Help me fix my Hafler TA1100

[Flatpicker]

Ok, one channel is bad and the other is good. I went through the circuit as best I could. All DC is the same for both sides. All small signals look the same until you get to the output. I replaced all the MOSFETs on the defective channel, but no change. I'm betting it has something to do with the output protection circuitry which I know nothing about. Even worse, the entire thing is surface-mount except for the MOSFETs and the voltage regulators.

Anyway, here's the good channel: http://mywebpages.comcast.net/tamron1958/OutputGood.jpg

The bad channel: http://mywebpages.comcast.net/tamron1958/OutputBad.jpg

And the schematic: http://mywebpages.comcast.net/tamron1958/Schematic.pdf

Any help is much appreciated!

 

[SonsOfThunder]

Just to make sure here: You checked the signal at R93 and R96 and it looks similar at both places? You might check both sides of those resistors to make sure that one of em isn't burnt up.

The more I look at MOSFET output amps, the less I like them. The difference between the characteristics of the N- and P-channel devices is too far apart. Give me a good old BJT amp anyday.

Peace!

 

[Gus]

interesting circuit.

Drain legs going to the output? Nice push pull drivers for the gate cap of the output stage.


Because it effects both sides of the wave make me want to check the bias setup around Q18

Nothing might be bad maybe the turn on voltage for the output MOSFETs is different enought that you need to readjust it.

Can you place a current meter in the power supply leg and measure the Current with the input shorted to ground on both channels. JW4 bias check bottom right

Maybe the Vgs needed for AB operation is more with the replaced MOSFETS. It looks like RV1 adjusts this.

The gain is increasing with drive that is why I am thinking bias

 

[Flatpicker]

Thanks guys! I'll check it out when I get home.

[quote author="SonsOfThunder"]...Give me a good old BJT amp anyday.[/quote]
I'm starting to agree. I think I'll go ahead and re-cap my old Harmon-Kardon 730.

 

[SonsOfThunder]

I see what you are thinking there Gus, but I don't think that bias is the issue, unless the bias pot has gone open. RV1 adjusts the bias current certainly. Some component is broken, I don't think this circuit is so sensitive to need matched outputs.

Measuring bias at one point like that seems rather a dangerous practice for production. Much smarter to check voltage across the 0.1 resistors, that way you know the current on EACH o/p device.

I know you guys don't believe me, but that @lesis RA-100 with some simple mods is a really decent amp, especially as cheap as you can buy them. And they make more power than advertised too.

Peace!

 

[Gus]

Yeah they are FETS and you need to match them IMO. It is not that hard set up a circuit with a drain R for about the current wanted with a regulated power supply and to measure/calculate the drain current vs gate voltage at a few points.

 

[Flatpicker]

Ok, R65, R66, R93, and R96 are ok, so that's out.

By the way, the output signals are with the amp running full tilt into 4ohms just below clipping.

I guess that's why it's called a Hafler, because only half of it works, yuk yuk...

 

[AnalogPackrat]

You may be on to something there--I have an old Hafler DH-110 pre with a dodgy ganged pot which leaves it with only one working channel. I should fix that thing up and sell it...prices have gone up since I bought it used. :roll:

A P

 

[Flatpicker]

Ok, this was a weird one! I emailed Hafler yesterday and a tech emailed me back and said to remove C21 and C22, and put a .47uF between pins 6 and 7 of U2B, and to duplicate the mod on the other channel before it goes bad as well. I only had one .47 lying around, so I gave it a try on the defective channel and it WORKS! Looks like a design flaw, huh? :thumb:

(I'd still like to know how this thing works.)

 

[Gus]

I was going to call you and talk about the electros used for the servo section using two back to back is not a good idea for a servo. The servo is a parrallel part of the feedback. SMD electro suck IMO measure some and compare to a good thur hole cap.

 

[Flatpicker]

I see what you mean now, but for me, hindsight is always 20/20. I'm amazed at how you know all this stuff!

So, did the electro's just go bad? It was working when I got it.

 

[Rafita]

Im also have this amp driving my ns10 , recently the left channel begin to distort the sound (Sometimes ) like if has a loose conection . i tought it was the monitor , but since i noticed a different frequency response (pink noise ) on each speaker i think its the Amp , a friend of mine recomend me a crown d75 , he had used it for 15 years and never blowd a speaker or twetter

 

[Flatpicker]

Well, now you know how to repair it. Just refer to the schematic I posted a link to and follow the procedure I wrote in the earlier post. Works great when you do that! :green: The chassis is easy to take apart and everything is marked well on the pc board.

By the way, the tech that emailed me from Hafler was as nice and helpful as anyone I've ever dealt with. They get my vote!

 

[Flatpicker]

Someone at Hafler would help you repair it - they are really nice!

 

[Jerz3455]

I FIXED MINE TOO!!

SOLUTION:

RECAPPING/REPLACING ALL 14 10uf 16v EL CAPS.

AND MAKE SURE THE WELDS ARE GOOD AND CLEAN.

I ran into an issue with one channel where I had gotten the bass back, but i still had the distortion thing going on. So I rechecked all the caps, and found that I hadn't soldered one of the legs of the C21 cap.

After rectifying that, I was good.

so.
+1 on the recaps

Hafler TA 1100

 

[bwalrave2002]

I was able to fix my two TA1100s recently, and thought I would share pics for those who are less electronically inclined (like me) and wanted more info.  Hopefully this helps someone!

You can solder the leads of the .47uF cap to the pads for C21 and C22 nearest U2B - those pads have traces back to pins 6 and 7 of U2B.  I used a .47uF from Digi-key - four total for $7.44.  Part ECW-F4474HL, P12199-ND, CAP FILM 0.47UF 3% 400VDC RADIAL.  Overkill, but great specs.  Well, great as far as a mechanical engineer can tell.

It's a pretty easy fix.  Screwdriver, soldering iron, and some patience and you're good to go.  My TA1100s sound as they used to - clean and clear with no clipping, even at loud volumes.

 

[bwalrave2002]

And a second pic...