Downscaling Zen v.9 for use as headamp

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Lately, I completed the Zen v.9 (without fb) and I was amazed by its sound, when played through a pair of Fostex FE167 in Voigt pipes.

I was thinking of using the same amplifier topology to drive my Sennheiser HD600 headphones. I have already seen the "Zen Headphone Amplifier" project in Headwize, but I thought of using the LU1014 power JFET as an output device, since I have a couple of them. Power requirement for the headphones is minimal (a couple of hundreds mWs), so what is needed is downscaling the zv9 circuit, with lower voltage and current and, probably, lower power MOSFETs.

Since my knowledge doesn't extend so far, could I get a little help from someone more expertized? I know there are a few here.

Thank you.
Regards,
Evangelos
 
Hi Hannes,

Thank you for your reply.

Actually, I think this is not a current (trasnconductance) amplifier with its low output Z of just 4 ohms. Zen v.9 (wired as a current amp without feedback) has an output Z of 80 ohms and that's what make the Fostexes sing (at least I think so and so Mr Pass says).

Any other thoughts on rearranging zv9 for lower voltage/current?

Regards,
Evangelos
 
Hi,

sorry I see you mean something really different :)

There's not much you can do then; reduce rails to +-15V maybe.

By the way, if an output impedance of 80 Ohms makes your Fostex sing, then maybe that does not happen with your headphones as they usually have impedances like 100 Ohm (speakers 4, 8 or 16 Ohms) - so that would be entirely different.

However, no need to stop if you like it ;)

Have fun, Hannes
 
hello.
if i remember correctly the hd 600 has an impedance of 300 ohm.
i think for listening to loud music you need an output voltage of 15 to 20v pkpk or so (clean,undistorted)..........
perhaps you can reduce the rails as far as possible for that output.
reducing the bias current is not so easily done because i think this will give you increasing distortion..............i would try it a little bit.
the output cap (10000uf) can decrease to 470uf or 1000uf.......
greetings........
 
mjf said:
hello.
if i remember correctly the hd 600 has an impedance of 300 ohm.
i think for listening to loud music you need an output voltage of 15 to 20v pkpk or so (clean,undistorted)..........
perhaps you can reduce the rails as far as possible for that output.
reducing the bias current is not so easily done because i think this will give you increasing distortion..............i would try it a little bit.
the output cap (10000uf) can decrease to 470uf or 1000uf.......
greetings........


Indeed, HD600 has an impedance of 300 ohms.

I will try lowering supply voltage and see whether is worthy.

My other question is whether I can raise the output impedance of the amp. Maybe by increasing the value of the output to ground resistor (R7 - 100 ohms in the original schematic)?

Regards,
Evangelos
 
Kenev
Dont reduce R7 it is only used for pulling the other side of the cap to be at ground level. Instead you should increase it to something like 2,2k-4,7k.
The 100 ohms was probably chosen to be adequately quick pull-down effect, and not loading the output too much, when driving an 8 ohm impedance. When you go for 300 ohms, the 100 ohms are way too small and would steal most of the output wattage and just turn it into heat.
So reduce output cap by times 10-30 (1.000uF is fine) and increase R7 by the same factor.

It is 'comme il fait' to have 100-120 ohms output impedance on headphone amps, primarely for short circuit protection of the amp, when inserting/pulling out the Jack.
You should basically just insert a 100/0.5W ohm resistor i series with the output. With the very linear impedance of Sennheiser HD600, it should not degrade the sound quality too much.

However: If you are a daredevil and know that you are not gonna do live-connect/disconnect, or if you mute the input before doing so, you can actually omit adding this 100 ohm resistor to the signal path.

To reduce the Idle current increase R4/R5/R6 and R1/R2/R3. There should be 0.65 volts over R4, and between 1.1 and 1 volts over R1 - choose a standing current and use ohms law to calulate the resistor values. Tell us if you need help.

I have a Sennheiser HD600 too :D and sorry Hanes, but these do need some voltage amplification. I would say 4 times amplification of CD level output is adequate, just to have a decent range on the input volume pot. So just keep R16 and R17 on the same level.

I too think that transistors generally sounds at their best at collector or drain, but be aware that the signal is inverted. With the Sennheisers I think there is an easy way around this though: The small plugs that go into each 'can' should be able to connect in reverse ( + to - and vice versa) :D ...I didn't try this on mine so I am not 100% sure.
 
Nrik,

Thank you for your recommendations. Actually that was exactly what I was thinking about, increasing,not decreasing the value of R7, in order to increase the output impedance.

