Unofficial: Schiit DIY Coaster Amp
Jul 10, 2018 at 10:19 AM Post #61 of 326
Actually it is very funny.
Now it seems like a pity party.

Anyway just curious about the tubes used. Im not very familiar on tubes but in Jason’s post he said something about making sure that the tube matches. How do I check on this? There are other websites that I looked into but they are very generalized. So with this specific tube to the coaster amp, how does one check that the tubes match?
 
Jul 10, 2018 at 10:32 AM Post #62 of 326
Also I must say that splitting the IC socket/holder to make it as a tube socket is a brilliant idea!
 
Jul 10, 2018 at 10:33 AM Post #63 of 326
There is a gentleman on here out of the east coast with a wood turning lathe, I may have to get with him sometime on a round container for a Coaster amp, I searched all over for wooden products and it was hard to find one with the 3.5 inch or bigger diameter. The input jacks take up a little bit of room but a small file or chisel could be used to set those into the cabinet side.
 
Jul 10, 2018 at 10:38 AM Post #64 of 326
There is a gentleman on here out of the east coast with a wood turning lathe, I may have to get with him sometime on a round container for a Coaster amp, I searched all over for wooden products and it was hard to find one with the 3.5 inch or bigger diameter. The input jacks take up a little bit of room but a small file or chisel could be used to set those into the cabinet side.

A friend of mine and I discussed either 3D printing or repurposing some sort of hockey puck shaped case (preferably clear). What other case ideas have been discussed/demonstrated in the main Schiit forum?

My friend wanted to add the coaster to a Raspberry Pi to make an all in one ("only one man would dare give me the raspberry!"). I have about 8 or 10 Pi's around the house.
 
Jul 10, 2018 at 10:38 AM Post #65 of 326
There are measurements you can make for matching the tubes and I follow Jason on that, but listening is a factor too. I have one set in my test jig that is so bad it is obvious. I bought five tubes and found one good matching pair, then I bought ten and got five pairs so it is a matter of luck on what you end up with. The easiest way for me to do it is with a calibrated dual trace scope and signal generator. scope.jpg
 
Jul 10, 2018 at 10:42 AM Post #66 of 326
A friend of mine and I discussed either 3D printing or repurposing some sort of hockey puck shaped case (preferably clear). What other case ideas have been discussed/demonstrated in the main Schiit forum?

My friend wanted to add the coaster to a Raspberry Pi to make an all in one ("only one man would dare give me the raspberry!"). I have about 8 or 10 Pi's around the house.

I have mounted one in a magic 8 ball, tobacco tins, and a headphone rack. I will mount one in a scotch bottle at some point. You can see those photos in the main thread but it can take some looking. Later today maybe I can re-post some here. Beer cans could work, maybe something like a Fosters lager can.

Once I get done with basic builds I was considering stacking a couple of them and combining channels and power but I need to think it through some. It can be done with power amps.
 
Last edited:
Jul 10, 2018 at 10:51 AM Post #67 of 326
I should mention too, if you only build one Coaster amp, you can add sockets to it, test your tubes and just listen for best result, then when done, remove the sockets and solder the tubes directly to the board. I am keeping a test board just as a test board but I may be doing a lot more modifications than the rest of you and building one Coaster or ten of them is not an issue for me as long as I can get parts.

I talked to the guy who does MrSpeakers headphones, they are the first company to make 3d printed headphone cups as I recall. He directed me to a company that would do most any 3d printing for you but that was a while ago and I may have a hard time locating it again. Having the initial drawing would be the biggest part of it and that is not one of my skill sets at this time. Maybe making a wooden model and sending it to a company could work but setup is the expensive part.
 
Last edited:
Jul 10, 2018 at 11:00 AM Post #68 of 326
Hoping that if this diy picks up, Schiit can then provide an option to purchase a circle case for the coaster amp. ive seen photos of it in one of Jason’s posts.
 
Jul 10, 2018 at 11:05 AM Post #69 of 326
Hmmm I have not seen that but I may have missed it. If you find it copy and paste it here. A magic 8 ball is the perfect size for the Coaster amp, they are hard to take apart but they only cost $6.00 or so, it is a toy made by Mattel. 8 ball coaster.jpg patinated coaster.jpg

AI also did a copper frame then added patina but I like the looks of pure copper better, I am doing something with that soon and I will post a photo.

I am not sure about them doing a case, it would surprise me but I have been surprised before.

I need to go through my PM's and message several people about this thread who have contacted me asking about Coaster amps. Maybe some of those folks can help with cases.
 
Last edited:
Jul 10, 2018 at 12:21 PM Post #70 of 326
There are measurements you can make for matching the tubes and I follow Jason on that, but listening is a factor too. I have one set in my test jig that is so bad it is obvious. I bought five tubes and found one good matching pair, then I bought ten and got five pairs so it is a matter of luck on what you end up with. The easiest way for me to do it is with a calibrated dual trace scope and signal generator.

I have none of that test gear.

But I did just get an email offer for a free multimeter with the purchase of an oscilloscope.

https://www.newark.com/MarketingPro...X1618,29X1619,29X1620,29X1621,29X1622,29X1623

Very affordable, starting at $3.5K and going up to $14.5K!

I think the best I could do is listen and guess and/or play tones to headphones and measure with my $15 SPL meter from amazon. https://www.amazon.com/gp/product/B00ECCZWWI/ref=oh_aui_search_detailpage?ie=UTF8&psc=1

61ncnETCbdL._SL1023_.jpg


Not ideal. I'm suggesting anyone else try that, I'm just using to illustrate my idea. I bought it to ensure I'm not listening too loudly with my Jotunheim.

