Installed a new LM4765 power amp chip and .... nothing.  1kHz sinewave input, crank the knobs all the way and just get a small distorted signal at the output.  Hmmmm.

Ooops.  The LM4765 tab was supposed to be isolated from ground.  Sure enough, there is a little plastic isolation bushing lying on my workbench from the disassembly, it has been so many months since I took it apart I forgot the bushing.  So I killed the new amp chip.   Slap forehead.

Ordered another new LM4765, installed properly and .... it works!  Then I had to download a Line 6 Pilot's Manual to figure out what the buttons and knobs do, but anyway its all working fine now.   I had been so frustrated with this repair, I was ready to toss it.  

But now, I will keep it around for a while, it really has decent sound.  I would give high marks to the clean and twang models, the blues and crunch are OK, but not so much to the other two models.  You can access all of the amp's features from the front panel, by pressing various combinations of buttons, but that's not something you want to mess with on stage.  If you don't have the manual printed out, every time you change a model the knobs settings are reloaded from the model in memory, so looking at the knobs will not tell you where things are set.  To find out, you have to press TAP, spin the model knob to let it know what you want to find out, then spin each tone and effect knob one at a time until LEDs tell you to stop spinning, then you know where the model equals the knob setting.  So like a lot of digital amps, you really need the footpedal thing if you are going to do a lot of model changing.

While I was waiting for parts to fix this amp,  I acquired another Yamaha DG-80, it arrived on the same day that I got the Line 6 Spider working.  After fooling around with the Line 6 for a while, I really appreciate the motorized pots on the Yamaha, when you change the model the software runs the motors, turns the pots to the new settings.  The Yamaha is just easier to set up, easier to store your custom models in memory, it stores more, and basically they just did a much better job of modelling;  the Line 6 definitely has limited character, with the Yamaha you can find more of everything.

I installed the new CS4221 chip I got from Ed, and I now get analog signals all the way to the inputs of the power amp circuitry.  Hooray. The signal seems to respond to bass/middle/treble input knobs vs. input frequency, and amplitude follows the volume knobs as expected.   I have today ordered a new LM4765 chip and will report back when that is installed.  So far so good.

for bench testing circuits, an old lawnmower or motorcycle finned cylinder head works fine as a heatsink.  Bob Pease used a VW cylinder head when testing new chips.

guitardog,  I apologize that I have not been monitoring this site for some time.  The Spider was removed from my workbench and sat in the corner, I worked on other things.  Yesterday Ed (Edwardeffect1) called to let me know that a CS4221 chip is on its way to me (Thanks Ed!) so I probably should get back to work on this amp.

I traced out a lot of the Spider pcb last year while attempting to debug this amp. The Spider component arrangement does bear a slight resemblance to the National LM4765 data sheet Figure 2, from the output pins of the chip to the speaker.  The input and feedback is a lot different.  

Crystal recommended that the CS4221 outputs go through a 2nd or 3rd order Butterworth.  Line6 didn't bother, they just AC coupled the CS4221 differential outputs to the inputs of the LM4765.  On the pcb there are a few places where they could populate caps to form various 1st order RC filters, like they could have made a lowpass in the LM4765 feedback path.  However, on my pcb all of these capacitors are not even populated, so the LM4765 is just a gain stage.  

If you can make a differential audio signal with similar signal levels to the Crystal chip, here is how you could connect it into their circuit:

AUDL- couples through C28, AUDL+ couples through C27.  AUDR- couples through C19, AUDR+  couples through C18.  You can follow the traces to see which capacitor pin is the in, which is out, then you can lift the inputs or cut the input traces and wire in your signals. The signal output from each cap heads over towards the LM4765, going through a 10K resistor on its way there.

If it would help, I drew the power amp section of the schematic.

thank you Ed, that is a very kind offer.  I set the Line 6 repair aside for now, I have more junk...errrr....valuable equipment needing repairs.  So I am not in a hurry to purchase parts for the Line 6, I will wait to hear from you, if you have spares when you finish your repair.

a signal injecting probe is a good tool but not much use on the programmable digital components, I think you just got lucky Ed.  In this kind of circuit, you need to know if the chip did get all of its necessary signals, i.e., did it get programmed correctly?  If it did get told what to do, and then it doesn't work, then yes that chip is bad.  But I am not able to tell if its getting programmed correctly. 
In the Line6, injecting a standard analog signal, like a sine wave,  is about all you can do and that only gets you past the opamps in the pseudo-stereo preamp stage.  As soon as the CS4221 digitizes it, all you have are ones and zeros going to the DSP and then digital back from the DSP.  If that works, OK fine, you will then find some analog coming out of the CS4221 and going to the power amp stage.  I know about the power amp stuff, I saw it wasn't getting any input.  Maybe that is what died first, and maybe there was an electrical tsunami that zapped the CS4221 too, I have no way to tell. All I am able to tell is that there is activity (stuff happening) on the pins going to the CS4221 chip, I am not seeing anything like what I expect on the pins coming out of the chip.  So without anything more scientific, I am just guessing/hoping this its a dead chip. I am replacing the CS4221 and the power amp.

