Aleph5....it works!!!

Hello All,

finally finished my Aleph 5 . Well almost, but at least it makes some music and not even bad at all.
Tested the first channel yesterday evening and got about 11A of bias wich was a bit to much for the fuse I used (good thing those fuses). After looking for a while at my printed circuits I found I had made a little mistake while designing the printed circuit mirrored for the left and right channel
Well now the connections to the plus and minus supply don´t look so nice for one channel and I´ll have to make a new board.
The rest works fine, I get 32.5volts before the chokes from a 25v transformer with a 2.8A bias current. After the choke it´s 31.5 volts with a 10mV ripple. DC-Offset is 48mV left and 23mV right.
Temperature is a bit high with 70° C measured directly next to the fets but this will be allright when I´ve corrected the bias to about 2A (don´t really know why it´s so high, any suggestions?)
Output is about 30V at 8 ohms with symmetrical clipping, humm is only noticable with my ear next to the speaker.

I´ve only heard it on my small "upstairs" system with small dynaudio speakers but it sounds like I hoped it would. Very natural but detailed without any harshness, good bass (on this speaker). I´ll try it on my downstairs system when I´ve reset the bias and finished the casing (wouldn´t like my 3 year old daughter to test how well I´ve isolated the mains....)

Here´s a picture if the right channel,
 

Attachments

  • rechter_kanaal_2.jpg
    rechter_kanaal_2.jpg
    39.8 KB · Views: 6,929
Thank you Pete,

well actually it´s dutch engineering:cool:

I changed the bias today by changing R19 from 221k to 47k. Bias went down from 2.8A to 2.1 A I´ll think I put a trimmer in this position to get the two channels the same.

One question I still have:

Cause I want to adjust the current source gain to the right value (50%) I have to adjust R21. I understand how to measure this so that´s not the problem.
When I look at the patent and at the values used in Aleph4 and 5 the current source gain seems to be only 25%. Is this so or do I make some mistake (using the formula in the patent)

william
 
Looks very good indeed!

How big (small) is it? It seems like it's much smaller than other aleph 5's

Am I right if I say that I see two cooling fans?
Could they be the cause of the humm you discribed?
Furthermore, could you tell us a little something about the fans? Which did you use, do you hear them running etc. I'm rather curious because you dared trying something that's almost forbidden in hifi.

If so, maybe the humm can be eliminated with some filtering of the powerlines of these fans.

Andre
 
Hello Dre,

it measures 435mm width, 400mm deep and 180mm high so it's not that small.

The fans are Papt 8412NGL wich have 12dBA at 12volts. You can hear them running when everything is very quiet and you get nearer than 1meter. Since I haven't listened very much yet I can't say that you never hear them but I think it'll work fine. It should get even quieter when the case for the amp is ready and the sides and part of the top are closed.

Actually the hum is in the normal range for power amps and I expect it to get even less when closing the case and connecting it to ground.

william
 
Aleph 5... it works!!!

William,

Glad to see your Aleph 5 is up and going. It look great.

Tell me, how did you arrange for 2.8 amps bias with 3 pairs of output Fets?

I can only get 1.875 amps using 1.00 ohm source resistors with IRF 240 power fets.

If found the bias trim only useful to reduce the bias, not increase, as increasing R19 above 220K had little effect.

Did you use lower value source resistors or difference power fets?

regards

macka
 
macka,

that's a good question but I really don't know:rolleyes:

The only thing that's not standard is the BC550 instead of a ztx450.

My understanding of the circuit is as follows:

If you take away R19 the bias is set by the ca. 0,66v Vbe of the BC550 + the voltage drop over R15 x current through BC550. In my case this was about 0,9volts. So about 0,2mA current through the BC550. Hfe for this would be about 60 or so (calculated it more precisely but the notes are at home and I'm not)
Using a transistor with even less gain would result in an even higher bias current.

Adding R19 ads part of the current through the BC550 so the voltage drop over R15 gets less or even negative.

I don't think the power fets have a lot of influence here.

william
 
AC current gain setting

I finally managed to set the ac current gain this afternoon and found some interesting things.:)

First I removed R21 and measured ac voltage over the negative source resistors with an 7,27 voltage@1kHz over an 8 Ohm load.

This was about 288mV so I put in a trimmer and set the voltage to 144mV.

I measured the power output into 4 and 8 ohms directly afterwards and noticed that the amp clipped very early on the positive halve into 4 Ohms.

Power was 50W/8Ohms and 37W/4Ohms.:(

Conclusion was that the current source wasn´t doing enough so I halved the AC voltage from 144 to 77 mV and measured power again:

50W/8 and 91W/4Ohm wich is spot on. clipping was also symmetrical on 4 and 8 ohm loads.

