Amiga 500 Restoration: Retrobrighting I
In part one I described how I got and disassembled an Amiga 500 for a thorough cleaning. This part is about my first attempts to restore the grey/beige color the case had originally had, but over time had turned it an ugly yellow.
This is the second post in a series about restoring an Amiga 500 back to its former glory. Here are all of them so far. I’ll try to remember to update this in all related entry.
First Things First
Disclaimer 1: It has been quite a while since my chemistry lessons in school, and even though reading up on the details, combined with what I still remember, left me with some confidence of having roughly understood what’s going on, what follows might be woefully wrong, at is guaranteed to be imprecise. So by no means should you use this as the basis for any upcoming exams 😉. Also, I understand the links provided are not scientific publications, so they themselves might be wrong. If you know better (and can point me to the resources to prove it), I would be happy to hear from you.
Disclaimer 2: Hydrogen-peroxide and products based on it need to be handled with great care. Always wear rubber gloves and, even more important, protective glasses when handling it. I know this might sound like an exaggeration, but I am serious: H2O2 is acidic! Getting it on your skin will cause discoloration and burns. Here’s a picture of my arm where some of the stuff got stuck without me noticing.
After what could only have been a very few minutes, I felt a stinging pain and could see two burnt spots. It took quite a while for the pain to subside even after washing off the peroxide. It goes through your skin and starts to make gas bubbles underneath. So if by accident you get some on your skin, quickly rinse if off with clear water!
Getting this stuff in your eyes can cause irreparable damage and even blindness! You have been warned!
Why does plastic turn yellow?
There are quite a few theories out there why the plastic many older electronic devices — including the Amiga — were made of turn yellow over time. Not all of them do, some keep their color well, while others look quite grotesque. The one I bought from eBay looked like this when I got it (and yes, the photo is rather close to what it looked like in person).
As far as I could figure it out, external influences like heat and UV light gradually break up molecular bonds in the polymers comprising the plastic, leaving free radicals. The presence of bromide flame retardants, often cited as the main culprit, may or may not play a role in this process, but if my layman’s reading of this paper is correct, plastic can change its color regardless. In any case, the changed molecular structure changes the physical properties in several ways. For one, the material can become more brittle, but also, the wavelengths of the light it reflects shift. It seems, free radicals like yellow best, so that’s what they tend to reflect most 😉 (and yes, that smiley is there on purpose).
Fortunately, these chemicals processes can apparently be at least partially reversed by re-adding hydrogen for the free radicals to bond with. Unfortunately just soaking the plastic in water won’t help, because the hydrogen in water is very happily bonded (with oxygen) already. It’s not as simple as one might expect, but in the end, there are always two hydrogen atoms, and one oxygen atom, sticking together; hence: H2O). What's needed is “bachelor” hydrogen, ready to enter a bond with the yellow-loving free radicals.
A rather cheap transport vessel for extra hydrogen is hydrogen-peroxide (H2O2). While it looks similar to regular water, it is less stable and — if pure — usually decomposes into regular water over time. Given the chance, however, the additional hydrogen may also decide to connect with the free radicals in the plastic, bringing it closer to its original structure, and thereby color.
So put in very simple terms, getting plastic de-yellowed should work by just soaking it in hydrogen-peroxide, ideally adding some energy to encourage the chemical to split off the extra hydrogen to facilitate the reaction.
A Bucket full of H2O2?
Turns out, an Amiga’s case is surprisingly big, meaning that if I wanted to submerge it in hydrogen-peroxide, I would need a large enough container to put the top and bottom cases in, and also quit a bit of hydrogen-peroxide. This would be the ideal solution, because it would ensure the whole surface being in contact with the chemical, hopefully leading to a very even discoloration.
It certainly can be done (see this video on YouTube), but for my first attempts it didn’t seem very practical. Not only did I not have a suitable container at hand, but (in contrast to The 8-Bit Guy whose video it is) I also don’t live in Texas with an abundance of sunlight. Instead, live in somewhat cloudy Germany. Trying this shortly before Christmas also wasn’t exactly the ideal time of the year in terms of daylight hours.
So I postponed the liquid hydrogen-peroxide plan for a little while.
First tests: Bleaching the power supply with a standard hair bleaching product
I decided to perform the first bleaching experiments on a test object that is typically not the focus of a lot of attention: The power brick. I figured that if something went wrong and the results weren’t good, it would be the part I would least be annoyed by if it didn’t look perfect. It’s usually hidden under the table anyway.
So first of all, I took it apart (turns out, it was quite dirty on the inside, so it was also a welcome chance to clean it) to get the plastic parts ready.
The underside was considerably less yellow than the top, supporting the theory that exposure to light (or lack thereof) plays an important role in the chemical reactions described above.
I bought the strongest hair bleaching product I could find in a local store. Even though the brick would have fit in a large pot and could have been submerged, I decided to try out this equivalent of the “Salon Creme” used in the 8-Bit Guy video.
