Forums :: NEW! Web Resources :: Features :: Photo Gallery :: Vintage Radio Shows :: Archives
Support This Site: Contributors :: Advertise


It is currently Oct Sat 23, 2021 11:12 pm


All times are UTC [ DST ]





Post New Topic Post Reply  [ 85 posts ]  Go to page Previous  1, 2, 3, 4, 5  Next
Author Message
 Post subject: Re: Looking for two beta testers for AM Transmitter
PostPosted: Sep Sat 18, 2021 1:15 am 
Member

Joined: Mar Thu 01, 2018 1:30 am
Posts: 733
Ever tried PCB WAY?
Cheap! I got PCBs for $3 each from them. Chinese company doing great


Top
 Profile  
 
 Post subject: Re: Looking for two beta testers for AM Transmitter
PostPosted: Sep Sat 18, 2021 1:27 am 
Member

Joined: Mar Thu 01, 2018 1:30 am
Posts: 733
Oh, I remember your reply to a post of mine where you mentioned doing something similar to this attached circuit.

1250KHz frequency. Tuned circuit optimised for this.
Bd139 could be 2n3553 or TIP31.

An emitter resistor in series with modulator transistor should limit 100mW DC in.

I got 300 feet range with 6f wire indoors. Sound quality was FM like and I was kinda surprised.


Top
 Profile  
 
 Post subject: Re: Looking for two beta testers for AM Transmitter
PostPosted: Sep Sat 18, 2021 1:31 am 
Member

Joined: Mar Thu 01, 2018 1:30 am
Posts: 733
Schematic. It was an amazing TX.
RFC1 - 1.5mH axial, RFC2 100uH axial.

T2 buffers audio from LM386N-1 and eliminates distortion. It sounded like an FM transmitter Haha I don't know why.


Attachments:
IMG_20210918_055925.jpg
IMG_20210918_055925.jpg [ 129.82 KiB | Viewed 532 times ]
Top
 Profile  
 
 Post subject: Re: Looking for two beta testers for AM Transmitter
PostPosted: Sep Sat 18, 2021 5:35 am 
Member

Joined: Mar Thu 01, 2018 1:30 am
Posts: 733
Where's yours?


Top
 Profile  
 
 Post subject: Re: Looking for two beta testers for AM Transmitter
PostPosted: Sep Sat 18, 2021 10:41 am 
Member
User avatar

Joined: Jun Sun 22, 2008 2:03 am
Posts: 1059
Location: Burke, VA 22015
Dare4444 wrote:
Ever tried PCB WAY?
Cheap! I got PCBs for $3 each from them. Chinese company doing great

I've been using them for years. The thing I like the most about them is that they don't charge non-recurring engineering cost which used to be a headache for me. They are set up to handle thousands of orders everyday, small or big. The standard turn-around time is 2 to 4 days. One can follow the manufacturing progress on-line. The pricing per board depends on only its dimensions - not on the numbers of holes - and quantity. Usually it doesn't make sense to order just a few boards - ten boards cost exactly the same. For small quantity orders, I like to pick a number to amortize the shipping cost to about $1 per board. I was wondering how they can do so cheaply. My guess was that they must automate a lot of the steps. But after watching these videos of all the manufacturing steps, one can see that there are still a lot of manual labor involved.
https://www.pcbway.com/pcb-service.html?step=1#miao1

_________________
Binh


Top
 Profile  
 
 Post subject: Re: Looking for two beta testers for AM Transmitter
PostPosted: Sep Sat 18, 2021 11:32 am 
Member

Joined: Mar Thu 01, 2018 1:30 am
Posts: 733
It's China. Huge younger population. Even $10 salary a day for workers is okay for them. Machines are government subsidized.
The USA made Chinese government very rich and now they are undermining the USA. Large number of young people and cheap labor due to massive competition. Even india can't compete with China.


Top
 Profile  
 
 Post subject: Re: Looking for two beta testers for AM Transmitter
PostPosted: Sep Sat 18, 2021 12:15 pm 
Member

Joined: Mar Thu 01, 2018 1:30 am
Posts: 733
Built next am transmitter with a cd4053 and complimentary amplifier. Cd4053 digitally driven by 12V square waves.


