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


It is currently Apr Wed 08, 2020 4:03 pm


All times are UTC [ DST ]





Post New Topic Post Reply  [ 214 posts ]  Go to page Previous  1, 2, 3, 4, 5, 6, 7, 8 ... 11  Next
Author Message
 Post subject: Re: Microcontroller Based Signal Generator
PostPosted: Mar Mon 27, 2017 9:31 pm 
Member
User avatar

Joined: Feb Fri 25, 2011 8:00 pm
Posts: 962
Location: Dublin, Ohio
Thanks for the input, Colin. I also had another comment from someone locally who was a technician.
His comment was that the front panel mount is better for "upper shelf" instruments which he has and I know I have.
The side mounted connectors means the device is sitting on the bench among the other stuff like my handheld multimeter.
So the front mount was better for him because he likes the shelf above the bench for test gear.
I am concerned though since the device is looking like it will be 4" X 6", i.e. rather smallish.
And the rotary selector has to be pushing in to select a value. If the small box is sitting up on a shelf and it is pretty light,
pushing in the rotary may push the whole box thus requiring one hand to hold the box and the other to turn the knob.
If the box is on the bench sitting on its back, then pushing the rotary is easy since the bench is there to hold the device steady.

So does this alter your view or do you still like side mount connectors?

_________________
Website - http://www.louhaskell.com/


Top
 Profile  
 
 Post subject: Re: Microcontroller Based Signal Generator
PostPosted: Mar Mon 27, 2017 9:47 pm 
Member
User avatar

Joined: Aug Wed 26, 2009 1:50 am
Posts: 1603
Location: Mesquite NV 89027 (from Coventry, UK)
That's a good point about shelf mounted instruments. I hadn't considered this as I don't have a shelf, so it doesn't affect me, and the side ports just looked tidier. However, I am not averse to having front ports. My existing signal generator and frequency counter are both configured like that, so it won't be a problem. Neither will the weight (or lack thereof), as my unit will be on the bench and I can hold it down while pushing the button.

Thanks again for the work you are putting in on this project.

Colin


Top
 Profile  
 
 Post subject: Re: Microcontroller Based Signal Generator
PostPosted: Mar Tue 28, 2017 3:52 pm 
Member
User avatar

Joined: Feb Fri 25, 2011 8:00 pm
Posts: 962
Location: Dublin, Ohio
Continuing on the path for the all-in-one PCB.....
I have updated the schematic, primarily adding in an attenuator for the RF output. The simple Hi-Lo approach seems to be what most instruments use.
I also have made a PCB design showing everything fitting into a 4.2" X 6" PCB with connectors on the side as discussed.
I really like the idea of getting everything onto the circuit board so that assembly is easy and putting everything into a case is simple with no point-to-point wiring needed for front panel controls.
I have also been working on a case. I have a friend who I use to work with who is in the "Maker" community here.
He has access to a laser cutter and can do acrylic and other material. He said he would help lay everything out and work with me on running a job to make the custom box.
Will need to see how much each one would cost but it looks like maybe $10-15 apiece.
I have to now get a spreadsheet going for a Bill of Materials so I can keep track of that and see how the cost is going.
I certainly have spent a lot on trying different things as part of the "cost of design". But I want it to be a solid piece of gear that will work well.

Here is the new schematic link:
http://louhaskell.com/data/generator/generator-schematic-v1.4.pdf

And the new schematic and PCB design screen shots.
On the PCB view, most components are show outlined in yellow. The ones outlined in white are the front panel items. This includes the LCD at top.
The next row down has the on/off switch, the rotary selector, and the attenuator switch.
In the third row are three pots for modulation level, sine wave output level, and RF output level. (The square wave comes out at 5V fixed.)


