D. I. Y. Watch Timing Machine.

[CWRNH]

13 hours ago, LittleWatchShop said:

0805 and above are pretty easy...especially if you use solder paste.  I have done it.

Engraving the board is a pain, it takes a long time.  Soldering is not as big of a deal but I can have it made and have most of the components soldered on pretty reasonably.   There are several other boards I am having made at the same time for various other projects so it will be convenient.

[LittleWatchShop]

2 hours ago, CWRNH said:

Engraving the board is a pain, it takes a long time.  Soldering is not as big of a deal but I can have it made and have most of the components soldered on pretty reasonably.   There are several other boards I am having made at the same time for various other projects so it will be convenient.

Never done that, but I can believe that it is very slow.  I have done lots of boards using transfer paper and etching.  They go pretty fast but a little messy.

[JohnR725]

4 hours ago, CWRNH said:

Engraving the board is a pain, it takes a long time. 

Did I miss the picture of the board? Then what you mean by it takes a long time hours days weeks? Then a little more detail would be interesting?

[CWRNH]

46 minutes ago, JohnR725 said:

Did I miss the picture of the board? Then what you mean by it takes a long time hours days weeks? Then a little more detail would be interesting?

I shared some pics earlier of the finished machine, I may not have shared pics of the actual board..  When I take it apart next time I will try to remember to take a couple of pictures of it. 

  The board I made has a few less components and uses all regular, cheap, through hole components to make it easy to solder and use the parts I have on hand.  The new design adds a regulator, battery level indicator, and a signal flasher as well as uses high quality, mostly smt components, and has added several test pads, some adjustable components, and few other changes but is very similar overall.  I tested the whole thing on test boards and everything works so the new design should work well when built all together on one board.  I added dip switches so I can shut down various parts of the circuit for testing.  Once all the bugs are worked out and the values are hammered out for the additional parts, I will have the (or several) final board/s made for my own use and might make it available to others in some form if anyone happens to be interested. 

Finding all the components with correct type, size and values took nearly as long as creating the design.  Many hours of research...... 

I tried using normal photoresist methods and it works but not as well as engraving with the cnc.  I have done quite a bit of work with photoresist and it is excellent but not optimal using what equipment and resources I have available at this time.  Screen printing is also a method I have used for many other purposes but it is not satisfactory for this particular application for much of the same reasons. 

Using cnc makes design and production easier, things can be changed easily without as much post processing.  You can just make a change, update the code, and send it to the machine to cut a new version in a new blank. 

The cnc engraving produces perfect traces and excellent resolution but is slow, now that everything (as far as settings) is figured out and the fixtures are made, the engraving time would be around 1.5 hours per side.  If I used my large cnc mill with a high speed spindle instead of the cnc router I could cut the cycle time down to about 15-30min per side. (still slow). 

Then after the board is routed you still do not have plated though holes or vias so it is a little more of a pain to assemble. And no solder mask. (smt components would be challenging without mask). After routing, I plated the board with tin so the copper wouldn't tarnish.   The solder and Flux I used was very cheap and flowed horrible.  With good quality solder the assembly would have been cleaner and very quick. I have over 30 years experience welding, soldering and brazing in the automotive and watch/jewelry industry but not much experience soldering surface mount components on pcbs, (nor do I really want to practice). 

Set up and figuring out settings (feeds, speeds, fixtures, drills, mills, angles, etc) took all day with a couple failed attempts.   I teach cnc programming and machine processes (and have a machine shop here) so I am not new to that aspect but still it took a couple tries to get everything optimized.  Now that all this is done I can create them pretty quickly and with great results.  But, Buying several already made and mostly assembled with surface mount components is much easier.  Cost per board is cheaper without assembly, however having it assembled and adding the cost of the components still is reasonable for what I am doing. 

[CWRNH]

Just now, CWRNH said:

I shared some pics earlier of the finished machine, I may not have shared pics of the actual board..  When I take it apart next time I will try to remember to take a couple of pictures of it. 