I have not seen problems with the motorized pots.  I think that maybe this is because the pots are not usually being yanked around by human hands, instead they are usually moved smoothly by the motors.

for broken power  (fat) traces, I scrape off the photoresist with a scalpel type X-acto to expose clean copper, then solder some Solder-Wick across the break.  Smaller traces I use solid wire, typically 30AWG wire-wrap wire to repair 8mil to 12mil traces.  Clean the board with a good flux remover like Techspray 1631, then lock the wire down with superglue.  If its a long wire, tack-solder one end, then route it by bending with tweezers, after it goes a small distance you use a tiny piece of Scotch tape to clamp it down, then continue routing it, continue with Scotch tape until it ends up where you want it.  Then superglue it down between the pieces of Scotch tape.  After it dries pull off the tape, you have a nice tidily routed hackwire.

Sometimes when removing a surface mount component, the old part yanks a pad off the board.  Check the part datasheet, usually this happens at pins which have no function (NC), so the pad has no trace to take away the heat.  This is no problem, bend the NC pin up and solder the new part down with those pins unsoldered.  But sometimes chip removal yanks off a good pad and some of the attached copper trace.  For some packages, like SOICs or flatpacks, and even J-leads like PLCCs, this is easily fixable, scrape the remaining pcb trace and solder on a new piece of ww wire which ends up with a stripped end being in the location of the old pad, bend the pin of the new part up and place the part. Solder it down, finally rebending the pin down to land on your new wire, solder that down on top of your hackwire.  About the smallest I have done is TQFPs with 0.4mm pin spacing, you need about a 30X microscope to do it.  No coffee.

The DG80 power supply circuit board is not well supported.  If the amplifier is dropped, the weight of the transformer and other heavy components will crack the circuit board.  Next time you have the amp apart, add some insulated standoffs/braces to the power board.

Ed,  thank you for posting.  I had seen your thread but did not have any information to contribute.  But since you helped me, I did some more searching for Flextone and I found a site which has some additional Line 6 information.  I see Flextone is mentioned in various service bulletins.  Try here:

http://www.schematicsforfree.com/archive/dir/Audio/Musical/PRODUCTS/LINE%206

Hope that helps.

I have had mixed results with the faraway chip resellers.  I would like to ask if you would update us on whether you do receive your chips as promised.  I set the Line 6 aside, I will probably not work on it again until next month.

scoop

Line 6 Spider, no output.  Found some downloadable schematics but only for the later versions.
Here is what I have done so far.  Questions at the end.

There is some similarity between the early Line 6 Spider circuit design and the later ones.  At the block diagram level its a lot like the Spider II 212. The actual schematic and layout of the mainboard is different, although the circuitry has similar circuit topology and identical components for a lot of it.  The preamp, codec, processor, DSP, memory, and front panel interface have enough similarity that the Spider II 212 literature is helpful for debugging.  The power amplifier section is a lot different, the Spider uses one LM4765 as a bridge driver, and the Spider II has two LM3886s, one for each speaker.

The Cirrus Logic CS4221 codec is used in both amps.  In the Spider 1, they ignored the Cirrus Logic recommended Butterworth output filter, just connected the left outputs to one half of the LM4765 and the right outputs to the other half.  The II does have additional filtering before the output amps.  The Spider II has an op amp buffer for the headphones, I think the Spider just taps off the LM4765 outputs.

Line 6 service bulletin 018 mentions this amp.  It states that C24 and C52 were not installed on production units.  If they were installed, then look for "smoke/burning around LM 4765 Power Amp IC U3".  It notes that if the unit has smoked, then U3 is probably dead.

So OK, my amp is dead, and the pcb does have C24 and C52 installed, so maybe my U3 is dead.  These caps look like ceramic 1206s, they are just bypass caps in parallel with a couple of 50V electrolytics from the Vcc/Vee rails to ground.  I figure they were probably not rated for this much voltage.  What the heck, I unsoldered the LM7465 and tossed it, and removed those caps.

I should be able to confirm that at least there is audio signal going to the power amp.  I traced out this part of the pcb. The CS4221 data sheet specifies a minimum of 10K resistive and a maximum of 100pf capacitive load on its analog outputs, so I was looking for something like that.  Instead I find the CS4221 outputs are AC coupled through 1uF caps to the LM7465 circuitry.   The LM7465 is arranged as a differential amp, with a 620K feedback resistor and a 620K resistor to ground from the noninverting input, and 10k input resistors to the coupling caps.  The AUDR- signal goes to C19, AUDR+ goes to C18,  AUDL- goes to C28, AUDL+ goes to C27.  With the LM7465 out of circuit, there is almost no loading, so I patched in a resistive test load across these differential outputs.  I attached probes to these nodes.  Ooops, maybe I should not have thrown out that LM7465 chip, because there is no audio showing up at these nodes, i.e. the 4221 was not sending any signal to the power amp section.

