[tangent]

Quote:

I just discovered it is a 8.4V 200mAh

 
Okay, so what does its voltage measure unloaded and loaded, and does the loaded voltage change when you toggle the bias switch?
 
 

now I have the battery empty

 
I hope that's not literally true. You can damage NiMH batteries by draining them too far. 0V is very very bad. You shouldn't go any farther than about 0.8 V per cell, or about 5.6 V for the entire battery in this particular case. (8.4V = 7 cells for a NiMH.)
 
 

could I dare to plug in the wall wart

 
Do you have a big power resistor on hand? If so, you can put it across the battery contacts, then measure voltage drop across it to infer that the trickle charger is working. V=IR, so if you use a 10 ohm resistor, you'd expect to see 0.2V across it if you set up the trickle charger for 20 mA, as would be appropriate for a 200 mAh battery.
 
You'll want to use at least a half watt resistor in this case. Power dissipation is 0.4W, so a 1W or bigger resistor would be better.
 
Or, you could put your ammeter across the battery terminals to measure the thing directly, though if something is very badly wrong, you risk blowing up the meter or its fuse. The resistor test is less risky.
 

 I read 11.61k and 9.8k

 
It's a pretty crude test, so it's hard to say for certain, but that result doesn't jump out at me and scream "FAIL!". The actual values depend on details of your circuit build and, crucially, on how your meter works. Without a result down in the 100 ohms or less range, I think you've ruled out a dead short.
 
By the way, the 9.8k is your RLED, in all likelihood. (The result I was looking for is the 11.61k.) If so, and you used a 10k resistor, then either:
 
a) you didn't use a 1% resistor
b) your resistor is out of spec, or
c) your meter is el crap
 
 

just blue led 

 
Most blue LEDs are at least 3.7V, and some go as high as 5V.
 
You can test whether your battery voltage is too low by running from your wall power supply, which should be considerably higher voltage than your battery. If all the problems clear up on wall power, then you know you just don't have enough voltage to do what you're asking of the amp.
 
 

is there a method to know V of those with the meter?

 
Few meters do a good job of testing Vf of LEDs. Many meters don't test with enough voltage to light the LED at all, and some that will light the LED don't give the right measurement. Diode testers on DMMs are typically intended to test regular diodes only, not LEDs.

 

[tent]

Ok at the moment I measured voltage unloaded and it was 8.05V but if I turn the pimetav2 on the voltage will quickly go down (the battery was almost empty) so not really able to tell the voltage drop difference when turning BIAS on/off.. so definitely need to charge the battery first..
 
I tried to put an 1ohm 10W resistor (the available one) on the battery and without wall wart plugged it measured something like 0.2V but if I plugg the wall wart (12V linear, isolated) the voltage drops to just 0.02 and decreases.. so I disconnected the resistor... the voltage of the battery goes back up to something like 6.56V and stays there.. if I plug the wall wart like that in the wall it will rise the voltage of the battery but if I disconnect it it will quite obviously go back to 6.56V.. so now I'm not sure if I should/could let it be charging overnight like that or not.. it seems to behave ok.. but...?? (I've also tried a 33ohm resistor, 10w I also had lying around but also that one drops the voltage to 0.02V c.a. while wallwart is powered... should I wait more time?)
 
About the BIAS and RLED: not sure what you really want to measure, but I confirm my rled is 9.8k (I measured that too).. do you think it is too few and I should change the resistor to one that is more stick to 10k? (testing it before with the meter maybe)
In addition I'm using another meter and measured also from the other led pin to V- and I read 11.78V (9.79V still on the other led pin). So again, the meter seems ok, maybe my resistor is not precise enough and I should change it?
 
