Modérateurs: Modération Forum DIY, Modération Forum Installations, Le Bureau de l’Association HCFR • Utilisateurs parcourant ce forum: Aucun utilisateur enregistré et 27 invités

Réalisation : ampli à base de Tripath

Message » 11 Fév 2008 18:18

Bonjour à tous,

J’ai saisi l’occasion d’acquérir à petit prix un Kit d’une société Italienne qui commercialise des amplis numériques en kit monté à base de puce Tripath TAA4100A.

Image

Pour ce kit, il ne reste qu’à mettre un transformateur torique (160 VA 2x18 Volt) et toute la connectique (RCA, borniers HP et interrupteur). Il faut bien sur un boîtier pour y loger tout ce petit monde.

On peut soit s’en servir comme ampli intégré en branchant les RAC avant le potentiomètre (ALPS) ou soit comme ampli de puissance en branchant après le potentiomètre.

La puissance est de 50 watts par canal sous 8 Ohm et le double sous 4 Ohm mais la puissance subjective est importante.

J’ai fait les premiers testes d’écoute sans mettre dans le boîtiers (évidemment DIY) et le résultat sur des Eltax Monitor MKIII est très intéressant. Ce qui m’a surpris c’est le niveau de basses (de qualité) qu’il arrive à extraire des Eltax qui ne descendent pourtant pas très bas.

Le médium est très informatif et chaleureux à la fois et les aigus ne sont pas aussi lissés que sur l’Acoustic Reality (module IcePower de chez B&O) que j’ai testé.

La différence avec ce que j’ai déjà testé en numérique (Acoustic Reality EAR 102, T-amp et Flying Mole CA-S3) est assez grande car on a le coté charpenté (densité des timbres) sans être sombre, la belle scène sonore, la maîtrise des basses, le niveau de détails et l’aération sans le coté lissé du registre aigu et sans la perte de matière au niveau du médium. C’est sûrement la typologie du schéma avant la puce Tripath et les superbes selfs à air maison qui donne ce son proche de celui du tube.

Image

De plus la dynamique est, comme avec la plus part des ampli numériques, au rendez-vous.

Je ferai un CR quand j’aurais fini l’intégration dans le boîtier que j’ai dessiné. Je suis parti avec l’hypothèse de faire l’ampli le plus compact possible.

Le prix de revient est d’environs 350 €. Pour ce prix, j’ai un bon ampli numérique mais sans le coté froid et sans aucune dureté dans le registre médium/aigu et une bonne réserve de puissance.

Voici les plans.

Image

Image

Image

Je complète par des photos. La première photo correspond à la première écoute avec le boîtier en forme de table.

Image

Et le boîtier sans la façade qui est en cours de réalisation ( faite en jotoba).

Image

Image

Image

Image

j’ai fait des essais prolongés ce week-end.

Le système était composé d’une Cayin CDT17 et de Dynaudio S1.4. Coté musique, j’ai pris plusieurs styles musicaux très variés (du classique au métal indus).

Le premier constat est que ce petit ampli arrive à tenir les S1.4, ce qui n’a pas une mince affaire (85 db sous 4 Ohm). Avec les Dynaudio, il faut quand même mettre le potentiomètre à plus de la moitié pour remplir obtenir un bon niveau sonore dans une pièce de 40 m² .

Le son est dynamique et charpenté avec une bonne bande passante. Le registre aigu est ciselé et détaillé avec des timbres plus que corrects (l’étalon du jour était le Classé CAP 151 MKII qui à des timbres superbes de véracité, après l’Audiomat Solfège Référence, c’est ce que j’ai entendu de mieux au niveau du haut du spectre). Sur cette version d’ampli numérique, il est plus plein et moins détaillé de ce que j’ai déjà entendu en numérique mais sans gommer trop de détails. Il n’y aucune surbrillance et aucune tendance au chirurgical (ce que je n’aime pas).