Top
 Profile  
 
 Post subject: Re: Looking for two beta testers for AM Transmitter
PostPosted: Sep Thu 23, 2021 12:46 pm 
Member
User avatar

Joined: Aug Wed 26, 2009 1:50 am
Posts: 1580
Location: Monroe GA 30656 (from Coventry, UK)
Tom,
I am afraid I am not in a position to be one of your testers right now, sorry. I just moved from Nevada to Georgia and am in total chaos. It will be a little while before I will be set up to build stuff, but will definitely be interested in putting one of these together.

Colin


Top
 Profile  
 
 Post subject: Re: Looking for two beta testers for AM Transmitter
PostPosted: Sep Thu 23, 2021 4:18 pm 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 4959
Location: Texas
Colin Ames wrote:
Tom,
I am afraid I am not in a position to be one of your testers right now, sorry. I just moved from Nevada to Georgia and am in total chaos. It will be a little while before I will be set up to build stuff, but will definitely be interested in putting one of these together.

Colin


Good to hear you weigh in, Colin. I think the three of us that are testing these will discover what flaws need to be corrected for a run of the boards. Then I plan to make a few boards and maybe parts available at cost for anyone here who wants to build one.


Top
 Profile  
 
 Post subject: Re: Looking for two beta testers for AM Transmitter
PostPosted: Sep Thu 30, 2021 2:34 am 
Member
User avatar

Joined: Jun Sun 22, 2008 2:03 am
Posts: 1059
Location: Burke, VA 22015
Hi Tom,

I got the boards I ordered on Sunday. From testing one of them, there are a couple of observations I'd like to make.

The synthesizer works like a champ. It provides a clean clock across the whole MW band and doesn't seem to exhibit any crosstalk problems. A mean to series terminate and lowpass filter the reference and divided clock is provided but it's not needed. Your tuning aid circuit works very well too.

The very high gain MPSA18 transistor swiches softer the MOSFETs and works well but there is some sort of secondary oscillation at the antenna output just before the output circuit is completely tuned. This oscillator is shown as fast amplitude jitter on an analog oscilloscope. The 2N7000 and ZVN2110A MOSFETs do not have this problem at all. Since these two allow the tuning aid voltage to vary monotonically while the antenna is tuned, they are preferred over the BJT if this detector voltage is used for automatic tuning.

Contrary to what you have observed, the LM386 doesn't heat up at all when its output is bypassed by a 0.1uF capacitor through a 10 ohm resistor. The average DC current to the modulator is about the same with or without the bypass capacitor.

With no audio input, the LM386 output DC voltage is about 6.5V (about half the supply voltage). This DC operating point is not typical but varies from samples to samples. It's too high for the modulator to be used as a legal transmitter. It needs to be around 3V. One way to do this without reducing the supply voltage is to insert a 100 ohm in series with the choke. Another way is to increase this resistor to 220 ohm (for my sample) and parellel it with a 120uF (or larger) capacitor. The latter allows more than 100% modulation well before the audio amplifier output is clipped.

The 2mH choke is way too large. Since the switching modulator behaves for the audio envelope like a second order lowpass filter where the choke inductance contributes to one pole, the choke has to be reduced to less than 150uH for an acceptable audio bandwith. Even with this lower choke, the audio still needs to be pre-emphasized to compensate for the high frequency the roll off and maintain a uniform modulating index.

No high quality choke is needed. A cheap molded inductor works fine since we need to drop the modulator's DC operating voltage with a series resistor anyway. A 150uH inductor typically has a SRF of several MHz, well above the MW band.

When a loopstick is used for the antenna matching coil, there is no need for an antenna or ground for short range. I got clear reception with a battery operated radio within a room (15 feet) with practically no hum.

My offer still stands. A second opinion for my board is always appreciated.

_________________
Binh


Top
 Profile  
 
 Post subject: Re: Looking for two beta testers for AM Transmitter
PostPosted: Sep Thu 30, 2021 4:25 pm 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 4959
Location: Texas
Thanks, Binh. I'm glad your boards are working.

I, too, noticed that secondary oscillation with the BJT as the antenna was approaching resonance. Although it doesn't affect the sound or range or anything else, I'm sure it produces harmonics that some could consider objectionable. And as you observed, the mosfet does not exhibit this behavior.

I looked at the scope trace with a sine wave to estimate the frequency of that oscillation; it's about one-twentieth the frequency of the output and is a little variable. I was wondering if it might have been related to multiples of 2 as in the counter somehow, but it doesn't seem to be related to that. Beyond that I haven't given it much thought as a 2N7000 works fine and still keeps the parts count as low as possible.