Image

Image

_________________
Website - http://www.louhaskell.com/


Top
 Profile  
 
 Post subject: Re: Microcontroller Based Signal Generator
PostPosted: Mar Tue 28, 2017 7:25 pm 
Member
User avatar

Joined: Feb Fri 25, 2011 8:00 pm
Posts: 962
Location: Dublin, Ohio
As stated earlier. here is a first cut at a Bill of Materials. This is not perfect for sure, but is likely 95%.
Some items need to be checked for size to make sure they have the right footprint for the PCB.
Also, some components may be cheaper somewhere else. e.g. Arduino Nano, LCD, and the DDS modules. If you are willing to wait, you can get this cheap from China on Banggood, Aliexpress, or various vendors on eBay. But you do take a chance on eBay with compatibility of some things. Lots of fakes out there.
Also, I am not good with the knobs chosen. Right price point, but would like better designs.
I had hoped to keep it under $100 so was happy with the total. ;-)

Image

_________________
Website - http://www.louhaskell.com/


Top
 Profile  
 
 Post subject: Re: Microcontroller Based Signal Generator
PostPosted: Mar Wed 29, 2017 12:01 am 
Member

Joined: Jul Wed 22, 2009 3:07 pm
Posts: 815
Hi Lou,

Thanks for sharing your design!

Two comments on the circuit.

If the encoder is a mechanical one, you would need some kind of debouncing for the contacts.
While playing with a mechanical encoder on the Arduino, I needed to put 100n caps to the ground for the encoder to work correctly.
Better draw it out explicitly in the circuit, people just might buy the plain encoder on eBay (cheaper), without that small board (that small board might include those caps, but it is not clear).

Depending on which DDS module you have, the module might include a 3.3V crystal oscillator chip (the small size gold plated chip). That chip runs way too hot on 5V. People have used the DDS module on 3.3V.
An alternative solution could be just running the osc chip on lower voltage, adding two Si diodes in the supply line within the module.
DISCLAIMER: I have not fired up my DDS modules yet, just read about other people using them.

Great work, keep it up!

Peter


Top
 Profile  
 
 Post subject: Re: Microcontroller Based Signal Generator
PostPosted: Mar Wed 29, 2017 3:37 am 
Member
User avatar

Joined: Feb Fri 25, 2011 8:00 pm
Posts: 962
Location: Dublin, Ohio
The rotary selector driver software uses software based debouncing by using a finite state machine approach to screen for and ignore bouncing of the contacts. After reading over many articles, I have seen what others have seen and it has been very reliable; no need for capacitors.

I looked up the module and it is stated as good for 125MHz operation at 5V and 110MHz at 3V. I noted the crystal as you mentioned. It has markings as "0328FCE 125M000 VI X1226". I tried to Google that and got nothing. I have noted noted any overly hot operation but then I have not been looking for that. I may leave it on for awhile and see what I get. Any additional info or pointers to details of the problem would be helpful.

Thanks,
Lou

_________________
Website - http://www.louhaskell.com/


Top
 Profile  
 
 Post subject: Re: Microcontroller Based Signal Generator
PostPosted: Mar Wed 29, 2017 4:08 am 
Member

Joined: Jan Thu 01, 1970 1:00 am
Posts: 2393
Location: Saskatoon
I'm not sure, but the 3.3 volt oscillators may only be on the AD9851 version of the module. These ones will have a 30 MHz oscillator, because the AD9851 has an on-chip 6x frequency multiplier to get a clock reference of 180 MHz. I have one of these, and I used the trick of putting two 1N4001 diodes in series with the +5 volt supply, and it worked fine. It's also necessary to use 1k or 2k resistors in series with the module inputs so as not to overdrive them with the 5 volt logic levels.

Anyway, it's definitely something to watch out for.


Top
 Profile  
 
 Post subject: Re: Microcontroller Based Signal Generator
PostPosted: Mar Wed 29, 2017 11:32 am 
Member
User avatar

Joined: Feb Fri 25, 2011 8:00 pm
Posts: 962
Location: Dublin, Ohio
For Peter and Bob:
Good update on that oscillator. This board does use the 9850 and not the 9851 chip so maybe Peter's comment applies only to that board. And yes, the level shifting is a problem with 3V devices. In this design, the LCD and onboard IL9341 controller work at 3V. That is handled in this case by the 74LVC245AN level shifter chip and the 3V power supply parts. In my early prototype, I had used the diodes and resistors approach. But, as I said, I will run the board for awhile and see if it heats up.

