Roland KC-880 Keyboard amp...no output, no service doc's

[nevetslab]

One of my clients sent over a Roland KC-880 Keyboard Combo amp, with no output. I have everything coming thru fine on the headphone output, so as far as getting that far, all is working. I pulled it apart, which requires removing everything....control panel assy, speaker grille, both 12" speakers to find the power xfmr mounted by itself, along with the power amp assembly. It uses three TDA8920BJ Class D amps (heat-sink mounted power amp IC's). Loads of Radial mounted Time Delay fuses, all intact, and on the bottom side of the board, receiving the control signals from the control panel there's a number of SMD xstrs and diodes. I found Q4, SM code LA, which read out as a BF550 PNP SOT23 part, which had failed. Closest I had to that was a BC856, also a SOT23 part. slightly higher voltage/current (-80V/100mA) vs the -40V/25mA part. Installed that, it being the only part I found defective on that assembly.

With it plugged in prior to disassembly, I was only seeing 30W power consumption @ 120V, where the spec sheet calls out 98W...no idea if that's true, first time I've seen one of these.

Put it back together, crossed my fingers and powered up. Still only drawing 30W, and again, no output from the speakers. Everything on the control panel is SMD land. I did just notice the Master Volume pot took a hit, with the control shaft knocked out of axis by 10 deg or maybe 15 deg. I had previously checked the four gang pot for continuity, and all four decks read nominal, as did rotating the knob yield variation in DCR.

I don't have the service manual for this...just the block diagram from the owner's manual.

Anyone have service doc's on this?

KC-880_OM.pdf

Thanks

[nosaj]

How about working backwards with signal input? speakers then power amp? i understand easier said than done, but what else do we have at the moment?
nosaj

[nevetslab]

As luck would have it, only having found that one defect on the power amp, which is mounted to the floor inside the speaker chamber, with the pwr xfmr separately mounted there, I sealed it back up, taking that chance that I had found what was preventing the power amp from running. Bad decision, as that wasn't the only problem...just the one I found. Their harness cables are too short to remove the control panel for probing. And, the bench space to have it all in pieces is presently more space than I have, without adding a temporary work space on wheels, which I guess will be the next step.

Don't Class D amps need a load on them for testing? Or is that just switching power supplies. If I don't need loads on the four power amps, then the work space requirement is easier. I'll have to see if I have enough clips to clip onto the 0.110" and 0.205" female fastons, which push onto each of the spkrs and tweeters. I do have my smaller fx 8 ohm/200W dummy load, so don't need the speakers, though they immediately tell ya if you hit pay dirt in restoring order.

There's three SMD op amps on the Headphone/Master Volume PCB, where one of the two control cables connect to. That board has a multiwire cable connecting to the main control PCB assy, but, neither end of it are able to be unplugged....hard-soldered. RATS!! Not so serviceable in design.

[Jon Snell]

Class D do not require a load to work.
Check your supply rails are there, and then look at the type of amplifier, look up service data on the specific chip and check they are all oscillating.
It may be that with no input signal, there is no oscillation because it is in sleep mode.

[glebert]

Can you locate where the "stereo link" network is? I would check and see if you have signal at that point. Are there protection relays going to the speakers? Interesting that there are three class D amps, the block diagram implies that there would be four.

[nevetslab]

No Protection Relays on the outputs of the Class D amps. These TDA8920BJ Power Amp IC's are dual channel, so two of them (driving the 12" woofers) operate in bridge mode, while the other operates in stereo mode to drive the two tweeters.

This coming week, I'll pull it apart again, and connect the control panel assy to the power amp/pwr xfmr assemblies and see what is and isn't working. We have a Roland Artist Relations office at1 the CenterStaging complex, so I will stop by there to see if I can get any assistance in acquiring the service manual for this unit. I know we have loads of Roland Keyboards and V-sessions drum kits in our inventory (that they own), so hopefully I'll get some assistance.

Onward..........

[glebert]

Just to rule out the simplest thing, double check that the MUTE isn't engaged or some problem with that circuit.

[nevetslab]

The single Mute switch thus far doesn't make any difference, and....all that circuitry is in SMD land, and not so mechanically accessible (PCB muiti-deck board stack), so the service docs are really becoming necessary in this pursuit. When I get the assemblies back out of the cabinet, I'll see if there's anything at the jacks. There's no physical stereo Link switch, even though we see one in the block diagram. Probably one of the jacks accomplishing that. I've already tried exercising the jacks (one input, two line output) with open circuit plug, with no change).

[g1]

I think as far as the mute, glebert was referring more to the 'mode' function of the power IC itself, rather than the switch. Datasheet attached.