Le médium apporte son lot de surprise car il est informatif et chantant avec notamment une couleur tonale très humaine. Je veux dire que les voix sont charnues et ont ce coté chaleureux que les montages à tubes peuvent apporter (dans le bon sens du terme).

Les basses sont à l’image des bon amplis numériques : dynamiques, percutante et sèche (quoique avec lui, elles ont un peu plus de chaleur et de matière qu’habituellement).

La scène reste entre les enceintes mais à une bonne extension en profondeur avec notamment un très bon étagement des plans sonores. La présence des interprètes est aussi très bonne.

On a fait des essais sur des Titus 202 et le résultat est encore mieux qu’avec les Dynaudio car il arrive à extraire un niveau de basses sans commune mesure avec la taille des enceintes. Les voix sont bien présentes et chaleureuse et l’aigu ne déborde pas du tout et présente un ciselé et une précision qui apporte à la scène sonore une bonne extension.

A l’écoute, je croyait entendre un ampli à tube (son charpenté et beauté des timbres) avec le petit plus du numérique (basses profondes et dynamisme).

Les points faibles sont qu’il n’y a qu’une seule source et l’absence de télécommande mais pour ceux qui ne sont pas fainéant et qui n’ont qu’une platine CD, il présente un très bon choix.





A venir les photos de l’ampli fini.

Cordialement

Amra
Dernière édition par Amra le 11 Fév 2008 20:15, édité 1 fois.
Amra
Membre HCFR
Membre HCFR
 
Messages: 533
Inscription Forum: 02 Mar 2006 20:47
  • offline

Message » 11 Fév 2008 19:15

Eh bin.... commentaires dithyrambiques!!
SoulMan
 
Messages: 1877
Inscription Forum: 04 Juin 2004 10:23
Localisation: Toulouse
  • offline

Message » 11 Fév 2008 20:16

Image

Image

Image

Image
Amra
Membre HCFR
Membre HCFR
 
Messages: 533
Inscription Forum: 02 Mar 2006 20:47
  • offline

Message » 12 Fév 2008 9:22

je vais revendre mon audreal duo 2.0, je ne lui trouve pas autant de qualités :)
tRg-audio
 
Messages: 1421
Inscription Forum: 30 Sep 2005 17:09
  • offline

Message » 12 Fév 2008 12:09

Bonjour

Il est possible d'améliorer encore grandement les possibilités de ce petit module.

Par exemple: En changeant les condos d'entrées, et en lui fournissant une meilleur
alimentation.

Les inducteurs sont vraiments excellents.
hautcoeur
Pro-Fabricant
Pro-Fabricant
 
Messages: 358
Inscription Forum: 13 Fév 2007 10:46
Localisation: Québec
  • offline

Message » 12 Fév 2008 12:50

Je me demande ce que ca vaut comparé au module de 41Hz qui utilise la meme puce mais moins chere (55e sans alim)

En tous cas sont étonnant tous ces petits amplis class D.

@+
Avi
melavi
 
Messages: 1950
Inscription Forum: 03 Fév 2003 18:02
  • offline

Message » 17 Fév 2008 15:07

salut

existe il des module tripath a tres forte puissance 500w et plus ?

si oui ou les trouver?

a+sylvain07
sylvain07
 
Messages: 6582
Inscription Forum: 16 Avr 2004 18:48
Localisation: paca
  • offline

Message » 17 Fév 2008 16:43

Bonjour,

Tu peux trouver des Kit de forte puissance chez 41Hz.

Amra
Amra
Membre HCFR
Membre HCFR
 
Messages: 533
Inscription Forum: 02 Mar 2006 20:47
  • offline

Message » 19 Fév 2008 15:12

les modules tripath a forte puissance son annonce 1000/8ohms mais tiennent t'il les impedances de 4ohm voir en dessous?


a+sylvain07
sylvain07
 
Messages: 6582
Inscription Forum: 16 Avr 2004 18:48
Localisation: paca
  • offline

Message » 19 Fév 2008 15:50

Le mieux est de poser ta question sur le forum de 41hz...
...et de nous donner le réponse ici ensuite! ;)
SoulMan
 
Messages: 1877
Inscription Forum: 04 Juin 2004 10:23
Localisation: Toulouse
  • offline

Message » 19 Fév 2008 16:04

Bonjour,

Le AMP8 supporte jusqu'a 2 ohms. Mais ne semble plus en production.

