Aleph 5, start up problem.
Hi, sorry to disturb you with this type of question but I couldn't find any answers that suits my questions in the old threads.
I have been struggeling with a start up problem some time now. At power on/off it takes about 1-2 seconds for the DC voltage at the output to be stabilized.
First I thought that the problem were caused by decoupling of the zener diode that sets the voltage to the current mirror in the input stage. Removing the decoupling took away some of the problem but not all, the start up time were about 10 seconds from the beginning.
Looking at the feedback path in the schematic in the link above, the 100k resistor that is in series with the 220uF capacitor and 10k to ground, the time constant for this path is abround 1-2 seconds, could this be the problem?
As output devices I use one pair of FB180SA10. Could that be the problem?
Since the turn on time is so slow, there is no "bump" in my speakers, the speaker cone is only slowly mowing, first back and then back to it its centerpoint.
What happens when you turn on your aleph, how long time does it take before they are stabillized? Is 1-2 seconds a problem for a speaker?
If you use a coupling cap in the feedback network of this
circuit, it will hold a charge if there is offset when the amp
has just been turned off, particularly if there has not been
a load. When you turn the amp back on, the feedback will
swing the output to remove this charge. The bigger the
cap, the bigger the charge.
You can put some resistance, say 47K in parallel with the
cap to bleed it off, or you can wait longer between turn-off
and turn-on. Also, maybe you want to have a bleeder
resistor across the power supply, say 4.7K @ 3W per rail
This presumes that the DC offset of the circuit is nominal
with the cap bypassed altogether. You should adjust the
current fed to the input diff pair to minimize this. Also, as you
note, putting a cap in parallel with the zener diode reference
for the input bias current source will allow for some thump.
Further to Mr. Pass's suggestions, it may be a problem that the power supply takes too long to settle down.
Hi, thanks for the input.
I measured the DC level before I started the amp today, 24 hours after I turned it of, and it measured 2.4V, the feedback capacitor needs to be discharged in other words.
I haven't tried your suggestions yet, want to make sure that I understand them correct first.
In the attached schematic I have added the R26=47k resistor to discharge the capacitor in the feedback path. My only concern is that the DC gain will be approximetly 2 and thereby an increased DC level at the output. You are mentioning that the bias current should be adjusted (i.e. R3 or change zener value of D1) but I don't see how the current adjustments could improve the DC level at the output (if I'm not using a potentiometer at the sources that could trim the currents individually in the diff pair, maybe that is what you mean).
What is the reason for adding bleeder resistors, except for discharging the power supply slowly?
I'm using a CLC network in the power supply but I couldn't afford so big inductors and capacitors that it should be a problem, calculations and measurements also shows that there shouldn't be a problem.
This is great stuff.
Many years ago I used a FB180SA10 MOSFET in the design of a DC motor speed controller.
A few weeks ago I stumbled apon Gamut ($$$) amplifiers
and they use a single MOSFET instead of matching up transistors.
I do have some questions concerning this amp design.
1) Is there a reason why the unbalanced Zin has been choosen to be so low ???
2) Have you tried using a matched pair of p channel JFET's for the differential pair ???
3) Is there a reason to go with a p channel MOSFET for the current source on the differental pair ?
Or would a low noise bi polar be fine for this application ???
|All times are GMT. The time now is 10:38 PM.|
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright ©1999-2014 diyAudio