Regarding lowering idle current, I'm not sure if this will be correct, since this will alter the operating point of the JFET and, I think this will reduce the transconductance of the JFET, thus lowering its output impedance. I think that Papa has carefully chosen this operating point for this JFET. Anyway, I can try this and see what comes out.

Regards,
Evangelos
 
Kenev,
Just saw your thread.

With the DOA follower amp, it's about "the best thing out there since sliced bread" for any phones from the Grado RS-1s, K170 AKGs and the Senn 650s.

I'm currently working up a somewhat different pcb to EUVL's super compact board and it's designed to match up to a Salas Shunt Regulator (+/- 24V @ 250mA load) with the EC Design Charge Transfer/ Cmultiplier basic supply.
Not the cheapest thing, and a bit power hungry with a total consumption of about 75watts both channels (including both the supplies).
However, even without the fancy input ccts, this Taylored DOA follower must rank up there with some of the best headamps ever built.

The fancy input ccts include an input buffer with a Lightspeed vol control, a version of the variable cross feed (Meier, etc) - then it's onto a combination of the HA-1 filters and some basic tone controls that are presenting some difficulties.
 
Well, I'll post a couple.

This is a balanced (circlotron) version.

Not published because it only works with rewired headphones (4 wires). And I guess not many are prepared to mod their expensive phones for a balanced sound. Also this particular version also has a slightly lower bias (at 200mA), and hence a bit less dissipation.

The small logpot on the left is the variable CCS for the LDRs for volume control in shunt position. Series R's are 10k Caddocks.

The DAO TCS is the next best thing, in circuitry terms.

;)


Patrick
 

Attachments

  • buffer assembly s.jpg
    buffer assembly s.jpg
    89.3 KB · Views: 2,064
I see your usual very neat, high quality project finish. Very nice indeed.

Taylored DAO- sorry 'bout that - and Balanced, eh? - and with the included LDR vol control -
Perhaps this might be a good place to see this "upgrade" of the original - A finished ccts, perhaps?

I did have a look at that idea about the 3rd amp for the gnd point in Headwize - pseudo balanced - a vague memory of "M3" ? It seems to have gone now.

Haven't heard from Steen in a long time now - do you know what's happened to him?

I thought both Hannes and Scott were going ahead with this one, too. Ah well...

My build was a pretty rough lash-up version with the perf board and it's still out there getting some "field testing", apparently!!
It did come back to be fitted with Salas' Shunt Reg and this is a bit of a shocker with 160VA pwr trannie of 60v CT at 1.2A for both channels, regged down to +/- 24v @250mA load (h.amp) - so, the h.amp sinks about 12w/ch and Supply another 24w/ch! Total about 75watts ( both ch).
The heatsink runs well above 50*C, but sounds better when really hot.

No pics, sorry - it's gone again - will do a proper build when the
pcbs are done.

Is there are commonly available smd substitute for the 301 - a higher voltage version of the BF862, perhaps?
 
> I see your usual very neat, high quality project finish. Very nice indeed.

This one is actually just a quick and dirty mock-up.

> Perhaps this might be a good place to see this "upgrade" of the original - A finished ccts, perhaps?

Balanced means no Gnd connections to the headphone, which means 4 wires, two for each speaker.

People who understand how the circlotron works should have no problems converting the DAO to this balanced version. People who don't should not really attempt, because of all the added complications like floating supplies, ..... etc.

> I thought both Hannes and Scott were going ahead with this one, too.

I have not heard or seen anything....

> Is there are commonly available smd substitute for the 301 - a higher voltage version of the BF862, perhaps?

For what? I did not remember specifying them.

If you can live with TO-92, then 2SK223 is quite high voltage. If not 301 SMD is easy enough to get actually (e.g. Digikey).

Any why the hell would you want to burn double the power at the PSU than the amp ?

:confused:




Patrick
 
Sorry Patrick,
... meant the 2sj103 (50v, 10ma P), not the J301.


The Salas shunt reg just sounds better with a lot of current - I tried running it at the usual 20% of the 250ma amp's load current down the shunt (50ma) and it sounds okay, but screwing up the CCS to 600ma does make a noticeable difference.

The extra voltage loss comes from the CT system with it's Cmultiplier.

None of the above is probably really necessary, as Steen found that a good power supply works very well indeed, but I wanted to check out Salas' regulator and also see how much difference that John B's CT system made in practice. Pushed the boundaries a bit, eh!

On the next one, I might just try out a pair of 24v battery pack system.

Perhaps your amp might find a home over at Headwize, etc?
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.