TSA8 amp troubleshooting

[e093116]

Hello,

I have an LS3 with a TSA8 amp that sounds distorted/muffled with little high frequency output on one channel.

I believe the problem is on one of the voltage amplifier boards, as when I swap positions of the boards within the amp, the problem moves to the other channel. To me, this implies that the signal inputs, output transistors, stereo crossover, and speakers/wiring are in good shape.  I do get great sound out of the "good" channel/board in either installed position within the amp.  I did recap the electrolytic capacitors in the amp.  I'm excited to be able to hear this thing with both channels working properly...I think it will sound great, and my wife and kids are getting anxious to listen to the new records we've purchased recently!

The boards are "-4s".  I was able to adjust the bias on both boards, in each installed position, to the specified less than 40 mV (I set them around 39 mV).

Unfortunately, I don't have the equipment to perform the test/measurements specified in the amp schematic note 2.  However, I'd like to perform the DC voltage checks specified in note 3.

Is a typical 1/4W 560 ohm resistor acceptable to be used for this test?  I would remove the input wire from the circuit board, then connect this resistor across the input and ground connections on the circuit board (leaving the ground wire connected), connect the negative lead of the meter to the same ground location, and then connect the positive lead of the meter to the location of interest on the board.

I'd probably test each board individually in the "left channel" position, since access is limited to the other board by the AVC board on top. 

Any comments/concerns on the proposed test setup, recommendations on components to carefully inspect/test, or other things to consider would be greatly appreciated.

Regards,

Bill

 

[Ron Rich]

Hi Bill,
On those, I generally just check components individually--usually a dual diode has failed- after caps of course--
I think the 1/4 watt resistor should be fine--- Ron Rich

[Ron Rich]

Aug 3, 2017 16:13:57 GMT Ron Rich said:

Hi Bill,-
On those, I generally just check components individually- caps 1st., of course--
I think the 1/4 watt resistor should be fine--- Ron Rich

[robnyc]

Bill, it is difficult to be specific as to tests without a signal generator. I'll give soem thoughts:

"Is a typical 1/4W 560 ohm resistor acceptable to be used for this test?" -That is fine, but all it does is turn on the first transistor. It has only slight effect on the DC readings of the transistors immediately following it.

You can localize the deficient area to either the preamp vs. voltage-drive stage on each card by sampling the audio signal at the gain pots near the front of the amp. At these points the signal will be taken ahead of the volume control.

For these tests I suggest grounding the AGC at it's time constant which is the 100MFD on the AGC board. The will help in removing any effect the AGC diodes might have on the tests.

Since you have two identical preamp boards I suggest comparing the resistors and transistor C-B-E resistance readings. Remove the boards for this.

Given that you changed components on the boards, it is good to do a visual comparison to check for any mistakes...you've probably done this already.

You will probably find that the deficiency exists in the first two stages. if you are lucky, it will be a transistor with failing junction, but I have recently had failed components in the packaged circuit, though these affected EQ balance not DC voltage.

There are a number of sites offering test tones in downloadable files. In the absence of a generator these can be played on a computer and fed through the amp for response comparison. Bear in-mind that all tone settings on the comp must be 'flat' or off and level should not exceed approx 5 millivolts at the amp input (the 560 ohm resistor) You should see substantial bass boost and roughly equal level at the aforementioned gain pots. These should be set to the same resistance reading or turned all the way 'up" to effective remove them from loading the signal.

"I was able to adjust the bias on both boards, in each installed position, to the specified less than 40 mV (I set them around 39 mV).' -That was typical of what was specified for the old germanium finals in many different amps in that era and is now, due to internal P-N leakage, way too high. I have several TSa series in service and run them in the 10-12mv range and the SHP's in the 1-2mv range. All that bias does here is keep the transistors turned on during the crossover interval. If you aren't hearing crossover distortion from a 'cold" amp -the bias is high enough. Heatsinks should idle at ambient temp.

RobNYC

[robnyc]

Aug 3, 2017 16:13:57 GMT Ron Rich said:

Hi Bill,
On those, I generally just check components individually--usually a dual diode has failed- after caps of course--
I think the 1/4 watt resistor should be fine--- Ron Rich

Ron, forgive my being contrary (ok, don't )....but if the defect moves with the pre boards, it isn't going to be an AGC diode defect.
Rob

[Ron Rich]

Rob,
Yep--you are correct--I caught that as I hit the send button--thought I had changed that post ? It posted it twice, never deleting the error--Oh, well, it's that kind o day !

[e093116]

Ron, Rob,

Thanks for the timely suggestions.

I was hoping to get after this last weekend, but a charity walk/fundraising event, 2 acres of mowing, and a clutch replacement on a buddy's 66 Mustang got in the way of what I wanted to do!