[nevetslab]

I think as far as the mute, glebert was referring more to the 'mode' function of the power IC itself, rather than the switch. Datasheet attached.

Good point....I hadn't investigated that nor went thru the document. I was hoping the one SMD PNP xstr I found that was bad, replacing it with similar part would have restored order, it being the only defective part I found on the power amp assembly. When I get this unit disembowled again and set up, I'll have a look at that. Thanks!

[nevetslab]

I stopped off at the Roland Artist Relations office at our rehearsal studio complex to request a copy of the service manual for the KC-880, so hopefully that will go thru.

[nevetslab]

Last week, before Christmas, I had run into the chap who had been doing our Keyboard Maintenance here at CenterStaging for years, then moved into doing studio bookings before moving on to other career opportunities a few years ago. He's now one of the production managers at AuntieM here in Southern California, who's one of my steady clients. He had the Roland KC-880 Service manual, and sent that to me. After digging into the document today to try and pinpoint the probable circuits causing the muting of the power amps, I found I had made a mistake when I looked up the two-letter code on the SOT-23 xstr case: "LA" The two selections was an RN1422-A NPN part, the other a BF550 PNP part. I elected to use the PNP part, and put the unit back together. It didn't work. (So much for an educated guess).

Now, with the schematics, I find Roland has made liberal use of this RN1422-A part, which is a 50V/800mA NPN SOT23 part, with the addition of a bias network ahead of the base inside the xstr case. That was the only part I found that had failed on the power amp/power supply PCB assy. They have that same part used as a Power-On MUTE circuit on every output line, as well as circuits ahead of outputs thru-out the system.

RN1421-RN1427 NPN Xstr_SOT23.pdf

That 'failed' part was Q4 in the circuit below. I tried to load the S/M for this KC-880, but the file size is just under 18M, so loaded that portion of the Power Amp PCB's control circuits that talk to the Power Amp IC's mute circuit, along with the P-ON_MUTE control buss that goes everywhere in the product. I was able to order the part from Mouser, so, I will find out if this part cures the problem after the part comes in next week.

KC-880_sm_PA-PS_Pg2.pdf

[nevetslab]

Yesterday the replacement RN1422A NPN surface mount parts arrived, so I pulled the unit completely apart again, replaced that wrong PNP part I had installed and put this correct part in its' place. This time, I connected everything up on the bench, though with only one woofer and tweeter connected, due to lack of bench space. Powered up, but still had no OUTPUT from the power amp section. Checking my connections, I found I had forgotten to plug the signal input cables to the three Class D IC power modules, so corrected that. Still no output, and still seeing the same 28W AC Mains being drawn. I was seeing DC offset voltages on the outputs, though that on the woofer connected, both sides were sitting at -11VDC, and on the tweeter, 0.7VDC. The other outputs were -6VDC on the tweeter, and on the other BTL-wired output, -1.7VDC and 8.6VDC, both no load. The supply voltages applied are +/- 27VDC, well within the nominal operating range for the IC's. The MODE voltage was 4.4VDC, also in the range for it to be in OPERATE mode. But, still no output.

I followed the applied signal, and had LF and HF signals on both channels of the Master Volume pot, then on the output of the buffer IC's that drive the output cable connector, and saw same signals at the output connector.

I have no way to check DC current feeding the three Class D IC modules, as all are PCB traces, and on the bottom side of the PCB. The pins on the 23-pin IC's are too close together to get safely at any of the connector pins. I saw PWM oscillation signal at Pin 1, though didn't verify the frequency nor magnitude yet. I'll have to remove the PS/PA PCB assy from the chassis and get it flipped over and insulated so I can do further probing on the back side of the PCB.

I did find it odd that on the other RN1422A NPN xstrs that have that bias resistor network ahead of the base, testing these Toshiba parts I just received I got NO readings in semiconductor Test Mode on my Fluke 8060A, while I did get proper readings on the other parts I found available on one of the PCB's....though it read like 1.99V instead of the typical 0.6 to 0.7VDC to indicate valid semi junction.. But, with the new part installed, I have the correct MODE voltage of 4.4VDC, so I know that part is working.

I didn't see anything in the TDA8920B Data sheet under their various Protect modes to explain why nothing is working at this end of the system. I still have full system functions at the headphone outputs, and see signals going to the power amp assy. This product thus far as NOT SERVICE FRIENDLY!!

[g1]

Q4 on the schematic you posted is RN1442, not 1422.

[Jon Snell]

Have you got both mute and standby voltages when requested?
Have you got oscillator out on pins 16 and 21? VSS and VDD ok? Check you have audio on pins 8 & 4.
The TDA is an extremely stable chip that shuts down when there is a problem; over volts, under volts, over current etc.