DIYHiFi.org

OK, let's party...

A couple of typical questions I received in the meantime:

1. Is it available yet?

You can place an order already, and in a few days an online order will be possible as well. However I’ll need a few weeks to set up a reasonable lead time. For now you better consider 15-20 days.

2. What does the kit comprise?

Built and tested board and encapsulated mains transformer. If you want preamplifier, you need a chassis and sockets and that’s it. (And switch if you want more inputs.) You can buy less, but I doubt you want to pay the export fees in that case, as these are the same fees for any shipment, and they make small orders pointless.

(Yes, unbelievable country, 4 billion dollars foreign trade deficit in the first four months of this year and they charge for export... OK, we'll talk later...)

3. What about PCBs?

They go different route and these are shipped from the USA. That's why they are not always available.

4. What kind of supply is it there?

Schottky diodes, followed by double common mode RC filter, and regulators. I think these ones are what they call shunt regulators.

Very exciting, Pedja - - a very nice topology for line stages... I'd like to see more about how you'd use it as a transimpedance stage (I guess I don't fully understand that topology).

It looks from the picture like your PCB has Alps potentiometer and muting relay built in as well. Is that the case? What else is on there?




By the way, I populated your high-current regulator board before leaving for holiday and I should be able to test it with an amp in September... I'll let you know.

Basically a “transimpedance” stage is synonym to “I/V converter”.

Most of industrial applications of our days use opamps to perform this task. Resistor only I/V would be the opposite approach (pioneered by Audio Note).

There is a third way too, where you use an active circuit to buffer the DAC output current, and then perform I/V conversion at the output of this active circuit. The “circuit” can be a single transistor, or more complex circuit, to supposedly rise above some shortcomings of active device(s) used. Such a stage must provide a low input impedance, and high output impedance. By default, it is common base / common gate / common grid stage. Additionally, the stages like this achieve the highest bandwidth for given active part (that's actually the bandwidth usually claimed in the datasheets). That’s why they are regularly used in RF devices. (Jocko is a pioneer in use of common base stages for I/V, and I’m sure that his RF background had something to do with this.)

This topology is what I've been using for about five years now. The Module R can be used this way too. You only have to ground (in AC terms, not necessarily in DC terms) the gate of the first JFET and to use its source as an input.


Yes.


Rectifier, RC filter, regulators, the circuit which general topology is displayed at site... potentiometer, and relay that you noticed... dunno... that's it I guess.


Great, waiting forward to it.

So, nothing new on Earth, but for those rather visually oriented, here is what a transimpedance stage is about, in comparison to more known architecture.

So, the first option includes resistor I/V conversion, then the active element (here JFET but the same applies to both bipolars and tubes) which does the opposite task (it is a transconductance stage, which means that its drain current is changed with respect to change in gate voltage), and then we again have a resistor to convert this drain current to voltage.

Apparently the two first tasks can be skipped by using transimpedance stage, and performing I to V only once.

Hi Pedja,

Use a two core cable,
connect the jFET source to DAC output.
connect the jFET gate to the DAC audio ground.

Am I interpreting your transimpedance diagram correctly?

Presumably a twisted pair would do for this.
What are the limits on the connecting cable?
Does it need to be treated as a terminated transmission line.
Does it need to be short or can it be longer, define please.

Yes.


Not really. You have to manage the DC voltage of JFET source to suit the DC output offset voltage of the DAC. But the gate of JFET indeed always has to be at DAC audio ground in AC terms.


I'm usually trying to make it as short as possible, because this is the audio signal.

PS: I'll be mostly absent for a couple of next days, so my apologize if my answers don't come quickly... I should be regularly back on Monday.

The first, I have to confirm that the Module LAB is now available and online ordering is possible.
http://www.audialonline.com/diy/modulelab/orders.php

And, news for those of you leaning toward transformer coupling... Another revision, which comes also very soon, will include a couple of changes that make this stage suitable for use with output coupling transformers instead of the capacitors. This however applies to use of Module LAB as a buffer only. The gain stage introduces quite an offset which regular coupling transformers simply can not stand.

... And, regarding the gain stage, if you don't need a gain you will probably omit it...?

I tend to agree with it, however, did you know this yet...? The distortion of drain follower and source follower tend to cancel each other.

Hi Pedja,

Can it be configured as a headphone amplifier?

Hi,

I didn't design this stage with such a purpose in mind. Nor did I ever actually tried this, but I don't think this JFET is able to drive headphones.

Revision A of the Module LAB is coming.

Revision A addresses request I got from some you who wanted to use the stage as a buffer only but complained about use of the output capacitor. So, the resistor is added into the source of signal JFET to drop the output offset voltage. A bias current of the buffer can be trimmed for minimal output offset. The cap thus can be omitted or it can be replaced by coupling transformer.

This figure shows also another change, a cascode has been removed from the gain sage.

I like this - - an offset trimmer is a good idea. Did you remove the cascode specifically because you've added an offset trim?

This fine offset trimming is possible only if buffer only is used. The gain stage shifts the gate of the buffer to about +12V (I assume V+ = 24V when gain stage is used) so the output voltage will be also about +12V.

The reason I removed a cascode from the first sage is... Firstly, a cascode indeed improves linearity, and this was confirmed by measurements. Subjectively it wasn’t that clear improvement though, but my guess was that the way it affects the tonal balance may work well in some systems. However, after some time both my neighbor and me found that we prefer this stage without cascode, and findings were relatively consistent in both of our systems, and varying some other parameters and elements. (But who knows, maybe once I'll figure out how to make a cascode to work better at this place...) So, for now, instead of putting this cascode to the board and leaving the users choice to use it or not, I removed this part and found more useful to use the space to put the footprint for another JFET in parallel with the first JFET (needed when the stage is used as an I/V).

That's curious because I arrived to the same conclusions a couple of years ago. I didn't like the sound even though I tried hard to convince myself that cascoding was needed. It did sound better without cascoding.