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


It is currently Oct Wed 20, 2021 9:28 pm


All times are UTC [ DST ]





Post New Topic Post Reply  [ 84 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: Oct Fri 08, 2021 12:38 pm 
Member
User avatar

Joined: Jun Sun 22, 2008 2:03 am
Posts: 1059
Location: Burke, VA 22015
Richard, thank you for your measurements to verify my simulation. I'm not surprised at all of the discrepancy. The simulation is only as good as the accuracy of the models. I'll go with your final values. Using 47uF for C3 pushes the -3dB low cut to 20Hz. I use 27uF for C3 because I have lots of them. The basic circuit approach can also be easily adjusted for flat audible response and sufficient HF cuts as shown below.


Attachments:
LM386_flat_audible_resp.png
LM386_flat_audible_resp.png [ 32.5 KiB | Viewed 456 times ]

_________________
Binh
Top
 Profile  
 
 Post subject: Re: Looking for two beta testers for AM Transmitter
PostPosted: Oct Fri 08, 2021 2:12 pm 
Member
User avatar

Joined: Jan Fri 02, 2009 11:32 pm
Posts: 745
Location: Equinunk PA 18417
bb.odin wrote:
Using 47uF for C3 pushes the -3dB low cut to 20Hz. I use 27uF for C3 because I have lots of them.
You're right, and upon reflection I've changed my mind. 27uF is a good value that gives a bit of useful sub-audible filtering.

_________________
-Richard


Top
 Profile  
 
 Post subject: Re: Looking for two beta testers for AM Transmitter
PostPosted: Oct Fri 08, 2021 7:17 pm 
Member
User avatar

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

With all these modifications, I think this will be a pretty good modulator. To make it for more user friendly, I'd like to add a mean to automatically tune the antenna, an easier way than the DIP switch to set the carrier frequency, and some sort of audio compression, the same way Richard did. I'm aware that these added features will likely increase cost and that's not what Tom wants.

In an effort to keep the cost as low as possible, I'm going to digitally process the audio since a processor is needed for carrier frequency setting and automatic antenna tuning. In addition the functions of the reference oscillator/divider and programmable divider chips will be done by the processor. A low-pin-count processor capable of performing those functions costs less than $2. A cheap chinese-made linear stage for the ATU costs less than $5. A cheap LCD display and push-button switches will add another $5. Another op-amp for pre-emphasis and anti-aliasing filtering costs less than $1. Subtracting the cost of the DIP switch, two ICs and sockets, the added cost would be around $10 to $11. However the modulator will be better than the TH5 and I think it's worth the additional cost.

Here's a conceptual block diagram for me to go forward from this experience. Sorry for going off topic.


Attachments:
Conceptual_LowCost_AM_2021-10-08.png
Conceptual_LowCost_AM_2021-10-08.png [ 235.73 KiB | Viewed 400 times ]

_________________
Binh
Top
 Profile  
 
 Post subject: Re: Looking for two beta testers for AM Transmitter
PostPosted: Oct Sat 09, 2021 3:28 pm 
Member
User avatar

Joined: Jan Fri 02, 2009 11:32 pm
Posts: 745
Location: Equinunk PA 18417
This sounds like a separate, brand new project with its own thread! I'm excited to see where you take it and I admire your enthusiasm that it will only add $11 to the cost. :mrgreen:

At the risk of dragging us further away from Macrohenry's thread topic, I have two initial thoughts.

1. Audio signal to noise ratio using 12 bit ADC and/or 8 bit DAC seems limiting. Under ideal circumstances an 8 bit DAC leaves noise at only 48db below max audio peaks. That is very audible and might be even worse depending on the DAC. If you're thinking to use an ADC built into a typical ARM processor, noise and linearity concerns are even stronger. Also, in order to service an uninterrupted audio data stream the processor can only work on other tasks between audio samples. Depending on how much math is being done on the audio samples, this could become restricting. I am not familiar enough with the actual math of digital audio filters to do "back of the napkin" estimates. In the pro audio world, these things are often still done using FPGA's in order to guarantee audio processing and data flow remain uninterrupted. That approach would be overkill here.