Regarding the 0.1 capacitor bypass from the output of the LM386, when it heated up my LM386 it was being bypassed directly to ground, essentially shorting out the some frequencies of the audio. In your posting above you mentioned bypassing it through the 10 Ohm resistor. Would you please clarify that?

Regarding the tuning indicator, there's a trade-off between total brightness, which makes it easier to see during the day, and apparent sensitivity, which is achieved with less brightness using a higher resistor value to the LED. In other words I have found that small changes in tuning are more visible when the LED is not as natively bright. So a possible upgrade would be to put a 10K pot in series with a small protection resistor and the LED. I'm curious if you think that provides enough of a benefit to make it worth trying to squeeze one more trimmer pot onto the board. Or does it work satisfactorily enough to leave that part of the circuit as it is?

Would you additionally confirm you tested yours with the 150 microhenry inductor? I'd be very happy if we could get away with one that small. I used 2 mH because that's what I gleaned from a lot of circuits like this that are on the web. Even with that heavier filtering, the audio already sounds much like FM. But smaller inductors are more compact and cheaper.

This morning I received an update from my order from Futurlec. They just shipped the parts early this week via DHL. I'm disappointed because I ordered them on September 12th. They had said I would have them in 2 weeks but it looks like that's not going to happen. Once I get everything in I'll ship them out immediately to the two beta testers. They'll have only one set of parts, but I will make the parts list available for anybody that wants to take you up on your offer of boards.


Top
 Profile  
 
 Post subject: Re: Looking for two beta testers for AM Transmitter
PostPosted: Sep Thu 30, 2021 7:19 pm 
Member
User avatar

Joined: Jul Fri 18, 2008 10:02 am
Posts: 1975
Location: near ST Louis Mo 62002
The .1 uf and the 10 ohm resistor are in series to form a Zobel filter .

_________________
Mark
(Conti the brain damaged robot )


Top
 Profile  
 
 Post subject: Re: Looking for two beta testers for AM Transmitter
PostPosted: Sep Thu 30, 2021 8:18 pm 
Member
User avatar

Joined: Jun Sun 22, 2008 2:03 am
Posts: 1059
Location: Burke, VA 22015
Hi Tom,

The secondary oscillation with the BJT only occurs as the antenna is near resonance but disappears when the antenna is at peak resonance. The BJT is perfectly fine but one has to carefully tune the antenna to maximum resonance.

The LM386 output on my board is bypassed with a 0.1 uF capacitor to ground through a series 10 Ohm resistor. This is the same approach given in the data sheets for a different purpose. It didn't cause my LM386 made by TI to heat up at all. Maybe the reason is that the switching transistor collector is shunted to ground by a 560 pF capacitor through a 0.1 uF decoupling capacitor. After discovering that the modulator input power is way too high (~ 200mW), I inserted in series another 220 Ohm resistor paralleled by a 120 uF capacitor to downshift the DC operating point to about 3V. A side great benefit of this parallel RC is that over-modulation becomes possible even though the LM386 cannot swing rail-to-rail. See my up-to-date modulator circuit below.
Attachment:
MWAM2_Modulator.png
MWAM2_Modulator.png [ 14.56 KiB | Viewed 320 times ]


I tested my board with a 1.4mH choke with a sinewave for the audio input. I monitored the envelope with an oscilloscope as the sinewave frequency was varied. At 440Hz, the envelope looked ok. Above 1kHz, not only the waveform quickly got smaller in amplitude but also looked disappointingly distorted. I then changed the choke to 470uH, 220uH and finally settled to 150uH. There was definitely less distortion at high frequencies for smaller chokes. The envelope still got smaller as the frequency got higher. As I already mentioned, pre-emphasis is still necessary to maintain a uniform modulation index across the audio bandwidth to improve the SNR at high frequencies. You have to take a grain of salt from what you saw on the web as far as the high choke value. When the modulator is perfectly tuned, it behaves on a first order like a series LR circuit where L is the choke and R is the equivalent DC load. If the short antenna is shunted by a large capacitance for tuning, R is the DC equivalent of the total series resistance in the circuit which could be small. Let's say R is 10 Ohm and L is 2mH. The time constant L/R is 0.2 ms which approximately corresponds to 800 Hz lowpass cutoff. In reality this cutoff frequency is lower because the tuned modulator behaves more like a second-order lowpass filter for the envelope (from an analysis done in a PhD thesis). So decreasing the choke inductance by a factor of 10 definitely improves the high-frequency audio response. Another way to widen the audio bandwidth is to tune the antenna with just the matching coil and as small as possible shunting capacitance. The ground loss resistance comes into play (since the antenna is no longer shunted by a large capacitance) and helps to lower the time constant of the modulator.

