The AM Forum
March 29, 2024, 11:20:27 AM *
Welcome, Guest. Please login or register.

Login with username, password and session length
 
   Home   Help Calendar Links Staff List Gallery Login Register  
Pages: 1 [2] 3 ... 5   Go Down
  Print  
Author Topic: ARC-5 works great on AM...  (Read 93462 times)
0 Members and 1 Guest are viewing this topic.
WB3JOK
Member

Offline Offline

Posts: 637



« Reply #25 on: February 02, 2015, 12:06:27 PM »

Strange goings-on... in crowded band conditions with the sensitivity restored, I noticed it did not seem adequately selective. After much frustrating experimentation last night and this morning, initially assuming the crystal was broken, I finally discovered that with the IF shield can removed from the assembly, the filter is quite sharp over the nose although (not surprisingly with a single crystal) the skirts and stop-band aren't very good. Still better than original.

But when I reinstall the shield can, the selectivity goes away considerably as I slide it over the frame!  Huh It's gradual, not "switched" so I'm positive there is no short circuit to the can. Perhaps the loss of Q wasn't an issue with the designed 20+ KHz IF BW (and that's with all the stages in series, so each can is probably even wider) but it's a big problem with a narrow crystal filter.

I think I may just remove the original coil and replace it with another T50-43 toroid. To resonate the original fixed 180 pf (plus 10 pf nominal from the air trimmer), my calculator tells me 16.7 uH are needed and that's only six turns. A trifilar winding is easy to make and keeps the phasing trimmer too.
Logged
WB3JOK
Member

Offline Offline

Posts: 637



« Reply #26 on: February 02, 2015, 04:42:00 PM »

Removed the original ceramic form air-core coil, and replaced it with the proposed trifilar-wound toroid. That took care of the shield-can sensitivity! It's still interestingly less selective with the can in the radio,  although the effect is less pronounced now, and I can't figure it out.

This is a simple circuit - the output comes from the plate of the 12K8 mixer to the IF can, and its small plate current goes through the primary. Then the output from the can just goes directly to the grid of the 1st IF stage. Nothing there to load it down, and with a crystal filter the load impedance is the primary determinant of selectivity (which is in the can). I'm not putting in enough signal to cause the primitive AVC to affect the gain, either.  Huh
Logged
WB3JOK
Member

Offline Offline

Posts: 637



« Reply #27 on: February 02, 2015, 08:13:36 PM »

After pondering further it appears that the Z presented to the mixer plate was probably far too low, given an estimated ESR of the 2.830 MHz crystal of 300 ohms and a 680 ohm load resistor through the new 1:1 input transformer. A quick use of an online pi-network calculator showed that 17 pf input and 169 uH series (20 turns on a FT50-43 toroid) would match 10k(?) ohms down to 1k ohm at 2.830 MHz. So I installed that under the chassis by simply lifting the plate lead from the 12K8 (pin 3) and using a handy 8.2 pf NPO disc (to allow for tube + wiring capacitance).

I didn't sweep it on the signal generator or check the sensitivity again (that 8640B is heavy and it's on another bench!). However, subjectively it appears quite adequate when connected to my OCF dipole. Selectivity will never be close to "battle mode"-ready without a four- or five-pole ladder filter, but that's a project for another day. I listened in on a few QSO's, checked the time on CHU (7.850 MHz AM), and called it good.  Cool
Logged
WB3JOK
Member

Offline Offline

Posts: 637



« Reply #28 on: February 03, 2015, 11:26:52 PM »

Here's the current version of the circuit in a .pdf document. I think it could be improved with the addition of a second crystal at 2831.5 or 2828.5 KHz instead of the "phasing" trimmer, but probably wouldn't be worth the effort since the skirts still would not be very steep and the filter nose doesn't seem excessively narrow as it is.

There's plenty of room inside the IF can for more circuitry once the original bulky ceramic coil form is removed (T50 toroids are very compact). However, four or more HC-6/U or HC-33/U crystals would be hard if not impossible to fit, and I'm not sure if 2.83 MHz crystals come in the small HC-49/U package.

