My first DIY modules - starting a Kosmo format build

[juggle]

Ok…video tomorrow. Got to talking to a friend and “lost” my evening.

But…did do something else today.

Had to go to the store for some stuff. Took a look at inkjet printers while I was there. I know…I hate inkjets. But…so many people seem to be loving them for making PCB negatives so I thought I’d take a look. Most were out of stock…what was left was actually the cheapest option. A Canon MG2522 for $35…quite a few of them actually. I’ve had good luck with canon inkjets over the years. I loved the BJ-100 when it originally came out and until just a year or two ago I still had a BJ-200e that I still used because it was rock solid reliable and ink was dirt cheap for it. But given the difficulty of using a parallel port printer anymore I finally pitched it after stripping it for parts a few years back (I was hoping it would have suitable rods for me to build a 3D printer around…it didn’t) I also have a Pro9000 my wife won in a photography contest about 12 years ago…they don’t even make drivers for modern versions anymore but windows will still use vista drivers so I can still print to it. Only problem is ink for it is crazy expensive and is pretty much dry each time I go to use it. I also have an old Canon all in one (that uses the same inks as the Pro9000) but I only use it as a scanner since it has a nice sheet feeder.

Anyway. Point is the $35 Canon followed me home.

But I’m not impressed. Results don’t seem to be as good as my laser

I tried it on a transparency first - not expecting good results. And got what I expected. It looked ok when it came out. But close inspection showed it wasn’t very opaque - and…the ink never dried. I did see an article about someone who uses this to their advantage by sprinking toner powder over it and then using heat to set the powder…but…that just sounds crazy messy and like just laser printing is a better approach.

Next I tried my vellums. It worked…but even with the settings cranked as high as they could go it’s just not as dark as my laser:

The inkjet transparency is the lower right corner…you can see how it smeared even 30 minutes after printing.

So here’s a better comparison without that but with both vellums on laser and inkjet…and laser on two transparency options:

So…the two laser transparencies look like they have the highest contrast. The laser Staedtler vellum looks like it’s denser than the laser strathmore (the strathmore is more UV transparent than the Staedtler though) Both inket vellum prints look less dark…but the staedtler looks better than the strathmore.

The “new” to me Koala laser transparency looks better than the generic transparency I was using before. It’s also thicker and didn’t seem to warp like the cheaper stuff did. So I may have to give tahat a try.

So I may have to try the Koala transparency with the laser…and my thought of trying the staedtler for my negative but the strathmore for my diffusion on the light might be the best bet.

But…I also got a string of UV LED’s so I’m going to build an exposure grid tomorrow which should let me get shorter more even exposures without the diffusion. So maybe that along with the Koala or Satedtler will give me the best results.

Though…I need to finish building these two Yusynth boards I did etch already. I’m waiting on the IC’s for one though so no rush there…the other I didn’t have the right trimmers but new ones will be here tomorrow so that will probably get built at least!

[Farabide]

I know I mentioned Big Clive before but he’s specifically covered getting the best from an inkjet for PCB. His UV setup was a nail polish dryer from the pound shop. Keep going.

[juggle]

Yeah, watching his videos was part of what inspired me to give the inkjet a try and I did use his setting suggestions. He seemed to have best luck with pigment inks…but inkjets that can use pigment inks are not cheap (the cheapest I found started at $300) so that’s not a very realistic option.

I’m not sure if this inkjet doesn’t have the same resolution as my laser or if it’s just that it doesn’t print as dark…but those test prints were the VCF filter with the noisy bitmap and the inkjet prints of them came out “cleaner” as even cranking the contrast it still didn’t print the noise. But looking at the prints on the lightbox side by side it’s obvious that even in the full black areas the inkjet just didn’t get as dark.

However - there’s a possibility that the inkjet could be better for stacking negatives to get more density since there’s no heat and in general inkjets tend to be more accurate and repeatable than lasers. But I’d really like to avoid stacking. And the inkjet can’t print on the transparencies and stacking vellum is a big pain due to it only being translucent.