Regarding the tuning indicator, the only modification is to add an anti-parallel diode at the base for protection against large reverse voltage. You already have a 4.7K shunt resistor in the emitter circuit to divert the current from the LED. Use a lower value if the LED is too bright. Here is my implementation of your circuit. The gimmick antenna coupling capacitor is implemented as a 0.2"x0.2" square pad 1/16" over the antenna pad with FR4 dielectric material. The values in the circuit were obtained from a simulation which gives uniform response at frequencies above 1 MHz for a maximum antenna peak voltage of 150V.
Attachment:
Visual_Ant_Tuning_Aid_1.png
Visual_Ant_Tuning_Aid_1.png [ 8.34 KiB | Viewed 320 times ]

Here is a picture of the built board. I installed pin sockets for some components to quickly change their values.
Attachment:
20210929_153719-1.jpg
20210929_153719-1.jpg [ 537.69 KiB | Viewed 275 times ]

_________________
Binh


Top
 Profile  
 
 Post subject: Re: Looking for two beta testers for AM Transmitter
PostPosted: Oct Sat 02, 2021 2:21 pm 
Member
User avatar

Joined: Jun Sun 22, 2008 2:03 am
Posts: 1059
Location: Burke, VA 22015
Adding treble boost to the LM386 turns out to be quite simple. It only requires bypassing its pin 8 to ground with a 1 uF capacitor. This gives a moderate 6 dB boost from 100 Hz to 5 kHz. There are a couple changes of values that need to be made but a board layout modification is not necessary. I haven't tested this yet but a simulation shows a not bad emphasis curve in the frequency response.


Attachments:
Amp_with_Treble_Boost_1.png
Amp_with_Treble_Boost_1.png [ 15.34 KiB | Viewed 248 times ]
Amp_with_Treble_Boost_Resp.png
Amp_with_Treble_Boost_Resp.png [ 18.39 KiB | Viewed 248 times ]

_________________
Binh
Top
 Profile  
 
 Post subject: Re: Looking for two beta testers for AM Transmitter
PostPosted: Oct Sat 02, 2021 3:38 pm 
Member
User avatar

Joined: Jan Fri 02, 2009 11:32 pm
Posts: 745
Location: Equinunk PA 18417
https://antiqueradios.com/forums/viewtopic.php?f=12&t=334000

In 2018 I proposed a similar idea for cheap and easy pre-emphasis. I used 374 ohm and 33nF (R1/C2 in my schematic), values arrived at by trial and error. The resulting pre-emphasis curve is satisfactory. I noted some LM386 audio instability which I attempted to tamp down via another R/C negative feedback path (R5/C1 on my schematic). Stability was improved but I don't recall that it was ever perfect under all conditions, and I'd abandoned LM386 by then anyway. Looking back now, if I had included a 10 ohm/.05uF series R/C from pin 5 to ground, as shown in TI's datasheet, the instability might never have occurred.

_________________
-Richard


Top
 Profile  
 
 Post subject: Re: Looking for two beta testers for AM Transmitter
PostPosted: Oct Sat 02, 2021 4:58 pm 
Member
User avatar

Joined: Jun Sun 22, 2008 2:03 am
Posts: 1059
Location: Burke, VA 22015
Hi Richard,

Glad to see you drop into this discussion.

I tried your treble boost scheme in the simulator. I was not happy with it because it didn't give me the desired nominal gain. In the end I decided to parallel the internal 1.35 kOhm resistor with a resistor in series with a large enough capacitor to give me the midband nominal gain (around 40 dB). Bypassing the lower leg 150 Ohm resistor with a capacitor boosts the high frequencies. By changing this capacitor value, I can move the starting point of the pre-emphasis curve at will to compensate the modulator's time constant.

I haven't experienced instability problem with the LM386 - maybe I have followed the TI's recommendation. In general, it's a good practice to decouple a large capacitive load from an amplifier's output with a small resistor to avoid any potential instability. With the simple mean to halve the amplifier's DC operating point while allowing AC to pass, it's unlikely that the amplifier overloads at high modulation index (its output is several volts above ground at 100% modulation).