Any suggestions from those more skilled in RF circuitry would be welcome.  Smiley

* ARC-5 crystal filter.pdf (18.23 KB - downloaded 372 times.)
Logged
WB3JOK
Member

Offline Offline

Posts: 637



« Reply #29 on: February 04, 2015, 05:04:29 PM »

Started on the transmitter since I got the missing 1626 tube in the mail today. I made a temporary wiring harness to my screen mod/power supply. While tracing out the schematic which I'd misplaced sometime since '04, I discovered that I'd made a PC board for the power supply components! (see pics). Somewhere I've got the Eagle files if anyone needs them. Basically it's 210 volts regulated for the oscillator, 150 volts regulated for the screens (fed through an improvised modulation transformer, actually a Triad N-48X 115v 15VA isolation transformer). Standard '60's ARRL Handbook 12AX7 speech amp driving a 6AQ5.

I put a 50 pf doorknob in series with the output and it loaded up nicely into my Bird dummy load at about 70% of the roller inductor. Dial is only off a KHz or two, not bad after 72 years. It's overmodulating a bit on the scope (displaying 20 v/div into 50 ohm load), but it's hard to take a cell phone pic and get the audio gain adjusted right at the same time  Wink


* P02-04-15_13.43.jpg (180.61 KB, 1280x960 - viewed 807 times.)

* P02-04-15_13.44.jpg (223.4 KB, 1280x960 - viewed 817 times.)

* P02-04-15_15.41.jpg (148.26 KB, 1280x960 - viewed 786 times.)
Logged
WB3JOK
Member

Offline Offline

Posts: 637



« Reply #30 on: February 04, 2015, 05:15:43 PM »

Regulating the oscillator voltage near 200V (two OB2's is about 210V) is the prescription for stability. There's amazingly little drift even from cold, and the eye tube opens at 7000.314 KHz with a FT-243 crystal marked 7000.00 in the socket (which could be the error of a 40 year old crystal itself, there's no trimmer and I haven't put the counter on it).

The Bird is only accurate at VHF which is why the needle is barely moving. Actual unmodulated carrier power is around 10 watts  Cool

Anyone got a (black) cover and a bottom plate to finish it off?  Smiley


* P02-04-15_14.56.jpg (213.62 KB, 1280x960 - viewed 827 times.)

* P02-04-15_14.58.jpg (112.69 KB, 1280x960 - viewed 780 times.)
Logged
VE3LYX
Member

Offline Offline

Posts: 769


Crystals are from the stone age


WWW
« Reply #31 on: February 04, 2015, 07:48:42 PM »

I am missing a bottom plate for my RX but then it was give to me at Durham hamfest  so I better not whine! Was pretty banged up but works fine now. gladyou gotthis tx up and running.
don
Logged

Don VE3LYX<br />Eng, DE & petite Francais
WB3JOK
Member

Offline Offline

Posts: 637



« Reply #32 on: February 05, 2015, 02:54:50 PM »

Having resumed work on my '04 screen modulator design, I'm having to fine-tune it a bit.
First, there seemed to be a lot of hum on the output for 50 uf of HV filter capacitance. About 20 volts of ripple on the +600 key-down. Hmm... something I forgot about the "economy" configured power supply. The effective filter capacitance for the HV output is not only the HV output filter, since it's effectively in series with the LV filter (center-tap) capacitor  Tongue
So I really had 50 uf in series with 22 uf, or about 15 uf! Adding 33 uf across the LV cap took care of much of the ripple.

I also found that, although I was regulating the PA screen voltage with an OA2 at 150v, that basically gives full carrier output with 600v on the plates. I must have tried to "rig" that back in '04 with a 56K dropping resistor and a 10 uF electrolytic cap feeding the modulation transformer. Doesn't give a very stable carrier output. Jumpering out the resistor (with a milliammeter) showed about 4 ma resting screen current, increasing on voice peaks, and a larger carrier output - but no upward modulation at all  Huh

I think this is telling me to use a lower screen voltage... If I can find a spare 0B2 (108 volts) I'll just plug it in place of the 0A2 and see, but that may be too low. Hope I don't have to resort to those newfangled Zeners Cheesy

Edit: found an OB2, plugged it in. 2 ma resting screen current, 13W unmodulated carrier with decent looking upward modulation only if the variometer is set to full coupling. At lower couplings, although the unmod carrier does decrease, so does the upward modulation and the waveform really deteriorates. I don't remember enough about tetrodes and screen modulators to figure this out, but I'll run it the way it looks best on the scope.

I also found enough room behind the transmitter front panel to mount the Centralab 50 pF doorknob once the old antenna relay was removed. Rather than hack up the panel for a coax connector, I rigged up an RG-58 pigtail to the binding post with a PL-259 on the other end.