I am going to try one more thing. I ordered a pack of inkjet semi-transparent film for silkscreen printing like Clive suggested and had good luck with. It was that or inkjet specific transparencies…and since the transparencies were almost as much as I paid for the printer I figured I’d try the semi-transparent film designed for UV exposure. The inkjet isn’t completely out of the running…but right now it’s not looking good for it. At least not with a cheap inkjet and I’m in no rush to buy a more expensive one.

I could try my nice Pro9000…but that would mean moving a bunch of stuff that’s stored on top of it

I did just check the specs…and the cheap canon is lower resolution:

Color: Up to 4800 x 600 dpi
Black: Up to 600 x 600 dpi
Resolution may vary based on printer driver setting. Color ink droplets can be placed with a horizontal pitch of 1/600 inch at minimum.

While my pro900 has much better specs:

Black: 4800 x 2400 dpi
Color:4800 x 2400 dpi

So I might try the pro9000…but it’s not very realistic for people to own a printer of that level. And it still uses dye inks…though higher quality dye inks that may be denser. It’s also got 3 of it’s 8 ink tanks reporting as in need of replacement and since they’re like $20 each…I’d rather not have to replace them

[juggle]

Ok…gave the pro9000 a try. Would have been smart to do that before buying another printer. Oh well.

Results are…better than the cheap printer. But still not as dense as the laser. I do see more detail (more of the noise in the print is visible like it is with the laser) But the density is still a notch down from the laser even with everything cranked to the max. There’s also some bleeding - it appears cranking things all the way resulted in more ink than the paper could absorb. So more ink isn’t the solution here. The ink itself just isn’t as dense.

Really need pigment ink for inkjet to work well I think.

[juggle]

I really need to pick up a decent recording interface so I can get better audio recordings. But…here’s a little clip of the 808 Kick:

Sound Sample of my M4TM 808 Kick

The Tone and clipper are subtle, not sure if they even come across in the audio here. I can hear a difference in person, just not much of one.

Also, I’m not sure if the pitch is acting as intended. Looking at the circuit I’m not sure it is. Seems like it should just go from low to high…but I’ve got low in the middle with both extremes getting higher pitched.

Oh - and I really need to spend some time to figure out the drum mode on my sequencer and get the triggers wired up so I can do something a bit more interesting than just feed my LFO into it. (yeah, I could do something with a VCA and the clock divider to get a bit more interesting…but…I didn’t feel like mucking about with that.)

[juggle]

Finished a few projects to help with the modular building:

Micro drill press for drilling PCB’s. Works better than I expected. The cheap $30 drill motor and chuck from Amazon are more accurate than I anticipated. I did make a few design changes from the original design I found on Thingiverse:

1. The 8mm smooth rods I found were 100mm long (and actually measured 103mm) so I made the frame a bit taller.
2. I added 608 bearings on the axle so the threads don’t turn against plastic.
3. I tightened the holes for the smooth rods - I’d rather they be a tight fit than have any slop.
4. I added holes and nut traps for 2 M3 screws at the front of the lid to clamp the the smooth rods in place - with the tighter fit this helps pull it together.
5. I added a clip to hold my chuck key on the side.
6. I redrew the lever handle and attachment so they have printed threads and the M8 all thread just screws in place.
7. I drew a printable base - but then just used it as a pattern to make one from wood with some printed feet.

I’m pretty happy with how it came out. The springs I have in there are a bit stiffer than I’d like but they’ll do until I find some better ones. It’s considerably quieter, more ergonomic, and most importantly more accurate than my big drill press. So big win.

I also made a new UV LED array for exposing PCB’s

I should have thought the wiring through a bit more carefully:

If I had flipped the direction of every other row I could have made one of each wire shorter and had them lay flat tested inside of each other instead of looping over on the ends. Oh well…It still works.

This is based on the design from http://www.johncon.com/john/PCB/#exposure but expanded from 10 to 12 strips to fill my printing frame. I lined them up as evenly as I could be it’s not 100% perfect mostly due to the LED’s not being consistently placed on the strips. Still- close enough and holding a piece of paper 1" away (the distance it’s designed to be used at) makes for VERY even illumination.

