomorrow, so if im able to go, i'll get some proper OA91 or 1N34s!
Rang Geoff at Castle Electronics today, my FT-857D is repaired and ready. RF preamp transistor, set of IF filters and a few PIN diodes needed changing. Learnt an interesting trick from Geoff on using both antenna ports to see if the front end streering diodes are leaky. Should have the rig back mid week (being sent to work) so finally back on air HF mobile!
Thursday, 19 September 2013
Transformer Troubles
Well, much Rf work was put aside for some time, whilst the show was got out the way, and the FT-857D was sent of to Geoff at Castle Electronics to be fixed. Also during this time, I myself underwent a 'gentlemans procedure' and all thoughts of radio were off for a few days.
The show was a pretty good success. Operating predominantly on 40m SSB, we made a couple hundred QSOs. 2m FM showed itself as suspected to be pretty much a lost cause. Bob M1BBV and Steve G7TAO both came to help out, and their assistance was very much appreciated, as was the dual band handie from Steve!
So now I can get back to a bit of RF engineering. The Wispy 10m WSPR transceiver is first up for completion. All the transmit side is now built, as can be seen below
The show was a pretty good success. Operating predominantly on 40m SSB, we made a couple hundred QSOs. 2m FM showed itself as suspected to be pretty much a lost cause. Bob M1BBV and Steve G7TAO both came to help out, and their assistance was very much appreciated, as was the dual band handie from Steve!
So now I can get back to a bit of RF engineering. The Wispy 10m WSPR transceiver is first up for completion. All the transmit side is now built, as can be seen below
But, its not yet operational. With no AF input, im still seeing lots of output, but also lots of harmonics. AF input doesnt seem to have any affect. Although the spectrum analyser says the 28MHz signal is strongest, the Marconi 2955 doesnt register it, but instead shows a few tens of mW at about 43MHz. I suspect that theres a problem with the RF transformer in the mixer. The PA is obviously taking current, as its getting hot, perhaps a bit too hot!
I think the transformer must be the culprit, and i'll take it off and have a look.
Further to the show station, the thought occurs that since were held in the grounds of the Deaf Trust, we should be demonstrating at the station ways in which the hobby can be enjoyed by those with hearing problems. So, im starting to investigate digital modes, with the idea of operating a digital station next year. It seems either old fashioned RTTY, or PSK31 are the modes to choose.
Subscribe to:
Posts (Atom)