Now I just have to dig out my homemade 28V T/R relay and maybe I can put the station on the air!  Cool

Logged
N2DTS
Contributing
Member
*
Offline Offline

Posts: 2307


« Reply #33 on: February 05, 2015, 03:48:28 PM »

For screen modulation, you need to over load it a lot, way past the point of maximum output.
For good audio, its an adjustment of the resting carrier power (screen voltage) and loading.
One way to do it is to put full screen voltage on the tube, tune up at full power output, then turn the screen voltage down to 1/2 the power output or less, then modulate.
On my rigs, I just increase loading till I get between 100 and 150% positive modulation.
A peak reading watt meter is your friend.
Power output with screen modulation will be very low, between 25% and 50% of full power ratings.

You will also need enough voltage swing to modulate fully.
I never had luck with transformer type screen modulation, but I suspect the transformer has to be a good match...
Logged
WB3JOK
Member

Offline Offline

Posts: 637



« Reply #34 on: February 05, 2015, 04:09:36 PM »

Thanks, that confirms what I have been learning by experiment  Wink
A tune/operate switch to apply full screen voltage might be a handy addition in future.

There's definitely enough voltage swing, since I can easily see overmodulation in the negative direction with the audio gain turned up. Not sure if a "real" modulation transformer would be any help, except maybe to improve the highs, compared to the 1:1  (115v 60 Hz isolation) transformer I used that was in the junkbox  Grin

I may need to add a few turns to the output link to get overcoupling since I can't go beyond 100% on the variometer. But that requires taking apart the variometer assembly... So maybe I'll just leave it alone.  Cool
Logged
VE3LYX
Member

Offline Offline

Posts: 769


Crystals are from the stone age


WWW
« Reply #35 on: February 05, 2015, 06:20:18 PM »

I had mine on air today twice. I have been wondering about eliminating (bypassing perhaps) the roller and feeding the antenna directly off the variometer coil alone. 
don
Logged

Don VE3LYX<br />Eng, DE & petite Francais
N2DTS
Contributing
Member
*
Offline Offline

Posts: 2307


« Reply #36 on: February 05, 2015, 07:18:11 PM »

Yes, positive modulation is all about loading.
I have a bypass switch but never use it except when I plate modulate the rig.
I just tune everything for maximum power output then increase the loading till power falls off lot and I get lots of positive peaks.
I had worked out the static screen voltage in the past, and to some degree it works as a power output adjustment.

Its not unusual to run 250 watts carrier and have 1200 to 1300 watts pep.
4X would be 1000 watts and 200 watts over that is not unusual at 80% negative modulation.
If I turn the screen voltage up (that feeds the modulator) I get mostly more carrier power past about 300 watts.

So tubes (they are all different) have an operating point they like to get good modulation.
Most tubes seem to like a lot of plate voltage to screen modulate well at a good power level.
Logged
WB3JOK
Member

Offline Offline

Posts: 637



« Reply #37 on: February 05, 2015, 07:32:58 PM »

Don, from what I've been able to figure out from many ARC-5 articles (especially those by ARC-5 guru Dave Stinson), the transmitter is designed to match a very short wire which has considerable capacitive reactance and a Z of about 5-12 ohms. This is a long way from the holy amateur grail of 50 + j0.

The simplest solution to minimize harmonics and other undesirable output is to put a 75 pf (for 80m) or 50 pf (for 40m) transmitting cap in series with the output post and your feedline. That gives the highest power output at around 60% of roller inductor turns. It seems that a 1:4 unun is more trouble for only very slightly cleaner results and possibly a few watts less output.

It occurred to me that I didn't have to take the variometer apart since I'd never run it at 0 coupling anyway... so I added two turns around the middle of the tank coil and hooked it up in series (trial and error to find out which direction to wind in). That proved to be too much. So, one extra turn works just right. Now I get the cleanest waveform at about 5/10 on the variometer.

Meanwhile, I've been tweaking the modulator circuit to reduce hum pickup, increase low frequency response, etc.  I must have picked some of my parts values based on what was handy from the junkbox at the time.  Wink I drove the mic input from a good GR audio oscillator and to my surprise discovered (although my speech amp is flat as far as the 6AQ5 driver's grid) the output at the mod transformer (6AQ5 plate) has a peak response at 2 KHz (falling off sharply above that, which is OK) but steadily decreasing as I lowered the frequency and below 400 Hz was way off and starting to show distortion. Naturally the RF modulated waveform was the same. I'm not going to put this thing on the air yet!