EDIT: I ran the 9850 DDS module for 4 hours today and the crystal was barely warm to the touch. So, I am going to suspect that the problem was with another board that uses the 9851. Good to be aware though and easy to verify so, keep those comments coming.

_________________
Website - http://www.louhaskell.com/


Top
 Profile  
 
 Post subject: Re: Microcontroller Based Signal Generator
PostPosted: Mar Fri 31, 2017 3:54 pm 
Member
User avatar

Joined: Feb Fri 25, 2011 8:00 pm
Posts: 962
Location: Dublin, Ohio
Some more design work and updates.
1. Since I have been using a module for the rotary switch, the issue of pull-up resistors was done on the module with some 10K resistors. I will be using a bare switch on the circuit board, so I tried to use the internal pull-ups on the Arduino to no avail. Did not work well. Therefore I added 10K pull-ups to the design - R19, 20, 21.
2. I had designed the input circuit for the AM alignment case where the input signal off of the top of the volume control is about 0 to -5V. Since the Arduino needs to sense 0 to +5V, I shifted that up with the U2.2 op amp by 5V to measure it. For the FM alignment, the "S-curve" will be about +/- 2.5V , i.e. an AC signal. That needs to be level shifted up 2.5V. So, I built a variable reference supply using the PWM output from pin 10 into a beefy low pass filter (C18 and R18) to get a solid low ripple signal. So now when measuring AM, I can set that to 5V and for the FM S-curve measurement, I can set it to 2.5V. The updated circuit is shown below.

Waiting now on tubes so I can fix up my new Zenith H724 chassis and try this thing out on FM alignment. Pictures to share soon.

Image

_________________
Website - http://www.louhaskell.com/


Top
 Profile  
 
 Post subject: Re: Microcontroller Based Signal Generator
PostPosted: Apr Sat 01, 2017 12:15 am 
Member
User avatar

Joined: Jan Mon 18, 2010 2:13 am
Posts: 20340
Location: Dayton Ohio
Lou, this is a very impressive project and one I had thought of some time back, but without any programing experience, would never be able to do what you've done.

As I said, I'm quite interested in this project and will try to keep up with you on it.

Thanks for sharing your work!

-Steve

_________________
Radio Interests
-Zenith
-Sparton
-Pre-War FM
Consoles and floor models, the bigger, the better!


Top
 Profile  
 
 Post subject: Re: Microcontroller Based Signal Generator
PostPosted: Apr Tue 04, 2017 7:13 pm 
Member
User avatar

Joined: Feb Fri 25, 2011 8:00 pm
Posts: 962
Location: Dublin, Ohio
Steve: Thanks for the comments. I have indeed put a lot of work into this project. I could have just built a one-of-a-kind and cobbled it together, but I wanted to make it a little more professional and able to be built into something decent looking so that others could build one easily. I have received so much help from so many on this website and forum. So I wanted to be able to use the skills I have to give this project to any who wish to use it. I hope it turns out to be something people will want to build.

I still have a way to go before I commit to a circuit board. I ordered what I had hoped would be the final parts list but a few items were physically not well suited to either the circuit board or the case. So will need to find replacement parts for those. I also wanted to build the "final" prototype so everything was fully tested. Some changes have come from that so far. Biggest item was the need for a heat sink on the 5V regulator. That sucker got hot and then I realized it was dissipating over 1 watt so that makes sense. See photo below for the latest layout on the prototype board. It is not pretty but seems to work.
Next big test is the FM alignment. I have recapped my Zenith H724 chassis and am now waiting for 2 tubes in the mail.

Image

_________________
Website - http://www.louhaskell.com/


Last edited by louhaskell on Apr Fri 07, 2017 3:48 pm, edited 1 time in total.

Top
 Profile  
 
 Post subject: Re: Microcontroller Based Signal Generator
PostPosted: Apr Tue 04, 2017 7:56 pm 
Member

Joined: Mar Sun 26, 2017 6:51 pm
Posts: 31
Looks like you are using the Arduino as the timebase (aka clock) for the DDS? If so, any possiblity of adding provision for something a bit more accurate? I'm thinking like an on-board OCXO, or at least provision for one...or provision for an external 10 MHz reference? For narrowband work you might want really good accuracy.. Just a thought.