  The board I made has a few less components and uses all regular, cheap, through hole components to make it easy to solder and use the parts I have on hand.  The new design adds a regulator, battery level indicator, and a signal flasher as well as uses high quality, mostly smt components, and has added several test pads, some adjustable components, and few other changes but is very similar overall.  I tested the whole thing on test boards and everything works so the new design should work well when built all together on one board.  I added dip switches so I can shut down various parts of the circuit for testing.  Once all the bugs are worked out and the values are hammered out for the additional parts, I will have the (or several) final board/s made for my own use and might make it available to others in some form if anyone happens to be interested. 

Finding all the components with correct type, size and values took nearly as long as creating the design.  Many hours of research...... 

I tried using normal photoresist methods and it works but not as well as engraving with the cnc.  I have done quite a bit of work with photoresist and it is excellent but not optimal using what equipment and resources I have available at this time.  Screen printing is also a method I have used for many other purposes but it is not satisfactory for this particular application for much of the same reasons. 

Using cnc makes design and production easier, things can be changed easily without as much post processing.  You can just make a change, update the code, and send it to the machine to cut a new version in a new blank. 

The cnc engraving produces perfect traces and excellent resolution but is slow, now that everything (as far as settings) is figured out and the fixtures are made, the engraving time would be around 1.5 hours per side.  If I used my large cnc mill with a high speed spindle instead of the cnc router I could cut the cycle time down to about 15-30min per side. (still slow). 

Then after the board is routed you still do not have plated though holes or vias so it is a little more of a pain to assemble. And no solder mask. (smt components would be challenging without mask). After routing, I plated the board with tin so the copper wouldn't tarnish.   The solder and Flux I used was very cheap and flowed horrible.  With good quality solder the assembly would have been cleaner and very quick. I have over 30 years experience welding, soldering and brazing in the automotive and watch/jewelry industry but not much experience soldering surface mount components on pcbs, (nor do I really want to practice). 

Set up and figuring out settings (feeds, speeds, fixtures, drills, mills, angles, etc) took all day with a couple failed attempts.   I teach cnc programming and machine processes (and have a machine shop here) so I am not new to that aspect but still it took a couple tries to get everything optimized.  Now that all this is done I can create them pretty quickly and with great results.  But, Buying several already made and mostly assembled with surface mount components is much easier.  Cost per board is cheaper without assembly, however having it assembled and adding the cost of the components still is reasonable for what I am doing. 

Correction....  I did use high quality resistors and capacitors for the pcb I built, but the rest of the components are generally inexpensive and common...  I also used good quality cables and connectors. 

[LittleWatchShop]

For doing just a couple of boards, cnc seems to be a great solution.  I only do small boards using the transfer/etch method and I only do single-sided.  These are only for test jigs and stuff like that generally.  If I want a real board I just send it out for fab.  There are lots of houses that will do your assembly for you...I think you have figured that out already.

Bouncing around in my head is the idea of building a complete system running on a Raspberry Pi to do the processing and interface.  I would probably a challenge to put it on an Arduino (well...lots of time).

 

[guidovelasquez]

3 hours ago, CWRNH said:

 I teach cnc programming and machine processes

Thanks. Your experience is useful to us. Your project is ambitious, but we already see why. In a mountaineering movie I heard the following statement: "If you have the tools and the skills, it is possible to do what you want." Keep feeding the forum with your ideas. We are awaiting your results.

[JohnR725]

3 hours ago, CWRNH said:

The cnc engraving produces perfect traces and excellent resolution but is slow, now that everything (as far as settings) is figured out and the fixtures are made, the engraving time would be around 1.5 hours per side.  If I used my large cnc mill with a high speed spindle instead of the cnc router I could cut the cycle time down to about 15-30min per side. (still slow). 

I was under the impression people doing watch repair have to be patient and 15 to 30 minutes per side seems pretty darn fast to me? I assume that includes drilling all the holes?

Then just in case anyone else in the group wants to follow in your footsteps software you're using? Or did you custom write to us

3 hours ago, CWRNH said:

And no solder mask.

Solder mask no problem at least according to this link.

https://www.bantamtools.com/blog/make-your-own-pcb-solder-mask-with-a-desktop-cnc

2 hours ago, LittleWatchShop said:

Bouncing around in my head is the idea of building a complete system running on a Raspberry Pi to do the processing and interface.  I would probably a challenge to put it on an Arduino (well...lots of time).

Another solution would be to find one of the 32 bit arm processor boards because they can be quite inexpensive.

