Well the time has come to pour the clear casting resin. I have been putting this off for some time but, it all turned out well in the end, as you can see below :)
The Finished "Crystal Cmoy"
Burr Brown OPA2107 Operational Amplifier
Wima MKS 0.68uf 63v Input Capacitors
Nover 470uf 50v Power Capacitors
Total Weight: 150g
If you have not been following the previous posts regarding this unusual headphone amplifier/Pre amp project, here they are...
Although admitably not the best quality, I decided to use these particular jacks sockets from an old sound blaster live sound card, due to the fact I could seal them properly to stop an ingress of resin.
(The below picture has been Mirrored)
(The below picture has been Mirrored)
Pre Casting tests.......
Resin was also dripped around all the connections pins around the bottom, to ensure an airtight seal.
To elevate the circuit within the casting, I soldered a couple of wire riser's on the virtual ground running down the middle of the amplifier.
Making the Mould ....
I pondered for quite some time regarding the design and materials for the mould. In the end I decided to use a 1.5mm thick card. When cut with a craft knife, it left a very clean and flat edge, which aided accuracy a lot.
I realise that there are better ways to create a mould, such as using silicone. The aim is to get the sides as square and true as possible, as this is a one off project card seemed ideal.
Next, I designed the mould templates in EaglePCB then, using double sided tape stuck the print out on to the card to be cut.
The folding lines were scored with the craft knife, to assist bending. The card had to be scored on the opposite side of the fold.
When it came time for assembly of the mould, each corner was tacked in place with super glue till all parts of the mould were together as one. At this point, I ran more super glue around the entire length of each side. After this had completely dried, a second run of glue was applied to ensure the joints were completely sealed.
As the brand of Super Glue foretells... "One Shot" That pretty much sums this whole process up, it's a one shot deal, it will either work or it won't .
Four or more things could happen here to scupper the end result ...
- The curing fails = Fail
- The end result is cloudy = Fail
- Bubbles = Semi-Fail
- Resin leaks into the power and audio sockets = Uber Fail !
It could all end up like a Jurassic mosquito, trapped in amber, waiting for another civilisation to dig it up!
A different type of "volume"
I estimated, water would be a similar density and weight to the resin.
Unlike a lot of resin casting products, where the ratio of resin to hardener is 1/1, the resin to catalyst ratios are worked out using a combination of weight to percentage of catalyst (1% (10ml) -2% (20ml) per KG).
The 1-2% value of the catalyst arises from the thickness or volume of the cast. A thinner or shallow cast produces less heat during the chemical reaction, so more catalyst is needed. A large casting produces and retains a lot more heat so, less catalyst is needed to cure.
150ml = 143grams -2 grams for the cup (Pure water is roughly 1gram per 1ml.... I used tap water)
Now! I have an idea of the weight to volume of the cast, I can narrow down the catalyst to resin ratio
- 150grams Resin at 1% = 1.5ml of Catalyst
- 150grams Resin at 1.5% = 2.25ml of Catalyst
- 150grams Resin at 2% = 3ml of Catalyst
In the end, I went with 1.5%.
I thought it might be nice to make a couple of "Input labels" to improve the appearance of the sockets.
Using Blue Tack and clear tape, the three sockets were plugged, fingers crossed...
With the resin mixed, I had to ensure I poured slowly and close to the mould, so not to encourage air bubbles.
You can see in the below picture, there is a dome of resin rising above the mould. This is to allow for shrinkage as the resin cures.
The mould was then covered, to stop any debris or dust entering the cast.
Here you can see the start of the catalyst reaction, the ambient room temperature was around 17c
4 Mins later
8 mins later
At this point, the surface starts to gel, it shows as dimpling of the surface.
It got quite scary, that's 108c! I was waiting for a capacitor to explode, I knew the big capacitors where rated at 125c but, the red Wima's where probably 85c at best :) .
With such a high temperature, the resin had fully cured within 60mins.
On reflection, I would have probably gone with less catalyst, maybe 1% to slow the curing and reduce the temperature .
Breaking the mould ....After leaving the cast for 24 hours, the first thing to do was to belt sand the top, so it was flat to the mould. I then had a point of reference, for squaring all the other sides.
Clamping the belt sander squarely in the vise...
After some wet sanding, I was left with the basic shape.
The previous precautions when sealing the sockets, worked rather well and no resin entered the Jack socket cavities. There are a couple of tiny air bubbles but nothing that can really be seen.
The only way to totally eliminate air bubbles would have been to use a vacuum chamber or dome since. (Having thought about this, it may well have forced resin into the air cavities)
One tip if you have a vacuum chamber or dome, would be, to just vacuum the resin after mixing before the pour, the mixing process does introduce some small air bubbles.
Using the vice again, I clamped my router with a makeshift platform.
Here I knocked off the sharp edges, which would be prone to chipping.
The bearing on the router bit follows the flat side, cutting an chamfer around all the edges.
To polish the surface I used P600 then P1200 grit wet and dry paper, dipped in water.
I found that T-CUT or Brasso made an excellent buffing polish, it literally shined the surface from a dull finish.
After an amazing response to the project, a few people commented on different ways to elevate the circuit in the resin (other than using thin wire during casting).
I had thought about this for some time, there is an upside to my method in the event of a fault, I can check the +/- rail voltages, also it was for pre-casting alignment reasons.
Although the circuit will no longer be serviceable once cast, it will give me an incite into what may have gone wrong, by checking the the virtual ground (the wire stands) against the negative and positive power jack connections.
Here you can see the 12vdc split -6/+6 voltages
HOT OR NOT !
I had a lot of comments regarding heat dissipation.
Here are the results at 12vdc (-6/+6) playing music at above normal levels for 60 minutes
- The Meter on the right is measuring ambient temperature of 16c
- The Infra red thermo-meter is measuring above the IC chip at 18c
- Even when running at 18vdc the temperature only varied by 1c
I already knew the circuit would not produce any significant heat, before I began. If this was a concern, I would have embedded a small heat sink on the top of the IC, revealing itself on the top surface of the casting.
Why would anyone need a headphone amplifier ?
Although there is no metal shielding, as you would have in a conventional chassis/PCB. The amplifier exhibits no unwanted noise or RF interference, as you may associate with an open chassis design. It is dead silent, even though it is next to my mobile phone and WiFi router.
Electronic engineers have been encapsulating or potting electronics in resin's for decades, usually for vibration dampening or moisture control, I just I decided to make it look presentable :)
There maybe some concerns, regarding the capacitors in the event of polarity reversal.
If your using a manufactured power supply, such as a wall wart or power brick, the jack has a positive centre, this is not really an issue.
In the event of catastrophic failure capacitors are built with a fail safe to release pressure. On the end of the capacitor the cap is scored, therefore weakening it. This in turn stops the capacitor building too much pressure.
As a safety precaution, pilot holes could be drilled as near to the capacitor ends (not into!) as possible. This would act as a weak link or escape valve for any build up of pressure
A diode could also be used, to prevent reverse polarity.
What would I have possibly done differently...
- Use less catalyst,
- Change the White power LED for Amber/Blue or Red,
- Maybe use an IC socket flush with the top of the Resin so the Op-Amp could be changed for different sound signatures.
I hope this answers a few questions :)
Rupert Hirst © 2011