GSSL Finished - Sounds bad :-(

[Hulk]

I maybe have found the problem but can I make accurate measurements on resistors while they are solderen onto the PCB or do they have to be unsoldered ?? I measure on each leg on each side of the resistor.

The 1M resistor in circuit the sidechain only shows 815k
The 1M8 resistor on the 4:1 ratio only shows 697k
The 3M9 resistor on the 10:1 ratio only shows 672k

These measurements are witho out the power turned on!

 

[gyraf]

You can't measure the resistors when they're soldered in. You need to lift one end to measure.

Is the wire link on the control pcb in place?

Jakob E.

 

[bcarso]

Hulk, it's hard to get accurate readings of large value R's in circuit even with power off. So I doubt that those are the problem.

If you have a situation where one R has a capacitor (especially a film or ceramic C) in series, you may be able to accurately check that value, but in general anything else connected will reduce the local impedance. Other R's in parallel will do it, and semiconductors may have enough bias depending on your ohmmeter to begin to conduct and also cause a lower reading.

Worse, you may have a still-charged cap from when the power was once on, that is slowly discharging with the power off and can give rise to wildly erroneous readings, including "negative" resistances.

Another remark: resistors are rarely off by much unless damaged. There is the odd excessively noisy one, but rarely ones way off of stated value.

To see if you are getting some loading effects from other components in the circuit with the power off, try reversing the leads of your ohmmeter. If you get a far different reading, then safe to say neither should be trusted.

 

[Hulk]

ok thanks bcarso

 

[DrFrankencopter]

Double check the polarity of the diodes in the sidechain...you'd get some pretty unpredictable results if one or two of them went in backwards (like gobs of even order distortion).

Cheers,

Kris

 

[Hulk]

I'm pretty shure I have the sidechain diodes in the right way but I have taken a few pictures for you to see maybe you'll spot something wrong!!

 

[Hulk]

You can also see the pictures here full size for more detail!!

 

[Hulk]

hhmm I think I'm going to make a new PCB with new parts - I have looked for shorts, measured and checked the schematics a hundred times with no results :?

 

[gyraf]

Are your tantalums at the control pcb oriented the right way around?

If you have access to an oscilloscope, try tracing the sidechain: the summer, the rectifier, the timing, and finally the offset added by makeup/threshold..

Jakob E.

 

[Hulk]

Gyraf aka. falkeøje (Hawk's eye) :green:

I had turned every singel tantalum the wrong way :shock: when I soldered them on the PCB I thought the little mark on them was a + symbol but it's a - symbol :roll:

But.... it still sound bad Would the tantalums be damaged be this little stunt or anything else for that matter??

 

[gyraf]

The tantalums could very well be damaged.

The longest pin (before cutting) is the positive.

Jakob E.

 

[Svart]

yes, tantalums rarely survive reverse biasing, depending on voltage of course, I would try new ones..

:thumb:

 

[Hulk]

I'm just going over the SSL PCB once again and I have unsoldered 2 of the sidechain diodes and tried to test them using my multimeter - I'm using the position on the multimeter with a diode symbol on it (like the diode IN4148 symbols on the PCB) and when there's a connection the meter gives out a bussing sound but no matter what direction I test the IN4148 diodes I cannot get the bussing sound from the meter - I know that sound really retarded but shouldn't I be able to get a buss :green:

 

[gyraf]

Normally no buzz for diodes, but a reading of some 600mV in one direction, nothing in the other direction..

 

[Hulk]

Thanks again Gyraf!!

I just tried to measure the diodes I got no voltages in one direction but when I measured the other direction my brand new multimeter send out a really loud "bang" and big fat puff of smoke and now it's dead :shock: What the h... was that???

 

[gyraf]

...A case of repair-under-warranty?