[CWRNH]

6 minutes ago, JohnR725 said:

I was under the impression people doing watch repair have to be patient and 15 to 30 minutes per side seems pretty darn fast to me? I assume that includes drilling all the holes?

Then just in case anyone else in the group wants to follow in your footsteps software you're using? Or did you custom write to us

Solder mask no problem at least according to this link.

https://www.bantamtools.com/blog/make-your-own-pcb-solder-mask-with-a-desktop-cnc

Another solution would be to find one of the 32 bit arm processor boards because they can be quite inexpensive.

I used "tina" for my first schematics and simulation. 

I used Fusion 360,eagle, and easyeda, as well as some others for design, settling with easyeda for the board design and fusion for the case. 

I wrote some of the gcode by hand, mainly for the holes, and cutting out.  

I use Mach3 and mach4 for my cnc controller software on my cnc machines.  These controllers use gcode so I had to convert the files.  I bought coppercam to do this, it works well but is frustrating at times. 

Yes everything was done on the cnc router, a very inexpensive and small one at that.  But it worked well enough. 

I used all carbide tools and went very conservative with the speeds/feeds. 

I found that programming for 20 degree v cutter while using a 10 degree cutter for the actual work gave best results. 

I used hand sanitizer as coolant to eliminate dust and mess, worked well and cleaned up easily.  Very little finishing was needed afterwards. 

I did all this on my dining room table which caused my wife to have some anxiety so I would not advise anyone to follow exactly in my footsteps...

I purchased the components all over, from several places, which was frustrating. 

I used a couple different scopes, a decent power supply, a multimeter,a frequency generator, several other tools, and tons of different software for testing. 

I also read every post here and on other forums, I poured over tons of articles and technical papers, and used Google quite often.  I sought advice from people that know what they are doing on a couple of occasions also. 

Now I am working on the microphone and the stand for it.  I have several other things I want to build that will utilize the amp.  Be patient, I only have very limited time to devote to this hobby/obsession.....  I will share what I can, when I can. 

Thanks for the help! 

[JohnR725]

31 minutes ago, CWRNH said:

I bought coppercam to do this, it works well but is frustrating at times. 

There's a lot of free software out there of course that can have its issues. Like the link below you might find that interesting.

http://flatcam.org

34 minutes ago, CWRNH said:

I used hand sanitizer as coolant to eliminate dust and mess, worked well and cleaned up easily.  Very little finishing was needed afterwards.

That's definitely an interesting solution to the dust problem.

34 minutes ago, CWRNH said:

cnc router, a very inexpensive and small one at that. 

I assume you mean one of the Chinese ones under $300 like the 3018 for instance?

35 minutes ago, CWRNH said:

I found that programming for 20 degree v cutter while using a 10 degree cutter for the actual work gave best results. 

What about using Carbide End Mills versus the V cutters?

 

[CWRNH]

23 minutes ago, JohnR725 said:

There's a lot of free software out there of course that can have its issues. Like the link below you might find that interesting.

http://flatcam.org

That's definitely an interesting solution to the dust problem.

I assume you mean one of the Chinese ones under $300 like the 3018 for instance?

What about using Carbide End Mills versus the V cutters?

 

I have flatcam and tried it a couple times in the beginning...  not great in my experience but works (I guess) and is free I believe..

 

A lot like this one but I built my own controller and use different steppers...

 

End-mills work but leave a bad bur all around that is difficult to remove.

[CWRNH]

If there is interest, and if it is okay with @svorkoetter and wlysenko, I can create and share finished CAD designs of their designs with the files and machine code, with the BOM and Placements info. 

After I finish my own latest board, I will have several extra boards left over, because there is a minimum order...  I may be willing sell a few (my version) if someone feels the need to have one to use with WOS instead of trying to figure all this out on their own to make one themselves.  The boards I am having made won't have all the components installed but will be much easier than building the entire thing from scratch.  WOS is awesome, making a decent amp that provides decent results, is a bit involved....

[CWRNH]

For fun here is an amp that can be used to drive a regular speaker from the WOS amp....

[CWRNH]

5 minutes ago, CWRNH said:

For fun here is an amp that can be used to drive a regular speaker from the WOS amp....

I need to check the values, after posting this I realized this may have a couple resistors wrong.  If anyone wants to build it message me and I will get you the correct values.