"Amp 8 is similar to AMP2 but has doubled output FETs. It is intended for high power amps with low impedance loads, 4 to 8 ohm bridged, 2 to 4 ohm stereo, woofer amps etc. It is designed to handle currents of at least 2 x 30A."

Attention ce kit contient des composants de surface. Donc difficile à souder.

Selon le forum de 41hz
----------------------------------------------------------------------------
"The "Standard solution" with the AMP2 TK2350 boards is a 2x300W into 4 ohm speakers with a 2x45V power supply, or 2x60V for 8 ohm speakers. There is margin to increase the voltage input and power output but some modifications or changes in components may be required. The basis for this is the triple weight copper used on the PCB, allowing high currents and low losses, and the space on the board for using higher rating components.
The AMP2 board is based on the Tripath evaluation board but is also based on a schematic from Tripath for a bridged very high power (>1000W) output board. The main difference between the the bridged high power board and the standard evaluation board is the use of double (8) power MOSFETs instead of four. In the layout of the AMP2, space for double MOSFETs has been included. Below is a discussion on how a very high output version of the AMP2 can be implemented and the components that may need changing, compared to the standard solution.

IMPORTANT: The solutions suggested here may not all be fully tested! Therefore, you should yourself test and verify suggested changes.

MOSFETs

First we could (should), as mentioned above, probably use double power MOSFETs. If you read the rating for your MOSFETs, at a first glance this should not be required. With a 60V supply into 4 ohms, you should have a maximum average peak current of about 15A and there are plenty of MOSFETs rated higher than this. However, in reality, the current rating is for a very well cooled device and at some well specified operating conditions. A de-rating is almost certainly required in real life applications. Also, in real life applications the thermal dissipation can be the limiting factor. Double MOSFETs will run at half the current, and therefore one quarter of the conductive loss and heat dissipation than a single MOSFET. Of course the heat conductive surface is larger for two MOSFETs than one, thereby improving the cooling.

An ideal switching MOSFET will switch on and off infinitely fast and have no resistance in the on state. In reality, the switching takes a few nanoseconds and during this time, there is a loss and heat dissipation due to the voltage drop over the resistance. The MOSFETs also have some residual resistance, Ron, when turned fully on, causing more losses. Typical power amp MOSFETs have a resistance in the order of 0,05-1 ohm.

The gate on the MOSFET is the "switch-button". It has some capacitance so it takes some energy, Qg, to flip the switch. The energy charge needed to fill the capacitance can be in the order of 10-200 nC per cycle. Therefore some current needs to flow in and out of the gate during the switch cycle. This causes some losses and causes a time delay in the switching. The driver circuit that delivers the current/voltage to the gate also has some losses. One of the two chips in the Tripath chipset is mainly a MOSFET gate driver. The chip is designed to drive MOSFETs with gate charges of up to about 100 nC. Unfortunately, using double MOSFETs also doubles the gate charge and gate losses and increases the driver chip temperature. Therefore, when using double MOSFETs, they should preferably have a gate charge lower than 50 nC. The AMP2 board is prepared for mounting a small heat sink on the driver chip if required. Tripath recommends this in case the gate charge is high, i.e. in the order of 100 nC. Tripath also claim that the chip can be used with gate charges up to about 140 nC if the driver chip is well cooled. Tripath say they have successfully used dual STW34NB20 that have a gate charge of 60nC with this chipset. Possible alternate MOSFETs with lower Qg include the ST Microelectronics TP19NB20 and the ON semiconductor FDP2670

The two desired characteristics, a low Ron and a low Qg, are usually opposing requirements. MOSFETs with a low Ron usually have a high Qg and vice versa. A compromise has to be found. A third important characteristic of a switching MOSFET is a short switching time. More about this can be found below, under BBM mode.

