2013. március 9., szombat

Beginning of the 8th Note attenuator

At this time when the attenuator is ready I finally have time to make a lists of things I wanted and maybe have some thoughts on the resuls ts somewhere. And this is the place.

The idea of the attenuator turns back to 2012 summer when I decided to build a headphone listening system from scratch. The one thing I learned before that I would need a passive attenuator, the best I can make.
But I already had some rules back then:

It need to be relay based stepped attenuator
I believe this is the ultimate passive attenuator. My research on passive attenuators proved me a high-end DIY system must have a stepped attenuator and the most advanced of all is the relay based logarithmic resistor ladder attenuator.

It need to be small
Small size gives more options on layout of the enclosure later, and having the shortest possible tracks always has a benefit in audio. I also had plans with this exact size, as it matches the ODAC in dimensions. Small size means less manufacturing costs, and I want to stick to SMT parts.

It need to have very good resistors
In the selection of SMT resistors there aer many options, and many manufacturers. There are the Vishay foil resistors, but those would be rather overkill and expensive. I had to find a resistor which is available in SMT and it is already trusted in audio circuits. That was the Susumu RG series. There were other options I wanted to give a try but I needed a selection of resistor values in small quantities and this was not easy. As I wanted to source a few of the resistor sets I chose the Susumu.

It need to have two inputs for start
I am 99% sure I do not need more inputs. One needed for the main DAC and still one left if someone just want to plug in an Ipod. It is not hard to add inputs externally if it would really be needed.

It need to draw little power
Many relay based attenuator powers the relays all the time. The use of monostable relays does not allow power saving. This is why I chose bistable relays. These relays only need power to switch, and only for about 2ms of time. The driver circuit and logic is a bit more tricky though. But is was not a problem.

It need ... how many steps?
It took a while to find out how many steps are optimal and I found that 128 steps are the way to go. Having much more steps, for example 256, you will miss many steps when you turn a knob, because it will be so close together.
Imagine a volume knob, with any diameter. A knob usually can be turned from 0° to 270°. You probably can turn it exactly 1° and no more if you try hard but not likely. Most people will be happy to set only every second degree, and the real word example showed exactly the same. Also the lowest diameter knob you have, the less control you have on the small movements. Turning a 20mm knob slowly is much harder than turning a 50mm knob. With 128 steps I was almost unable to turn a 20mm knob only one step further... which was about 2°. Maybe with that small knob 64 steps would be more comfortable.
I like bigger volume knobs and I am sure many do. I tested 128 steps on a 48mm knob and I was more than happy with the result. So 128 steps are the way to go!

It need to have Potentiometer for control. It is a must!
Why we like potentiometers?
The advantage of using a potentiometer is if we turn it to a specific position we get a specific value. It feels natural. This is something You can only achieve with a potentiometer. And people like it. I like it too. Period.

How about alternative controls?
I do not think there is place for turn up and turn down buttons on a front plate. We are used to knobs for too long. The other similar solution some circuits use is a rotary encoder. It is kind of a digital potentiometer, but it has no end position and it may have multi turns until you reach the last volume level. That is annoying because every proper hi-fi potentiometer knob has a position mark! You may have hard time to find one without that. If it already has a position mark it would be silly if it would not tell you the volume, wouldn't it?

But there is nothing against implementing an infra-red remote control. Especially if I already use a micro-controller. Having some additional features on unused pins is good. The most known remote control protocol is RC5, so I used that. I noticed some implementations use a wast amount of calculation time to decode it. To prevent this I implemented a new method to decode the RC5 signals. This code is the most efficient RC5 decoder yet!

It need to have an onboard power regulator
The digital parts have so low power requirements an onboard regulator would not generate too much heat if any. I chose LM317 because I was already familiar with. I would never put a switching regulator (buck controller) near an audio signal.

It need to act like a good old volume control
A real solution need to be intuitive, and simple like a knob. I don't think it needs fancy displays either. So I did not design one but the controller may be supporting an external display circuit.

I think the above are essential probably to most people needs. When I designed the attenuator I faced some problems and had to find several solutions. But I always followed my rules above and I think it turned out very well.

There are some interesting subjects I will write about more like the new behaviour of the controls I implemented (now called LULO) to make the potentiometer and the remote control friends.

Nincsenek megjegyzések:

Megjegyzés küldése