Hafler XL280 mods gain/low frequency cutoff problem

[HaflerFreak]

Ok guys, I got something for you to chew on...

Popped the top and plugged her in. Here's what I have so far:
AC in 122.5
B+/B- = 62.5 (ref 65)
B++/B-- = 72.0 (ref 75)
voltages are stable
DC at inputs 1 and 2 = 0.0 (ref 0)

BAD Side:
Base voltage at Q7/Q8 is -15.53 (ref -15)
B voltage at Q5/Q6 is 15.49 (ref 15)
voltages are stable
GOOD side:
-15.06 at Q7/Q8

That's all I got for now, gotta go to work -
Please reply if you think this 1/2 volt discrepency is anything to worry about...

[Stormrider]

I am more or less a beginner here... But, i own three Haflers. Two P230's and a P125. I have done some troubleshooting, and repairing on these amps.

Have you checked the dc offset on the bad channel?

If the bias is jumping around on the bad channel, check the voltages around the transistor near the bias pot. Its Q9 for my Haflers, but the XL280 appears to be differant in areas.

Also, the main check for diodes is to see if they are short circuit. If you read low, or no reistance on a diode replace it.

Good luck!

[HaflerFreak]

When I had everything apart I replaced all 4 1N4003's in the power supply with brand new, throughput matched glass passivated diodes.

I replaced the 2 1N4003's at the bottom of the 'bad' side (D5 and D6) and I checked the 1N4148's and the 1N5245B 15V zeners for shorts when I had all the caps out and had open circuits all over the place.
No problems were found, but I did not remove and check the diodes (except the 1N4003's) nor did I replace them.

One thing I have already noticed is that my main voltages, 72 and 62.5, are proportianately low with 122.5 AC. As far as I know, this is my transformer producing low voltage?? AND that voltage on the bad side is 15.5... High, hmmmm, 15 volts reference?, hmmmmm - does this sound like a leaky 15 volt zener or just a difference in tolerances in components?

I will try to get you guys more volt readings tonight, but some of those specs are just a bit to dangerous for me to try to get to.

Can anyone suggest a specific reading that would be of help?
(besides the DC offset that was suggested)
How exactly would I measure DC offset - is that just voltage at the speaker output, I think that's what I've gathered from my readings.

[HaflerFreak]

Hey,
I know some of you out there were waiting for this, but I'm guessing this thing is only worth its good side at this point.

Here are my voltages from the back of the board with JFETs noted.


I guess this is gloom and doom from what I hear. They call these the 'unobtainiums'. I don't even see XL280's on eBay except rarely and for outrageous prices.

I would really like to hear from a kind soul on this forum who has a working XL280 driver board to sell, I think they even have a trading post here, don't they?!

I have heard these things are in Germany for a high price, how do I get a set of FETs!!

[mlloyd1]

haflerfreak:

if you have dead or dying jfets needing replacement, you are in trouble.
they are indeed unobtanium. maybe you can get these from borbely audio in germany, but i'm not sure. if so, they will definitely not be cheap. not too attached to your first born are you?
one potential solution is to use selected 2SK170 and and 2SJ74 as replacements. These are more easily sourced, but not quite the same spec-wise. The 2SK170/2SJ74 have less transconductance (bad), less capacitance (good) and more noise (bad) than the 2SK147/2SJ72 pairs.
There are places where you can buy matched pairs of these.

good luck ...
mlloyd1

[HaflerFreak]

Hi mlloyd1,

I'm glad to see you back on this thread, I need your help! I have a question for you (or anyone) that I bet you can answer.
I emailed erno borbely at borbelyaudio.com about a set of these JFETS, here is what I got back from him:

Hi Scott, yes, I still have a few and I can match them. I cannot
remember if Harry used BL or V groups in the 280. Does it say in the
manul?
Please see prices on my homepage. Min. Order is Euro 50. Please fax your
cc details incl. 3-digit sec. Code and exp. Date.
Best, Erno.
---------

Well, did the Hafler XL-280 use BL or V groups?