It should be noted that the voltage rating for the MOSFET should be significantly higher than the rail to rail voltage to allow for transients. For high voltage supplies 200V MOSFETs should be used while 150V MOSFETS may be used for amplifiers with a supply voltage of +/- 50V, maybe +/- 60V after evaluation. Tripath have made a technical note, TN16, on selecting switching MOSFETs. The TN16 is slightly outdated but is still useful and can be found on our downloads page. It should be noted that 150V MOSFETs can have better Qc / Ron and switching times than 200V MOSFETs.

The AMP2 board allows the use of MOSFETs is size TO220 as well as TO247.

Gate resistors

A resistor is placed between the driver chip and the MOSFET, in order to limit the maximum current from the driver chip. A low value resistor will give fast switching but may cause thermal overload on the driver chip. A high value resistor will give slow switching, increasing switching losses.

 Lower resistor value Higher resistor value
Switching Fast Slow
Distortion Lower Higher
MOSFET losses Lower Higher
Driver chip losses Higher Lower

Tripath recommend a gate driver resistor of about 15 ohms for MOSFETs with a gate charge higher than 70 nC and 22 ohms when the gate charge is less than 70nC. With double power MOSFETs, half of this value can be used. For the AMP2 board, 15 ohm resistors are used with STW34NB20 MOSFETs. With double STW34NB20, 10 ohm gate resistors could be used. This is well in line what Tripath use for the Evaluation Boards. There is also a protection diode in parallel to the gate resistor. For double MOSFETs you need double gate resistors and double protection diodes and the AMP2 board is prepared for this.

BBM mode and shoot-through

For each channel there are two power MOSFETs; the high side (positive) and low side (negative). Normally, one of these would be open, letting current flow through the MOSFET through the loudspeaker and back to ground. These should never be open at the same time as this would mean the positive and negative rail voltages would be short circuited, there will be a current shoot-through causing over-current or damage. Perfect MOSFETs would switch on instantly. In reality there is a short time delay when the MOSFETs switch on and off. The shorter this switching time is, the better. Typical audio devices have a switching time of 10 to 100 nanoseconds. Therefore a slight time delay in the driver circuit can be used to prevent shoot-through. This is referred to as a Break-Before-Make (BBM) time. A short BBM time should give lower distortion, while a too short BBM can have serious side effects in terms of shoot-through. On the Tripath chipset the BBM time can be set by a jumper to 120 ns, 80 ns, 40 ns or 0. For most applications, 80 ns should be used. Using 120 ns may be feasible for woofer amps for increased shoot-through safety margin at very high power. A 120 ns BBM will give higher distortion but mainly above a few kHz so this could be used for subwoofer amplifiers. 40 ns can possibly be used for very fast switching MOSFETs. It is recommended that this is tested at full power as the time delays are temperature and current dependent. The AMP2 board is prepared for using an alternate filtering solution for preventing shoot-through even with short BBM timing. At the time of writing, this has not yet been tested.

Source-drain bypass capacitors

The original Tripath has source-drain capacitors on the low side MOSFETs that prevent output ringing. For a dual MOSFET solution, these capacitors should be used for all the MOSFETs, both high and low side and the AMP2 board is prepared for this.

Power supply capacitors

There power supply capacitors on the PCB. These handle the high ripple current from the power supply to the MOSFETs. Tripath use two 330 uF 63V capacitors rail to ground and two 33 uF 150 V capacitors rail to rail. The AMP2 component kit uses four 220 uF 100V capacitors rail to ground and two 33 uF 160V capacitors rail to rail. Please note that very high ripple current and very low ESR value capacitors are required. For high power applications, four 470 uF 100V capacitors rail to ground and two 150 uF 150V rail to rail are suggested and will fit on the AMP2 PCB. In addition, the regular power supply capacitors rated at least 10.000 uF per rail, preferably 20.000 should be used.

