[L0rdGwyn]

It's a beautiful looking amp in this picture of the completed amp. Are the feet still missing?

I don't like feet on my personal amps, but I will add them for future builds. I might even add them to this one.

 

[Xcalibur255]

Talk of switches are one of the few subjects that will automatically trigger a de-lurk from me.

https://www.amazon.com/Twidec-Rocke...s=lighted+toggle+switch&qid=1628793422&sr=8-4

Wanted something like this on one of my amps so badly, but it's actually quite hard to find quality ones in the correct ratings for some reason. This link is basically for the photo.

 

[Xcalibur255]

BTW, congrats on the build completion Keenan. /re-lurk

 

[UntilThen]

I don't like feet on my personal amps, but I will add them for future builds. I might even add them to this one.

Oh why don't you like feets on your own amps?

My amps are arriving, probably today.

 

[L0rdGwyn]

Oh why don't you like feets on your own amps?

My amps are arriving, probably today.

How exciting! I like the flexibility to use different types of isolation.

 

[L0rdGwyn]

I'm so very close to being able to test my DAC! The stage is set, just waiting for the go-ahead for some custom toroidal transformers, hopefully will be able to finalize the order today or tomorrow...

Sowter I/V transformers arrived today.



4-layer PCB layouts and routing are done. Here is the digital supply board - I've included some additional length / mounting holes on the route as the USB module will be on standoffs above it, this will keep the connections between the two PCBs as short as possible.



And here is the DAC board. As I've said before, two of these will be stacked as it is dual mono.



Got some Korean and Japanese made PCM56P-K to try out. DIP-16 socket I am using lends itself to being able to remove the chips non-destructively, so can choose the pair with the best measured performance.



Over time, I have amassed a collection of spare Antek toroidal transformers from various projects - turns out I have the one I need for my tube output stage already, noice.

Sort of unusual decision for me, but I am thinking about keeping the tubes inside the chassis as opposed to top-mounting. Assuming all performs well, this will be a 4.25" thick chassis, don't really want to add additional height on top for the tubes for space-saving reasons.

Anyway, just need to get these custom mains / isolation transformers ordered and I will pull the trigger on the PCBs and parts.

 

[UntilThen]

Sowter I/V transformers arrived today.

My eyes lit up at the mention of Sowter. Huge respect for them now with what I'm hearing from Odyssey. What's an I/V transformer?

 

[L0rdGwyn]

My eyes lit up at the mention of Sowter. Huge respect for them now with what I'm hearing from Odyssey. What's an I/V transformer?

It's a bit technical but here is what it is!

Many of these old R2R type DAC chips are "current output" types, meaning rather than outputting a voltage that can be amplified directly (by a following preamp, for example), the output is a small AC current. This current needs to be converted to a voltage signal that can be amplified. Many of these chips have built-in circuitry to perform this function (often times opamps directly on the chip), but better performance can be achieved using a dedicated current-to-voltage (I/V) stage.

There are lots of ways to do this, the two major categories are active and passive.

Active I/V stages typically use opamps, transistors, or even tubes to perform the I/V conversion and simultaneously amplify the signal.

Passive I/V stages often use resistors to perform the I/V conversion, then the signal is amplified to line level by a following amplifier stage. The less common passive approach is to use a I/V transformer, less common due to the need for specialty transformers and expense.

So, these are I/V step-up transformers. They convert the AC current output of the PCM56 DAC to a voltage and simultaneously provide some voltage gain across the transformer. The signal on the secondary of the transformer will be amplified by an additional tube gain stage to reach 2Vrms output with a suitably low output impedance.

Transformers have other advantages compared to resistor I/V - they have a built-in low-pass filter due to the transformer bandwidth which will remove high-frequency quantization noise from the output signal (if using a resistor, an additional low-pass filter must be added in circuit). Another advantage is galvanic isolation between the DAC PCB ground and the secondary ground - keeping the grounds separated will reduce injection of noise from DAC circuitry into the output stage ground.

Low noise is a top priority, so we'll see how we did when I throw this thing together.

 

[triod750]

I almost understood this explanation, or I might have done. I wonder if I will remember it? It transforms simple current to wonderful music with the help of a tube. Simple as that.

 

[Galapac]

Got some Korean and Japanese made PCM56P-K to try out.

I have heard of those…they are the DAC chip itself, right?

 

[L0rdGwyn]

I have heard of those…they are the DAC chip itself, right?

Yup, those are the DAC chips, an oldie but goodie, highly regarded for tonality among the old R2R chips, good measured performance as well.

There is an 18-bit version, the PCM61, which is a drop-in replacement for the PCM56, although testimonials point to the PCM56 being the better performer subjectively.

 

[Tjj226 Angel]

It's a bit technical but here is what it is!

