Improving the B1…

If you’ve been following the blog for some years, you’ll likely know that I am a big fan of the Pass B1 buffer. I’ve built several versions and my own board layout has been one of the most popular designs shown here. It’s also almost 15 years old now (!) – oh how time flies!

The reasons for my fascination with the B1 are many: It’s a simple and elegant design; it sounds very good; and the lack of extra gain is very often a big plus because many modern signal sources have a high enough output level as-is. However, the single-supply original version of the B1 requires bulky/expensive/poor-sounding capacitors in the signal path, and so a DC-coupled version with a dual-rail supply that doesn’t require capacitor coupling seems an obvious next step.

There have been a few of those (for example the “DCB1”) but they’ve been more complex and so in my view don’t have the same appeal as the original. However, it turns out Nelson Pass also did a ”B1 R2” version with a complementary JFET pair (SK170/SJ74). I think I saw this some time ago, but I was a bit shocked to learn that it is actually much older. Since it’s possibly even more simple and elegant than the original B1, it was the perfect project to restart doing PCB layouts for the winter season (as described here).

The result is one of the cleanest PCBs I can remember doing, and fortunately it also works as expected. The version shown here has an onboard volume pot to make it a complete preamp, but there is also a version without the pot that should be even more versatile (and usable e.g. as a class D input buffer).

I don’t need this right now, but as it is a very versatile design it’s a good one to have in the arsenal for future use. Project files coming soon, probably just after X-mas 🙂

Dead on Arrival?

Isn’t that what “DOA” normally means? Fortunately in this case nothing has died (yet!) and “DOA” instead stands for “Discrete OpAmp” 🙂

The background is that I have noticed that quite a few clones of various Bryston amps have started cropping up over the last year or two – not least the excellent BP26 clone I have built myself. On one of my PCB-searches I was looking into these clones and found a small dedicated headphone amp that looked interesting. However, looking at the PCB layout I honestly wasn’t particularly impressed and a small “I can do this better” thought started forming at the back of my mind…

Some more digging revealed that Bryston is one of the (few) companies that publish schematics of their older models which would explain the surge of clones – these schematics are obviously a great resource. I started looking through the schematics and found that a lot of these products are build around the same “DOA33” discrete opamp building block. That’s obviously interesting since Bryston’s commercial products tend to be well-regarded and that gave me the idea to do a version myself, but with the modular “opamp” concept intact. As a replacement for an integrated opamp this is probably a stupid idea, but as a way of building compact and modular discrete amplifiers it’s not so bad I guess.

I started doing some layout work and eventually ended up with an all through hole design of around 45 mm squared that I was quite happy with. As a “mini-evolution” of that design there I also did an SMD version (with SMD passives) which I still need to test. The idea being that it can be smaller (I could have done more here), that SMD passives can actually be had better/more accuarate (and for less money) and that it would be possible to get the SMD pasives mounted professionally and relatively cheaply by one of the PCB companies that also offer assembly service. Then it would be possible to have a drawer full of “building blocks” and then simply mount the leaded parts (mainly matched transistors) when needed. I didn’t get to that just yet, but I might explore the idea again later.

To accompany the DOAs, a series of “baseboards” would be required to provide space for the external components required. The baseboard shown here (and the only one I have done so far) is the most simple circuit to build a basic preamp, i.e. similar to a CMoy-style amp with an integrated opamp, and is probably most accurately described as a “test bed”. to confirm that the opamp circuit works. The baseboard just includes the various gain setting resistors, some PSU caps and an output capacitor to get rid of any DC on the output.

What I am looking into next is one of the more “advanced” circuit variations that Bryston describe, namely a headamp (with a DC-servo to get rid of the impractically large output cap for headphone impedances), a RIAA (with two modules per channel) and also a specific preamp with the volume pot isolated between two gain/buffer blocks (similar to Bryston’s own designs) etc.