2. I'd consider using a single programmable clock rather than a crystal osc + 4046 PLL + programmable divider. Could a PL611 replace them all? I know that you know of other clock generator options.
edit: PL611 part is a completely wrong-headed suggestion, what was I thinking? SI5351A is often used, still available and cheap.

I do like your idea to do some initial audio gain ahead of the ADC. That could be program-dependent rather than fixed gain amounts. I considered doing that in my transmitter last year, but eventually decided it was too ambitious for that project. You are doing that and taking it to the next step of digital processing.

_________________
-Richard


Last edited by richfair on Oct Sun 10, 2021 1:16 pm, edited 1 time in total.

Top
 Profile  
 
 Post subject: Re: Looking for two beta testers for AM Transmitter
PostPosted: Oct Sat 09, 2021 6:22 pm 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 4956
Location: Texas
Wow, that is a lot of additional capability for $11. Coincidentally there is a thread in the Clubhouse by a fellow that wanted to get an AM transmitter and he's getting one with much of that capability. However he's ending up with an FM transmitter but he's paying nearly $150 for it.

https://www.amazon.com/Retekess-Wireles ... B088KLF8DJ

My mind organizes into four tiers of AM transmitters.

Barebones hobbyist- must accept the limits of LC or fixed frequency crystal oscillators - works out of the box- no additional settings- can sound good enough for very many folks

Simple hobbyist with synthesized frequency- choose any desired frequency and preemphasis then leave it- as we've discovered this can sound really good

Advanced hobbyist-digital controls and displays with audio processing and auto antenna tuning - this level is attractive to those who more frequently change settings.

Professional-includes all of the advanced but built like a tank and probably more parts, more advanced modulator, etc.

The value proposition in any of these tiers will be up to the individual paying for it. It's nice that we are providing our members options.


Top
 Profile  
 
 Post subject: Re: Looking for two beta testers for AM Transmitter
PostPosted: Oct Tue 12, 2021 7:46 pm 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 4956
Location: Texas
PARTS UPDATE:

DHL online tracking shows no activity since October 5 when it was processed in Singapore. DHL customer service says the delivery window is Oct. 13-18. In essence they said not to contact them unless it doesn't arrive then.


Top
 Profile  
 
 Post subject: Re: Looking for two beta testers for AM Transmitter
PostPosted: Oct Tue 12, 2021 10:16 pm 
Member

Joined: Jan Thu 01, 1970 1:00 am
Posts: 2775
Location: Saskatoon
Unless you opt for one of the guaranteed delivery times, companies that ship from around the world generally wait until they have a full load, in order to keep costs down, thus the vague time frame. I order stuff from Asia all the time, and I've learned to be patient, and expect it when I see it. Always good to have several projects on the go, so that while you're awaiting parts for one project, you can work on another.


Top
 Profile  
 
 Post subject: Re: Looking for two beta testers for AM Transmitter
PostPosted: Oct Tue 12, 2021 10:32 pm 
Member

Joined: Mar Thu 01, 2018 1:30 am
Posts: 732
Wow! A lot of work. Congratulations to the designer for this excellent circuit. What's the modulation% ? I'll have to read this thread carefully. Meanwhile, I posted a modified Vectronics 1290k transmitter with 4 transistors on my thread about MW AM transmitter designs. Looks complicated though :( where the circuit on this thread is pretty straight forward. I'll check all the discussion here very carefully as audio quality is being discussed too.


Top
 Profile  
 
 Post subject: Re: Looking for two beta testers for AM Transmitter
PostPosted: Oct Tue 12, 2021 11:06 pm 
Member

Joined: Mar Thu 01, 2018 1:30 am
Posts: 732
bb.odin wrote:
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:
The attachment MWAM2_Modulator.png is no longer available


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:
The attachment MWAM2_Modulator.png is no longer available

Here is a picture of the built board. I installed pin sockets for some components to quickly change their values.
Attachment:
The attachment MWAM2_Modulator.png is no longer available


Whoa! I've never tried this, series resistor with parallel cap at the output of lm386. Were you able to modulate upto 100% then? I'm curious. I've never had any instability issue with lm386. Why? Check this attachments, usually the 100uH RFC followed by a 100n cap to gnd is the first thing connected to pin 5 of LM386. I also isolate the supply to pin 6 with RFC and caps plus I put a 47p - 330p cap directly across pin 2 and 3. 20+ years, not even once it oscillated, got hotter, etc. And I also keep the gain no more than 50 (1k + 10u between pins 1 and 8 ) . Gain of 200 is bad for stability too. Do not connect 100n directly from pin 5 to ground! The chip will start heating. The 100uH RFC should be the first component connected to its pin 5.


