Development Blog

3rd December 2016
Got the idea! Figured out the concept. Programmed the first code for my Arduino Uno and made a crude circuit on a breadboard. Since this board is intended to be able to display basically any colour, just as a rainbow, I decided to name it "Bifrost" after the Rainbow road from Nordic Mythology.

4th December
Needed to learn some electronics. Read and Youtubed about EAGLE to design circuits and boards. Learned and poked around a little with it and designed some small stuff just to get a grip. Started to design the circuit from my breadboard and the first layout of the PCB.

6th December
First PCB done! Tweaked the software to be ATTINY85-compatible.

7th December
Ordered the first test-PCBs (v1.0).

January 4th 2017
Received the first batch. Noticed a flaw in my design. I forgot about pull-down resistors for the FDD and HDD lights. This left them open-ended to read and so those lights flickered randomly. Added these resistors to the design. Also forgot to actually add +5 to the ATTINY85 onboard. DOH! Latched power from the PWR_LED into the design since that is always on anyway.

January 7th
I am very unsure if the PWR_LED lead from the Amiga can support the draw for this board. I don't want to fry mine or someone elses Amiga! People need to be able to buy a safe product. I dont have a reliable multimeter to check it and datasheets dont give me the data I need. Transistor in AMIGA that supplies the current is rated for 200 mA, and each LED is drawing at most 60 mA/ea + the ATTINY85 @ max 20 mA. That is exactly 200mA all combined. Cutting it a bit too close for comfort imho!

January 8th
Designed an additional "Power Board" to latch over the Floppy power connector and pull current from there instead. Updated the Bifrost board to include a separate +5V lead.

January 9th
After some testing of the v1.0-boards they appear to be fine. However, the LEDs highest setting is EXTREMELY bright! I added brightness throttling to the software to draw less current because my concerns about drawing too much power.

January 15th
Ordered the second round of test-PCBs (v1.1) + power board.

February 12th
Recieved the v1.1 boards!

February 13th
Testing underway! All looks good! I dont really care much for the additional power board I designed that is "needed" to run it. I made the decision to actually invest in a proper Multimeter tool! I always wanted one anyway and there is no time like now when having one for a project like this could be crucial.

February 15th
Finally got to measure actual current draw of my design. I use the software as-is with throttling the LEDs because even though throttled they really are bright enough. Put all lights to all colours (R+G+B) on all three LED's and full "supported" brightness. And good news! The whole circuit draws no more than 50-60mA! Way less than 200mA that the Amiga transistor is rated for. :-)

February 17th
Updated the Bifrost board design to include a solder latch for using the PWR_LED current or a separate current from the power board. Kept the option of the power board in case I want to remove the throttling of the LED in my code.

February 21st
Ordered the third round of test-PCBs (v1.2).

March 17th
v1.2 boards arrived. All testing looks good, both with and without the power board. I decided to go with the throttled LED's to keep the board down to drawing a maximum 50-60mA in favour of loosing the additional power board. A new design flaw was noticed. I still need a pull-down resistor for the PWR_LED so it does not mimic the Amiga Filter being turned on and off. So silly to not add it while doing it for the other two LEDs. Updated the design for v1.3.

March 18th
Quite confident after testing in bench and in Amiga now. This is actually working very nicely! Using my own Amiga for it and all seems fine. Starts to look for a supplier of a small run PCB and partial assembly. Asking for quotes.

March 21st
Got some quotes back and I decided for a place. Discussed a lot with the factory to get all sorted. Revised the board and removed the solder latch/bridge and the separate +5V entry as I am confident they will not be needed. Board is now at v1.4 and deemed worthy to be the first official release!

March 26th
Ordered the v1.4 boards!

March 28th
Started designing the LED plastic top to channel the light up and to make it look like a proper LED. 3D-printed it to use for myself.

March 29th-April 6th
Redesigned the top in various stages because of fit issues and light channeling issues. 3D-printed plastics does not let much light through even if transparent plastic is used, but I managed to circumvent this in the end by making it hollow.

6th April
Want to build a website for the project, however simple it may become. Found to be available. Perfect! Fooled around a little with some designs.

10th April
Recieved the v1.4 boards!

11th April
The happy face was not staying for long ofcourse. While testing now I noticed a new design flaw in my board, however wierd because it only appears for certain colour combinations(!). The ATTINY85 can have hiccups and hang when the Amiga Filter is turned on because less current momentarily flows through to power it and a disruption can occur (even tho it is well within the voltage spec for the circuit). This is such an easy mistake but luckily its an easy fix too as I can just latch a capacitor onto the board, and I found a good spot for where to place it.

