THD distortion HELP!!!

[nicholas1113]

Hi:
Can anyone help me by giving me advise on how to perform an actual 0.006% THD of the circuit Blameless cause I am stuck on the circuit. I found out that my circuit THD　is 0.63 and it is still high enough. I hope someone can give me some advise on how to reduce THD?
Is there any other issue caused the circuit THD to be so high?

Below is the diagram of Blamesless circuit and THD Spice Log and FFT graph Diagram.

 

[wahab]

first try is to remove collector of TR12 from ground, and connect it to collector of TR4....
a more radical step is to remove TR12 itself and only use one trans in the vas...

 

[wahab]

this amp doesn t exactly hold its name...
it work well if you remove TR12 and R30, and then connect the
base of TR4 directly to collector of TR10....
C3 must be kept as it is from collector of TR10 to collector
of TR4...
beside, it will eliminate the strange behaviours we can see
in the spectral analysis...

 

[Steve Dunlap]

Try setting the value of the bias pot, PR1, to 211 before you do any other changes. When I run a sim on this circuit using models for the transistors he shows (except outputs, I didn't have those) I get THD of 0.004%.

 

[jcx]

actually 1st try is setting up the sim and verifying you can meassure low thd by looking at the input sinewave distortion

look for my (jcx) posts containing "plotwinsize" or "fft" keywords for LtSpice settings for high resolution distoriton measurements

 

[nicholas1113]

thanks WAHAB for the advise .. i try to cut off the TR12 . but it still the same of THD 0.6 something . I wondering y like this.

 

[wahab]

it seems that dunlap got it right, it must be crossover distorsion...
try what he says and set the power transistors idle current to 5O mA..

 

[nicholas1113]

diyAudio Member

Steve Dunlap's Avatar

Join Date: Sep 2008
Location: Here

Try setting the value of the bias pot, PR1, to 211 before you do any other changes. When I run a sim on this circuit using models for the transistors he shows (except outputs, I didn't have those) I get THD of 0.004%.

I've try to tune until 211 and this also seem to be the same. do u mean ur models all the same ? I using the
BD 139 npn for the medium power trans Vceo=80V Icrating =1.5
QMJE15030 npn for the power transistor Vceo=150 Icrating=10
QMJE15031 pnp for the power transistor Vceo=150 Icrating=10
QMJL4281A npn for the power transistor Vceo=350 Icrating=15
QMJL4302A pnp for the power transistor Vceo=350 Icrating=15

 

[tomchr]

Check the bias currents by running a DC sim. If everything checks out to reasonable values, then run a transient sim with a sine wave input. If that looks good, then I'd probably run an AC sim of the open-loop gain (AVOL). If AVOL looks decent, then start looking at THD.

It is of course possible that you've already done all of that and now you're stuck on the THD. But I would find it odd if you didn't notice any oddities along the way.

I remember running THD sims on Self's Trimodal amp (blameless with class A/AB switching) back in the mid/late 1990'ies using PSpice and getting THD figures on the order of 0.001 %. Better than I could measure in the lab.

What does the THD look like at 1 kHz? How about 100 Hz?
Are any of the internal stages slewing in your 10 kHz sim?

~Tom

 

[rdf]

look for my (jcx) posts containing "plotwinsize" or "fft" keywords for LtSpice settings for high resolution distoriton measurements

And 'timestep'. It looks like an LTSpice artifact.

 

[nicholas1113]

Hi Tom

Thanks for your advice .
Ya definitely it got changes when i tune it until 1Khz. the result of Total Harmonic Distortion is 0.014849%
And if it tune to 100Hz the result of Total Harmonic Distortion is 0.092889%
But in the Dougles Self Book there is mentioned that in 10 Hz it will produce 0.0006% to 0.0011%

 

[nicholas1113]

Thanks Tom

Yes i try to tune the Input to 1khz and the result of Total Harmonic Distortion: 0.014849%
while i try to tune it to 100Hz and its Total Harmonic Distortion: 0.092889%

There Dougles Self book mentioned that the 10hz should be lesser than 0.0006% to 0.0011%. and i dunno y my THD is totally so high. Does it affected by the model of LTspice ?And i using

BD 139 npn for the medium power trans Vceo=80V Icrating =1.5
QMJE15030 npn for the power transistor Vceo=150 Icrating=10
QMJE15031 pnp for the power transistor Vceo=150 Icrating=10
QMJL4281A npn for the power transistor Vceo=350 Icrating=15
QMJL4302A pnp for the power transistor Vceo=350 Icrating=15

 

[nicholas1113]

Try setting the value of the bias pot, PR1, to 211 before you do any other changes. When I run a sim on this circuit using models for the transistors he shows (except outputs, I didn't have those) I get THD of 0.004%.

Hey Steve,
I able to do the THD until it is 0.004% . I using the medium power transistor of BD179 Vceo=80 I crating is about 3A.
And it someting I attached below

 

[Steve Dunlap]

Hi nicholas1113,

Glad to have helped.

 

[rdf]

nicholas1113, you might like Blackman-Harris filtering on your FFT.

 

[mightydub]

see the diyAudio LTspice wiki for advice on turning off compressrion, setting the timestep, and THD measurement

Common Issues Encountered By New Users - diyAudio

It is rough and needs to be clarified but if you're this far along you should be able to understand.

 

[nicholas1113]

Hi rdf:
nicholas1113, you might like Blackman-Harris filtering on your FFT.
__________________
"My loony bun is fine Benny Lava!"

I am not sure wat mean how to fliter the FFT . Can teach me ?

 

[rdf]

No problem. Blackman-Harris 'windowing' instead of filtering would have been clearer on my part. Do this:

 

[nicholas1113]

Oh i see it .. Thank you anyway. But for the Vbias Multipler , i using the BD 179 transistor to function it but I get an advise that the BD 139 is more easily to mount it is it ?

 

[nicholas1113]

Hey rdf.
I have done by following what u have told me and the result is simply like tat.. there are even more noise compare to previous 1 .