DG300b Headphone Amp Build

It has been a long time coming but over a year since I purchase a board set from Neurochrome I am just about ready to start building. Most of the parts are here with only a few, relatively inexpsensive things to come.

First of all a big thank you to @tomchr (Tom Christiansen of Neurochrome), Pieter Treurniet of Tribute Transformers and @MisterRogers for all the help I have had so far. They have all been extremely helpful and I probably would not have committed to the project without them taking the time to walk me through some things. This is the first time that I have attempted anything like this and I am hoping that the knowledge and experience here will help me make good choices and complete the build successfully.

I originally planned this for use with the HD800 but last year I acquired a pair of Hifiman HE1000 original version and, unless I suddenly find the cash to upgrade to the latest SE version, that is what I will be using. Clearly a bit of a change from a 300ohm to a 34ohm driver but if anything I suspect this will match up better. It will be headphone only to start, I can always go back and add in speaker terminals later if I wish. Rest of the system is a Dell Precision laptop via USB to a Holo Spring level 2, currenlty feeding a Jotunheim via XLR to the HE1000s. I will be enclosing the Spring in plywood to match the amp and am toying with the idea of incorporating an Allo Digione Signature into the enclosure with the Spring as a better source but that is/might be another project.

TThe DG300b will be built/modified with the help of @tomchr as follows:

Use the balanced input via Jensen IPTs so XLR only. This is the standard setup.

Driver valve is 6n6p.

Power valve is 6c4c (sometime called 6s4s). This is a 2a3 (actually more like a 6a3) type tube on an octal base and 6.3V filament, otherwide electrically identical to 2a3. Very well regarded on diyaudio and much cheaper than a good 2a3. For example, if you were to rate all the available 2a3 type valves on a scale as Poor - Adequate - Fair - Good - Excellent then I get the impression that the 6c4c is a Good and costs about 1/10th the price of the Excellent stuff. Time will tell if this is true.

For volume control a Tribute AVC will be used. After the Jensen IPTs the amp is single ended so a 2-gang AVC will be wired in post IPT but before the gain stage.

To avoid the expense and give flexibility for speakers at a later date the OPTs are 4-8-16ohm. Having contacted Pieter of Tribute to purchase the AVC I got incredibly lucky and he offered to make the OPTs for me. He didn't have time t make the power transformer as well so that was made by AE Transformers in the Netherlands. I am hoping that this will be superior to the Classictone that @tomchr specifies.

Coupling caps will be the Miflex KPCU in specified 0.22microF. These are one of the few things yet to arrive, with stock due at Hifi Collective in the UK in early June. I seriously considered the Jupiter caps but they are a fair bit more expensive. Maybe I will try them in the future.

Most resistors will be Takman metal foil (no experience, just read lots of good things about them) and a few Audio Note tantalums in key positions.
​

More to come.

 

Things yet to be decided:

Upgrade any other capacitors? There is a power supply board which feeds a HV regulator (21st Century Maida) and three filament regulators (one for each output valve and one shared by the gain valves)and then the driver board. When pushed @tomchr suggested maybe upgrading C14, C15 which are located on the driver board where the power comes in. He also warned against messing with the Maida: "I’d leave the Maida regulator alone. Mounting capacitors there daddy longlegs style is likely to end in tears." So bearing in mind the designers warning, is there anything to be gained here or am I just being seduced by boutique parts? In case anyone is wondering, the part listed on the BOM for C14 and C15 is a Panasonic 10microF, digikey part number P13671-ND.

Wiring. I have been given a strong recommend for Duelund DCA16GA tinned copper multistrand wire. I will aslo need something for connecting up the input XLRs, the AVC and the power valve sockets, perhaps Duelund solid copper at 26 or 20 guage. Any thoughts?

Shielding. Basic plan is for a 6mm aluminium top plate and an aluminium back plate (thank you @MisterRogers ), possibly perf plate for the base. The plate will rest on a stack of baltic birch plywood sheets cut to the same size as the top plate and internally routered out to provide space for the boards etc. A sub-base for the various boards will be built underneath the 6mm top plate to reduce the number of holes to be drilled. Is it worth lining the inside of this plywood wooden box with perf plate or other metallic material to make a faraday cage of sorts?

Similarly, is it worth getting a copper braid to cover the hook up wiring (or using shielded microphone cable for hook up)?