I’m currently printing some spacers that will attach it to my printing frame and hold it 1" from the boards to be exposed. Hopefully this will give more consistent exposure results and let me go back to <1 minute exposures. I’m hoping laser printed transparencies or even laser printed staedtler vellum (which gives the most contrast but blocks a lot of UV light) will work consistently with one layer now. Going to take a few hours for the spacers to print so I’ve got time to get a board or two prepped for test exposures!

[juggle]

So the new UV exposure array works great.

I messed up on my corner spacers just a bit though - I decided to try using heat set inserts in them so I could screw the array to them. Mostly because I got an assortment of heat set inserts off Amazon and was looking for an excuse to use them. Three came out great…but I got sloppy on the 4th and somehow missed that I was missing the hole entirely. Whoops. So I only had three corners supported tonight but it still worked great:

Basically just a chunk of 1/4" MDF cut to the same size as my contact printing frame. Then I made some spacers sized to sit on the corners and hold the LED’s so they were 1" away from the glass. The glass is inset inside the frame about 5.5mm so the spacers aren’t quite an 1" to get the distance right:

The LED’s aren’t as bright as the 6 in my UV flood…but they give a MUCH more even illumination. And with that many of them they don’t need to be as bright.

I did a test with 10 second intervals:

That particular piece of board is getting pretty beat up because I’ve stripped it and re-used it so many times for exposure tests. So it was a little hard to evaluate as there were flaws in the copper itself that were easy to confuse with flaws from the exposure.

But…40 seconds I see absolutely no leakage through the masked areas and the fine details look great. 50 seconds I see just a tiny tiny bit of leakage through some of the mask and things look just a tiny bit more defined.

This is about what I expected since the site I got the idea for this exposure rig from came up with 49 seconds as an optimal exposure. His laser may make slightly denser negatives than mine. (This was with the test pattern printed on a Koala brand 4.5mil transparency) I could probably go with 49 and be ok…but the slight bit of bleed I see at 50 worries me. 45 is probably a safer bet…but I went with 43. And was really happy with the results:

Yes, I’m back to the Yusynth ARP design that I tried my very first time. With 43 seconds of exposure in the new rig it looked pretty light…but once it was developed it darkened up nicely. I then put it back in for two more exposures to harden the mask…which may have been a mistake.

The yusynth logo and name didn’t come out great…but that was mostly due to the photoshop work I had to do to clean this one up (remember my initial post about this particular design from yusynth being a bitmap instead of vector like his others.) Honestly the biggest issues with this board are still due to the artwork. I cleaned it up in photoshop…but it still isn’t as nice as proper vector artwork and it shows. Some of the pads are a bit thin in places as a result. But I got rid of most of the noise.

Etching went super quick again. The 1:2 HCL/H2O2 mix works really well and fast warmed up just a little bit.

Now, the reason I don’t have a photo of the final board and why the double exposure to set the mask may have been a mistake. Start to finish it took me under 30 minutes from hitting print to a fully etched board. The process is definitely coming together! But…once it was etched I had to strip the mask.

It’s been in the heated developer for an hour now and is still not 100% stripped. It’s about 95% there but some of the edges are still hanging on. Stripping the mask is still the slowest most inconsistent part. So I won’t get to build this one up tonight. I wasn’t really wanting to drill it tonight anyway. Since Fiberglass dust is so nasty I don’t want to drill it in the house unless I have a good vac going and my respirator on…and I’m not running the vac after my daughter is asleep So it will be stripped by morning and I’ll drill it so I can try to build it up. A little worried about some of the traces just due to how the art came out…but should be able to make it work.

[juggle]

Ok, pulled it out and cleaned it up on my way to bed so final update for tonight:

While I waited on it I finally started building the Steiner filter:

Unfortunately I may not be able to finish it tomorrow…there are a few resistors I may not have. At least I don’t have them in my 1% resistor collections. I’ll have to dig out the old 5% bin/bags tomorrow and see if I can find what I need. That’s just no fun since they’re not labeled or sorted so it takes forever to find any given value, but I do have a lot of oddball values in there so I may get to finish it tomorrow.