Finally I change the two leds putting there some "old" leds found in my parts bin, that looked to me being of the ones of the good old day's.. green in color.. now when I turn on the pimetav2 also led2 shines.. ok.. led1 now makes very few difference in brightness when turning BIAS switch on and off.. still does tough.. but the proböem atm is that everything fades quickly since battery is exhausted.. so I think I need to concentrate further on how to charge (or better dare to charge) the battery...
 
tent:wq

 

[tangent]

Quote:

I tried to put an 1ohm 10W resistor (the available one) on the battery

 
I meant the amp's battery terminals, not the terminals of the actual battery.
 
 

do you think it is too few and I should change the resistor to one that is more stick to 10k?

 
The only reason I brought it up is to get you to indicate whether this is a trustworthy meter or not. It affects how much I'm going to second-guess your measurements.
 
Anyway, forget the test we started off with. As I said, it was a crude test. The only way for it to have incorrectly failed to indicate a short is for it to be actually broken.
 
I think the real problem is the discharged NiMH.
 
I don't want to tell you to charge it in the PIMETA until I hear whether you did the 1 ohm resistor test across the battery itself or — as I intended — across the amp with only the wall supply powering it.

 

[tent]

ok, maybe I'm missing then what you mean by testing the battery terminals at the amp... do you mean measuring difference between V- and V+? Or just on the "batt" 4 pins plug? I really think there is no difference between measuring it at the battery terminals o at the pins.. anyway I've now charged the battery using an external charger and I've measured absolutely NO difference in voltage (still at the battery terminals) when switching BIAS on or off... so is this clarifying maybe some doubt? for the rest I'm willing to retry the 1ohm resistor test as soon as I understand how to do it. Anyway looking at the schematic of the charging part it really seems simple and failure proof, so since I do not see any evident error there it should be able to charge the battery imho.. would it help if I post some photos?
 
tent:wq
 

 

[tent]

I also wanted to point out again that now that I'm using two green "normal" or low voltage leds the problem of the green led1 to change brightness is not there anymore..
So now for example I have full battery (giving 9.24V when amp is off, 9.04V when on and going down slowly, like 0.01V each 10seconds), and the led1 is same brighness with both BIAS settings, led2 is clearly on when BIAS on and off when BIAS off.. I also managed to set 1V trough RBLIM (1k) rotating the RBIAS, so seems good news and problem solved on this side..
 
But problem still is this V- to OG being different than V+ to OG (-5.53V vs 2.55V). I also tryied to change the OPAG to see if something changes.. but nothing..
 
So could we say it is not a problem of some sort of short but something else instead?
 
tent:wq
PS: (I also have 11.05V and 9.78V still between led1 legA and legB and V-, but you confirm no need to change that rled 9.8kohm resistor, right?)
 

 

[tent]

ok sorry for posting so much but I dared to make some further investigation step: I tried to measure the difference between V-/V+ and OG also with the amp powered with the wall wart (12V) and indeed some further odd result comes out: -11.75V vs 2.55V (curiously enough the 2.55V is same as with battery only).. I really start thinking something is wrong in the virtual ground rail splitter maybe?
 
tent:wq

 

[tangent]

Quote:

I really think there is no difference between measuring it at the battery terminals o at the pins.. 

 
That's right. It's simply easier to alligator-clip a resistor to the amp's battery terminals than to clip it to the header pins.
 
I just want to make sure you're not connecting the resistor across the battery itself, either outside the amp or in parallel with the battery installed in the amp. We're trying to test the charge circuit separately, without the battery getting in the way, confusing things.
 
 

I've now charged the battery using an external charger and I've measured absolutely NO difference in voltage (still at the battery terminals) when switching BIAS on or off

 
 
Good.
 
Quote:

led1 is same brighness with both BIAS settings, led2 is clearly on when BIAS on and off when BIAS off.

 
Also good news. Now the question is, will the problem return with the blue LEDs reinstalled, indicating it wasn't the battery after all? I think not, but if you're happy with green...
 
 

PS: (I also have 11.05V and 9.78V still between led1 legA and legB and V-, but you confirm no need to change that rled 9.8kohm resistor, right?)

 
Yes, and no need to keep repeating this crude test. There's no more juice to be squeezed from that stone.
 