Attachments:
IMG_20211013_033307.jpg
IMG_20211013_033307.jpg [ 449.69 KiB | Viewed 304 times ]
IMG_20211013_033856.jpg
IMG_20211013_033856.jpg [ 424.52 KiB | Viewed 304 times ]
Top
 Profile  
 
 Post subject: Re: Looking for two beta testers for AM Transmitter
PostPosted: Oct Wed 13, 2021 4:21 pm 
Member
User avatar

Joined: Jun Sun 22, 2008 2:03 am
Posts: 1059
Location: Burke, VA 22015
Hi Dare444,

At the risk of hijacking Tom's thread, I have a few comments about your recent ones.

As I already stated before, over-modulation is possible with that parallel RC circuit although it's better to avoid it. Its real purpose is to drop the DC operating point for power limit compliance. The added bonus is that the low-end clipping is handled more gracefully. When the modulation goes over 100%, the drain goes negative and therefore the drain and source (or collector and emitter) reverse role.

Regarding the amplifier output decoupling, you just reiterated what had been discussed many times here and somewhere else. A choke is good but a lower-cost small resistor is also adequate. Power supply decoupling with a choke or ferrite bead is a good practice in mixed signal designs but somewhat overkill here. Split ground plane and power supply rail, in the spirit of low cost and low part count design goal, are equally effective.

A high gain amplifier is not necessarily unstable with sufficient compensation. Latest one with pre-emphasis, although with very high gain, is compensated to have enough phase margin and never goes to a phase reversal. Even if it does, its gain margin will be positive and large enough and therefore not a chance it will oscillate at RF.

In conclusion, what Tom has here is a solid design, the result of many hours experimenting and testing. Let's wait and hear what the beta testers have to say.

_________________
Binh


Top
 Profile  
 
 Post subject: Re: Looking for two beta testers for AM Transmitter
PostPosted: Oct Wed 13, 2021 11:21 pm 
Member

Joined: Mar Thu 01, 2018 1:30 am
Posts: 732
I need to study audio processing now. Following this thread closely as digital audio processing is being discussed.


Last edited by Dare4444 on Oct Thu 14, 2021 3:25 am, edited 1 time in total.

Top
 Profile  
 
 Post subject: Re: Looking for two beta testers for AM Transmitter
PostPosted: Oct Thu 14, 2021 1:33 am 
Member
User avatar

Joined: Jun Sun 22, 2008 2:03 am
Posts: 1059
Location: Burke, VA 22015
Dare444,

You already posted that circuit in your own thread. I don't feel it's appropriate to hijack this thread any further. For that reason, I shall refrain from making any more unrelated comments. Sorry.

_________________
Binh


Top
 Profile  
 
 Post subject: Re: Looking for two beta testers for AM Transmitter
PostPosted: Oct Thu 14, 2021 3:31 am 
Member

Joined: Mar Thu 01, 2018 1:30 am
Posts: 732
bb.odin wrote:
Dare444,

You already posted that circuit in your own thread. I don't feel it's appropriate to hijack this thread any further. For that reason, I shall refrain from making any more unrelated comments. Sorry.


I erased. See. I'm not familiar with digital audio processing or the modules that are being used. Need to learn more from you. No one is hijacking this thread. My previous post about LM386 RF isolation was related to this thread only. It's an overkill to use RFCs at pin 5 and 6 but Tom may make use of my suggestion if he encounters any instability again.


Top
 Profile  
 
 Post subject: Re: Looking for two beta testers for AM Transmitter
PostPosted: Oct Thu 14, 2021 3:48 am 
Member

Joined: Mar Thu 01, 2018 1:30 am
Posts: 732
I'm still looking for the schematic.