I had resisted building this kind of modulator until now, thanks to Tom. It's more suitable for fixed frequency or narrow band operation (such as cell phones where efficiency is important). For operation in the MW band which spans almost two octaves, you're not going to get the components' values perfectly right. As a result there will be detuning which causes sideband imbalance resulting in envelope distortion. So unless the components are chosen perfectly for a carrier frequency, this modulator will never be high fidelity (expect a few percent of distortion). There are four variables involved, two of which are fixed (matching coil and antenna) and one is unknown (ground loss resistance unless some measurement is done). The only variable left is the choke inductance. I have the formula for choosing the choke with finite choke and infinite Q series LC. Since we don't have infinite Q, I'm trying to find a choke value which keeps detuning as small as possible across the whole MW band. If one observe the collector/drain waveform on a scope, it will look like a perfect half-wave rectified sinewave. This happens when the transistor turns on with zero drain-source voltage and the loss in the transistor is minimal.

_________________
Binh


Last edited by bb.odin on Oct Sat 02, 2021 5:36 pm, edited 1 time in total.

Top
 Profile  
 
 Post subject: Re: Looking for two beta testers for AM Transmitter
PostPosted: Oct Sat 02, 2021 5:12 pm 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 4959
Location: Texas
UPDATE FOR BETA TESTERS:

I'm STILL waiting on parts ordered Sept. 12. Here's the latest from Futurlec. Now I know why I like Mouser and DigiKey so much, though more expensive...


Attachments:
Image4.jpg
Image4.jpg [ 121.64 KiB | Viewed 228 times ]
Top
 Profile  
 
 Post subject: Re: Looking for two beta testers for AM Transmitter
PostPosted: Oct Sat 02, 2021 5:53 pm 
Member
User avatar

Joined: Jun Sun 22, 2008 2:03 am
Posts: 1059
Location: Burke, VA 22015
Macrohenry wrote:
Now I know why I like Mouser and DigiKey so much, though more expensive...


If you factor in the high DHL shipping cost and the long shipping time, it probably costs about the same. For small order less than 14oz, DigiKey will ship it first class which is not very expensive. I'll stick with local distributors.

_________________
Binh


Top
 Profile  
 
 Post subject: Re: Looking for two beta testers for AM Transmitter
PostPosted: Oct Sun 03, 2021 1:12 am 
Member

Joined: Mar Thu 01, 2018 1:30 am
Posts: 733
Macrohenry wrote:
UPDATE FOR BETA TESTERS:

I'm STILL waiting on parts ordered Sept. 12. Here's the latest from Futurlec. Now I know why I like Mouser and DigiKey so much, though more expensive...


Damn. They used to be 2-3 weeks, makes sense when you are ordering in bulk. Cost savings add up.


Top
 Profile  
 
 Post subject: Re: Looking for two beta testers for AM Transmitter
PostPosted: Oct Sun 03, 2021 2:44 pm 
Member
User avatar

Joined: Jan Fri 02, 2009 11:32 pm
Posts: 745
Location: Equinunk PA 18417
Kudos to Macrohenry for this latest project. I wanted to say HI.

Binh, I like your makeshift Spice model for LM386, and keeping more AC voltage gain while boosting audio upper frequencies makes sense. It will be a simple and useful modification (if desired) to Tom's version.

A percent or two of transmitter distortion is, in my opinion, perfectly acceptable in this case. The value of inductor placed between the LM386 and "psudo class E" modulator (L1 on Binh's schematic above) seems to me therefore not to be very critical. This is my conclusion based on observation not calculation. 2h is very much too large to maintain respectable audio response, as was pointed out, but I've not seen or measured a practical difference in transmitter performance when the inductor is varied from about 100uH to 600uH. The inductor's dc resistance makes some difference, of course, but the L value or Q doesn't really seem to matter much. At least I didn't see much. I went so far as to wind a variable inductor (about50uH to a few hundred uH), using a ferrite stick, and while adjusting could find no worthwhile performance difference while transmitting at a fixed broadcast frequency. I don't have access to rf spectrum or other rf analysis equipment, but I can measure carrier frequency, relative signal strength, and detected audio characteristics. And I have good ears. I didn't see or hear any worthwhile differences of the modulated carrier as this inductor's value was changed, all else being the same.

This little transmitter should perform like a champ!

_________________
-Richard


Top
 Profile  
 
Post New Topic Post Reply  [ 85 posts ]  Go to page Previous  1, 2, 3, 4, 5  Next

All times are UTC [ DST ]


Who is online

Users browsing this forum: No registered users and 16 guests



Search for:
Jump to:  


































Privacy Policy :: Powered by phpBB