 
Quote:

further odd result comes out: -11.75V vs 2.55V (curiously enough the 2.55V is same as with battery only).. I really start thinking something is wrong in the virtual ground rail splitter maybe?

 
That is indeed very odd.
 

would it help if I post some photos?

 
Probably. High-res pics of the board, both sides, please.

 

[akselo]

Hi
 
I am having some serious trouble with my Pimta v2 build and was wondering if some of you could give me some hints on how to troubleshoot. The amp worked fine for quite a while and sounded great until suddenly the left channel died. 
 
I used the following comonent values: 
 
R1-R7: schematic values
R11=3.09k
RLED=10k
LED1: Vf=1.8 v
 
C1=0.47 uF
C2=2*680 uF
C3 and C5=0.1 uF
C4=6.8 uF
C6=10 pF
 
OPALR: BB OPA2132
OPAG: BB OPA134PA
 
Vol.pot.: Alps RK127
 
Measurements discribed in the troubleshooting guide:
 
Vg+=11.61 V for all opamps and buffers
Vg-= 11.51 V for all opamps and buffers
V+-= 23.12 V for all opamps and buffers
 
Offset Vlg=4.6mV
Offset Vrg=3.9mV
 
Input voltage= 23.8 V
Supply voltage(no load)= 23.89 V
 
Quiesent current Iq=117 mA
Expected Iq= around 50-60 mA
 
The bias circuit is not active for the moment.
 
Headphones: AKG K702
 
I don't know what I was thinking when I selected the value for R11. This is way to low so I'm afraid that I've fried one or more buffers as the current limit is set way to high. Is there any way to check this or do I just have to get new ones? Could this have caused damage to the opamps as well?
 
I read somwhere on this forum that the volumepot has some metal parts underneath that can cause a shortcircuit so I decided to desolder this. I'm a little bit afraid that I ruined it during the process as it only reads 42.8k and 42.5k on the lowest setting. Isn't this supposed to be 50k? Should I replace this or is it okay?
I also discovered that I used 35 v caps for C2 and 50 V caps for C1. Is there any disadvantages to this?
 
Here are links to a few pictures of it:
https://www.dropbox.com/s/5dn6jonrjdw5aup/_DSC1938.jpg
https://www.dropbox.com/s/ymvcowat66l2nai/_DSC1936.jpg
https://www.dropbox.com/s/39ddnk0v6ibd8dc/_DSC1935.jpg
 
Any help would be greatly appreciated.

 

[tent]

Quote:

 
That's right. It's simply easier to alligator-clip a resistor to the amp's battery terminals than to clip it to the header pins.
 
I just want to make sure you're not connecting the resistor across the battery itself, either outside the amp or in parallel with the battery installed in the amp. We're trying to test the charge circuit separately, without the battery getting in the way, confusing things.
 
 
 
 
Good.
 
 
Also good news. Now the question is, will the problem return with the blue LEDs reinstalled, indicating it wasn't the battery after all? I think not, but if you're happy with green...
 
 
 
Yes, and no need to keep repeating this crude test. There's no more juice to be squeezed from that stone.
 
 
 
That is indeed very odd.
 
 
Probably. High-res pics of the board, both sides, please.

 
I've attached some pics.. but let me know if I should upload them in some separate image hosting, zoom in some specific place or take it out from the cables.
 
 

 

 

 

 
But still wondering how it could be that just on V+ there is a fixed voltage value whatever I use 8.4V or 12V.....
 
tent:wq

 

[tangent]

Quote:

R11=3.09k
...
I don't know what I was thinking when I selected the value for R11. This is way to low so I'm afraid that I've fried one or more buffers as the current limit is set way to high.

 
Possible.
 
The best argument against it, though, is that it's typically the right channel that gets damaged due to [un]plugging headphones while music is playing, since the ring contact in a TRS plug is right next to the sleeve (ground) lug, so the right channel is momentarily dead-short connected to ground when you [un]plug normally.
 
