Sounds of Amp5 (built, run ~20hours)

[longitude]

I just built amp5 and run it for tens of hours,
Soprano, violin and tenor are VERY good and clear^^
HOWEVER, when play the music with many instruments such as symphonies,
I find it fuzzy and not so bright.
Bass is ok in amp5.
But since many review claimed that tripath chips have good dynamic and vivid presentation.
I don't know why my amp5 is not so good at symphonies (even YAMAHA AV amp sounds better...)
Any suggestions for improvements?
Mine is standard vision with only two 10000uF bulk capacitors,
will it help to use four of them?
Or how about the input capacitors? Can Changing the input capacitors improve the amp5 to play symphonies better?
Or I should run-in my amp5 for more hours?
I am really satisfied with other type of music,
but really like to listen to the symphonies with my amp5,
PLEASE HELP~~~

[Mr T]

I have found that polypropylene input capacitors improve A LOT the mids and highs.
Paralleling the caps is important, this reduces their inductive behaviour.
Try two 2,2µF in parallel on each input.
Of course you'll have to bypass the original circuitry (a wire is OK) and you'll need to fiddle a bit to install them. I use to solder them right on the input RCA sockets and then run a twisted rigid wire to the board inputs.

You may try these, they are really excellent, easy to find and relatively cheap (250V series is fine)
http://www.vishay.com/capacitors/polypropylene/

Mr T

[Reverie]

I second Mr. T's advice.

Also, Wima MKP are worth a look, and are generally slightly more compact than the Vishay in my experience, although Vishay have more 'fame' in this use.

www.wima.com

I'm using the MKP2 in as my output caps.

Also, you might consider changing the value of your input caps, although I'd usually only advise this in extreme situations. I personally like a midway point of 2.2uF (note - that is an Amp4 value, it could a totally different affair for Amp5).

Good luck! :)

[thomaseliot]

quote:Originally posted by longitude
[br]
I don't know why my amp5 is not so good at symphonies (even YAMAHA AV amp sounds better...)


That can't be.

Poly input caps improve timber and clarity of instruments, not dynamics and complex sounds.

It looks like your problem is on low power and/or low efficiency speakers or low level source.

I think minimum requirements are:

- 2V rms source level (if you used standard components for input gain)

- 30VDC for rails (four 10000uF caps are far better than two for complex orchestral sounds)

- more than 90dB efficiency speakers.

What do you have?

[longitude]

quote:Originally posted by thomaseliot
[br]
quote:Originally posted by longitude
[br]
I don't know why my amp5 is not so good at symphonies (even YAMAHA AV amp sounds better...)

I think minimum requirements are:
- 2V rms source level (if you used standard components for input gain)
- 30VDC for rails (four 10000uF caps are far better than two for complex orchestral sounds)
- more than 90dB efficiency speakers.
What do you have?


Ha...Thank you all for replying^^
It's really good to get advices.
The followings are my equipments:

source level: Pioneer DVD-525
I know it is not a perfessional CD player...
However, accroding to some reviews such as:
http://www.pcavtech.com/play-rec/dv525/index.htm
It has pretty good audio cd playback performace.
I also installed a 50k 23-steps SMD type attenuator in my amp5 for volume control.

I have 32.7VDC for rails,
And listen to your advice, I just installed 2 more 10000uF caps on the board.

My speaks are two usher HT-1 (89dB, 4ohms)

Looks like many people change their input capacitors,
I also plan to do that.
But since the recommanded capacitors are all big...
Is it ok to install them on the other board beside and connect them to the main board using wires?

[cgguy2]

Some people mount the caps where the volume chip would go. I ended up zip tying mine to the bottom of the case next to the board (via a few drilled holes). I did this because it was easy and it will also eventually let me clip the leads very short. If I put the caps on the board (with my case layout) I'd need to take a much longer path. As it is it's a very short path from pot/switch to the board. Some people have argued that the minimal benefit you can get from better caps is killed by the noise picked up by long leads.

You're using two HT centers as your mains? That's interesting. What's the secret here?

[longitude]

There is no secret...
I have usher HT-1*5 for home theater.
Using the same speakers seems more balance to me.

So...the best solution of input capacitors is using not so big but better caps (if I can find one...WIMA MKS can fit, however it is polyester type...I can't find any Polypropylene type caps with 5mm lead separation, any recommendation?) and fit them to original holes on amp5 PCB...right?

[V-bro]

You can perhaps put an axial cap straight up and use shielded cable for the longer lead...

[longitude]

Also...I know that long leads of input capacitor will pick up the noise,
and leads should be as short as possible.
But, can someone give me a upper limit of the length that noise pick effect is negligible?
1 cm? 2cm?