Current sense resistors

The current overload inputs of the chipset senses the voltage over the low ohmic current sense resistors. Nominally, 0.01 ohm, 2W sense resistors are used. This allows an average current of 20A per side before the current sense resistors are thermally overloaded. It would be possible to use 0.005 ohm sense resistors and/or use a higher power rating but this is probably not required.

Current overload resistors

The current overload peak limit value where the chipset will shut down, can be set by two resistors in the 0805 size. The data sheet on the TK2350 chipset includes information on how to calculate the value for these resistors. Note that the power supply capacitors are placed after the current sense resitors so current may flow between these capacitors and the MOSFETs without being seen by the current sensing devices.

Over voltage sense

Over and under voltage may need adjustment for high supply voltages. With carefully adjusted voltage sense resistors, supply voltages up to +/- 65V can be used. The board uses four resistors in the 0804 size for over/under voltage sensing. Note that the chipset may be damaged by voltages exceeding +/- 70V even with a correctly selected over voltage sensor resistors. The over voltage sensors ensure the power MOSFETs are swotched off at over voltage voltage conditions but the chipset may still be damaged by excessive over voltages.

Modulator gain

The modulator gain resistors set the main voltage gain of the amplifier stage. For optimal performance, the modulator gain should be matched to the rail voltage. Please see the chipset datasheet for information on how to set the modulator gain. Please also note that the AMP2 board uses two resistors in series for setting the modulator gain. These resistors voltage divide the modulator feedback signal so that no high voltage signal is not present near the low voltage chip, while still maintaining a reasonable current. One of the two resistors is a 0805 size resistor and one is an axial 0,25W resistor. Four of these dual resistors are on the board, in total eight resistors.

Output filters

The output filter is a second order filter. A high current 11 uH inductor is the first filter component. The default size is a T106 size toroid. For high power applications, the size may need to be increased to prevent overheating. A T130 or even T157 size core may be used. The AMP2 board has space for up to T157 size toroids and this size is recommended for very high power applications. In most applications the toroids will be the tallest components on the boards. In the output filter there is a 0,1 uF and a 0,22 uF capacitor per side. These capacitors should be of high ripple "switching" type. Standard plastic capacitors should not be used. While the value of these do not need changing for a high power application, there is space for using other packages on the board if required. There is also a 3W resistor in the output filter that may need to be substituted for a 5W resistor for high power applications.Â

Summary

Here is a summary of component changes that could be made for a "Very High Power" solution compared to the standard solution for the AMP2 board

Four more MOSFETs, preferably a type with low gate charge
Four more gate resistors and diodes. The resistor should be half the value of the 4 MOSFET version, probably 10 ohm but this is dependent on what MOSFETs are used.
Six more anti ringing caps (220 pf) with their associated 16 ohm resistors for the MOSFETS
Larger power caps. Four 470 uF, 100VÂ rail to ground, two 150 uF 160V rail to rail
Two larger power resistors 20 ohm rated at 5W (instead of 3W)
One heat sink for the driver chip
Two larger toroids, T157-2 size
Changed gain settings. Four axial 0.25W resistors need to be replaced
Possibly also current sense resistors with a higher rating
Changed over current trig level resistors. Two resistors size 0805 need to be changed
Changed over voltage resistors. Four size 0805 resistors need to be changed.
We may publish a "Very High Power" version application and a component kit for the AMP2 board, based on the guidelines found above either on request or when this has been fully tested by us or someone else. "
----------------------------------------------------------------------------


Je travaille présentement à un bloc mono Tripath de 1000w 8 ohms avec nos propre modules.
Pas facile lorsque l'on arrive dans ces puissances. Beaucoup de contrainte électroniques et thermiques. Ces modules ne seront pas vendus au DIY dans l'immédiat.