You're certain you didn't swap the channels? I've done that before. If you did, too, that would help me believe this is the explanation.
 
I didn't figure out the channel swap in my amp until I used the amp with a computer in a FPS game, and realized the sounds were coming from the wrong direction.
 
 

Quiesent current Iq=117 mA

 
So is the left channel buffer hotter than the right?
 
Try a spit test first: lick your finger, tap the top of the chip with your fingertip, and guesstimate how long it takes the spit to evaporate. Repeat a few times. If it evaporates faster on BUFL than BUFR, BUFL is considerably hotter, and you've figured this out without burning your finger or breaking out the thermocouple.
 
 

Could this have caused damage to the opamps as well?

 
No.
 
I don't mean the op-amps couldn't be damaged, but they can't be damaged purely because R11 is too high. Something else would have had to happen to them.
 
 

the volumepot has some metal parts underneath that can cause a shortcircuit 

 
MMmmmmayyyybe. I've just tried to reproduce it here, and it's not happening.
 
In any case, it doesn't explain your high Iq.
 
When you reinstall it, it wouldn't hurt to be sure you've got a bit of an air gap below it. Or, I suppose you could Dremel off those outer "feet" surrounding each PCB lug.
 
 

I'm a little bit afraid that I ruined it during the process as it only reads 42.8k and 42.5k on the lowest setting.

 
That's lower than what I get here (more like 47k) but the absolute value of the pot doesn't matter much.
 
What matters is that it give the proper voltage divisions, and that these be well-matched. So, put a lowish voltage across both gangs of the pot (the outside lugs) and measure the voltage at the center pin relative to one of the outer pins. It doesn't matter which way you measure, as long as it's consistent throughout the test sequence. The voltage source is ideally an adjustable bench supply so you can dial in something easy to work with like 1.0V or 10.0V, but any small battery will suffice. It could also be a test tone generator.
 
For each measurement, compute the decibel difference: dB = 20*log(V1/V2). If it's consistently under 3dB difference except down near the bottom, the pot's fine.
 
 

I also discovered that I used 35 v caps for C2 and 50 V caps for C1. Is there any disadvantages to this?

 
Nope.
 
For audio path caps like C1, you'd ideally want a very low voltage, since they'll never see more than about 5V peak-to-peak, but you can't find film caps that low in voltage tolerance. Even 50's are a bit hard to find. 63V, 100V, and 200V are more common for the low end of a cap line.
 
 

Here are links to a few pictures of it:

 
I don't have any comments of consequence, but here they are anyway.
 

• Nice soldering job, clear pics. Thanks for not making it hard on us.
• What are the turret pins you're using for the I/O connectors? Are you using a mating connector, or soldering to them?
• Why did you populate RCCS? IC1 isn't populated, so you must not be using the trickle charger. Also, you've used a 1/4W resistor here, so it's probably too small to be used with a trickle charger anyway.
• Next time you use those Vishay-Dales, you might consider bending the leads so the values are face-up. You don't have to worry about this with stripe-coded resistors, but with these... I also like to make sure resistors oriented the same way have their labels all facing the same way, too.

 

[tangent]

Quote:

I've attached some pics.. but let me know if I should upload them in some separate image hosting, zoom in some specific place or take it out from the cables.

 
The current pics zoom in far enough. I was a little weirded out when I saw them...there are no IKEA PIMETAs!
 
The haphazardness of your build worries me, but I don't see anything specific that could cause your problems. It just makes me wonder how solid it all is. (Bent over caps, LEDs flying around on stilts, melted caps, flux residue...)
 
Check underneath that white wire that obscures the area between IC2 and C3-. There could be a solder short there, but I can't tell with the wire in the way. It could explain everything.
 
I assume those green SMT caps across R4 for some kind of bass boost? Are they at least PPS film? If they're ceramics, they're not going to be doing your sound quality any favors.

 

[akselo]

You're certain you didn't swap the channels? I've done that before. If you did, too, that would help me believe this is the explanation.