I may end up more or less “full circle” back to the BP-26 design I originally came from, but then at least I will have had some fun along the way 😀

• 

• 

• 

• 

• 

Project files: The HF Line Stage…

Like the NE headamp this is another ancient amplifier design by a (now-defunct) Danish magazine – “High Fidelity” – that I wanted to try out. I’ve therefore redone the original PCB layout and since it is a bit of legacy design for a fairly narrow audience, I am publishing the files directly in case anyone else wants to give it a go 🙂

Is it an exceptional design? – probably not. However, while I haven’t tested this in a big setup I have to say that I was pretty impressed with the clarity and level of detail on offer when I did a quick test on my desktop, so it’s very likely that I’ll try putting the prototype boards in a case at some point.

USB-C experiments – part 2…

As I wrote about a few weeks ago the USB-C power supply standard might open up a few new possibilities for DIY designs. Having recieved some PCBs now, it turns out I was right… 🙂

Two of a kind…

If I showed off every blank PCB I’d bought over the years I’m not sure this blog would feature much else. Also, it would give away just how many projects I really plan/start up but never finish and I don’t really want to do that either 😀

However, here I’ll make an exception because these two boards were both bought from Aliexpress on the same day, they arrived on the same day (from two different vendors) and they are actually the same basic design – a John Curl-designed JC-2 discrete preamp.

Obviously the boards look far from identical, but the amplifier sections are the same. What’s different is the power supply and a few selected parts and implementation choices. On the large board, the power supply is a dual-mono discrete regulator. On the small one it is a pair of 78xx/79xx regulators shared by the two channels. On the small board there is an onboard volume pot but no output caps, on the larger board it is the other way around etc. OK, to be fair the large board has a pair of higher-biased output transistors added, but again, the input stage topology and parts choices are essentially the same so these two have a lot in common.

Apart from being interesting comparisons, both to each other and to the balanced JC80, I am always intrigued by these simple discrete circuits – what can they really bring in terms of sound quality?. The JC-2 has received quite a lot of discussion (scrutiny…) on diyaudio with even John Curl himself chipping in, and I’m sure there are ways to evolve and improve the circuit. However, that’s not really what I aim to do with a pair of ready-made boards.

Plans for the boards? Well, yes and no… The large board is probably going to be a stand-alone preamp eventually. With the discrete regulator and the space for good-quality caps etc. I think it must showcase more or less how much it is possible to squeeze out of the relatively simple JC-2 design. The small board was attractive for a different reason though – the onboard PSU arrangement means that it can be powered directly from an ICEpower module (or a similar class-D amp), so if it sounds any good, here’s an easy and compact way to build an integrated amplifier – hopefully it sounds good too 🙂

Starting a JC-80 clone…

As some regular readers will probably know, the fact that I have been doing DIY audio for a long time now means that I do have some “exotic” parts left from the times where audio-grade parts were more readily available than they are today. I suppose I could just try to sell off these parts at ridiculously inflated prices, but as the reason I originally bought them was that I wanted to build stuff I generally I try to do that instead.

During one of my (countless) online searches I found a discussion about this clone of the JC-80 preamp (original schematic here) and that piqued my interest. The design uses the well-known (but now exceedingly rare) 2SJ109/2SK389 dual-JFETs for the input stage. I have a couple of pairs of those left over that I have been thinking about using for something and this design seemed like a good candidate. Apart from being discrete and balanced – both big positives in my book – the design has one other unique “selling point”, namely that John Curl apparently has stated that it’s the best preamp circuit he’s ever designed. That meant I ended up deciding I wanted to give it a go (even if it’s maybe just “one of the best” 😀 ).

The boards I used are from jimsaudio (also on ebay). My original plan was actually to design my own board because I was not that impressed with the layout of this one from looking at the pictures. However, since it is very much a “dead-end” design with obsoleted parts I decided not to bother with an own redesign and ordered a ready-made pair of boards instead. It seems like COVID-19 hasn’t fully loosened its grip on global postage though – I ordered these boards in the beginning of December and it’s barely a week since they arrived.