I wonder what the phase noise will work out to be.... it will be interesting to see.


Top
 Profile  
 
 Post subject: Re: Microcontroller Based Signal Generator
PostPosted: Apr Tue 04, 2017 10:00 pm 
Member
User avatar

Joined: Feb Fri 25, 2011 8:00 pm
Posts: 962
Location: Dublin, Ohio
The Arduino is not being used as a timebase for the DDS clock. Each DDS module has it's own 125MHz crystal. Using the internal divider which is settable in software, the resolution is about 1Hz so pretty accurate. I measured these with my digital frequency counter and they are pretty much spot on.
For more info, see the data sheets on the AD9850 and details of the module as spec'ed in the parts list.

_________________
Website - http://www.louhaskell.com/


Top
 Profile  
 
 Post subject: Re: Microcontroller Based Signal Generator
PostPosted: Apr Fri 07, 2017 3:46 pm 
Member
User avatar

Joined: Feb Fri 25, 2011 8:00 pm
Posts: 962
Location: Dublin, Ohio
And now for FM.....
As I mentioned I got this Zenith H724 chassis with a destroyed cabinet for $15 (and $10 shipping) off of eBay. It was stock full of bad tubes, so many that it was "no coincidence" ;-( Have to send off a comment to the seller. Needed one tube to order from Bobarino that I did not have so good to go. Recapped it and found the AM was real spotty and scratchy.... some of those bad fixed caps in the IF cans I expect.
The alignment instructions for the H724 did not include sweep alignment so I found general instructions for another Zenith a did a couple years ago for a friend. I first did the basic peaking of the discriminator can and then moved to the S-curve using the external settings on the home-brew generator. Set 10.7MHz, 450KHz sweep and a fast rate for the scope trace. Peaking of the primary and secondary was straightforward as I had seen in a YouTube video from our friend bandersontv. See first photo. Then I put it on internal and hooked up the home-brew input to the discriminator diode. See second photo. Not as good as I had hoped but usable I believe. The scope is still better. Two comments on that. There is no sensitivity on the input. I adjusted the signal strength of the incoming signal to adjust the height of the display. Secondly, I had a residual 0.5V DC component on the signal. I have the input set to zeroed in the center so I could adjust the offset reference voltage in the input circuit if I needed to. But that is another pot in the box.. would need place for that. Final photo is peaking of the 2nd IF on the scope. As expected , putting that into the input gives nothing since it is a high frequency RF signal and the ADC could not hope to keep up with it. I obviously need an RF probe with the little diode in it. I do not have one but I know they can be built easily.
On another note, I got some cheap cables, I think from China, for the connections between the generator, the scope, and the radio. Since they have BNC connectors, I just "assumed" they were shielded cable. Nope, I got noise couple in unless I arranged everything very carefully on the bench. That is just real crappy.

All in all, I am OK with the results. Still need work on the internal display for FM. Also, need to put some final tweaks on the physical board design, check everything about a million times before I send the board off for production. It is not the $20 loss I hate to suffer if I screw up, but the month lost in schedule due to shipping from China.

Image

Image

Image

_________________
Website - http://www.louhaskell.com/


Top
 Profile  
 
 Post subject: Re: Microcontroller Based Signal Generator
PostPosted: Apr Sat 08, 2017 10:22 pm 
Member
User avatar

Joined: Aug Wed 26, 2009 1:50 am
Posts: 1603
Location: Mesquite NV 89027 (from Coventry, UK)
Looks promising. If you had an RF probe, would that allow the peaking of the 2nd IF without the use of the 'scope?