R21 was 330ohm for the left and right channel.

So somehow 25% seems to be the right value for setting the voltage over the negative source resistors.

william
 
how it sounds

Hello all,

after a few weeks of listening and waiting for my Aleph to burn in (or out)I must say that I´m not totally satisfied with the sound.
Compared to my other amp there´s a bit of space/ambient information missing, bass doesn´t go as deep and it is a bit (note a bit!) forward in the upper frequency range.

What it does very good is to display textures and fine details, pace and rythm in the music.

Since I´m quite shure that the potential of this amp is higher than this I would like to do some mods and see if things get better:

1. change the 3 caps (220microF) from panasonic FC to Black Gate (any experience anyone?)

2. rewire the power supply with normal copper wire (not silver plated as it is)and bypassing it with some smaller caps

3. experiment with the setting of the current source. I´ve now chosen a setting so the amp will make 90watts at 4 ohm (like in the specs) but this means that r21 is about 330 ohms instaed of 453 and the current source does 75% instead of 50% when working in an 8 ohm load at 10watts.
With 50% I only got 37watts into 4 ohms!

4. change the input to the volksamp60 configuration and go a bit easy on the preamp (although the other power amp also only has 12k or so)

Does anybody have some more ideas?

william
 
Aleph 5 .... it works

Willam,

I played around with my Aleph 5 quite a bit before I got peach's and cream.

I agree with you proposed plans.

The problem may relate to wiring of the power supply and the current share set up

Suggest you scap the idea of 90 watts into 4 ohms unless you have too have it.

Optimising the current share for 4 ohms will be at the detriment of quality into 8 ohms according Mr Pass and my tests comfirm this.

Make sure all wiring to the boards is as short as possible from the power supply and if you are using seperate power and mains boards run seperate wires to each from your filter caps.

Ideally the 392R resistor for Q1 should be close to the output according to Mr Pass. I ran a seperate cable to the filter caps for my Mark Finnis boards, this seemed to boost the highs quite a bit.

Also, try local bypassing the fron end with 100uf black Gates or similar.

Hope this helps

regards

macka
 
Hi Macka,

I´ll try the current gain setting first and see what happens. If the power into 4 ohms is not enough I fear I´ll have to up the bias current. And go for 70° heatsink temp :eek:

Don´t understand how the 90W@4ohms spec was reached though.

I already have seperate wires for + and - going to the main board and these are decoupled with 470microF panasonic FC. The board is 70mu copper and R392 is connected with a very short track to the - supply
I´ll change the 4mm^2 silvered wires for + - and 0 for some 10mm^2 ofc copper ones (I know these are not too thin but I happen to have them)

Boosting the highs doesn´t seem a good idea at the moment, there´s enough of them already;)

william
 
ac current gain

so I had a go at it last night.

I set the current gain to 50% wich gave a value of 523 ohms for R21 on both channels.
Power into 8 ohms stayed the same, into 4 it´s now 37 watts down from 90watts@75% current gain.
I had a look at the patent again and now I finally understand how the power of an Aleph can be calculated:)

I haven´t listened to it yet but I would like to know if somebody can give me an explanation why the sound gets worse if the current gain is higher than 50%.

Would this mean that the sound get's better with a current gain lower than 50% (with the final result a normal se amp)?

william
 
Hmm.

Mr Pass wa very patient with my same question... it is difficult to work out with formula because you need to know the transconductance of the fets and add this the the value of the source resistors .. all of which are in parrellel (see the patent).


It is easier to measure and trim but you must derive the current from the voltage measured across the current sense resistors and the source resistors (all in parrellel) and then calculate the ratio.

From what I recall when I did this the values in the circuit for the Aleph 5 were spot on. If you dc bias or supply rails vary appreciably from the stated values you may need to check this.

I am not sure why the sound varies with differing setting of current gain but 50/50% sounds best to me.

best regards

macka
 
Hmm.

Mr Pass was very patient with my same question... it is difficult to work out with formula because you need to know the transconductance of the fets and add this the the value of the source resistors .. all of which are in parrellel (see the patent).


It is easier to measure and trim but you must derive the current from the voltage measured across the current sense resistors and the source resistors (all in parrellel) and then calculate the ratio.

From what I recall when I did this the values in the circuit for the Aleph 5 were spot on. If you dc bias or supply rails vary appreciably from the stated values you may need to check this.

I am not sure why the sound varies with differing setting of current gain but 50/50% sounds best to me.

best regards

macka