When I got the boards I warmed a little – they are 2mm thick, gold-plated and with 2 oz copper (which you can clearly feel when soldering) and generally of very high quality, so at least that’s something. On the other hand they do have some idiosyncratic touches that are very common with these Chinese designs – not sure why. The small caps have a 7.5mm pin spacing for no obvious reason – 5mm would have been much more common. Likewise, the large axial caps are just supply decoupling and could easily be a smaller box-type footprint and so on. All in all these are minor issues, but the do seem to be entirely avoidable to me.

As you can see I have nearly finished putting the boards together. The power dissipation of the JC80 is around 20W per channel so heat sinks are definitely required and just mounting the boards to the bottom of a case doesn’t seem to be enough to keep a finished amp cool. Instead I found a suitably large angle profile on ebay that I can use for mounting brackets and which has also just arrived. Unfortunately the seller didn’t cut it to my specs as instructed so I have to do that myself, leaving me with a bit more mechanical work than I originally anticipated here.

Assuming the boards work as they are (and hopefully they do) that then means I can start looking at case designs and think about how to eventually test the claim that this is the best John Curl design ever 🙂

• 

• 

• 

A well-balanced preamp…

Another intriguing (and completely unnecessary…) preamp design is this one. As you can see it is pretty compact and two-in/one-out balanced with onboard XLR-connectors.

The design is by Bruno Putzeys (of Hypex/Grimm/Purifi fame) and it’s discussed and explained in this article. To be honest I’ve read the theory a few times and I’m still not sure I get it, but with Bruno’s credentials it’s not a big leap of faith to imagine that it’s something worth trying 🙂

I’ve noticed this design – usually called the BPBP (= Bruno Putzey’s Balanced Pre) – in the past, especially as there have been a few board group buys before, but for some reason I didn’t bite when I had the opportunity. However, I recently spotted a board for sale with SMD-parts populated (also a group buy offer on at least one occasion) and since this was what I wanted I grabbed it.

Obviously having the SMD parts populated meant there wasn’t much work do but sometimes it’s also nice to do something that is straightforward. I also managed to find a pre-made rear panel layout file, so even that part is expected to be quite easy 🙂

As my board came with the SMD-parts pre-populated the opamps are the default LM4562/LME49720, while the regulators are (perhaps not surprisingly) the low-noise Hypex HxR12. I bought a cheap pair of these regulators at some point so when I finally got my hands on a pre-board it was obvious to use them here.

The two on-board inputs are relay-switchable which is really great – it means that this pre is much more likely to find a place in a permanent system at some point when it’s done.

The offboard volume pot is implemented a gain control for one of the amplifier stages. The good parts of this choice are highlighted in the article – you can use a linear pot which typically has better tracking and even for the balanced amp you only need two decks, meaning that it doesn’t impact channel matching in the same way as a 4-channel pot.

The downside for me is that the volume pot is part of the feedback loop and ideally there should be no wires to it. However, the board-mounted pot is too much of a compromise for me so (once it arrives) I’ll be starting out with an Alps Blue on short wires and then possibly look into DIY’ing something else based on one of the relay-based designs available.

Once the volume pot is here I’ll try out the preamp in a more “realistic” setting than the bench, but I do expect it to sound quite good. It will also be interesting to try out if there are any specific synergies when paired with the another Hypex amp – who knows? 🙂

• 

• 

• 

More PeeCeeBees…

After a bit of waiting for the postman to show up I now have a couple of new projects to look forward to…

As the title alludes to, it’s a preamp in the “PeeCeeBee” series developed by diyaudio user “Shaan” where I’ve previously showed the small power amp design called the V4.

“What’s another discrete preamp” you might ask, but Shaan’s circuits are normally very well thought-out and solidly designed, so that alone makes the design interesting I think. The second thing that makes it interesting is that the “PCB-pre” allows adjustment of the feedback between the preamp stages via a potentiometer that can be front-panel mounted. This means that it is possible to adjust the amount of harmonic distortion the preamp adds which directly affects the sound signature.