12th April
First official board (v1.4) is running perfectly in my own Amiga. I will leave it in there for the moment so it gets used and I have the possibility to play around with it whenever I feel like it.

13th April
Started to think about molding the LED top instead into one piece as this will make it easier to channel the light from the LEDs. Started to design an inverse mold to 3D-print.

15th April
First inverse mold done and printed. Started creating the second half.

17th April
Started making the pitch video. Found a nice piece of music to go with it from Epidemic Sound. I decided to keep it simple and just do it in Movie Maker.

19th April
Printed the second half of the inverse mold.

22nd April
Fooled around with some more designs and worked on both the pitch video and the website.

25th April
Assembled all the printed parts of both the inverse mold halves.

30th April
First mold's first half was poured into the inverse mold.

1st May
The first mold's second half was poured into the inverse mold. The first half hardened and trimmed clean.

2nd May
Mold from yesterday have hardened. Trimmed clean. Poured the first round of plastics into it. Excited!

3rd May
Ok, that mold was a letdown. Plastic has hardened but mold was not good. Too much spacing between halves made it too thick to use. :-( Will try the mold again a second round but try to apply some more pressure to keep them tighter together. Thinking Plan B to be a new mold that will be "one-half" only. Also thinking about a hybrid mold that consist of both a 3D-printed object and molded plastic material. Thinking opaque encasement to shield LEDs from spillover light from other LEDs and fill it with transparent plastic.

4th May
Going with the white hybrid. That will reflect and funnel light better. Making 5 different 3D-models to go with the two compound plastic for the molding. Poured and now have to wait 24h... Worked more on the presentation video clip. I am not versed in this stuff at all so I'll consider it good enough.

5th May
Still not sure about the plastic top. I have a few ideas I will explore. I will probably have to adjust the dimensions of it to get it done right. Noticed also that Case revisions have different width for the tops. Found 15mm on earlier cases and 16mm on later cases. Since I want this to fit all computers I will have to go with 14,5mm. Adjusted the 3D mesh model to reflect this.

6th May
Finished most parts of the project website.

7th May
Finished all parts of the project website. I'll consider it good enough. Officially published the Pitch Video and Demo Video on Youtube. Uploaded the webpage to my hosting site. Dang, now I am nervous! :-S

19th May
Made a new 3d mesh for the LED tops. 3D print hybrid with a mold from the original LEDs for smoothness. Also made a split wall to simulate the original double LED design. 3D printed a prototype and poured a mold for it.
Improved the Development Blog section of the webpage from text-only to a formatted html-page.

20th May
Looks like I have misjudged the shipping costs for this item. The actual cost I can get away with is 10eu on average (for registered mail), and I am charging 6eu. Thinking about adjusting but then a unit would be really pricey for the user that want one. I don't want to charge a higher total to be honest.

28th May
Been away on work for a while so haven't been able to focus on this the past week. However, the new LED top looks stunning. Today it had hardened enough for me to test it. It is crystal clear right out of the mold, so I would need to sand it down just a notch to spread the light better. I hope I will have time for that tomorrow. But now its really on the right track and it fits perfectly in the LED slots. Good progress!

1st June
First day of the month where I plan to ship out my boards has arrived! Now I am just going to do one more experiment with the LED tops. It looks really good already and to be honest I would normally say its good enough. However, they do leak a little light, mostly probably because I opted to print the LED top "chassis" in white PLA. It would be awesome if it didn't leak any light at all, so now I am just going to print it in black instead and mold it and then I will evaluate which one to use in the final product. Have to wait for tomorrow for it to dry before I test!

19th June
No updates for a while again. I have been busy pouring molds and creating the LED tops for the initial batch of boards. Hopefully I will start sending them out soon (this week), so keep your eyes out for when I contact you regarding yours!

24th June
On track now. Today I assembled and tested enough Bifrost boards to cover all pending orders. Now I need to just finish up enough LED tops for them. Unfortunately I have only had one mold so far, and each top takes 24h to harden so it has really been a bottleneck. Today while doing the boards I actually made two more molds for the tops so now I can speed up the last of the production and be better prepared if more orders come in. Looking to ship all orders on Monday (two days from now) provided nothing happens with the new molds so they are not usable.

26th June
First batch of Bifrosts have gone out now. If you placed your order before 13th May it is now on its way and you should have received a tracking number. If you ordered after this date, I am only waiting for freshly molded LED tops to dry up and I will ship the rest as soon as I can. Boards have all been built and tested so there is nothing saying that the rest can't ship this week. I am working towards making that happen at least. So just hang in there for a little while longer! :-)

27th June
Yesterday a kind user posted a heads up for this project on the german Since then there has been a great influx in orders. Since I am making these to order I suddenly have a backorder for a few weeks to cover. This will not affect already placed pre-orders so for those that pre- ordered: Don't worry! :-)
Now since I have started shipping I thought it would be proper to create a brief unboxing video. So I did that and uploaded it. I also added a "Shipping Status" overview to the front page where people can see the status and development of their orders quickly.

