advertisement


Headline clone - advise needed

C

cubeasic

pfm Member
Hallo,

I found the schematics to the naim headline headphone amp in a polish forum. Unfortunately I don´t speak polish, so I´m bringing this over here. Also, it seems that almost anything from naim has been given the "pinkfish treatment" here, except their headphone amp.

http://img854.imageshack.us/img854/8464/21877f0c70.jpg

Guessing from the polish forum posts, I think it should be the naim headline 1, not the actual version 2 of the amp. The headline 2 is supposed to be more stable under low impedance loads.

But, after comparing component values (mostly resistors) and routing of the schematic to photos of the headline 2 and found them to be nearly identical. Also the sketched power-supply section (which I did not compare in detail) seems to fit more to the headline 2 than the headline 1. The obvious difference between version 1 and 2 are the black hats on the "input stage"-transistors, if I´m not mistaken, to keep them in thermal equilibrium.

Version 1:
http://img266.imageshack.us/img266/6791/5c38ed96f6.jpg

Version 2:
http://img545.imageshack.us/img545/8961/1000x500pxllb117364enai.jpg

Side by side:
http://img692.imageshack.us/img692/1576/nahaoliveblk2.jpg

Anyway, I am in the middle of making a self-etch-pcb to the schematic and have some questions:

(1) I have 32 ohm headphones, so I would like it to be compatible, does anyone have some comments on the headline 2 or 1 low impedance load issue? I didn´t find any difference in the schematic and pictures of the HL2, the thermal coupling of the two transistors per channel seems doable, and the power supply is also not too complex.

(2) Speaking of power-supply: the headline is powered by a single regulated 18-24V feed. Each channel has split local regulators (about 12-13V) for input and output stage. The output stage is supplied by a simple LM317 circuit, I guess to give low output impedance. The input stage is fed by additional regulation after the LM317. I can only speculate that this is also some variation of a gyrator, but I´m not sure. It seems more complex than a teddyreg, so I ask myself at which level this regulator type might be? Apparently it has a higher output impedance and is better suited at the input stage.
I think here can be modded the most: single transformer windings for each stage, regulators of choice. But which ones?

(3) Components: the transistors ZTX384 and 214 can apparently be substituted by BC550 and BC560 and the output transistors ZTX652 and 752 by the 653 and 753 (I found this in a thread to NAP250 transistors). Is this correct?
Is the small ceramic cap across base and collector 5.6pF like in the 321-cards? (It is "39pf?" in the schematic). Other caps are tantalum, polystyrene (470pF and 220 pF) and the big output elkos (1000uF). I guess the cap across the "central transistor" is 180 uF? (It has no units in the schematic). It seems to be on of the tantalums.

All in all, I am confident that I could knock something together, but some of the details are still not clear to me. Any help is appreciated.

Thank you
Florian

EDIT: here are some higher resolution pics and schematic:
https://www.dropbox.com/sh/w25y4ztaof6exf9/VvrZf_qm_A
 
The headline is fundamentally the standard Naim amp stage scaled-down to lower voltage rails ...so:

1) No need to change anything for 32ohm headphones; it will deliver bass down to about -3dB at 5Hz into such loading with the 1000uF output cap shown. HF input filter is about -3dB @ 72Khz as drawn, a prudent measure.

2) Given the small total current draw and reasonable voltages, yes a regulated raw supply rail is easy and recommended. Yes, the input stage uses a gyrator to knock-down supply noise further. This simple approach works well. You can leave out the components circled with a question mark on the linked schematic - they relate to the relay not shown, not required for a copy)

Simple upgrade: change that 8K2 value in the front-end current source for 2 x 10K, and tie the mid-point of these two resistors to 0v with 10uF or more (cap type uncritical)

3) The transistor substitutions you suggest look reasonable; an alternative output pair to try would be standard BD139 / BD140. Yes, 39pF would be a good start for the Miller cap, something like 56pF would virtually guarantee stability if in doubt; 5.6pF nowhere near enough in this context. You need some test gear to be able to test such substitutions properly (input HF square wave, watch for overshoot on a scope)

The cap across the bias multiplier will be surprisingly uncritical and you cannot make anything unstable by using too much here across the bases of the output pair. 0.18uF, 1.8uF, 180uF all will work (I expect that ropey schematic infers 180nF ; I'd try 1uF or more, film type for preference but not important)
 
Thank you martin, for that great reply. That covered almost everything.

If you allow some follow up questions:

Why use 2 x 10k on the current source mod, and not 8.2k, divided on two resistors (similar to the starfish)? I guess the actual value is also not really critical here?

And more importantly, I forgot that in the first post: How do I set the 1k trimpot at the bias multiplier? Should I check for some offset-nulling or bias setting somewhere?

For the Miller cap: I think - once finished - I could check square waves at work.
 
2 x 10K is a convenient value; you can use 2 x 4Kish, but you might want to use twice the added capacitance... This tail bias current value really is not to sensitive though - providing it is quiet.

I'm not party to the Naim bias setting, but if you do not have the means to test the distortion output at a fixed load, I would start by setting for 26mV measured across either of the 1ohm emitter resistors. It will be close enough.
 