Many of these old R2R type DAC chips are "current output" types, meaning rather than outputting a voltage that can be amplified directly (by a following preamp, for example), the output is a small AC current. This current needs to be converted to a voltage signal that can be amplified. Many of these chips have built-in circuitry to perform this function (often times opamps directly on the chip), but better performance can be achieved using a dedicated current-to-voltage (I/V) stage.

There are lots of ways to do this, the two major categories are active and passive.

Active I/V stages typically use opamps, transistors, or even tubes to perform the I/V conversion and simultaneously amplify the signal.

Passive I/V stages often use resistors to perform the I/V conversion, then the signal is amplified to line level by a following amplifier stage. The less common passive approach is to use a I/V transformer, less common due to the need for specialty transformers and expense.

So, these are I/V step-up transformers. They convert the AC current output of the PCM56 DAC to a voltage and simultaneously provide some voltage gain across the transformer. The signal on the secondary of the transformer will be amplified by an additional tube gain stage to reach 2Vrms output with a suitably low output impedance.

Transformers have other advantages compared to resistor I/V - they have a built-in low-pass filter due to the transformer bandwidth which will remove high-frequency quantization noise from the output signal (if using a resistor, an additional low-pass filter must be added in circuit). Another advantage is galvanic isolation between the DAC PCB ground and the secondary ground - keeping the grounds separated will reduce injection of noise from DAC circuitry into the output stage ground.

Low noise is a top priority, so we'll see how we did when I throw this thing together.

Just to make sure everyone understands, it is important to note that the IV transformer is not actually doing any of the current to voltage conversion. The resistor is still doing all the work.

The key here is that generally speaking, the resistor we have to use is a very low value. Something around 100 ohms or less is preferable. If you use a 100 ohm resistor as your grid resistor on a tube, the input impedance of the tube and your grid leak resistor will form a voltage divider. Much of your signal will be shunted to ground and you won't generate a huge signal which is why you normally need a high gain tube.

Why do you need a low value resistor? Because most dacs want a very low input impedance into the IV stage. A high value input impedance will cause distortion in the dac among other bad things.

If you use a step up transformer, you can use a much higher value resistor on the secondary. The transformer will "reflect" the resistors impedance back to the primary by the square of the turns ratio. In other words it will make a big resistor look like a little resistor to the dac.

The drawback is that a high quality step up transformer is very difficult to make and thus it is very expensive to buy a high quality one. They will also add their own distortion and other artifacts to the sound.

One of the drawbacks can be viewed as a benefit though. Since the transformer can never be perfect, it will filter out some high frequency noise the dac might generate. I believe this is part of the reason dacs built like the one lordgwyn is building sounds so analog.

 

[triod750]

So there might be a lot of trial and error behind a good sounding device. And then you put it in a chain of other devices where it isn't comfortable and then you 'know' that it is bad...

 

[L0rdGwyn]

I went ahead and put some feet on the OTL amp.



Here with Bendix 6080WB and Fivre 6C5G.



I am going to send this amp out on its mini tour this week to get some feedback. Good thing too because it is getting pretty freakin' cluttered around here! I have no space for anything, too many amps and parts...

Once this is out the door, can move forward with the other projects. These custom toroidal guys - who came highly recommended - are taking a very long time to confirm my order, so I am going to order parts for my DAC, I am dying to build it, will use off-the-shelf transformers if I don't hear back soon. My protoboard is vacant again for another design, will start working on the next amplifier too, it's going to be a pretty interesting one, stay tuned

 

[L0rdGwyn]

All is ordered for DAC - decided to move forward with off-the-shelf transformers for now, not terribly expensive and will allow me to prototype. If the custom transformers work out, then super. Custom route is preferred to minimize capacitance, and thus mains-coupled noise, but the off-the-shelf option should work well.

I will use a medical-grade toroidal with primary-to-secondary electrostatic shield and hum band as my balanced isolation transformer. Mains transformers will be itty bitty Talema 7XXXX series. All are spec'd just so - very little headroom, again to minimize transformer capacitance.

I made some PCB's for the Talema transformers, they are tiny, 1.7" x 1.7" area.



Here are the final DAC PCBs. White nearly doubled the price, so went with my usual black. Four-layer PCBs are pricey, but it is the best way.





All power supply electrolytics will be long-life low ESR Rubycon.

I have some nice 9-pin tag boards that I had planned to swap into my protoboard when needed. I think I will keep the tubes inside the DAC chassis, I have extra tag boards so will likely use a pair on the interior, like so.



Now the wait for all the parts to arrive, probably a week or so. In uncharted territory here with digital, but I think my design is sound, we'll see how it goes!

Got some quality listening time in this morning with the OTL while working on all of this stuff.

Visseaux 6J5G and Tung-Sol 5998.



I've been on a bit of a techno kick, really like this compilation from Oliver Lieb.