[tangent]

Quote:

Originally Posted by espressogeek /img/forum/go_quote.gif I see 20v on each op amp socket

From where to where? Pin numbers?

Quote:

With the op amps in the sockets I see 9.5v at each op amp.

Same pins as before?

If so, does the supply voltage at the PPA amp board's V+ and V- input pins also sag comparably?

Quote:

I measured the current draw in this scenario and I see about 9.8ma of current at the V+ terminal

That doesn't make sense. If the supply voltage is being dragged down so far, it's got to be a much bigger load than that. Like, several hundred mA.

And if the implied voltage sag above isn't happening, then you should still be seeing more like 100 mA here.

Quote:

please excuse the 100k in positions R6L and R6R.

I may see fit to excuse the values, but I can't excuse those long leads. You've gone and created loop antennas in the feedback loop. I won't say this is certainly causing the problem, but it isn't helping.

You can just take R6 out, and the amp will still work. Gain will go way down, but as long as you're using unity-gain-stable chips, this should be fine.

Quote:

ImageShack - Hosting :: cf001541np8.jpg

I don't see much solder up on the top side. It's possible you have some bad joints, since the solder evidently didn't flow very much. You might apply a little liquid flux to anything looking suspect and reflow it. If you don't have flux, just add a little more new solder to the joint.

 

[espressogeek]

Thank you for your help and reply! I have removed the offending R6 resistor on both sides of the amp. I will replace it properly when the new ones arrive. I am using the LME409710's btw.

I still have the hot'ish left side output transistor issue. I went back and hit up all of the pads on that side with rosin and solder and I now see a good bit more solder up top. I will hit up some other ones later this afternoon when more time is permited.

I'll try to be very clear on my measurements. The voltage "sag" this morning might have been a bad reading on my part. Here is what I see now.

Measuring from V- to pin 7 (V+) on each op amp, LME49710 op amps installed.
Left - 19.05v
Ground - 19.08v
Right - 19.08v

Measuring the input at V+ and V- - 20.01v

We can ignore the current readings because I did some test and the meter has a problem with this apparently. I have a friend bringing over a known good meter after work so hopefully I can get some current numbers. Any other things I should look at? I really appreciate all of the help.

 

[espressogeek]

Update: I now have a working meter . Ok , so I have connected the meter in line with the power supply and amp and when I hook it up I immediately draw 125mA. This climbs up to around 157mA before I unhook it because the transistors in the left channel get quite hot.

I just checked some more and realized that it is not just the left channel but the following.
Q34L , Q24L, and Q34G. The other three are fine.

When I look at the voltage across the output resistors I see the following.
R34L - Begins at 170mv and climbs along with the temp of the transistor to about 203mv before I unplug the amp
R24L - Begins at 170mv and climbs along with the temp of the transistor to about 203mv before I unplug the amp
R34G - 25.5mv (I'm sort of miffed about this, I would have thought it would be higher considering the temp)
R24G - 25.7mv
R34R - 22.3mv
R24R - 22.2mv

UPDATE: I did some checking between the ground and right channels an the afflicted left channel. I noted that the R13 resistor was 200 ohm on the ground and right and 100 ohm in the left. I have now swaped out all three, just to be safe, with 100 ohm parts. My current draw to R24/34 R and R24/34G is around 23mv now. The left side is still screwed. I have also replaced the Q34L and Q4L parts thinking this might resolve the issue. I only have one spare BD139 for Q24. I would replace it if there was a good explanation for doing so. Does anyone think this might be the culprit? How about Q23/33L?

I swapped the op amps between the ground and left channel. That produced no change in the symptoms.

I'm going to keep poking around this thing and searching. Any advice is appreciated.
Thanks

 

[StanRex]

Ok, I ve bought a hakko 937 clone, some 0.5mm solder, flux pen

and soldering the ad8610 got a lot easier!

I have 3 mounted on browndogs. I ve tried switching the PPA on. And well, Q24R gets real hot. And there s a bad smell floating in the air. I think something s burnt ^^

now, I m thinking it s either :
_ a bad joint from me
_ q24 got killed when I had tried doing the power supply test with the AD843
_ q24 got killed because the browndogs arent too securely attached in the sockets. If I try to put the PPA upside down, the adapters will fall down from the sockets. And maybe the bad contact between the ad8610 and the sockets created some kind of short. That s also kinda problematic because I can't test voltage at the V+ and V- pins of the opamps, since I m pretty sure I ll make them move when touching them

I'll try to put a few pictures ASAP.

edit : just checked and I get 21.9 between pins 4 and 7 of the ad8610. Considering I m using a 24V tread as a ps, looks ok to me

edit 2 : apparently my left channel is working : I can adjust the buffer bias; however, when I try to do the same for the right channel and ground, I just get 0mV at the output resistors pins.

edit 3 : pics





ps : there s quite a bit of flux left. I tried cleaning with 99.7 isopropyl but there are some spaces between pins I just cant access

edit 4 : I ve cleaned up the remaining flux with a toothbrush . It looks better now. Didn't update the pics though

edit 5 : and yes I m planning to replace C5 with boxed MKP caps

 

[espressogeek]

Seems to be the week for hot output transistors . Sorry to hear about your issue.