Back at the CS4221 , I checked the clocks.  Sampling frequency is the same as LRCK, the Left/right clock, it is 31.197kHz. SCLK measures 1.9966MHz, MCLK measures 11.979MHz.  So MCLK is 384Fs. Because there is a 47Kohm pulldown on the SDOUT pin, the 4221 wakes up in the clock master mode, it generates SCLK and LRCK from the MCLK input.  Also, it defaults to LRCK = MCLK/256, it has to later be programmed to change it to 384, so this indicates that the microprocessor was able to communicate with the 4221.  On the pcb, the ADC signal is the output from the 4221 DAC to the DSP chip ADC input, and the DAC signal is the DAC data from the DSP chip to the 4221's internal DAC.  The DAC signal from the DSP does show a small amount of activity suggesting that the DSP chip is trying to do something, you can see some changes in the serial data when wiggling various front panel knobs.  In particular, turning on the tremolo causes the MSB to shift back and forth.  The ADC signal never changes, you see a couple of bit transitions during the left data sample and a similar couple of bits during the right data sample.  Nothing affects this pattern, it does not respond to input signal level, or whether the input is plugged or unplugged.  The analog signals to the 4221 are fine, the clean signal is on pin 17 and the boosted signal is on pin 20.

A lot of circuit functionality can be tested using the built-in microprocessor.  I have done the reset, checked EPROM version, I have run through all the tests which can be done through the front panel without the additional Line 6 test fixture, everything tests OK.  However, without the test fixture, apparently I can not run any audio ADC or RAM tests.  The front panel tests only exercise the accessory TLC542 ADC which is used to measure the front panel pots and to detect whether the guitar or headphones are plugged in, it does not test the ADC and DAC in the codec nor the ones in the DSP chip, it does not test the DSP memory.  In these amps, signal level (volume) is controlled digitally within the 4221 codec. 

I intend to try confirming that the uP is attempting to change the signal level when the knobs are twisted.  If I rearrange everything in my shop, I could get this mess over to one of the digital workbenches where I could hook up a logic analyzer and do some more probing.  I have tried just looking at the serial communications between the uP and the 4221 with a scope, but I think it must be very asynchronous,  does not update the 4221 control register except when the front panel buttons are touched, and my scope can't catch it.  For now, all I can say is that it seems too quiet on those pins, I expected more activity than what I am seeing.

Now to the questions/requests.

a. Do you have or know where I can purchase any of the CS4221-KS parts?
b. Have you debugged and repaired one of these amps?  What would you look for next?
c. Do you have a Spider text fixture, or the schematic for a test fixture?

the cardboard tube Silvertone reverb should have an inductive driver transducer and piezo receiver transducer.   One side of the driver should be grounded.  Follow the wiring back to the reverb driving circuitry, check those transistors in the drive circuit.

I got a 1464 schematic from Musicparts.  They have the 1465 schematic reprint for $15.

http://www.musicparts.com/products.asp?Company=Sears

The reprint I got was fairly readable even though it was all A size. It included all of the Sears technical manual, not just the schematic pages of the manual.  The parts list helped a lot, because there are times when you can't identify a component, either it has unreadable markings or the schematic is unreadable.  So it helps when the documentation tells you, for example, that C9 is a .047mfd mylar 200V cap.

I have owned a couple of Stages, I still have one Stage and one Pro, and I have repaired quite a few of them for friends.  Every one of them had dead or dying caps and diodes in the power supply circuit, a couple of them had dying/dead/broken power resistors in the output stages, and two of them had broken input jacks.

Start out by replacing the rectifier diodes and filter caps, just do it.  Then, before powering it up again, take a good look at the cement power resistors, you may find one has come unsoldered, or it has a broken lead.  I would unsolder one lead of each big low-ohm power resistor and confirm its resistive value (measure with a precision ohmmeter) or just shotgun all of them with new resistors.  Inspect the bottom (solder side) of the pcb with a magnifier and bright light, you are likely to find a bad solder joint and possibly a cracked trace.  If a lead or pcb trace opened where high current was flowing, you could find a burned or cracked trace.  High current traces need a lot of copper, if you find cracks then scrape the mask away on both sides and form a bridge across the crack with wire, or you can make a tidy high-current bridge by soldering a piece of solder wick across the crack.

Still, the 185 is one of my favorite Fender amps from that era.  It does have too many knobs and switches, making it difficult to find that perfect tone, you have to play around with the settings for a while to find it.   But its worth fixing, its 35 years old but it will sound great 35 years from now if you just take care of it.