I don’t have the packaging anymore, but as far as I remember it did not specify the exact concentration of hydrogen-peroxide anyway. I figured that even the strongest one would probably not exceed the product shown in the video. I was a little concerned, because the list of ingredients contained some additional stuff, including ammonia, perfume and others, but that’s exactly why I picked the power supply as a “guinea pig”. One box of this product sells for around 3€ or 4€ , so it wasn’t a big investment either.
I mixed the ingredients as per the instructions, producing a white creme that didn’t even smell too bad. I then laid out the case parts on plastic wrap and started applying the creme generously. Once done, I wrapped the plastic film around to prevent the mixture from drying out.
For the very first attempt, I just let it sit over night, without any light or heat added. I wanted to find out, how much of a reaction I would get without any additional energy. After about 12 hours I unwrapped and rinsed both parts to see… nothing. At least there was no obviously discernible difference between before and after. I won’t rule out that letting this sit for much longer would have produced some change, but I was not patient enough to try. So instead, I repeated the whole thing the next day — with a fresh batch of the white creme.
This time, once both parts were covered and wrapped, I put them into the oven, heated to around 50°C. I deemed this a safe temperature, because I figured a power supply could reach it by itself on a summer day. Apparently my oven is not designed to maintain such a low temperature constantly. I had put in a thermometer next to the plastic parts. Turns out, the actual temperature and the one displayed by the oven itself were quite different. It only reached 50°C on average when the knob was dialed to around 65°C, oscillating by about 10°C in either direction.
The part in the middle is the trapdoor cover for the memory expansion port. It was put in for comparison. It showed almost no yellowing in the first place, and I wanted to see what effect the bleaching would have on original color plastic.
I left it in there for a few hours, regularly checking the temperature and looking for obvious changes. Very quickly the creme increased in volume, getting foamy — similar to shaving foam. Over time, bigger gas bubbles formed. After about 4 hours the bubbles started to disappear. It looked as if the remainder of the creme was beginning to dry out, turning into a something I can only describe as “crisp looking". At that point I decided to take the parts out of the oven.
Even through the plastic film, I could feel that the creme had indeed begun to turn into tiny crystals. I assume, the plastic wrap was not a good enough seal against the constant stream of warm air. So the water in there eventually evaporated. It’s not easy to see in the photos, but I tried highlighting the effect:
I rinsed everything with warm water, using a brush to remove all sticky residue.
(Alert readers might have noticed that I am not wearing gloves here. This is because I had forgotten to take a photo earlier and posed again, but only after having thrown the gloves away already).
Even before the parts had completely dried, the effect was already obvious. I had put a small piece of scotch tape on the side of one of the pieces to make the before-and-after comparison easier.
So the bleaching had obviously worked, which can be seen even better in the next photo, with the mouse as a reference. However, the creme drying out had the ugly side effect of getting an uneven, quite blotchy overall appearance.
Lessons learned (so far)
After the two attempts with the hair bleaching creme, I came to the following conclusions:
- There is obviously enough hydrogen-peroxide in the cheap hair product to counteract the yellowing.
- Additional energy is needed to achieve an effect in a more reasonable amount of time.
- UV light is not the only source of said energy — moderate heat works, too.
- Letting the hair creme dry produces crystalline particles. That causes ugly spots and an overall uneven result. Some of that might also be attributed to uneven application of the creme.
- The trap door cover hardly changed at all, so the effect on “original” color plastic appears negligible.
In general I'd call this first try a partial success, but there is clearly room for improvement. It was a good idea not to start top case right away. I would have considered a result like this a degradation, not an improvement. So more experimentation was needed. I decided the hair bleaching product was not very well suited for my needs. I don’t know if the same would happen with the all such products, but at least this one's formula seems to produce the dry crystals.
Bleaching the mouse — A
bucket pan full of H2O2!
I decided to let the power supply alone for the time being. Instead I replicated the 8-Bit Guy’s setup with liquid hydrogen-peroxide solution on the stove top to “cook” the key caps and the mouse, which I had taken apart and cleaned in the meantime, too.
Some googling suggested hat hydrogen-peroxide solution could be bought in pharmacies in concentrations between 3% and 30%. Apparently most of the stronger hair bleaching products contain around 12%.
I went to three different pharmacies. In the first two, without only very vague explanations, they offered maximum concentrations of 3% and 6% respectively. Only in the third one did they explain to me, that around the beginning of 2017 there was a change in regulations, requiring sellers to document who bought the stuff and for what purpose, for anything 12% and higher. To avoid the paper work, most pharmacies just stopped selling these concentrations to anyone but business customers. I have no idea why they didn’t just tell me in the first two, but even if they had had it on sale, the prices would have ranged somewhere between outrageous and ridiculous (30€ per liter, at 6%)!