In my observation, leads of the input caps of "Red Wine Audio Signiture 30" are VERY VERY long...but seems no problem in this machine...why?

I have two 3.3uF 100VDC and 3.3uF 400VDC metalized polypropylene caps and plane to install them.
What is the difference between 100VDC and 400VDC caps using as input caps of amp5?
The original shipped caps are 50VDC type.
If there is not much different in performance,
I think I would like to install small 100VDC ones.

[Reverie]

This quote from Jan on the support section might help:

'Note that unlike the input caps, the leads of the output caps must be as short as possible or they will not be efficient.'

Although it's not directly on the subject of input caps, he implies that the length of the input cap's leads isn't very important.

Also to note, if you go over to the 'Ground Breaking Success for TK2050' thread in 'Other Tripath Amps', you can see an internal picture of the Red Wine Audio Signature 30, with long capacitor leads indeed.

[V-bro]

I never noticed any xtra noise being picked up, and the 100V rating caps should be fine. Actually the voltage rating could be far lower as the caps won't have to deal with high voltages anyway. (the voltage rating becomes much more important if you use the caps in a loudspeaker filter.....)

[thomaseliot]

Hi Longitude,

better input caps are a good improvement, but would not help with your problem.

Yours is a problem of coupling of amp and speakers. Maybe 89dB (referred to drivers or to the whole speaker with its Xovers?) at 4ohm is not a good choice for AMP5.

My hint is to try first AMP5 with efficient 8 ohm speakers (to a friend of yours or to a shop): maybe you've not yet heard what AMP5 is capable of with very big orchestras.

[Reverie]

quote:Originally posted by longitude
[br]
If there is not much different in performance,
I think I would like to install small 100VDC ones.


I came across this issue when I chose my output caps. After asking the advice of an expert, his words were:

"I would recomend MKPs all the way, and the largest voltage that size will permit because they will give you a lower ESR and better ripple handling ( they are more robust )."

However, I can't quantify how significant this is.

[BddB]

Hi,

My DV-444 has a "line"-level of 200mV RMS on the analogue output.
If this is the case for 525 also, you might want to change your input stage gain.
Otherwise you'll only get approx. 1/10 of output gain.
/Cheers

[longitude]

To BddB,

THAT is a paradox to me >.<
Since I normally do not listen to the music that loud,
I choose input stage gain=1(general use) and power stage gain:Rmfb=8.2+1=9.2KOhm(my rail voltage is 32.7V),
and I even add a step attenuator for volume control.
If my DV-525 player has low analogue output,
and I change my input stage gain to 2.1,
eventually I have to attenuate more using attenuator,
then it means I input even lower signal to the amp.
I think this is not good for THD+N?

I am really confused about gain setting,
And like many people, I do not intend to use the amp for only one signal source,
so for now...I think I will stick to general input gain setting=1...

[longitude]

To thomaseliot,

I think 89db refers to whole speaker...since they print the value on the backside of the speaker...I don't know...
I will try my amp with speakers of higher efficiency, thank you!

Or maybe the problem is I use 2 parallel 150VA transformer to provide 300W input power?
But I doubt it, since I did not play it loudly and the problem happened in moderate sound level.

[BddB]

Hi again,

I still beleive the issue is lack of power, as symphonic music needs a lot of it.
As I see it, the input sensitivity to gain ratio needs adjusting, if you want to use the the amplifiers full potential.
I wouln't worry about the THD (most likely completely inadiable).

/Cheers

[longitude]

I will make a CD of sinusoidal sounds to measure the peak to peak voltage of my pioneer DV-525.

But I am still confused,
If originally the maximum input voltage after step attenuator say...0.2V peak to peak,
now I increase the input stage gain to 10,
but to listen to the the music in moderate sound level,
I have to attenuate the signal 10 times the original one to get the same volume.
So eventually the peak to peak voltage that I input to my amp does not increase at all...
Strange...

Maybe I get it all wrong...
Could someone explain it to me?

[V-bro]

I didn't build AMP5 yet, but I had to do the gain setting for AMP32 (AMP9 didn't need any)...

Are you sure you got the right values on the right places? Rinput should be lower ohms than Rfeedback if you want to up the gain (for instance 22K Rinput and 47K Rfeedback to go up...).

Also note that a passive attenuator allready decreases the gain, so you can decrease Rinput and increase Rfeedback to get the gain right...

[longitude]

I use 22K Rinput and 22K Rfeedback.
The signal gain is 22K/22K=1 (general use).

[thomaseliot]

quote:Originally posted by longitude
[br]To thomaseliot,

I think 89db refers to whole speaker...since they print the value on the backside of the speaker...I don't know...
I will try my amp with speakers of higher efficiency, thank you!