I was surprised because my improvised modulation transformer is a small 60 Hz power transformer, so if anything I would have expected poor response above 500 Hz or so, not the exact opposite... I may need to try some swamping resistors for impedance matching, or a different transformer, or maybe couple the 6AQ5 to the screens another way. Choke? RC?   Huh

Logged
N2DTS
Contributing
Member
*
Offline Offline

Posts: 2307


« Reply #38 on: February 05, 2015, 07:51:08 PM »

You could try the DX60/QIX modulator, no transformer, hifi, and it works like a champ for only 2 tubes from D104 to screen.
Logged
WB3JOK
Member

Offline Offline

Posts: 637



« Reply #39 on: February 05, 2015, 08:37:54 PM »

Thanks for the tip. If I can't get acceptable results with the setup I have, I might just rip out the 6AQ5 and transformer, and build that circuit in the same space!



* dx60_screen_modulator.jpg (82.07 KB, 800x390 - viewed 836 times.)
Logged
N2DTS
Contributing
Member
*
Offline Offline

Posts: 2307


« Reply #40 on: February 05, 2015, 10:20:27 PM »

That circuit really does work fantastic as long as you do not try to push it too hard.
With a line input, its broadcast quality, and it souns good with the 1st stage set up for a D104.
I built it as a standalone modulator to use on anything up to a pair of 4X150's at 100 watts out.
Logged
WB3JOK
Member

Offline Offline

Posts: 637



« Reply #41 on: February 05, 2015, 10:59:24 PM »

I've been in the shack experimenting some more... thought the problem was with the "modulation" xfmr, but when driving it directly with the GR oscillator, although I could not achieve 100% mod, the response was perfectly flat from below 100 to at least 3 KHz. I also experimented with a resistance load for the 6AQ5 and coupling it to the transformer with a large capacitor (thinking the DC component in the primary, about 25 mA, was causing trouble). No change. So it's not the transformer.

I got out a '66 Handbook and started reading the "Audio Amplifiers and Double Sideband Phone" chapter and the screen modulation section. My lights came on - "strongly recommended" negative feedback in the modulator is the answer! A simple .047 capacitor from the 6AQ5 plate back to the cathode resistor of the preceding 12AX7 stage flattened the response nicely, and nearly eliminated the <400 Hz distortion. But then I couldn't get 100% mod due to the reduced gain.

Anyway, now it's just a matter of cut and try until I find the right combination of two R's and one or two C's for the feedback path to give a reasonably flat speech frequency curve with enough gain to fully modulate the screens!  Roll Eyes
Logged
VE3LYX
Member

Offline Offline

Posts: 769


Crystals are from the stone age


WWW
« Reply #42 on: February 06, 2015, 07:27:28 AM »

I dont use the holy grail 50 ohms. I use it on a long wire and it works quite well I think. I did build a tuner and have used that to match the unit to my home brew 811A pair linear. I get sufficient power out to use it which is all I want from it. I myself don't desire super audio. I want it to sound like it did in 1943. I cathode modulate mine and have thre e units i use. Just plug in the key hole and hook up filament supply. Works well. Been watching a parallel discussion on the method here and don't understand all the fuss. It s very easy to do and works well. A morning with a soldering iron will get it done. No schematic should be required and there is nothing to discuss. It works period. Your screen mod I found interesting too. (I have build two rigs from scratch with screen. ) Two of my cathode modulators are built to use carbon mics and one is for a crystal mic. I have been using it lately because some folks don't like that bandits a 12 oclock audio. I do! However trying to fit in but sometimes it is hard to put a square head in a round hole. As long as my radio work reasonably well, aren't hurting themselves and aren't interfering with others I am content. I have enjoyed very much reading your thread here and admire you project whatever the outcome. Building is a lot more fun and instructional then BS and postulating. I see you also understand that. Anytime you want to try a contact give me a time a freq. It might be impossible for us but who knows. On a low noise day with good prop it might just happen. So far I have gotten out about 200 miles in normal average conditions on 80M . Anyway keep at it and keep posting. Tis most interesting.
don
Logged

Don VE3LYX<br />Eng, DE & petite Francais
N2DTS
Contributing
Member
*
Offline Offline

Posts: 2307


« Reply #43 on: February 06, 2015, 08:08:21 AM »

Feedback really helps in screen modulators, I added it to the DX60/QIX modulator.
There is no phase shift in that circuit so you can use a lot.
Maybe just change some values in the 1st stage to get more gain, higher plate resistor...
Logged
WB3JOK
Member

Offline Offline

Posts: 637



« Reply #44 on: February 06, 2015, 10:29:52 AM »

Going to play with it some more this morning, time permitting...