Fortunately I found an online shop that conveniently offered a “six-pack” of 1 liter bottles of 11,9% 😏 H2O2 solution for a very reasonable total price of 20€, including shipping.
I decided to start with the mouse, using the same temperature as for the power brick, just this time completely submerging the plastic in the solution on the stove top.
Controlling the temperature turned out to be even more challenging. To prevent the solution (and the plastic along with it) from overheating, I had to turn the stove to its lowest setting and move the pot way off-centre, so that only a small portion of it was actually receiving any heat from below.
With this arrangement — the extractor hood above set to full, and the kitchen window wide open — I waited, checking the temperature and the plastic color regularly. After the first hour, virtually nothing had happened, but little bubbles forming. However, by the time about two and a half hours had passed, the yellowing started to disappear visibly.
After 4 hours the plastic had changed its color enough for me to consider it “original” again, so I turned off the stove and took the mouse out of the hot liquid with some pliers. I rinsed all parts with lots of warm water.
Finally, see this before and after comparison. The first picture was taken right before the procedure began, second one right after rinsing. There is some moisture left between the mouse buttons, so it looks a little darker there in the photo — it actually isn’t.
The improvement was pretty exciting and got it very close to what I would consider original color. Turns out, though, that even after taking the parts out of the solution, rinsing and drying, the bleaching reaction did not stop immediately. When i looked the next morning, the mouse had become quite a bit lighter still, even lighter than ideally I would have wanted. It does not look bad, it is just not as beige as the Amiga originally was:
More lessons learned
With the mouse done, here are a few more take aways:
- The submersion method yields very even effects — just as expected.
- Pick as small a pot as you can — you’ll need less liquid to cover the parts.
- You want really good ventilation. Standing right next to the pot for a while made my eyes burn. My wife complained about her throat getting sore.
- Controlling the temperature is not trivial. This is a consequence of the small amount of liquid and the comparatively low desired temperature (in contrast to what the stove is designed for: cook food) Hence, a thermometer is a must have.
- The bleaching continues for quite a bit, even when the plastic is taken out of the solution. Factor that in and take parts out sooner rather than later.
- Don’t reach into the hot solution with your fingers 😉
- Rinse off with warm, not cold, water. I am not sure this is really necessary, but I figured it would be better than exposing the plastic to a temperature shock.
Armed with this new knowledge, I was confident it would also work for the key caps.
Bleaching the key caps
As laid out in part one, the key caps had already undergone a session of soaking and scrubbing in warm soap water. So they were ready for their bleaching bath.
The setup was exactly the same as for the mouse. Just enough hydrogen-peroxide solution to cover them all. However, due their shape and size, the key caps tended to float, so at least those closer to the surface were not fully covered all the time. To ensure they would still get the most even treatment possible, I used kitchen pliers to turn them all upside down.
Notice that I removed the metal braces from the larger keys — I did not want to risk corroding them. Some keys would turn around on their own, especially with the little bubbles forming, so I needed to slightly shake the the pot every once in a while to get rid of them and then ensure all caps were face-down again.
With the experience gained from bleaching the mouse I took the keys out about 45 minutes earlier, after just a little more than three hours.
In the pictures you can clearly see the special keys having a nice contrast with the letters and numbers. Here are the before-and-after shots, In the top right corner you see the floppy drive eject button, which I threw in, too.
Interestingly, similar to what I observed with the trap door bay, plastic parts that did not suffer from yellowing before did not seem as receptive to the bleach. You can clearly see this in this detail shot of the space bar.
I let everything dry on a cloth over night, before putting the caps back into the keyboard. There are only very few things to look out for when doing this. First of all, remember that there are two special (smaller) springs that go under the space bar, in addition to the center one which is the same size all others.
When putting back the metal braces under the larger keys, I noticed that they would not move as easily as before. This is because the bath in the soap water had washed away the grease that makes sure there is no noticeable friction. So I added some synthetic fat to lubricate the points where the braces a hooked into the white receptacles under the keys. For this I used the stuff that came with my coffee machine, because it does not damage plastic over time. WD-40 would not be well suited here. If you go through this procedure, make sure you use such a non-aggressive lubricant.
Lastly, just as some keys had takes more force to pry off the keyboard base plate, some also took more pressure to go back in. At first I thought be springs were not equally sized after all, when I saw several keys stand out about 2 millimeters higher than their neighbors. But pushing them down just a little more decidedly made them audibly click back into place.
I started with the larger keys, because that way you have enough room to work with the metal braces. When all keys were back in place, I put they keyboard back into the case to see how it looked. At that two things became very obvious: First, even though they had spent almost an hour less in the solution than the mouse, the “after-bleaching” was still significant. The regular keys had become almost white, with the contrast to the darker keys significantly less pronounced. But more importantly, it was now very clear that I would now have to come up with a good way to treat the top and bottom case parts 😯.
A report of how that went will have to wait until next time 🙂.