Though if I do finish it I may have to pull the megadrone and make it’s own case for it to make room in the rack

[Bitnik]

I’m somewhat awestruck. It’s like watching somebody interested in cars construct a classic model out of nuts and bolts and blocks and sheets of metal.

[Farabide]

Someone needs another rack I think.

[juggle]

@Bitnik

It’s like watching somebody interested in cars construct a classic model out of nuts and bolts and blocks and sheets of metal.

I tend to get a little obsessive / lost in the weeds on everything I touch

My sandrail:

I tore down to the bare frame, rebuilt the motor, rebuilt the transmission, upgraded the suspension, assembled my own EFI controller for the new motor…I even did most of the work making the suspension seats (I bent the steel frames, cut and glued the foam, installed the slings and covers) but that last part was mostly because I was working for the company that made because my own business was a bit slow at the time

I’m also into amateur radio and have always enjoyed building my own radios more than buying them.

I was into photography for a long time (even majored in it in college) and along with making my own pinhole cameras did some experiments with making my own emulsions and coating plates.

I even get that way with cooking. A few weeks ago I made beef stroganoff and while it simmered made my own noodles from scratch to go with it.

For a few years the only beer I drank was what I brewed myself. Which led to me growing my own hops (I considered growing my own grain but didn’t have the room!) I also make mead from time to time and have seriously considered getting my own beehives…but I REALLY don’t like getting stung

I really relate to the “How to Make Everything” youtube channel where he tries to make all kinds of things from the ground up. I just like getting into things I’m interested in and diving as deep as I can to really make them my own.

@Farabide

Someone needs another rack I think.

It’s going to have to happen sooner than I expected! Think I still have enough ply for one more…but this stand probably won’t support two racks so I may have to build something better…

[juggle]

So here I was thinking I had my process nailed down. Making boards I’m happy with in about 30 minutes (and then another 30-60 minutes to strip the mask…but…still pretty darn quick.)

But…I had ordered that special ink jet “paper” for screen printing. And it showed up today…so I gave it a test.

And…wow:

The top is my laser printer output (slightly scaled down in that test by mistake) the bottom is the cheap $35 printer on the cheap ($16 for 50) screen printing “milky translucent” paper.

MUCH more even coverage and blacker blacks. By far the best most even coverage I’ve got from any of my tests yet.

So…looks like I get to do at least one more exposure test and will be keeping the cheap printer for printing negatives. It was just a matter of finding the right media.

I printed a few other tests and tried this paper in the laser as well…but the inkjet on this “paper” wins every time for densest print.

Also did a little test to see how even the light on the new exposure array is. I put a piece of vellum up against the glass of my print frame and it’s amazingly even:

The areas that are darker is just because the piece of vellum is bigger than the frame so not flat against the glass everywhere. Anywhere it’s flat against the glass it’s VERY evenly illuminated.

If I move it just a little closer you can easily see the grid though:

So the calculations from that site to find the optimal distance seem to be pretty good. I’m guessing that with his calculations and how close things were last night that I can probably just go with his 49 second exposure with these inkjet printed tests and get great results now.

But I’ll probably do another test exposure to be sure.

And I’m still going to focus on building some of these boards before changing focus and etching more just yet!

[Farabide]

Would you care to collate your final process and equipment?

[juggle]

I was actually thinking about doing a new thread to summarize what I’ve learned now that I have a system down I’m pretty happy with.

I’m almost to that point…but want to test this new paper option all the way through first. Just don’t have another board I’m ready to etch just yet.

Just drilled the ARP board and thought I’d do a little where it started where it is comparison since it was the first board I tried etching:

The first attempt was a lot messier! And the traces were thicker because it was scaled wrong. Plus the cleanup I did in photoshop to get rid of some of the noise cost me some thickness as well. It’s definitely the trickiest one I’ve tried (so wasn’t a great option for a first test!) due to the way the PDF was prepared.

