One Aleph-X working, One to go

[Netlist]

Uli: Thanks...I'm not a big technical book reader

William:
I usually don't believe very much in burn in of components but in the case of my Aleph-X the offset voltage settled after a few days, say a week or so.
Don't know why or exactly what stabilized but the end result was better. I mean, in the first hours/days I had to adjust a lot and couldn't get the thing really stable. After a period these problems disappeared.

/Hugo

[uli]

Hi William,

did you load the outputs to gnd?

What you encounter is a too negative temperature coeffizient
in the ccs of the diff pair. Proper loading for the outputs is IMHO
33R or less each.

(or you implement a dc servo)

Uli

[wuffwaff]

Hi Uli,

yes I used around 100 ohms to ground (4 x 390 ohms): the thing stays stable after warmup. maybe I should use something else as a zener for the current source to get abetter temperature behaviour?

William

[Steve Fay]

William they sure look great, nice job. they just need alittle tweeking. Good luck.

Steve

[wuffwaff]

Hi,

I´m finally back home again and had quite some time to think about some changes I would like to make to change hum, and absolute DC behaviour.
One change I made just now was to make the feedback paths (wires) th same length and twist them lightly until they reach the circuit board. This helped a bit so now hum is only there with my ear against the woofer. From 10cm you can´t hear it anymore. If this is still too much for my other speakers I´m afraid I´ll have to shield my torobars and toriod.

On the subject of absolute dc offset it seems to me that it must be possible to make it a bit more temperature independant. Now it starts at 8V and goes to almost 0V within an hour. Leaving the top off or blowing some air onto the circuit board sees the offset go up again.
What I would like to try is to bias the zener a bit higher (from 3mA to 6mA) or maybe use a different kind of voltage reference.
Next up would be a lowering of the resistance from output to ground (now 4x390Ohms parallel to 5x or 6x) but I don´t really like it cause it will cost some output power.
If somebody has another method I would be pleased to hear it!

William

[wuffwaff]

Hi,

I tried two things since yesterday:

Upping the current through the zener from 3 to 6mA changed abs DC offset but not the way this changes over time so this is not the way.
Changing the output resistors to 5x390Ohms (78 ohm) had more effect. The offset now changes less from cold to warm and is below 1V after 20 minutes. Lowering the output resistors further will probably see an even better behaviour but I will leave it at this value for now.

The third possibility would be to change the 4k7 resistors to 3k9 or 3k3 but if I remember well this will also reduce the gain.

William

[Netlist]

Try lowering the output R's a bit more.
I just went trough Nelson's comments on this matter and 30 ohm would be a minimum. Lower will start burning too much power.
If that doesn't help try lowering the 4.7k's. I could have a look but IIRC I used 3.9k. Worked very well, I never used the output R's.
Also, but I'm not sure anymore about this one: there is some kind of optimal bias point that stabilizes the Abs. offset. Go higher or lower and it starts drifting again. Bear with me for this last tip, it's just something from vague memory and could be wrong.

Edit: I had a quick look at the 4.7k's and it looks like I used 2k.
Indeed a bit low but it works.

/Hugo

[wuffwaff]

Hi Hugo,

Thanks. I´ll try 3k9 and see what happens. Like I said, once warm the value stays more or less the same. It´s only the big difference between cold and warm that I don´t realy like. It starts at -6V now, is at -2V after 5 minutes and stays below 1V even after a few hours of idling.

William

[wuffwaff]

HI,

tried 3k9 and 3k3 for R46/47 without a lot of difference. Absolut Offset is below 1V after 40 minutes and then very slowly reaches 0V after heating up. I think I´ll leave it there for now. Diff offset stays at 12-16mV independent of these changes.

Changed C9 and C10 to 3n3 (from 4n7). Now I´ve got an almost perfect square wave on 6 Ohms, a bit of overshoot on 8ohms and a bit rounded on 4 Ohms.

Measured the gain for different loads and frequencies:
1kHz unbalanced in, 8Ohms: 19,81dB 2Ohms:19,29dB
10kHz unbalanced in, 8Ohms: 19,70dB 2Ohms:18,81dB
20kHz unbalanced in, 8Ohms: 19,64dB 2Ohms:18,43dB
100kHz unbalanced in, 8Ohms: 18,74dB 2Ohms:15,96dB

The gain for + and minus out wasn´t the same and always around 1.3dB lower for the minus out. I´m not shure what causes this..........maybe the unbalanced input?
Changing R46/47 from 4k7 to 3k9 didn´t change the gain.

Damping faktor is around 50 with these values.

William

P.S. still haven´t tried putting in C7 and C8

[jleaman]

How do they sound ? There big..