I checked and it's connected correctly but it might acctually have been the left channel. I decided to check again earlier this day, and now both channels are dead
 
 
 

So is the left channel buffer hotter than the right?

I tried the spit test, but didn't see any big difference between the two. All three buffers gets very hot and burns my finger if I keep it on there for more than a few seconds.  
 
 
 

What are the turret pins you're using for the I/O connectors? Are you using a mating connector, or soldering to them?

Here are links to the pins and connectors I used:
https://www.elfaelektronikk.no/elfa3~no_en/elfa/init.do?item=48-924-85&toc=0&q=48-924-85
https://www.elfaelektronikk.no/elfa3~no_en/elfa/init.do?item=48-924-93&toc=0&q=48-924-93
 
I think i will swap these for molex connectors as I get metal particles on the board everytime i connect/disconnect them from the board. Is it possible to make the molex plugs without a crimping tool?
 
 
 

Why did you populate RCCS? IC1 isn't populated, so you must not be using the trickle charger. Also, you've used a 1/4W resistor here, so it's probably too small to be used with a trickle charger anyway.

This was one of the first things I noticed as I started the troubleshooting I built this three years ago just as I started my power engineering education so I wasn't very good at reading schematics at the time When the amp failed I put it away and forgot about it until now. Too bad I suck at electronics. We have just had a look at power electronics like frequency converters, rectifiers and inverters.
 
Should I try to swap the buffers and R11s for new ones or do you have any other suggestions?

 

[tangent]

Quote:

Is it possible to make the molex plugs without a crimping tool?

 
Sure.
 
The advantage of the crimping tool is how it curls the ends of the contacts around so they kind of weave into the stranded wire you're using. If you use needle-nose pliers instead, you still get a decent connection, just it's a flat connection, so the surface area in common between the contact and the wire strands is lower. This means it won't be as strong and might be measurably higher in resistance. You can fix that by flowing a little solder in there, if you're concerned about it.
 
On the other hand, the tool I use — Molex part # 63811-1000 — is cheap enough that it's cost me about $0.01 per crimp now, and there's no reason to believe it won't work for thousands more crimps.
 
 

Should I try to swap the buffers and R11s for new ones or do you have any other suggestions?

 
You could try removing the buffers and then shorting them from pin 3 to 6 with a bit of resistor lead clipping. That turns the amp into something halfway in between a CMoy and a MINT. If it starts working beautifully, you know at least one of the buffers isn't a buffer any more. If the problem remains, you can put them back one at a time, measuring the amp's Iq each time to see if just one of them shoots the amp's current draw up unduly.

 

[MisterX]

I like GC/Waldom/Molex # W-HT-1921-P for KK pins but to each their own.
 
Trying to find the pin 1 indicators for the buffers on the .png image of the PCB is easier said then done and they're covered once the buffers are soldered in place....
 
http://tangentsoft.net/audio/pimeta2/bitmaps/brd-2.01.png
 
Maybe actual pictures of the top and bottom of the PCB would help?

 

[tent]

Quote:

 
The current pics zoom in far enough. I was a little weirded out when I saw them...there are no IKEA PIMETAs!
 
The haphazardness of your build worries me, but I don't see anything specific that could cause your problems. It just makes me wonder how solid it all is. (Bent over caps, LEDs flying around on stilts, melted caps, flux residue...)
 
Check underneath that white wire that obscures the area between IC2 and C3-. There could be a solder short there, but I can't tell with the wire in the way. It could explain everything.
 
I assume those green SMT caps across R4 for some kind of bass boost? Are they at least PPS film? If they're ceramics, they're not going to be doing your sound quality any favors.

 
I checked carefully and I really do not find any short or similar.. no contact evident between IC2 and C3.. what could I've done wrong?
(The green caps are very good caps for the C6 positions, Im using 12pf and I suppose it is correct to put them across R6 positions.. the bass boost instead also has good film caps, but are big yellow ones across R4.. my first test hearing tests are impressive tough!)
 
tent:wq