I'll report back with a better update when I'm able to spend some quality time with the amp...

Regards,

Bill

[e093116]

Ron, Rob,

I did my best to check the components individually between the two boards, and the only discrepancy I could find was that CR5101 on the "good" board measured 0.78V on the diode check function on a Fluke 87 meter (215K on resistance setting) and 1.15V/500K on the "bad" board.

It looks like this component is actually two diodes in series (based on the schematic). I have some 1N4002 diodes and when I measure two of them in series, I get the 1.15V measurement, when I measure one, I get 0.65V. 

So...if I didn't know any better, it seems as though the "good" board has a has "half" shorted CR5101 (in my limited experience/understanding)! 

I dialed down the bias to the recommended 10mV and swapped the board positions and sure enough the bad sounding channel changed with the board position. 

Any further guidance and recommendations would be appreciated.

Bill

[Ron Rich]

Hi Bill,
You may use two 1N4002, or two "signal" (IN914, or equal) diodes as a sub for that stabistor. If the low voltage 100 uf has not been changed on the upper board, and is "paper", change it ! Ron Rich

[robnyc]

Bill, that diode (CR5101)is directly involved with idle bias on the finals. I'd go ahead and replace the faulty diode. Have you tried taking the signal at the gain pots in order to see if the problem is ahead of the volume control, or after on the pre-driver section of that board?

Rob

[e093116]

Hi guys,

Thanks for the feedback.

I purchased a signal generator (HP 3312A) and oscilloscope (Tektronix 2224) yesterday, so I should be able to replace the diode and also do some better testing this weekend.

If I want to check the circled voltages on the schematic, how should I set the signal generator?  The schematic note 2 states ".0017, 1000 Hz".  Does the .0017 mean 1.7 mV p-p?  The signal generator is adjustable from 0-10V p-p into 50 ohms.  Is there anything I need to worry about in terms of impedance matching?

I understanding the part about shorting the 100 MFD cap.

One thing I need to get is a 5 ohm 50W power resistor to put across the output terminals.  The schematic shows mech connector pin 1 connected to signal ground.  Is this a connection I need to make with a jumper (I don't have the amp handy at the moment to verify)?  I'm assuming that I need to turn on the mute relay in order to load the power resistor to complete the measurements.  Is this as simple as putting a jumper across mech connector pins 1 & 3?

Finally, I'd like to get some lead solder when I order the power resistor.  Can you tell me what you use or anything you recommend?

Thanks again for your ongoing assistance.

Bill

[robnyc]

Ok Bill, lets see what we find here:

" Does the .0017 mean 1.7 mV p-p" -Correct.

" Is there anything I need to worry about in terms of impedance matching" -The amp must have a 560 ohm resistor across each input to turn on the first transistor and get signal through. I use a simple pad of 560 with a 100K in series with the gen out into the 560.

"One thing I need to get is a 5 ohm 50W power resistor to put across the output terminals.' -Not really. Any reasonable load from speakers or a resistor of approximately equivalent resistance will do. In fact, if you are just testing the first stages up to the volume control, no load at all is needed as long as the volume control is a min. I just use either a pair of 20W -8 ohm non-inductive resistors across whatever the top wattage is and a speaker with a 70volt tran tapped at 1/2 watt to let me know the generator is stable and going through.

"I'm assuming that I need to turn on the mute relay" -No, the relay does not need to be energized at any point for these tests unless you want to check the mute or AGC squelch in which case just temporarily short the "Reject" screw (white wire) terminal to chassis ground.

For this sort of work I still have a lot of Kester "44" (type QQs-571) but any general purpose solder for electronic work will do. Not acid core.

Good luck, I'll check back later.

RobNYC

[e093116]

Rob, Ron,

Thanks for the help on this one...I got busy and haven't been able to dedicate any time on this issue recently.

Earlier this week, my day job brought me out near an old ex-operator friend that has a lot of spare parts, on a whim I called and stopped by and picked up a "core" TSA8 amp for a nominal charge.  I bought a few other parts from him and we did a good job of winding each other up through the negotiation process, I think it was an enjoyable experience for the both of us.

Anyway, when I got home I snuck out to the shop and swapped over one of the VA boards from the core, and this solved the issue I've been chasing.

So in some ways I took the easy way out (more $$$ than time!) but at least I had the correct area of the amp identified, I just need to get better at troubleshooting problems down to the component level.  This is one of my goals with this "hobby"...I'll have a lot more spare time over the coming winter months.  At least I'm tooled up now for future endeavors, and I have a couple more project phonographs waiting.

Finally, I played a new record I got (Ace of Spades by Motorhead) at high volume to celebrate.

Regards,

Bill