Fame and fortune here I come!

Well OK maybe not, but it looks like I’ve made it to audioXpress magazine and surely that counts for something? 😀

I didn’t subscribe to their newsletter (I do now! 😉 ) so I found out a bit late, but they seem to have open-sourced the Borbely Hybrid Headphone-amp design (kudos for that) and referenced my post as one of the examples of the usage of the circuit. That’s really great and very much appreciated, but next time guys let me know in advance and I’ll provide you with a better board picture, ok? 😀

Since that discovery gave me time to reread the original post, I started googling for turn-on/turn-off protection and delay circuits and maybe it’s time to see if something can be done here? That the amp had dangerous levels of DC offset at turn-on and turn-off was more or less the only real flaw with an otherwise great and very interesting design. Google did provide some inspiration that I am going to look at and see if I can come up with something good. I was tempted for a moment to use a uC, but putting software in a Borbely-design seems inappropriate in some way 😉

Advertisements

Project files: The Borbely Hybrid headamp

What is it?
This is my version of Erno Borbely’s tube/MOS-FET hybrid headphone amplifier. As discussed in a previous post I found this design a while ago and made plans to do my own board for it at some point. For a long time it was just another point on a long to-do list, but one random Saturday a couple of months ago I just somehow got started and managed to finish it shortly afterwards. There are only minor changes compared to the orginal. Apart from what is described below, it is mainly space for RN60-type resistors for all positions and use of bigger (and more easily available) heatsinks.

Note: I don’t really recommend building and using this as-is (see explanation below), but I decided to publish the board files anyway because I think it is still an interesting circuit. I am definitely not ruling out trying to make it useable at a later state with some soft of turn-on/turn-off delay circuit,. but for not that is not a priority for me.

How big are the boards?
Each mono board measures 3.2″ x 3.6″ (app. 81 x 91 mm.)

What is the status of the boards?
The board file is version 1.1. I have built version 1.0 and it works as expected (note the rather large caveat mentioned below though!!), but I had swapped the labels on Q2 and Q5. I also corrected the footprint for the two CRDs to make it easier to fit axial types which are the most commonly available.

Does it use any special/expensive/hard-to-find parts?

  • Power FETs: I decided to make the boards fit 2SJ313/2SK2013 FETs because they are mentioned as substitutes in original article and while still a bit difficult to get (watch out for fakes!) they are still easier to find than the original 2SJ79/2SK216s. This means that this layout is not useable with the original devices because they have a different pinout. You would however (most likely) be able to use IRF510/IRF9510 or IRF610/IRF9610 pairs instead. I haven’t tried this myself, but at least the pinouts match. If anyone want’s to have a go at this I’d be happy to donate one of my spare sets of boards to “the cause” – just drop me a line 🙂
  • Small-signal transistors: I replaced the originals (2SA872/2SC1775) with BC550/BC560 variants that aren’t quite as low-noise as the original types, but much easier to get. If you prefer “audiophile-approved” (near) unobtanium parts at all cost, then I believe the 2SA970/2SC2240 would work as well if they are turned 180 degrees to compensate for the different pinouts.
  • Tubes: You can pay a fortune for branded ECC86/ECC88 variants or you can use something cheaper, i.e. chinese/russian equivalents. I chose the latter and bought some 6N23P-types from ebay. These seem to work fine, but note that the 6N1P which is another oft-recommended substitute for ECC88 will not work in this application. Also: One of the tubes I tried gave me a steady-state DC offset of around 1V that I could not get rid of whereas the other channel was fine. I replaced the tube and redid the adjustment and then it was fine as well. This could be a fluke, but I think you should buy more than a single pair of tubes just to be on the safe side.

Anything else I need to know?

  • Important: As described in some of the diyaudio-posts I read about this design it has one major flaw and that is serious DC-offset at power-up and power-down. Both my boards are after adjustment within +/-10mV when they are fully warmed up and stabilised, but during power up they both swing the output to very close to one of the supply rails (meaning 20V or so) and stay for such a long time that I think it would be fatal for headphones. There are a couple of solutions to this, either an output capacitor or a delay circuit of some sort. I don’t have time to try either at the moment, but if you do so please report back.
  • The original circuit is differential input. However, my layout grounds the negative input as standard to make the inout single-ended only. I normally wouldn’t make a simplification like this, however it avoids a very long and unsightly trace through the ground plane. Since I expect that SE-input is how most people would use it anyway this was an acceptable compromise for me (but of course it might not be for you 🙂 ).
  • When you mount components, note the resistor that is supposed to be mounted on the back of the board. This is much, much easier to do if you solder it before mounting the tube socket and not afterwards…

Downloads:
Download design files here

Related information:
See the original build article for more information about the design and a great walkthrough of component values etc.

Note: Always read the “intro post” for additional important information about my designs.

Another Borbely clone – almost…

You may remember this post about a clone of one of Erno Borbely’s design. Now there’s another one! …but unlike the last one, which was green, this one is red! 😀

It is actually a combo of amp and PSU from ebay seller jimsaudio. The amp is not really a pure Borbely design, but instead a Borbely JFET frontend mated to a (Walt Jung) diamond buffer output stage – which in itself doesn’t really matter if the sound is OK. Quite like the last one though, it is full of parts that are now pretty much impossible to find. However, 20 years of “collecting” parts can actually be helpful sometimes :D.

The power supply is also a “bastard” design – a tweaked version of Borbely’s take on the Sulzer regulator I think. However, Erno has definitely left his mark here as well – even the regulator uses some very hard to find transistors – and I can’t be the only one that winces a little at putting “unobtanium” devices like 2SK170BL/2SJ74BL JFETs in a power supply? OK, it was only one pair of each and I had an odd number anyway, but still it hurts a bit…. ( 🙂 )

Anyway, if the results are worth it then I can live with it and initial impression is definitely that it is. The sound is smooth and detailed without any noise or hum, even from just a simple bench test. Now all that remains is the usual waiting period before I get around to making the rest of the mechanical stuff for the casing – shouldn’t be more than a year or two 😀

PS: While looking around the internet I found another Borbely design for a tube/FET hybrid line stage running at just +/- 24VDC. It comes in two variations, one intended for line stage duty and the other one beefed up a little to drive headphones. The differences are small though, so it should be possible to put both variations on the same PCB. I am thinking that there could be a way to redesign this with transistors that are still available – maybe not 100% the same performance, but it would be worth a go I think 🙂