Si tu à un jour un ampli Tripath de 1000w ou plus, j'aimerais bien le comparer avec le notre.
hautcoeur
Pro-Fabricant
Pro-Fabricant
 
Messages: 358
Inscription Forum: 13 Fév 2007 10:46
Localisation: Québec
  • offline

Message » 19 Fév 2008 19:18

al'heure actuelle j'ai un lps2 qui bridget fait 600w/4 et tiens le 2ohms mais je vaudra plus de puissance et un damping factor meilleur il est de 350 ici.

a+sylvain07
sylvain07
 
Messages: 6582
Inscription Forum: 16 Avr 2004 18:48
Localisation: paca
  • offline

Message » 19 Fév 2008 19:52

Un peu indigeste en Anglais pour la majeure partie d'entre nous.... :roll:
breizheau
 
Messages: 15483
Inscription Forum: 23 Oct 2005 12:30
Localisation: Roanne (42)
  • offline

Message » 19 Fév 2008 20:34

Bonsoir,

Désolé pour l'anglais. Je l'ai mis tel que sur leur forum parce que:

1- Je suis pro et je ne voudrais pas qu'on puisse me prêter des mauvaises intention
face au discour d'un concurent.

2- même avec la meilleur volonté du monde, aujourd'hui je n'ai pas le temps de traduire tout ca. Et ca ne suffirait pas car il faudrait aussi traduire les documents de Tripath.
(En gros pour faire cet ampli vaut mieux de savoir lire l'anglais, surtout les éléments ayant
traits à la sécurité)

Sans le détail cela dit:

Que la solution proposé n'a pas été testé.
Que la Puissance théorique de plus de 1000w

Le circuit de pilotage des transistors tripath est fait pour des transistors à 150nc
mais peut 140nc si bien refroidis.

En gros pas si facile de choisir les Mosfet. Voir la note technique tripath tn16

Les resistances du circuit de protection tripath doivent être augmenté en conséquence.

Il faut doubler les mosfet et leur diode de protection.

Lorsqu'ils ont écris cette note, le systeme alternatif de filtrage au BBM n'a pas été testé.
l'ajustement des BBM doit etre fait à température et au voltage d'utilisation maximale.

En général la carte prévoit les emplacements des composants suplémentaires

Faire attention à bien choisir les condos de remplacement pour le rail d'alim et ajouter
des condos externes.

Les bobine du filtre de sortie et les resistance doit être amélioré en relation avec la puissance désirée.

EN RESUME
4 MOSFET de plus,
4 resistances et diodes pour la base
6 capa (220 pf) et leur résistances de 16 ohms associées pour les MOSFETS
De plus grosse capa d'alim. 4 x 470 uF, 100VÂ rail vers la masse, 2x 150 uF 160V rail vers rail
2 plus grosse resistance 20 ohm de 5W (a la place de 3W)
un disspiateur pour le chip de control
2 plus grosse bobine, T157-2
Remplacer les 4 resistances axiales 0.25W pour le reglage du gain
Changed gain settings. Four axial 0.25W resistors need to be replaced
Possibly also current sense resistors with a higher rating
changer les resistances de taille 0805

Voila ...

Notre amplificateur a nous est basé sur un circuit différent mais tiré du même
shéma de base qui est celui de référence de Tripath.

Le meilleur endroit pour trouver les infos est donc le schéma de référence et le
manuel du circuit tripath.

Hep! Attention quand même, Ce genre d'ampli peut-être mortel!
hautcoeur
Pro-Fabricant
Pro-Fabricant
 
Messages: 358
Inscription Forum: 13 Fév 2007 10:46
Localisation: Québec
  • offline

Message » 19 Fév 2008 21:13

quel est le facteur d'amortissement de ses modules?
a+sylvain07
sylvain07
 
Messages: 6582
Inscription Forum: 16 Avr 2004 18:48
Localisation: paca
  • offline


Retourner vers Amplis et Préamplis

 
  • Articles en relation
    Dernier message