I've got 10 tubes on order from E-bay.
 
Last edited:
Jul 10, 2018 at 12:27 PM Post #71 of 326
You can tell quite a bit just listening but I am willing to help if it comes down to it. I used to run my own electronics business so I have lots of equipment. There might even be apps that will show you signal level.

Virtual Bench might be a good one. Look for oscilloscope app and many come up.
 
Last edited:
Jul 10, 2018 at 12:37 PM Post #72 of 326
Jul 10, 2018 at 2:10 PM Post #73 of 326
How Do I Get Better Results?

No, this won’t be a missive about how you can swap out the capacitors for audiophile-grade products the size of a beer can, or use specially formulated solder with silver and gold to increase performance. Boringly, but fittingly, this is about a few engineering-y, measure-y things you can do to improve the results of your build.
  1. Use a tube cushion. The original Vali used 30 durometer Sorbothane pads under the subminiature tubes, as well as another piece of foam under the board to keep the tiiiiinngggggg (microphonic) noise down.
    • If you don’t have Sorbothane, use a soft rubber or foam pad under the tubes—you’re looking for something very soft. 30 duro Sorbothane feels a lot like bacon.
    • Different tubes have different levels of microphonics. Some never settle down. Listen while you match the tubes (below) and get rid of any perma-ringers.
  2. Match the tubes.These tubes do vary. Actually, quite a bit. This matters because the variation changes the gain of the amplifier. If you use two unmatched tubes, you could have gain that’s off by a couple of dB on one side. Lower gain tubes also tend to have higher distortion, at least in our measurements.
    • To match tubes, measure the gain of each side with a 250mV RMS sine wave input at 1K. The amp will put out something like about 1V RMS. However, one channel might be 1.08V RMS, and the other one 0.92V RMS. Swap the tubes until they’re closer (say, 1.08/1.02V RMS, or about 5%).
    • Consider adding pin sockets to the build to speed up tube testing. There are small machined sockets you can use for this purpose. They’re not perfect, but they are better than soldering and desoldering a half-dozen tubes. You can solder in the tube when you find the right one.
    • The 6418s seem worse than the 6088s in terms of variation, but the 6418s also have better measured performance in terms of THD.
  3. Adjust plate load for true midpoint operation.You’ll get maximum output from the amp when the output DC level (before the coupling capacitor) is 7.5V. This allows the amp to swing the rails (or pretty close).
    • I’ve provided a 10K recommended plate load (R101 and R201) for both 6418 and 6088 tubes, but, depending on the tube (see above, they vary), you may want to change the plate loads to get the output DC closer to 7.5V.
    • If the voltage is lower than 7.5V, increase the plate load
    • If the voltage is higher than 7.5V, decrease the plate load.
    • If you’re going to do this, change both loads to the same value, and don’t stress if one comes out at 7.7V and one comes out at 7.3V. That’s the tube variation again.
So, if you do all three of these steps, does it transform the amp into a super-high-performance product suitable for all headphones from CIEMs to HE-6s?

In short, nope. It’s still going to be a simple, limited, slightly noisy, high distortion, slightly microphonic little tube hybrid. But that doesn’t mean it sounds bad.

To go farther with the design, you’re moving into flying-parts-land. Adding feedback would lower the gain, noise, and THD. But that’s more parts, and it adds feedback, which some people like to avoid. Going to a Darlington output stage would improve performance as well, but that’s a huge change, and, again, more parts. Increasing voltage would be welcome, but then we’re getting well beyond what this was supposed to be—a coaster, which can also be built into a simple little hybrid amp.


Time To Build Your Own Coaster?

I hope you enjoyed this engineering odyssey, and learned a few things about the many decisions that go into making even the simplest product. Now, imagine the engineering decisions needed on something as complex as, say, an Yggdrasil (which has three different sets of firmware, a couple of dozen power supplies and voltage references, and both analog and digital sides to worry about. Or even a Freya, with its relay ladder, tube and buffer stages, and remote control.

Sounds like fun? Yep, it is.

Or at least we think so!


Here is part of Jason's last post from around July 4th, he talks here about tube matching and such.This is the last post on the Coaster amp and probably the most important one IMHO>
 
Jul 10, 2018 at 2:26 PM Post #74 of 326
To match tubes, measure the gain of each side with a 250mV RMS sine wave input at 1K. The amp will put out something like about 1V RMS. However, one channel might be 1.08V RMS, and the other one 0.92V RMS.

Hmm I guess that would require the use of an oscilloscope or an oscilloscope app maybe? I'm going to be learning a lot it appears. Thanks to all for putting up with my questions. If they get to be too much let me know.
 
Last edited:
Jul 10, 2018 at 2:32 PM Post #75 of 326
yes this part requires at least an o-scope or equivalent, as I said earlier, the biggest thing at first is just building the working amp. I can help with tube matching later if you like.

You also need a sound source but you can get that from the sound output of a computer I would think if you do not have a signal generator.

I grabbed some sorbothane on ebay and a three inch by three inch piece is way more than enough. It may get me through 8-10 amps.

https://www.ebay.com/itm/SORBOTHANE...573942&hash=item5670b40aa7:g:a9EAAOSwDlxU3TYd
 
Last edited:

Users who are viewing this thread

Back
Top