The biggest problem with this latest attempt at the ARP filter is the eurorack connector footprint. The vertical spacing is fine…but the spacing between the two columns is too tight. So I’m going to have to drill those oversized to compensate.

I think if I get one more good board out start to finish I’ll be confident enough in what I’ve learned to share the details in a dedicated thread!

[Farabide]

Wonder if you could use the eurorack connector and an ink pad to mark the center of each hole on your board to avoid the squeeze.

[Farabide]

Oh and thanks again. This is great stuff.

[juggle]

@Farabide

Wonder if you could use the eurorack connector and an ink pad to mark the center of each hole on your board to avoid the squeeze.

Not a bad idea. I’ve actually been using the 10 pin header and the DIL sockets as tests to make sure I’ve got my scale right on my prints. This one I was a little worried about as it looked like it would be close…I tried to compensate for it when I drilled the holes but apparently didn’t compensate enough.

@Farabide

Oh and thanks again. This is great stuff.

Glad you’re enjoying it! I know I’m getting a little burned out on etching circuit boards. I’m itching to get back to soldering up circuits, mounting them to panels, and playing with them

And I really need to get the trigger outs on the sequencer wired up. I’m just putting that off because it’s a dull one to wire - just 8 sockets and nothing else. Though…I also kind of want to add LED’s on them so I could draw up a custom board to handle the LED’s and transistors to drive them - except I’d have to make multiple small boards or order some larger blank PCB material as the only ones I have that big are dual sided.

Which is something else I will be trying soon now that I’ve got things working more reliably. I want to try doing @sebastian’s KOSMO version of the kassutronics ladder filter (I seem to be on a big of a filter binge lately!) I actually already printed the negatives and have the right size boards. I’m just nervous about trying to keep two layers lined up. I’ve seen a few methods for doing it and can’t decide which one to try. Plus my dual sided boards are bigger than needed so I need to trim one down for that. Maybe in a week or two…

[Farabide]

@juggle

I’m just nervous about trying to keep two layers lined up.

Early 90’s I ordered a £4000 memory upgrade for a single database server. I asked if an engineer would install the two chips. The company had their head engineer call me to say “go on, be a man!!!”
Don’t fear the f**k up @juggle.
But if it’s more grind than giggle; park it and make some tunes.

[juggle]

Mostly I just want to finish building a few of these modules before tackling the dual layer It’s just twice as much to go wrong with one board and I’m pretty sure I’ll mess up and get something backwards the first time. But I’ll be trying it soon since being able to do two layer boards will open up so many more options.

Unfortunately it looks like I may need to order some more parts before I can finish this Steiner filter

I haven’t dug for the 3 missing resistors (2 values) but since I have to order some parts for some of these others I will probably just wind up ordering a few while I’m at it.

The 3 - 1.5nf and 1 - 680nf caps I just don’t have. So I’m going to have to order those…and the 1.5’s need to be closely matched so I’ll have to order a few extras as well.

The two BC547 transistors I should have - but they need to be matched. And I just haven’t got around to matching transistors yet. Need to setup a breadboard for that since I need matched transistors for a few other boards in the works anyway. Will have to make an evening of it sometime this week.

The board soldered up well so far. Everything fit great, and with the nice big traces the lack of soldermask isn’t an issue:

I did do some repairs along the bottom since this was the board where the copper had some damaged sections. Some areas I just dragged the solder to make a bridge to help reinforce things. But where it was really bad I used a cut off bit of lead to help strengthen the tracks:

So parts on order to finish this…and hopefully a few other things as well

[eric]

@juggle

It’s just twice as much to go wrong with one board

Nah, much more than twice
All that can go wrong on one side can go wrong on the second, that’s already twice.
Then you have all the misalignments, the vias in a bad place (like just under the edge of a cap thats bigger you remembered…), etc…

But you’ve gone too far to stop now !

(And your reports are great readings, thanks.)