Aleph3 dumping factor and power

[JacekPlacek]

Hello !
I have two questions concerning Aleph3.
J have made own copy where:

1) Power supply is limited aproximately to +- 20.5 V.
2) Topology is similar to mini-Aleph so two IRF244 on the top of schematic diagram ( current source ) has been replaced by one IRFP044N and two IRF244 at the bottom has been replaced by
one IRFP044N. Current DC bias is the same as in original Aleph3 (currents sum of two IRF244 is about 2A) and I set to 2A.
Resistors 0.47Ohm in mosfets I changed to 0.33Ohm.
Current sense output resistors I limited to two 0.47Ohm resistors what is two times less than in original Aleph


As necesarry in concept of Mr Pass I set work conditions so that AC current flowing via output sense resistors was two times greater than current flowing via 0.33Ohm resistor in IRFP044N current source.


For me it sounds OK but
what bothers me is that some electrical parameters are not achievable in my circuit in comparison to Aleph3 service manual. Of course from some reasons they will be a little bit worse but
they appear much worse in my case.
It is said in service manual that dumping factor should be > 100. I made simple measuraments utilizing signal generator, osciloscope and multimeter.
First I connect 8.2Ohm resistor on output and set output amplitude to about 6V ( 12Vp-p). Next I add to the resistor next resistor to get total resistance about 3.4Ohm ( it is minimum value of impedance that my speakers can achieve ). Output amplitude decreased to 5.6V ( 11.2V p-p). The result is almost the same in total frequency range from very low frequencies to 20KHz. One can simply calculate that output resistance ( impedance ) is about 0,4Ohm what gives dumping factor 20 to nominal 8Ohm speaker impedance. It is at least 5 times worse then in Aleph3 service manual but it shouldnt. Almost nothing changed ( and as a matter of fact it shouldn't have changed) when two output current sense resistors 0.47Ohm I changed to four 0,47Ohm resistors like in original Aleph3.

Next matter is available power. I have impression that AC regulated Alaph current source does not work enough good in my
circuit. All seems to be OK when I try to get from amplifier no more than DC bias ( about 2A ). Osciloscope observations are very good in case of this measurements because one can see when signal is distorted or limited. Even with minimal load impedance ( 3,4Ohm ) signal is pure provided load current is no more than about 3A in peak which gives output voltage about 10Vp( 20p-p ). It gives about 14W output power without significant distortions which is few times less than written in
service manual for load 4Ohm and 2Ohm. If I try to increase output signal above 10V in peak( still for 3.4 Ohm load resistance) then positive part of sinus signal is very limited and distorted ( could not achieve more than 12V - positive part of sinus looks like sqare)
while negative part of sinus goes very low without significant distortions. Concluding , I have very big power limitations in positive parts of signal. I realize that I utilize during normal listening no more than about 2W of power where signal is pure then but I do not like if something does not work as it should.


Thanks in advance for any suggestions.

Best regards
Jacek

 

[jan.didden]

Hi Jacek,

I am no expert on this type of amp, but it seems to me that if you decrease the current sense resistor, you also decrease the error gain into the "active current source". This CS should automatically decrease its current during an output swing to increase the load current. It seems that happens not strongly enough. Maybe you should look into that.

Jan Didden

 

[jarek]

Yes, your current sources seem to be not active enough

 

[Nelson Pass]

I take you at your word that the active current source is
supplying 50% of the current, but it is important that you
confirm this to be the case, particularly as you seem limmited
to 3 amp output peaks, when you should get about 4 amps.

Also, the Aleph 3 operates at +/- 25 volts.

Referring to the damping factor test, I measure the voltage with
an 8 ohm load and then remove the load. The increase in output
voltage into the original gives me the approximate damping
factor. (the higher the damping factor, the more accurate this
method.) If you don't get quite 100, I wouldn't worry about it,
but you should do a lot better than 20.

 

[JacekPlacek]

Hi Jacek,

I am no expert on this type of amp, but it seems to me that if you decrease the current sense resistor, you also decrease the error gain into the "active current source". This CS should automatically decrease its current during an output swing to increase the load current. It seems that happens not strongly enough. Maybe you should look into that.

Jan Didden

Yes, your current sources seem to be not active enough

I can not agree with Your diagnosis.
Decreasing sense resistors do not decrease "error gain" because if one change sense resistors one has to change other resistor - this which leads signal to base of npn transistor to get Aleph active current source condition ( R114 on original Aleph diagram ) - AC current flowing via sense resistor is two times greater than AC current flowing via Mosfet creating this active current source. In my case decreasing sense resistor from 0.235Ohm ( two parallel 0.47Ohm ) to 0.117Ohm ( four parallel 0.47Ohm ) caused that I had to change R114 from 2K to 1K. Of course these values are different than in original circuit as a result I utilized 1 mosfet of other type than in original Aleph not two mosfets. Anyway even though sense resistor decreased
two times but R114 decreased too so sensitivity of current source
should not change too much.
My measurements showed that described wrong behaviour of my Aleph3 did not depend on whether sense resitor had 0.235Ohm or 0.117Ohm

Regards
Jacek

 

[JacekPlacek]

I take you at your word that the active current source is
supplying 50% of the current, but it is important that you
confirm this to be the case, particularly as you seem limmited
to 3 amp output peaks, when you should get about 4 amps.

Also, the Aleph 3 operates at +/- 25 volts.