I sketched a PCB for a 100x160 mm Eurocard for the main circuit without regulators:

http://img528.imageshack.us/img528/1845/headlinelayout.png

and with components:

http://img210.imageshack.us/img210/5672/headlinestuffing.png

I tried to separate power- and signal-ground.
One thing I´m not sure about is, whether the 1k output resistor to ground should go to signal ground or power ground. Depending on the headphone impedance, a significant amount of current (relative to the headphone return current) will flow through the 1k resistor...
At the moment, I plan for signal ground. Anyone has a opinion on this?

EDIT: I just talked to a friend, and he recommends the 1k output to go to power-ground. Apparently the amount of current is significant. So I will change this.
 
I powered the amp for the first time yesterday (from a lab supply), and: yeah, it worked!

Still, I had some trouble at first: one faulty BC560, and a solder bridge between pcb-lines...

One thing puzzles me though: I get an obvious hum when the pot is not turned in the top 80%. When I set the volume quieter, the hum increases with decreasing volume.
Usually, the input filter on the audio components I know is with RC filter with fixed values of R and C. But the headline apparently uses the pot to make the R-connection to ground, see here: http://img854.imageshack.us/img854/8464/21877f0c70.jpg
So the R to ground seems variable to me, what could be the cause of the humming.

However, the setup is quite experimental at the moment, so this can be another cause of the humming.

Pictures:
http://img18.imageshack.us/img18/9073/imag0093cet.jpg
http://img708.imageshack.us/img708/6963/imag0095fa.jpg
 
An interesting read! Given that the headline uses the standard naim amplification stage, is it correct that this is not Class A, please (which seems to be preferable and well viable for headphone amps)?
Many thanks,
Stefan
 
Revamping this old thread. Could we have the schematics of the Naim Headline headphones amplifier? Your images in the imageshack do not exist.
 
The Naim pre-amp stage is class A. The power-amps (like NAP 250 etc) are class B dual rail, totally different to this.
 
Thread resurrection shuffle...

As this seems to come up again and again.... virtually everything in hi fi amplification is class A other than power amps.
 
Just finished a headline clone:

oxqDly4.jpg


It sounds wonderful with my Audeze LCD2s. I have the PSU on board (but the PSU area can be cut away) and tried parallel secondaries with a 4 diode bridge vs. the traditional Naim way of using only 2 diodes (see here for details) and can report that the 4 diodes version sounds untidy in comparison. I added a snubber, a CRC and a choke to the PSU and used the 2x BD911 configuration of the later MK2 model.

Here is the hand drawn schematic mentioned earlier in this thread. I also found 2 other reverse engineered schematics online, see here and here. They all have some minor errors, so I carefully checked photos of the original PCB and corrected them. Also I added the mod mentioned in #2.

cubeasic said:
One thing I´m not sure about is, whether the 1k output resistor to ground should go to signal ground or power ground.
Click to expand...

The 1k is located on the other side of the connector: only if nothing is plugged in the 1k connects the output to GND.
 
Last edited:
Fantastic work. How Naim should have built it in the first place, with a built in power supply instead of the stupid NAPSC hanging off it. If you add an on/off switch I'll nominate you for the new head of NAIM.
 
:D:D:D:cool::cool::cool:
I planned to use a mains input socket with fuse and switch, or to fit the power switch on the front panel...
 
Here a short update on this project. 3 versions were built:

slJ1iuJ.jpg


Modified PSU section: a Mundorf as the second cap in the CRC sounded much better (had to wire the 2 existing caps in parallel for this and relocate the R of the CRC). As signal resistors I tried RC55Y, MRS25 and RN55/60/65, the latter sounded by far the best (sorry Arkless Electronics for that). The middle version was the best mix of size, cost and sound quality:

aUKbfB0.jpg


The left version was fitted into an old PSX case:

GhcwZ4y.jpg


I re used the power switch, LED, AC socket, and the 4pin DIN will be used as the audio input.

QRfqdXE.jpg


BTW some versions of the PSX came with a wonderful H+F transformer:

Z4uOzM1.jpg


500VA, 40VDC, perfect to power the output section of a power amp project :cool:
 
jpk said:
Just finished a headline clone:

oxqDly4.jpg


It sounds wonderful with my Audeze LCD2s. I have the PSU on board (but the PSU area can be cut away) and tried parallel secondaries with a 4 diode bridge vs. the traditional Naim way of using only 2 diodes (see here for details) and can report that the 4 diodes version sounds untidy in comparison. I added a snubber, a CRC and a choke to the PSU and used the 2x BD911 configuration of the later MK2 model.

Here is the hand drawn schematic mentioned earlier in this thread. I also found 2 other reverse engineered schematics online, see here and here. They all have some minor errors, so I carefully checked photos of the original PCB and corrected them. Also I added the mod mentioned in #2.



The 1k is located on the other side of the connector: only if nothing is plugged in the 1k connects the output to GND.
Click to expand...

Good Day
Dear it’s Excellent work!!!! You are best!!!

please see my Head Amps based on Naim components and schematics
https://flic.kr/p/2nh5mnu

And video of headamp works
https://www.instagram.com/tv/CXqY9GpALwM/?igshid=YmMyMTA2M2Y=
 
You must log in or register to reply here.


advertisement


Back
Top