 

[StanRex]

Well, I m thinking it could also be the buffer transistors. Not sure of anything though. But the good thing is I have spares

edit : I've also realized that apparently my meter sucks a bit too
I use the default 2.2ohms resistors for output

and I got something like 60mV and 7.5mA across them

I think Mr Ohm felt offended ^^

is there any way to find out if the issue is located at the output transistors or at the buffer transistors without having to desolder them all?

thanks in advance

edit : nvm, my voltage readings were ok and I wasn't properly measuring current. I still get 0mV at the output resistors of 2 channels, gonna try switching opamps see if the channels follow, and probably switching output transistors after that.

 

[espressogeek]

Poor Mr Ohm. I had a buddy bring over a Amprobe 37XR-A yesterday and for the cost I am quite impressed. I ordred one like it last night for around 120. I definitly think having a quality meter is worth the beans.

Unfortunately my situation went from bad to worse. I replaced Q23/33 and now I just smoked Q24/34L with a whopping 1.7 volts through the associated output resistors . I have no idea what is going on but I suppose a complete rebuild of that side of the amp is in order.

 

[espressogeek]

I have pulled out all of the suspect components on that side of the amp and clean it off. Do you think the board is still serviceable? What can I check to find out what is going wrong? I just dont want to populate this board again and end up in the same position.

 

[StanRex]

oh and btw, there s something I hadnt mentioned earlier. When I plugged in the amp and Q24 got hot, I had a short between IG and IR.

I suppose it could have done some damage?

 

[espressogeek]

What was causing the short?

 

[StanRex]

My stupidity : when I had soldered the input wires, the unbraided wires that I hadn't cut got pressed against each other

Yeah, I know, epic fail ^^

 

[JamesL]

I was about to bias my opamps into class A, but I noticed that my current draw was about 16mA over the sum of the draw of the components.
I have 20mA bias at the buffers, OPA627(should be drawing 7mA each), no class-A, and no LED. = 81mA

I'm reading the current draw at S1, and it starts off at 81mA, but steadily rises to almost 97mA before it settles down.

Is that big enough a difference to be worried about?
I've been running it fine for a few weeks with no problems, and it hasn't been getting excessively hot either.

 

[espressogeek]

Success! I rebuilt the left channel with new parts after posting the pictures and everything seems to be working perfectly! The darn thing plays music and sounds great!

The bass coming from this thing is AMAZING! It is super fast and really tight.

UPDATE: I biased the outputs to 60mv. After listening for a while I checked the right channel and it is somewhere around 16mv. I tried to bias it back up to 60mv and I can only get about 30mv out of it. Everything sounds fine.

UPDATE: I just checked resistance across R12. On G and L I am seeing about 360 Ohms resistance. On the right channel with the low bias I see about 140 Ohms. I reckon this is the issue. I suspect that I should just need to replace R12 on that channel and I will be good to go.

 

[tangent]

Quote:

Originally Posted by JamesL /img/forum/go_quote.gif I'm reading the current draw at S1, and it starts off at 81mA, but steadily rises to almost 97mA before it settles down.

This is normal. The PPA guide text saying you can figure buffer current draw just from voltage across the output resistor is very much an approximation. That's where most of the current goes, to be sure, but it ignores all of the stuff behind the output pair.

 

[slaikan]

Hi

Finished building the PPA v2, very very happy with it, no problems with the build which I'm thankful for since its only my third DIY project. First was the CMOY (recently destroyed the OPA2132 with reverse polarity... oops), then the Audiosector Premium USB DAC and now the PPA. I struggled more with the casing and getting everything to fit properly than bulding the PPA itself.

I used all the standard components off the Tangent parts list except for C5's are WIMA MKS4 250V 0.1uF as thats all I could find in a hurry at the time.

For those who have experience with different caps, should I replace them with WIMA MKS2 63V 1uF or maybe something else and how much do these caps affect sound quality?

I'm using Panasonic FM throughout, 220uF 35V caps on the input side and 3x 680uF 35V on the power supply side. Is there a significant difference with 'boutique' caps like Blackgates, Nichicon Muse, Elna Silmic etc. cause its quite difficult to get that where I stay, so it all needs to be imported.

What are you opinions of using CAT5e for hookup wire, thats what I'm using at the moment, I decided not to import fancy hookup wire and my perferred local supplier only sells hookup wire in 20 meter reels, per colour...

Thanks