Referring to the damping factor test, I measure the voltage with
an 8 ohm load and then remove the load. The increase in output
voltage into the original gives me the approximate damping
factor. (the higher the damping factor, the more accurate this
method.) If you don't get quite 100, I wouldn't worry about it,
but you should do a lot better than 20.

I realize that Your way of dumping factor measure "does not disturb" amplifier so much as described by me. I'm sure it will improve my previously measured dumping factor so I will do tests today and let You know how dumping factor changed.
Having opportunity to talk with Aleph creator I would like to ask You about gain of Aleph in open loop and closed loop. I did some tests and for 8Ohm load , total gain seems to be about 150. I accounted it
as follows:
1) I set output AC voltage to 6Vp ( 12Vp-p)
2) Gain from input to output was 10.14

Trying to establish gain of particular stages to account open loop gain ( we can not open it of course ) I did as follows:

3) I measured AC Ugs on output mosfet and AC output voltage on 8Ohm load resistor. Gain was about 32 so it gived me gm of output mosfet about 4S. It is like in theoretical characteristics of IRFP044N for current = 2A.

4) I measured AC voltage difference between gates of input differential pair. I measured AC voltage on drain resistor R108 which is output of differential pair. Gain measured that way was only 5 which is rather strange. Few days ago when I havn't assembled output mosfets yet I simply connected R104 ( resistor which closes loop )
to ground and measured gain of differential pair. It was 10, two times more than in closed loop. IRF company does not show characteristcs of IRF9610 for such small currents ( in Aleph3 it is 10mA which flows via mosfets in differential pair) so I could not compare my measurements to theoretical values.
Could You tell me what are mentioned values measured by You in
original Aleph3 ?

I accounted theoretically what should be gain of two stages of Aleph3 to get gain 10.14 in closed loop and it gaved me about 130 which seems to be near my experimental values. What do You think about that ?

Regards
Jacek

 

[JacekPlacek]

I take you at your word that the active current source is
supplying 50% of the current, but it is important that you
confirm this to be the case, particularly as you seem limmited
to 3 amp output peaks, when you should get about 4 amps.

Also, the Aleph 3 operates at +/- 25 volts.

Referring to the damping factor test, I measure the voltage with
an 8 ohm load and then remove the load. The increase in output
voltage into the original gives me the approximate damping
factor. (the higher the damping factor, the more accurate this
method.) If you don't get quite 100, I wouldn't worry about it,
but you should do a lot better than 20.

BINGO !

On Friday I decided to do some tests. I replaced R114 which leads sense signal from output to npn transistor by potentiometer. I started from "fixed" condidion that AC current flowing via sesnse output resistors should be 2 times greater than AC current of Aleph active source current. Of course at the begining there was so bad as I described before. Still decreasing potentiometer value
I was looking at osciloscope and multimiter. Long time nothing was changing but sudenly output voltage stared to rise improving dumping factor. For certain value I got situation that no matter whether output of my Aleph was loaded, loaded by 8Ohm or loaded by 3.4 Ohm , output voltage almost stayed the same.
I was able to get even negative impedance - loading output caused that output voltage was increasing not decreasing.
It was what I needed - big ( about 100 ) dumping factor. Also I was able to get enough output power for 3.4Ohm what was imposible before. For output sense resistors equal 0.117 Ohm ( 4 0.47 Ohm parallel resistors) I got that R114 should be about 600Ohm. I measured relation between AC sense current and AC Aleph active source current. It was of course not 2:1, it was 1.2 : 1. For output sense resistor equal 0.235Ohm ( 2 parallel 0.47 resistors ) I got that R114 should be about 1K and previous relation should be about 1.04 : 1.
Well, most important is that my Aleph3 works as it should

Regards
Jacek

 

[Nelson Pass]

The the ratio of output current divided by the current provided
by the positive current source is 1.2 to 1, then the gain is
set too high. Yes, it will deliver more current into lower
impedances, but it will also experience shutoff in the current
source on negative swing, giving rise to distortion.

I recommend that 2:1 figure - the AC current delivered by the
positive source should only be about 1/2 the total AC current
going to the load.

As to the open loop gain figures of 130 to 150 is quite close to the
40+ dB I typically see.

 

[JacekPlacek]

The the ratio of output current divided by the current provided
by the positive current source is 1.2 to 1, then the gain is
set too high. Yes, it will deliver more current into lower
impedances, but it will also experience shutoff in the current
source on negative swing, giving rise to distortion.

I recommend that 2:1 figure - the AC current delivered by the
positive source should only be about 1/2 the total AC current
going to the load.

As to the open loop gain figures of 130 to 150 is quite close to the
40+ dB I typically see.

Well, if I return to 2:1 my dumping factor will be poor ( measured Your method was below 40, measured by changing load from
8Ohm to 3.4Ohm was about 20 ) and of course available power
especially for low impedances will be poor too ( positive parts of sinus signal are very poor) as I described before.
My situation seems to be hopeless

I have one question. Did You measure how big is influence of h21 parameter of npn transistor that cooperate with Aleph active current source ? As a matter of fact I utilized BC 550C instead of MPSA18. This transistor creates voltage amplifier. Based on MPSA18 it seems to have a little greater gain because it has hFE a little more than BC550C. As a matter of fact we sholud compare h21 parameters but I in frequency range we consider it shouldn't have big meaning. On the other hand in other solutions ZTX450 is utilzed which has less gain than BC550C so it rather can not be reason of my problems.

Regards
Jacek