28th June
Today I only molded a few more LED tops and assembled 5 more boards. I have started to look around if I can extend this idea to other Amiga models. A600 is definitely doable. A500 _could_ be doable with some workarounds I have ideas about. I registered domains and just for good measure. But I cannot promise anything yet. ;-)

29th June
Today I sent out the second batch of boards, and the last of the preorders. Milestone achieved: All preorders shipped! :-) Other than that I just continued to mold more LED tops for remaining orders.

2nd July
For the past couple of days I have molded more tops, and today I made two additional molds to speed up production even further. This is still a bottleneck. Additional packaging boxes and processors ordered to top up my stock as I am running dangerously low on that. Orders keep coming in in a steady pace, which is really fun and gratifying. Yesterday (1st July) my Bifrost project got a small attention flash on Amiga news site, which is really cool. :-)

9th July
I'm in the midst of fulfilling orders. I am pushing new units to completion, but new orders keep dripping in and will not give me a rest. :-) It is a luxury problem, if anything. I just hope that customers doesn't get too tired of waiting, I am working as fast as I can. I thrive to give anyone and everyone that want a unit to get one. Today I tweaked the software of the Bifrost to be more "reliant" when detecting and working with Guru's. Upped the software build to v1.3.1.

11th July
All boards completed and tested ok for next batch, scheduled to be posted tomorrow (12th July). I am still waiting for additional packaging material and I sincerely I hope I will get it tomorrow so I can post everything. DHL don't let me down please!
I have plenty of LED tops in queue now done, so I added an experiment today. I have set one batch of LED top diffusers with a dash of whitening in it to see if that distributed light even better. Hope for the best, will find out tomorrow! :-)

12th July
Managed to get the whole batch 3 of boards out. In the nick of time, two minutes to spare until postoffice close. :-)

13th July
Aaaand, the experiment with the dash of whitening in the LED tops... I only added 2% white hobby paint to the epoxy mixture but even with that little it became totally opaque. What a lolfail. :-D
Oh well, atleast I know that now. Price of progress and development is that some of the stuff you try just won't work, but the one's that do work are all awesome as it advances the product even further.

15th July
Two of the LED molds are starting to show signs of wear. Still usable but only for a few more runs. Will make two new molds during the day and replace them. I am now totally out of microprocessors so I really hope my shipment of new ones arrives this upcoming week. This poses a slight risk of future orders being delayed now, tho this might be by just a few days.
I need to adjust the shipping price soon too, worked out the difference mentioned in this log at a previous day where I myself cover the "miscalculation" and realized I am assembling these for peanuts. :-/ I am soon at break even in terms of own money invested in this project and when I hit that I can and will lower the price on the board itself. Don't know when (or if) this will happen as it's directly related to the boards I can sell. Fingers crossed!

17th July
Delay threat avoided! Got a shipment of microprocessors as "predicted" today so the immediate danger of delayed orders is no more. :-) I fooled around a little with the code (as I still have about 600 bytes of flash mem to spare) and I added a Voltage Calibration and Saving feature. This is now part of, and software is therefore upped to, version 1.3.2. This is handy because most Amigas have a slight variance in voltage dips when guru and being able to save what is unique to the machine the Bifrost will be much better at avoiding "sporadic" dimming which can be a side-effect of a freshly power-cycled machine. Calibration is done manually by the user (by rebooting a recommended three times so Bifrost can read all values properly) and saving those values by holding the middle button. A brief rain effect on the LEDs will show that its complete. Green rain effect means you have saved your calibration to memory and its activated, and pressing it again removes and defaults back to "old operation" and a Red rain effect will indicate this. So it's a toggle thing basically. Enjoy!

28th July
Decided to not accept new orders from today because the queue has been building up a lot. I havent been able to produce boards as fast as orders have been coming in, so in order to get my own stress levels down and also be able to focus on all the great people that have already ordered one this was the proper thing to do. I will open it up again once it is more managable. So if you havent got your board ordered yet, dont worry, it will be back soon. :-)

9th August
I havent updated this section for a while. Actually there is not much to tell. I have just been building a LOT of boards and have had them shipped. So far never missed a deadline! Though I must admit it has been close a few times, lacking either this or that. Shipped batch 6 today, which I am quite happy about. Now its time to focus on the "last" batch before I will take a small building break for a few days before I start accepting orders again. Cheerio!