Top
 Profile  
 
 Post subject: Re: Looking for two beta testers for AM Transmitter
PostPosted: Oct Thu 14, 2021 4:42 pm 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 4956
Location: Texas
Dare4444 wrote:
I'm still looking for the schematic.

Thanks for your interest. On September 9th in this thread you asked about a schematic and that question was answered. The modulator has been a work in progress and the Sep 9 answer is still relevant.

In short, this device is in two parts, the synthesizer in the modulator.The synthesizer values have not changed and that schematic is still available on my web page referenced on Sep 9th. The modulator is simpler. Its values have been thoroughly discussed and the beta testers will be given options on those. After settling on needed parts a unified schematic will be available..


Top
 Profile  
 
 Post subject: Re: Looking for two beta testers for AM Transmitter
PostPosted: Oct Thu 14, 2021 6:20 pm 
Member

Joined: Mar Thu 01, 2018 1:30 am
Posts: 732
Synthesizer is there. Not the RF section. You wrote about it.. Isn't the mosfet getting hot if its connected direct to the synthesizer? I tried it with hef4060 and it got hot. A DC blocking capacitor was then needed. It wasn't class E though. Did you check modulation depth? 2n7000 modulates 100% ?? My big question.


Top
 Profile  
 
 Post subject: Re: Looking for two beta testers for AM Transmitter
PostPosted: Oct Thu 14, 2021 9:52 pm 
New Member

Joined: Apr Sun 28, 2019 12:17 am
Posts: 1
please put me down for ordering one of these.
I want one


Top
 Profile  
 
 Post subject: Re: Looking for two beta testers for AM Transmitter
PostPosted: Oct Sat 16, 2021 6:00 am 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 4956
Location: Texas
Dare4444 wrote:
Synthesizer is there. Not the RF section. You wrote about it.. Isn't the mosfet getting hot if its connected direct to the synthesizer? I tried it with hef4060 and it got hot. A DC blocking capacitor was then needed. It wasn't class E though. Did you check modulation depth? 2n7000 modulates 100% ?? My big question.


Here's the RF section. I think it's simple enough to answer your questions without needing a schematic. Note the 2K resistor is bypassed when using FET. The 10 nF cap is the DC blocking capacitor. To reiterate, a full schematic will be available after final parts have been selected. I scoped it to make sure 100% modulation was possible.

In this image the transistor leads are not as clearly labeled as I intended. The two right columns of letters depict the right column of terminals. The left column of letters goes with the left column of terminals. This was done so different FETS with differing lead configurations could be used more easily.


Attachments:
Image5.jpg
Image5.jpg [ 72.36 KiB | Viewed 149 times ]
Top
 Profile  
 
 Post subject: Re: Looking for two beta testers for AM Transmitter
PostPosted: Oct Sat 16, 2021 6:58 am 
Member

Joined: Mar Thu 01, 2018 1:30 am
Posts: 732
Whoaa! So 100% modulation is possible with a single mosfet like 2n7000? Years ago when I was experimenting with 2n3904 transistors, modulating a 30mW pushpull amplifier never got me more that 72% modulation. Modulator was a single lm386n-4. So was it due to the audio chip? Would I have got 90% modulation if I used parallel RC technique to make the voltage swing bigger? The 2n7000 is a godsend for part 15 homebrewers. Cgate is 50p, Or 2K Zin at 1MHz. Thanks for the info. I need to learn more from you. I was curious if a 2n7000 af amp could modulate the source of another 2n7000 for 100% mod? I liked the pushpull class C 2n3904 amp for its very high fidelity.


Top
 Profile  
 
 Post subject: Re: Looking for two beta testers for AM Transmitter
PostPosted: Oct Sat 16, 2021 7:42 am 
Member

Joined: Mar Thu 01, 2018 1:30 am
Posts: 732
...


Last edited by Dare4444 on Oct Sat 16, 2021 4:58 pm, edited 1 time in total.

Top
 Profile  
 
Post New Topic Post Reply  [ 84 posts ]  Go to page Previous  1, 2, 3, 4, 5  Next

All times are UTC [ DST ]


Who is online

Users browsing this forum: Mike Toon and 5 guests



Search for:
Jump to:  


































Privacy Policy :: Powered by phpBB