 

[LittleWatchShop]

C7,C3: 47nF (maybe you meant to use 47nF and not 47pF in all instances)

R4: 10 ohm

C1 is not necessary

C8 is not necessary--short pin 2 to ground

C9: 10uF recommended on datasheet

R5  is not necessary, short it

The layout does not look correct w/r to C7-R4 but it is hard to tell for sure from picture.

TTFN...will look at it again in the morning.

 

[CWRNH]

7 hours ago, LittleWatchShop said:

C7,C3: 47nF (maybe you meant to use 47nF and not 47pF in all instances)

R4: 10 ohm

C1 is not necessary

C8 is not necessary--short pin 2 to ground

C9: 10uF recommended on datasheet

R5  is not necessary, short it

The layout does not look correct w/r to C7-R4 but it is hard to tell for sure from picture.

TTFN...will look at it again in the morning.

 

Actually, I accidentally posted the wrong schematic.  The whole thing is wrong, I need to post the correct one, I will check over everything and repost.  SORRY!

[CWRNH]

8 hours ago, LittleWatchShop said:

C7,C3: 47nF (maybe you meant to use 47nF and not 47pF in all instances)

R4: 10 ohm

C1 is not necessary

C8 is not necessary--short pin 2 to ground

C9: 10uF recommended on datasheet

R5  is not necessary, short it

The layout does not look correct w/r to C7-R4 but it is hard to tell for sure from picture.

TTFN...will look at it again in the morning.

 

Here is the correct schematic, I will update the PCB too.

[CWRNH]

8 hours ago, LittleWatchShop said:

C7,C3: 47nF (maybe you meant to use 47nF and not 47pF in all instances)

R4: 10 ohm

C1 is not necessary

C8 is not necessary--short pin 2 to ground

C9: 10uF recommended on datasheet

R5  is not necessary, short it

The layout does not look correct w/r to C7-R4 but it is hard to tell for sure from picture.

TTFN...will look at it again in the morning.

 

The original circuit I actually used to create this schematic was one I found on the web, I just modified it to use with the WOS.  The wrong schematic I posted earlier was one I was working on and never finished.  Somehow I must have deleted the good one so I had to make a new one (which is posted in the reply just above).  Now I will redraw the PCB to match the Schematic.  Then I will post the whole thing again for review..  The schematic in the post right above this one matches the working circuit I am using on a board at this time. I have too many files going, I need to delete or archive all the old ones so these types of mistakes are less likely.....

[LittleWatchShop]

Take a look at this 386 board

[LittleWatchShop]

25 minutes ago, CWRNH said:

Here is the correct schematic, I will update the PCB too.

Much better!  C6 is still unnecessary. C2 and R1 should follow data sheet recommendation.

[CWRNH]

7 minutes ago, LittleWatchShop said:

Much better!  C6 is still unnecessary. C2 and R1 should follow data sheet recommendation.

Updated routing to match the schematic.  Not the greatest layout but should work...

 

[CWRNH]

14 minutes ago, LittleWatchShop said:

Take a look at this 386 board

MUCH easier!  LOL

Pretty much the same thing

[CWRNH]

So... Here are the updated speaker amp screen shots.  This should work fine but as LittleWatchShop pointed out, it is easier to just buy one from amazon...

The circuit below does work, probably should be updated as LittleWatchShop pointed out, but I am using the circuit on a board now and it preforms fine so far.  Mine is just not surface mount so obviously the layout is different on my working board but the schematic matches....

[LittleWatchShop]

There is some coupling from pin 3 to pin 2. Pin 2 is connected to ground with an internal 50k and the external 470pF.  If it is not causing an issue for you then leave it.  I would guess the coupling is on the order of 10pF, so it is greatly attenuated by the 470pF, and it is a HP filter with a pole way beyond frequencies of interest...so...yeah...probably OK as is.

[LittleWatchShop]

On 4/18/2021 at 6:04 AM, CWRNH said:

I am wondering if someone can help me find the data sheets for this. Made by TI, probably in the mid to late 80s. I have many thousands of them, All NOS, in the TI packaging.... I asked TI, they were no help, I searched the web, nothing...

I believe the mystery is solved.  It is a Unijunction transistor (UJT)

They make simple oscillators

Here is one oscillating on my bench.