​

 

Layout plan at this time. Nothing is finalised yet but I am trying to keep the power transformer as far away from the OPTs as I can, and also keep the power valves close to the OPTs. The images show the AVC in front of the driver board but if I can I will push the AVC under the driver board to keep the wiring as short as possible (the AVC will connect to the driver board close to the Jensen IPTs, the large white circles with 2 rows of 4 solder pads either side of the hole for the power valve socket), using an extension rod to connect to the front panel.

https://i.imgur.com/5x12lU2.jpg
https://i.imgur.com/p6Vy9NS.jpg
https://i.imgur.com/7MGGrKI.jpg
https://i.imgur.com/XiU0zt0.jpg
https://i.imgur.com/GryavCV.jpg
https://i.imgur.com/5Zexdq8.jpg

The XLR inputs will be directly behind the driver bard, again for shortest wiring, with IEC inlet on the other side of the back panel. So powersupply as far left as possible, signal as far right.

Undecided about the headphone 4-pin XLR, could be on top either in front of or in between the OPTs, could be on front beside volume control.

 

random thoughts, based on building a Tubelab SE running 45's with a transformer with 48 and 150 ohm output taps.

High impedance cans (HD800) are more susceptible to noise/hum then low impedance cans (he500). When I cased it, I had enough hum to be annoying on the 800's, so I'm working on fixing that now. But my wiring was less than ideal because the project was on pause for many years and I made changes when resurrected. Went from two boxes to one box. So now output jack is in front and output tx are in back.

Consider making a breadboard first. Wire and test with everything out in the open, will make it easier to find problems. In the case, its hard to troubleshoot, at least mine is.

Maybe wire the Tribute AVC's for balanced output, and not use the Jensens?

I also added a bunch of mini jacks on my top plate. Connected to ground, B+, B-, output bias. Will let me check and change bias if I swap tubes, etc.

Good luck
Be curious how you like your output tube

Randy

 

Regarding the headphone output, the OPTs are capable of being wired as 4-8-16ohm (in reality it is actually 4-9-16ohm). Currently wired for "8"ohm but can be changed for 16ohm easily. As I understand it there are various ways of doing this:

Leave the OPTs as "8"ohm. This is probably too low for the 34ohm HE1000 so put a power resistor in parallel with the headphones eg 12ohm resistor. This will give a load of around 9ohm (so a near perfect match) but will dissipate around 75% of the power in the resistor. The amp should do approx. 2 - 2.5W so the HE1000 would see something around 0.6W. I am sure that this would be OK but part of me grudges throwing away that 75% power... But it would also allow a 300ohm headphone to be used, giving a load of around 12ohm but with even more power burned off in the resistor.

Change the OPT to 16ohm. Now could probably run direct without a resistor (I am sure I have read that a load of up to double the OPT tap is perfectly fine so 34ohm is just about ok) and the HE1000 has a power handling of 6W so no problem there. No expensive power resistor, simpler build but there is a danger. I am told that it may trash the amp if it is used with no load on the OPT eg playing music and no headphones plugged in.

So, use OPT at 16ohm but still use a resistor in parallel to protect the amp. 34ohm resistor gets 50% of the amp power to the HE1000 for a near ideal load of 17ohm. A 100ohm resistor eg Mills MRA12, would get some 75% of the power to the 'phones for a load of around 25ohm. @tomchr thinks the higher load would actually result in lower distortion so this is looking attractive. Not so flexible for 300ohm headphones though, load in this case would be around 75ohm which is a lot for a 16ohm tap (or at least I think it is).
​

Is there anything that I am missing here or is it just a case of choosing one of these options? I am aware of the L-pad arrangement which has been used to allow speaker and headphones to be connected but at this time I do not intend do this.

 

Randy,

Thank you for your reply. I will have two small holes in the top plate to allow access to the biasing pots and will take the measurement points to the back panel and use DMM sockets.

The breadboard plan sounds perfect. I am sure that I will be terrified when I start testing things so the more space and access the better!

Regarding the AVC, it is 2-gang so ideal for a SE circuit. I already had the Jensen IPTs by the time I investigated the AVC and in consultation with @tomchr and Pieter of Tribute I decided to splice it into the existing circuit. In the Head-Fi build thread by Sceleratus there was a basic diagram of where this would be done. It seems fairly simple although I will confirm with Tom the best way to do it reliably.