Or maybe the problem is I use 2 parallel 150VA transformer to provide 300W input power?
But I doubt it, since I did not play it loudly and the problem happened in moderate sound level.


As Benn well noted your problem is input level. Though would not be wise to increase ten times input gain: the total amp gain would be 180, that is too much (normally it is 18). So you would use alternative 47K resistor in the kit to get 2 as input gain and 38 of total gain: this is still not optimal, but better than before.

The fact is that 2mV is too low. You can do three things: add a preamp, get a higher output CD player, add an external DIY DAC with spdif. The last choice would lead to best results.

[longitude]

Thanks^^

But still...I don't understand.
While I increase the input stage gain,
To listen to the music at the same sound level,
I have to attenuate more using volume control pot (stepped attenuator),
Doesn't it means the peak to peak voltage after input stage is still the same?
"x2 using 47K/22K resistors" but "/2 at volume control attenuator"

[thomaseliot]

quote:Originally posted by longitude
[br]Thanks^^

But still...I don't understand.
While I increase the input stage gain,
To listen to the music at the same sound level,
I have to attenuate more using volume control pot (stepped attenuator),
Doesn't it means the peak to peak voltage after input stage is still the same?
"x2 using 47K/22K resistors" but "/2 at volume control attenuator"


Sound pressure level out of speakers is not linear but logarithmic. With 1W you get a 89dB 1KHZ sound at 1 meter; with double power, 2W, you don't get 178dB, but only 3dB more: 92dB one meter away from speakers.

So using 47K to double the gain you get just a bit more dB than the noise of a fly (that is conventionally 1dB). With a 88K feedback resistor you get another 3dB more, but with a distorted 72 (4*18) total gain.

[longitude]

mm...sounds reasonable,
Although I still feel a little strange...
I think I need some time to digest these knowledge and modify my amp5,
Thanks, Thomas!

[V-bro]

There are pictures in the gallery:
http://www.41hz.com/forums/photo_gallery.asp

Maybe some people can give better advise after having had a look...

Maybe there is some interference problem also causing distortions? I've read a lot of times this is quite a critical issue with class D amps...Have you tried to encase the transformers, in- or maybe outside the amp casing? Or shield the amp? I'm no expert on this, but maybe someone can elaborate?

[andrewbee]

Hello Longitude,
Is this your amp5 http://www.41hz.com/forums/photo_gallery.asp?action=displayimage&photo=303&loc=/gallery/longitude/DSC09457r.jpg

If so, the wires with the clear insulation which are coiled which appear to me to be the speaker wires are too long. They should be only as long as necessary and they should also be twisted together (each channels 2 wires). They will act like antennae. Also, you may also want to shorten the black wires to the left which appear to be the input wires and finally the AC wires should be twisted together (the ones going to the transformer and from the transformer to the pcb.

All the wiring should be as short as possible, again, the speaker wires are the most important.

I have bridged my two AMP 5 boards and am very impressed so far with them.

[longitude]

Thanks!

The speaker wires now change to 4 channel Mogami speaker cable which wires are already twisted inside the PVC jacket and I will upload the new photo.
There are 4 AC wires, 1 to AC1, 1 to AC2 and 2 to PGND (0-24, 0-24),
How to twist them correctly?

[Fredrik]

Get a preamp, remove the attenator.
I recommend you to parallell inputs and output on 2 x amp5.
One to each speaker. Jan told me this.
This works excellent with my speakers. (4ohm).

http://www.41hz.com/Forums/photo_album_view.asp?cname=Main+Album&mid=714&cid=221

http://www.41hz.com/Forums/topic.asp?TOPIC_ID=832

[V-bro]

Funny, I just connected AMP9 in parallel and am very (!!) pleased with the sonic results! Here's some more on the scheme: http://41hz.com/forums/topic.asp?TOPIC_ID=970#4274

[longitude]

mm...I would like to try paralleling if budget is no problem...haha...but not now~
And I think I should have a better pair of speakers to match them,
I doubt I can hear any difference using my setup now~
Still...thanks to point out a bright future!

(Any good quality and not so expensive "DIY" preamp recommend? Which matches the class T amp better?Tube Preamp or solid-state ones?)

[Lulu]

I have ordered the "T-preamp (http://www.tnt-audio.com/ampli/t-preamp_e.html)" to offer it, but the "project" was canceled. I don't have opened it or tested it yet.

[The worst foe lies within the self]

[longitude]

But I think this kind of preamp (T-preamp) simply act as switch + volume pot.
It does not add any sound quality to the signals...

[Lulu]

Tube preamp could be your solution, indeed, as it as sound signature.

[The worst foe lies within the self]