I am assuming this ties in with another question that I had, which was: Do I match the 2SJ72 FET's to each other or to the 2SK147 FET's?

Price for this amp is no object for me up to about $200 (Then it's an object), hopefully that's enough euros

Also what do you make of my voltages mlloyd1?
Q1 and Q2 have 0 volts at the gate, almost 15V at the drain, but only 300mV at the source.
Does replacing the JFETS make sense given that most voltages are fairly close to nominal except those 22 volt FET Source voltages are all in the 1/3 of a volt range.

What would cause this? If not the FETs, what do I check next?

Obviously I don't want to buy new FETs if it could be something else.

Also, remember that this amp PLAYs like this! It doesn't have much bass in the affected channel - but it sounds almost fine otherwise - is that a JFET sounding problem?

[HaflerFreak]

Hi, sorry to post another reply so quick, I found this reference at the Borbely site:

Fortunately, JFETs are sold with much
narrower IDSS ranges, which makes life
easier in terms of proper biasing. The
2SK170 comes in three IDSS groups: the
“GR” group is 2.6–6.5mA, the “BL”
group is 6–12mA, and the “V” group is
10–20mA.

Question again is: Do Hafler XL-280's use the BL or V group when matching JFETs?

[HaflerFreak]

Oh yeah, I got those DC offset numbers.

GOOD channel = -4.0 mV (That's pretty darn good!)
BAD channel = -13.7 to -14.5 mV and not too stable. Still, -14.5 is within the realm of OK...

[Stormrider]

What are your sources? Whats your signal chain before the amp? Have you switched the left and right inputs to make sure its the amp channel itself? How about switching the channels the speakers are on?

No chance of getting an oscope into this? I wouldn't think the problem lies with the output transistors, as the amp is functioning, just not correctly. These are very robust devices and do not fail because of a small problem. If they do fail, you know it right away.

I would be looking at the passive components, especially any in filters, bypasses etc... For example, a bad resistor could send the signal through a large capacitor.

There is my .0002 cent.

[HaflerFreak]

Hi Stormrider,
Good questions, I have never really posted my testing procs.

My intial failure occurred after replacing the driver board electrolytics and tweaking the bias (yes, I know, that was bad).
The failure occurred while watching a scene from 'The Black Stallion' with some really pretty music.
I was listening to a variable pre-out source from my TV connected to the CD input of my DH-110 preamplifier. I had the volume at almost 11 o'clock on the DH110 (that's pretty high volume) and when I turned up the TV volume to about 4/5 of max - FIZZLE.
At first, it sounded a little like when I turn the amp off and let it drain the PS caps into the speakers - when they drain out there is a little distorted sound before they go totally dead. The side that fizzled kept playing but at the point it went from good to bad there was some distortion.
I turned it all off and let it cool before turning it on again.
When I powered up, my low bass output problems began - but no distortion.

TESTING: I have been testing with my Harmon Kardon HD7450 CD Player through my DH-110.
The first things I did were switch input wires and speaker wires. It was immediately apparent that it was not the preamp or source. It was NOT immediately apparent that it was not the speakers. The difference can be subtle with certain musical selections and I wound up with confused posts all over DIYaudio.com that I wish I could delete.

No, there is very little chance of getting this thing on a scope unless I take it to a shop. I don't have any techie friends

I can't tell you how happy I am to hear that comment about the output devices being robust.

Your last comment is, unfortunately, the exact same stuff that I have come up with.
It seems like a passive to me, and I am about to disassemble it from the heatsink again - but I really want a clue as to what to test or look for this time. I have never really tested any of the transistors or JFETs (except above voltages). I have already had this thing totally apart once, and I don't like to push my luck

I mentioned as a joke in one of my earlier posts that the one polycarbonate that I did not test was C13 and Murphy's Law would dictate that this is it, hmmmmmm.