T106-2 core vs.??? core Amplifier problem

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.

zox2003

Member
2004-11-23 3:19 pm
usa
Ok...I think that I might have very interesting issue here.

I have build two 200W /4ohm amplifier modules (using LM311+IR2010+IRF540Z) using core I found localy in store(unknown type but I think is ferrite) with success.
Amplifier carrier is on 300khZ,self-oscillating..tested on full power for hours (on speakers-3 way proffesional boxes) and worked great (great sound-bass,mid and highs were great).Carrier level was just 150mV.
I decided to get some samples of T106-2 core (which is a lot smaller then my core(4 inch diam.)) and try it.
I got it 2 days ago and finally tried it.
Results were not that good as I expected.
I used 2 x 20 gauge wire to make core (44 turns ) to get 40-50uH inductor....After I installed it I got ripple (carrier level) on output of 4Vp-p(old core was just 150mVp-p),offset voltage of 0.450Vmeasured using DMM (old core was 0.001V).
I found out that big offset voltage has something to do with amount of FB feeded back to comparator (I guess 3Vp-p vs. 150mVp-p) so I tried to reduce feedback but it did not work well.
Basically everything looks ok when you turn amplifier on but once you apply anything to input of LM311(even you do finger touch) it offsets Vout to 0.4V.Very strange....

Do you have any suggestion???

Regards....
 
You cannot just switch core (dimensions and type) and remain the same number of turns. As far as offset consurned you must be able to add some dc counter component to cancel it. When I was build iraudio2 amp opamps needed to be trimed for zero offset at out of the amplifier.

Anyway you should post your sch. also to get more accurate answer.

PS.

BTW it seems that you got 26uH not 40-50 as you thought so this is the problem.
An externally hosted image should be here but it was not working when we last tested it.
 
One leg of the comparator is seeing 1K at DC, the other is seeing 22k. The offset of the comparator (including switching time asymmetry) is being amplified by overall gain (23). This is basic amplifier design. You can find a fix in most schematics of class AB amplifiers.

You have zero offset upon power up because the circuit is not yet oscillating, it does not start until it gets some input signal. I though this was quite obvious too.

If you were getting 150mV ripple with the old inductor, the value was way too high. If you are getting 4V now, the value is way too low. I though that RF engineers knew the values of their inductors ;)
 

zox2003

Member
2004-11-23 3:19 pm
usa
Eva,

Thanks for reply....As far as value of inductor...I was doing this whole experiment over this past weekend so i did not have L meter at home...BTW...even at work if I need to measure inductance (BTW a lot smaller value then this inductor is-(range of up to 100nH max) I use Network Analyzer.RF HF design are a lot more sensitive and less tollerant on values then this amplifier inductor design.

But, thanks on all this informations regarding class D.You are very helpfull person in this field.I am wondering where do you live(secret location?).If you live at place I think you do then we might be originally from same location....:D

As far as DC offset----with other-original core I did not have a offset issue (with same part values).

So what is your recommendation to fix that problem?
What is "normal" ripple value?

Regards
 
44 turns on T106-2 should give 26uH which is a suitable inductance value. Wire gauge is not critical as long as it can handle speaker current without overheating. With +/-40V rails, 680nF output capacitor, and 300Khz idle operation this should produce a bit over 1Vp-p carrier residual. If it doesn't then there is something wrong with your circuit.

I bet the offset has been always there but at first you just didn't notice. You can try comparing one channel with the old inductor against other with the new (if you are enough scientific minded).
 

zox2003

Member
2004-11-23 3:19 pm
usa
Eva I have 470nF on output currently...I will just add some turns to reach same effect like with 680nF

Just to remind you that I loaded these modules to full power (with old core) and they worked beatifully.So I do not see what would be 'wrong' with circuit.....Maybe....
 

zox2003

Member
2004-11-23 3:19 pm
usa
OK...

Amplifier is fixed...I used Eva's advice and inserted 100uF capacitor in series with 1K that goes from pin 2 to GND (LM311) + I added variable resistor (5K) to offset adj. pins on LM311 (with this I was able to do precise adjustment).So far so good...I am attaching updated schematic here:

Thanks,Eva
 

Attachments

  • ir2010 rev2 new.pdf
    62.9 KB · Views: 170
zox2003;
I'm really glad that you've your amplifier working right, and that you've learned a new circuit trick.

But I do have a question for you, I also need to learn some circuit tricks myself: why do you have the 5 CD4069 inverters in series? Do you want to create a delay?
 
Inductors in class d are a minefield for the uninitiated.
I used power inductors to start with and they got very hot around 120 degrees.

Someone suggested a t106-2 core, I found this with 2 yards of SWG18 worked fine with my class d amplifier. It doesnt even get warm. I also found a little extra inductance didnt load the amplifier quite as much.

You are finding different levels of carrier on the output due to different inductances.
The inductance is basically a voltage divider with the output capacitor.
 
Someone suggested a t106-2 core, I found this with 2 yards of SWG18 worked fine with my class d amplifier. It doesnt even get warm. I also found a little extra inductance didnt load the amplifier quite as much.

You have to take into account that more inductance + less capacitance = higher output impedance, which means higher filter overshoot and higher distortion at low ohm loads.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.