Considering how much emphasis I usually place on system synergy I think this will be a good opportunity to put some of those thoughts to the test, because if the system matching doesn’t matter then I should prefer a specific setting regardless of the partnering electronics, right? 🙂

Anyway, even with the best will in the world I will certainly not be the first one to finish building the pre because I am short some parts that I need to order. Also I can see that Neal over at enjon.uk received his boards so far ahead of mine that I can’t really beat him to making a finished amp 😀

Also included in my shipment were a couple of PCBs for the V4H rev. 2 power amplifier. The V4H is the “bigger brother” of the original V4 with two pairs of output devices per side, but otherwise identical in design and topology. I did actually buy the original V4H PCBs and I was starting to working on them around the time when the preamp group buy was announced. However, when I discovered that the design was being updated to v2 I decided to wait a little and get a pair of those boards instead.

Like the original V4, the V4H is built around bipolar transistors for input and VAS stages, and then Exicon/Hitachi lateral MOS-FETs in the output. This is normally something of a “winning recipe” in amp design and I am looking forward to hearing what a “normal” class AB design can deliver compared to the class D offerings from ICEpower and Hypex that I have been getting used to over the last months…

• 

• 

Moving with the times?

I regularly complain about the declining availability of DIY-friendly parts – as do quite a few others if you look around. An obvious case in point is opamps, where (more or less) you can say that on one hand high-quality parts have never been better or cheaper, but on the other hand these parts have never been available in less DIY-friendly packages. Even the classic SO-8 SMD footprint is starting to look big and many newer devices are only available in smaller packages, some even “leadless” and/or “powerpad” packages that are very hard for the average diy’er to solder.

However, the many available opamps are actually obvious candidates for some experiments, especially with basic preamp/CMoy-style circuits where the sound quality and sound signature of the opamp really shines though. Some years ago I made a PCB design specifically for that and a few months ago I decided to revisit this layout and update it specifically for use with SMD (SO-8) opamps.

Of course you could use the old design with DIP-adapters, but fast opamps and DIP adapters are usually very bad for stability, so it was way better to create a new version of the layout intended for SMD packages instead. This naturally means you lose the ability to “roll” opamps (i.e. experiment with replacing them), but to most people I think that is an acceptable trade-off.

I haven’t really kept up with the development on the opamp scene the last few years, but what I would look at myself are the various OPA16xx duals. Here you can get both bipolar and FET inputs, “standard” high-quality opamps as well as “audio-specific” devices and more – and the prices are a small fraction of what you paid for top-of-the-range models such as the OPA627 and OPA2107 ten or fifteen years ago. AD also does a few interesting variants by to be honest here I am even less aware of what is available than for TI – I am sure diyaudio.com has answers if I ever need them.

The prototype here is made with an OPA1656 dual opamp and the initial impression is that I was on to something with this idea, because it does sound very good. I also have both the OPA1642 (FET-input) and the OPA1612 (bipolar input) in stock and I think these would be interesting candidates as well. Another interesting thing is that the power consumption is lower – in fact so low that without the LED my lab supply is showing a consumption on 0.00A for both rails (it’s not very sensitive). That also means that battery-powered versions would definitely be viable.

The download file for the new version is added to the original project file post because that’s way easier, so go and grab it from there if you want it 🙂

Harmonic experiments…

Not much to say about this one really – I finished casing my build of the Korg Nutube B1 preamp (as shown originally here). By using a built-in power supply based on a filtered IRM power module I managed to keep the unit compact and still avoid the external PSU which I quite like.

I’ve only done very quick listening tests with the amp, but considering how varied opinions I have read on the sound quality I am positively surprised. The post title is obviously in reference to the fact that it is possible to “tune” the profile of harmonic distortion that the amp produces which I haven’t really tried, but I actually would like to try it and get a sense of where my personal preferences lie.

I am also positively surprised that the amp does not seem to have a big turn-on/turn-off thump (as I have seen some people mention). Not sure why (maybe it’s the PSU?), but it definitely makes it more likely that I will actually use it in my system where “livability” is a major factor (even if it’s not an actual word…). Maybe I should do a “second generation” of my original B1-based office system?