Help needed with Tripath Board

[m.parigi]

I tried building a power amp based on the Tripath 0104 eval board and incorporating RLJones' and C. Brodersen's suggestions/tweaks.
My original goal was to build a no-compromise power amp, that's why I shopped around for the best components I could find.

Now that I've assembled evarything, though, I can't power up the amp without having it blow up in a few seconds.
I already blew up a board (the MOSFETs on one channel shorted out).

Now that I've gotten a replacement board, it looks like this is blowing up too. The MOSFET of one channel have already blown, and I've replaced them using the surviving ones from the old board. Once repaired, I've turned it on briefly (a couple-three seconds) before having to turn it off because something was smoking (I didn't have time to understand what).

I've already checked the power rails (without any load applied), and they are at +/- 75V. On the scope I don't see any big AC residue. The trouble is that I can't keep the amp on long enough to measure anything!

Is there anybody that can help me out with this?

On a different thought, I begin to suspect that I don't have all the necessary equipment/skills to troubleshoot this project. Does anybody know somebody that could troubleshoot this for me? I could ship him/her the amp...

Ideas, anybody?

[sonnya]

try raise "break before make timing" to the maximum!?

Sonny

[mrfeedback]

Hello All,
A method I use when running up rebuilt TV swichmode supplies and rebuilt audio output stages is to to power them from a power limited variable AC supply with the circuit arrangement as follows -

240 into variac, 40 0r Watt switchable lamp in series between the variac output the primary of an isolation transformer, all built onto a suitable board.

On turning up the variac output voltage, if there is a fault condition, the lamp will glow brightly and dissipate mostly all of the power, therefore not enough energy dissipated in the circuit under test to cause damage.

I have used this arrangement in my day to day servicing work for many years, and this has saved me from expensive grief many times - I highly recomend this for running up new constructions.

I hope this can help you,
Regards, Eric.

[fcel]

I have used similar method (but no variac or isolation transformer) of hooking up a light bulb in series with the amp under test and if there is a short somewhere in the amp, the bulb would glow very brightly ........ to tell you to pull the AC plug right away and thus little or no damage done to the amp.

[m.parigi]

Thanks Mrfeedback and fcel, but I've got a problem with the filter capacitors.

I've got over 140,000 uF in the power supply section, and by the time they are full discharged the damage's already done!

I don't understand how is it possible that by just changing a couple of capacitors and resistors following the footsteps of others that have tried it before me I can cause the board to fry.

The modifications I made were:
a) changed the input resistors and associated filter capacitor to raise the input sensitivity.

b) I included an RC network on the input for preventing the circuit from oscillating. This is suggested by both Tripath and Jensen Transformers

c) I changed the value of four capacitors to 0.22 uF and the value of the associated resitor to 22R along the guidelines of Tripath, for preventing the circuit to go into self oscillation when the load is absent or removed.

d) I added some bypass capacitors (0.22uF and 1000PF in parallel) to the chassis ground on each the output terminals, following the advise of Tripath.

Can anybody spot anything particularly dumb I could have made?

[nar]

Originally posted by m.parigi
I tried building a power amp based on the Tripath 0104 eval board and incorporating RLJones' and C. Brodersen's suggestions/tweaks.
My original goal was to build a no-compromise power amp, that's why I shopped around for the best components I could find.

Now that I've assembled evarything, though, I can't power up the amp without having it blow up in a few seconds.
I already blew up a board (the MOSFETs on one channel shorted out).

Now that I've gotten a replacement board, it looks like this is blowing up too. The MOSFET of one channel have already blown, and I've replaced them using the surviving ones from the old board. Once repaired, I've turned it on briefly (a couple-three seconds) before having to turn it off because something was smoking (I didn't have time to understand what).

I've already checked the power rails (without any load applied), and they are at +/- 75V. On the scope I don't see any big AC residue. The trouble is that I can't keep the amp on long enough to measure anything!

Is there anybody that can help me out with this?

On a different thought, I begin to suspect that I don't have all the necessary equipment/skills to troubleshoot this project. Does anybody know somebody that could troubleshoot this for me? I could ship him/her the amp...

Ideas, anybody?

You should check if the ouput board is delivered the good DC fotr the gates of your mosfets.....by running the front end alone and take your multimeter.
If you find more than 20 V / theoric Gate source voltage for the Fets,then it is normal you blow them.
I suppose you have enough heatsinking.....

[rljones]

If you look at my original post, or one of the follow-ups, I alluded to a similar problem. The 'smoking gun' was a burnt out cap on the output filter (C21 as I recall, but double check as I'm going from memory). Another piece to this puzzle was running without a load. If an 8 ohm load was attached, I did not experience this problem. At no time did I damage a Mosfet, however. This is why I then added a 470 ohm/5W resistor in parallel to the output load in order to always have some sort of load present. Once I replaced C21 with a greater voltage rated polypropylene, I no longer experienced any smoke.

I've built up two Tripath amps (one 103 and one 104) and they're both running at my home 24 hrs/day without problems. I only changed what I originally indicated on my original posts, nothing more or less.

[mrfeedback]

Hello RLJONES,

I've built up two Tripath amps (one 103 and one 104) and they're both running at my home 24 hrs/day without problems.

How much cost for the output stages, and what sonics resultants ?
WAF ?
FAF ?

Regards, Eric.

[rljones]

Eric,

I bought the boards directly from Tripath, as I recall it was about $400 apiece. Each board is a stereo unit. I then used a ParMetal chassis ($100) and built up a nice power supply for high voltage and one for the 5V bias supply ($200 or so for these). The cream was adding my usual Jensen input transformer (JT-11P), which runs around $150 for a pair. The amp total cost was thus about $800 to $900. (Again for details, please search this forum for a series of very complete postings that I did.)

The WAF and FAF was phenomenal: all like it. I kept the front panel very simple: 3 LEDs in a triangular shape: one red for power, one blue for signal and one yellow/amber at the top of the triangle for warning (distortion). The sound is very good, bettering many other amps I had on hand for comparison.

Robert

[km]

hi.

maybe i can help ; here is what we do;

use external power supply for the +5v and floating 12v (sitting on v-) , these supplies can be switched on before the +/- main supply.

then turn on the main supply gradually (with a variac), or preferably with a current limited vaiable power supply.

have a signal connected to the input and a scope on the output, if you like you can have a resistor of sau 1 kohm on the output.

you can use the setting for longest deadtime if you like but it should work with shorter settings as well.

the board has an undervoltage shutdown and will first start operating at around +/- 35v dc , with 4 ohm load and an error this could be very short , with our setup it should be safe unless you have a dead short in the output , if you do it will show at low voltage (i.e. before the amp starts operating)

our boards runs fine without output load and i doubt this will be cause for any problems (i have tested at least 50 boards), we suggest to use linear supplies for the mentioned +5 and +12 supplies , if you need more feel free to email me.

bye and good luck - k madsen - www.cadaudio.dk

$400 is somewhat more than what we charge