Edited for spelling.

 

I debated LPAD vs. customer OPT's - went the LPAD route with the best resistors I could afford. I'll be interested in your experiences.

 

Your exploits have had me noodling on building my next around 45's. I picked up an extra DG300 board set that will do a standup job.

 

With the L-pad there is a resistor in parallel and also one in series. I am guessing it is the resistor in series that is critical for the best sound quality? Or did you put Path Audio resistors in both positions?

 

In both positions.

 

So Tom and Pieter know a LOT more about this circuit and transformers respectively than I do.

But any tx will degrade sound, nothing is perfect. So adding a 2nd tx if you can get away with one cannot help, IMHO, can only hurt.

For context, you're asking about a few feet of wire in your unit. A tx has likely hundreds of feet of wire, and it has to couple the signal across the wire. There is a chance the tx does only a little harm, and is inaudible, so maybe try with and without to see.

Is the L and R ground in the AVC tied together? If they are separate, then it can be used for balanced. Otherwise, you could ask Pieter if there is a way to separate them so the tx would support balanced out.

 

Randy,

There are 3 tabs per channel on the AVC: In, Gnd, Out. Two channels so 6 tabs in all so I think your plan could be made to work.

That said, I still feel that I will keep the Jensen IPTs as well as the AVC. The IPTs will give me isolation that an AVC will not. Notwithstanding the fact that I am altering the original schematic by @tomchr by inserting the volume control, I am trying to keep the alterations to a minimum. Furthermore, when I was emailing back and forth between Tom and Pieter at no point did either of them suggest using the 2-gang AVC for balanced in place of the Jensens.

But I do appreciate the input!

 

The two capacitors in the circuit where the B+ and bias voltage enter the Driver board (C14, C15) are 10microF, 450V rated. The BOM lists a Panasonic item (CAP ALUM 10UF 20% 450V RADIAL). These are board mounted. Is it a good idea to replace these with something like the Clarity Cap TC polypropylene (10microF, 800V)? These would be mounted off-board but within a few centimetres of the board position. They would be mounted on a bracket to hold them rigidly in place

The alternative would be a board mounted electrolytic such as the Audio Note Kaisei.

 

Hmmm, others with more skillz (@tomchr) would know better, but I don't think a change there would make that much of a difference; assuming the PS is up to the task.

 

Interesting. It was @tomchr who, when I asked if any components would/might benefit from upgrading, suggested these two caps and a handful of resistors. He did stress that he is not a fan of boutique parts and is more interested in the electrical parameters.

 

Yea, that's why it's best I defer; I've built a lot of amps, but for the most part I know enough to be dangerous. @tomchr - I'd be interested in your thoughts.

 

That's why I am trawling for advice, trying to reduce that "danger" level.

 

So 23 months later the boards are pretty much complete lol. Clearly I'm a fast mover. I have some time off work in a few weeks so that will be when I start testing and find out if I have trashed anything.

Amp has XLR sockets on the back panel. Pin 1 of XLR socket goes to chassis, pins 2 and 3 go to inputs on PCB which will be approx. 15-20cm from the back panel. Then it's the Jensen input transformers for Bal > SE conversion before volume control and then the amp proper. Any advantage to using something like Belden 1800F cable here (shielded twisted pair) as opposed to twisting ordinary hook up wire? I think if I were to use the Belden (or similar) cable the shield would connect to chassis, not direct to Pin 1 (AES 48, the "Pin 1 problem" or maybe I'm completely misunderstanding some of the stuff I have read).

For context, chassis top plate and back plate are aluminium, most of the rest is baltic birch plywood. Power transformer and OPTs on top,PCBs underneath.

 

Funny - I was just thinking about you the other day mate, wondering how your build was progressing. Re: wiring, I'm a believer in using interconnect wire with shielding in HV tube amps. If your layout is more spacious, it may not be as critical, but why not use it?

How about some pictures

 

I will get some pictures once I get things mounted on a bread board.

Should be fairly spacious, the cabling from the XLRs to the PCB will be around 30cm from the centre of the power transformer which I believe is the one to get as far away as possible. One of the OPTs will be directly above but I don't think that's a big issue. Luckily the back panel you gave me has the XLRs exactly where I would choose to have them.