Top
 Profile  
 
 Post subject: Re: Microcontroller Based Signal Generator
PostPosted: Apr Sun 09, 2017 12:36 am 
Member
User avatar

Joined: Feb Fri 25, 2011 8:00 pm
Posts: 962
Location: Dublin, Ohio
Colin:
Yes, indeed, I will need to make or buy an RF probe and then put that into the input of the generator device. Then it would be rectified and present a wave that was a simple "hump" like we see in other displays. I thought about building that in but decided that should be inside a probe as other boxed do.

_________________
Website - http://www.louhaskell.com/


Top
 Profile  
 
 Post subject: Re: Microcontroller Based Signal Generator
PostPosted: Apr Mon 10, 2017 4:17 pm 
Member
User avatar

Joined: Feb Fri 25, 2011 8:00 pm
Posts: 962
Location: Dublin, Ohio
Hit a major milestone today - I sent off the PCB to be manufactured in China. I had fought with the software for days trying to fix a bug only to discover that the Arduino part spec had an error in it. It was in the library so I assumed it was correct. Not so much. I also added 2 1N4007 diodes into the input line from the wall wart. These serve to drop the voltage of a 9V input down to about 8V (this wall supply was putting out about 9.3V so the drop was to 7.9V). This significantly reduced the heat on the 5V regulator IC with the heat sink. That was a big help. I also noted that 7.5V wall supplies are available so if one of those is used, the diodes can be replaced with jumper wires.
I tried to design the board with both Vcc and Ground planes but that was too complex. There were too many disconnected floating islands in the design. I expect someone very experienced could repair that but I hope that my simple 2 sided board with wire paths will be fine. I made the Vcc and ground paths extra wide so that should help. And I have those bypass caps everywhere close to the respective ICs. I have included a final version 1.6 schematic and Gerber viewer output photo for a status update. Note on this view, only 1 side of the paths are shown.
I get 5 boards for $14 plus another $14 for air mail. The DHL cost was $28 for 3 to 4 days and I felt that was not worth the extra. If these boards indeed work OK, I will have extras to sell to others if they are interested.

The delivery will come in 2 to 4 weeks so not much will happen here now until the boards arrive and I can build and test a full system.
As always, send any comments or questions and I will see those and try to answer.

Image

Image

_________________
Website - http://www.louhaskell.com/


Top
 Profile  
 
 Post subject: Re: Microcontroller Based Signal Generator
PostPosted: Apr Mon 10, 2017 9:05 pm 
Member
User avatar

Joined: Jun Thu 12, 2014 3:16 am
Posts: 69
Location: Tarpon Springs, Florida
Hi Lou
I would like to be first in line for one of those extra boards.
Fantastic job!
Please let me know when a board becomes available.
Thanks so very much

Bill

_________________
"There is nothing common about common sense."


Top
 Profile  
 
 Post subject: Re: Microcontroller Based Signal Generator
PostPosted: Apr Mon 10, 2017 10:17 pm 
Member
User avatar

Joined: Aug Wed 26, 2009 1:50 am
Posts: 1603
Location: Mesquite NV 89027 (from Coventry, UK)
That is a really good price for the PCBs.


Top
 Profile  
 
 Post subject: Re: Microcontroller Based Signal Generator
PostPosted: May Wed 03, 2017 2:01 pm 
Member
User avatar

Joined: Feb Fri 25, 2011 8:00 pm
Posts: 962
Location: Dublin, Ohio
Today, at 3 weeks and 1 day after order, the printed circuit boards arrived. They look really great. From the appearance, the quality is amazing. Need to check the actual fit and finish now as well as my design, both electrically and physically. Although I was pretty careful with the layout and parts specs, I am sure there will be some problems. I wonder if anyone ever got a complex board right the first time.
AT $25 for 5 boards, I can afford to tweak the design and resubmit if needed. I just hate that 3 week delay.
Now to fire up the soldering iron.

Image

_________________
Website - http://www.louhaskell.com/


Top
 Profile  
 
Post New Topic Post Reply  [ 214 posts ]  Go to page Previous  1, 2, 3, 4, 5, 6, 7, 8 ... 11  Next

All times are UTC [ DST ]


Who is online

Users browsing this forum: No registered users and 4 guests



Search for:
Jump to:  


































-->


Privacy Policy :: Powered by phpBB