Here is the current circuit. (The odd-looking tube symbols are from a user-created library "radiotubes" for Eagle which I downloaded from Cadsoft's website). I used large interstage coupling caps to avoid too much low-frequency rolloff. Would you recommend different values of plate resistors for the 12AX7?

* ARC-5 screen modulator.pdf (20.07 KB - downloaded 397 times.)
Logged
N2DTS
Contributing
Member
*
Offline Offline

Posts: 2307


« Reply #45 on: February 06, 2015, 10:57:25 AM »

Well the higher the resistor (220K now) the more gain.
330k, or 470K. Try a 330 K.

The higher the resistance, the more voltage swing.


Logged
WB3JOK
Member

Offline Offline

Posts: 637



« Reply #46 on: February 06, 2015, 01:45:29 PM »

I tried another 220K in series with it... made no difference! I think the gain of the 12AX7 itself is the limiting factor.

Anyhow I determined with a substitution box that the optimal value of the feedback resistor is near 100K.
What I find strange is that the voltage at the grid of the 6AQ5 gets smaller with increasing frequency when the feedback loop is closed, but the plate swing is more or less constant. There is no high-frequency rolloff within the circuit except perhaps the inductance of the transformer primary.

The modulator output (and the RF envelope) shows some distortion above 75% mod when supplied with a sine wave of 250 Hz or less. Clean at smaller amplitudes all the way down to 100 Hz. Looks fine all the way to 100% from 350 Hz on up to over 2 KHz. I think maybe this is limitations of the transformer.

Speech with a hand-held "communications" mic, which won't have much output below 250 Hz anyway, looks good on the scope. I got it done just in time for the AM "Noon Forum" on 7290 to break up! Tried to make contact in case anyone was still on freq but with only 13 watts and at noon I doubt anyone could have heard me.

Now to get back to work on the 4x4-125A class AB1 amplifier I also built back in 2004 (in my profile pic) to go with this PWer. 250W of carrier should be a bit more audible (13 dB increase).
Logged
N2DTS
Contributing
Member
*
Offline Offline

Posts: 2307


« Reply #47 on: February 06, 2015, 03:34:58 PM »

I used the first stage of this circuit:

http://amfone.net/Amforum/index.php?topic=32889.0

Seems to give loads of gain out of a D104.
I did not use the 10 meg input resistor but went with 3 meg to cut the lows somewhat.
What is the voltage on the tubes?

If you think about the grid changing the current flow through the tube, a higher plate load resistor
should give more swing on the output as the voltage drop through a higher resistor should be more if the plate voltage is high enough...
Logged
WB3JOK
Member

Offline Offline

Posts: 637



« Reply #48 on: February 06, 2015, 05:26:08 PM »

That's pretty much the standard speech amp circuit... looks about like mine, too!   Wink

Regarding the increased voltage swing, if increasing the plate resistor from 220K to 470K, won't the 1M grid resistor of the next stage load it down proportionally more, so there's less swing across the next grid? Anyhow I have plenty of audio gain, driven from an old Shure CB10E hand-held mic.

I am surprised at how critical the adjustment of loading (both the variometer and the roller inductor) is. If tuned for maximum power output, the modulation is severely distorted with barely any upward peaks. With just a little more loading, which only drops the carrier slightly, the scope shows a normal pattern.

Is this true in general of screen-modulated transmitters? I wouldn't dare put this one on the air without a monitor scope!

This afternoon I tried again to join a couple of ongoing QSOs barefoot, but no luck. Maybe tomorrow morning...
Logged
N2DTS
Contributing
Member
*
Offline Offline

Posts: 2307


« Reply #49 on: February 06, 2015, 06:04:30 PM »

Yes, that is normal, and even more loading will give a lot of positive peaks.
Scope is nice, mod monitor is nice, pep watt meter is nice, all three allows one to fine tune things very nicely.
You could use any ONE method.

I love the way screen modulation works, it can be VERY HI FI (if no transformers are used) along with nice positive peaks.
Since I liked the way it worked, I just built up something that would run it at 300 watts carrier output.
14 watts should do ok on 40 meters with a good antenna...
I have made contacts with 5 watts.
Logged
Pages: 1 [2] 3 ... 5   Go Up
  Print  
 
Jump to:  

AMfone - Dedicated to Amplitude Modulation on the Amateur Radio Bands
 AMfone © 2001-2015
Powered by SMF 1.1.21 | SMF © 2015, Simple Machines
Page created in 0.063 seconds with 19 queries.