Rowe R87 works perfectly, then forgets where home is.

Greetings Jukebox friends,

I have been lurking in the shadows of the Phono user group and read many of the topics. I am hoping someone might be able to provide me advice on a jukebox problem that I can't seem to resolve. I have worked on quite a few machines of all eras and am pretty handy with jukebox technology.

My problem is a Rowe R87 that came to me non-functioning. The carousel just ran and ran. After much troubleshooting, I found that an LM339 quad comparator IC on the motor control board had a bad section that was preventing the CCC from recognizing home. Once replaced, the juke appeared to operate flawlessly.

Or so I thought.

As long as I manually set the carousel position to Home (99) and then switch the machine to ON, it works fine. It plays every record in the order selected, it sounds great, etc. No problems at all. After it plays the last record, the carousel does not return to home, but remains where it is. This is key. For the sake of argument, let's say it played record 104 last.

When I turn the juke off and then on again, the carousel rotates 360 degrees, past home and stops where it last stopped (in my example, record 104)

Now, If I select record 101, it thinks that it is at the home position and will advance 1 position and pick up 105, and every record thereafter will be off by those same 4 positions until I switch to Scan mode and manually move the carousel back to 99. Then it will operate fine again.

I thought the problem was the NiCad battery. The original battery looked fine (no leakage) but I replaced it with a 2.4V NiCad cordless phone battery. I've checked the battery after leaving the machine off overnight and the battery seems to hold at around 2.1V.

What keeps pointing me to a memory problem or the battery is that when I turn the juke off and on for short intervals, the Favorite record display keeps showing the most frequently played record. But if I switch the juke off for several hours, it forgets the most frequently played and defaults back to #100. However, the problem with the carousel being offset by several positions happens any time I switch the juke off then on.

Any thoughts? It's a really tough machine to troubleshoot without the schematic for the CCC!

Thanks for any advice you can provide.

Mark.

Hi Mark,
I am only somewhat familiar with that model--however, it would be my assumption that the memory chip is starting to fail--
It is normal for the carriage to stop at the last selection played and remain there till the next selection is made (IF--the battery is good-only). Are you 100% sure that the carriage was "aligned" as per the service manual, and that you do not have excessive "play" in the sprague set-up ?
As for forgetting the most popular---If the battery is good, it would also indicate memory chip problems--did you fully charge the new battery ?? ( they take at least 12 hours to fully charge). Ron Rich

Thanks Ron,

No I have not left the machine on for 24 hours, but I can do that. I have the service manual and can check the sprag adjustment. Nothing leads to to believe this is out of adjustment because once I manually scan the carousel to position 99 and switch the service switch to On, the machine plays each and every selection perfectly and in the selected order. It has never missed, not even once.

But switch the power off-and-on, even for just a moment, and unless the carousel happened to be at position 99, the machine will pick incorrect records as I described. In fact, switching the service switch from ON to SCAN and back to ON is all it takes to cause the problem to happen.

One thing I might try is to remove the new battery and see how it works without one. Obviously it will forget the favorites, but it should also forget the carousel's last location and reset to Home (pos'n 99). I'd be okay letting it stay that way if it didn't create other problems. I don't need to know the most/least popular record or how many plays have been made.

I haven't looked to see what the machine uses for RAM or how hard the replacement chip might be to find.

I will say, this is not the most attractive jukebox, but it sounds great. I feel good that I saved it from the trash heap.

Mark.

Hi Mark,
Without batteries, some--- of these CCC's will not function--others will do just fine, for awhile--then "mess up"--I do not know why, but I think it's something to do with the re-charge circuit ? Whenever I have experienced CCC problems with this series Rowe's, I have always just exchanged CCC's with one that has the Datasync update in it ( Look Dad--no battery needed or installed !!).
One thing you might do, if not already done, is remove all traces of the "foam" they used to back the board--make sure it's still insulated!!
Ron Rich

Ron,

Well it was worth a try. I clipped out the replacement battery and it acts the same way. Of course the most/least popular isn't working now, but the other problem is unchanged. I guess that puzzles me even more.

Mark.

Mark,
That's typical of CCC problems ! Ron Rich

Amazing! I fixed it. After much time studying the Home input into the CCC, I decided that the transistor that acts as the gate putting the home signal on the data bus didn't have enough gain. There is a factory modification where a jumper wire feeding that transistor was replaced with a 220 ohm resistor. For fun, I shorted the resistor out to boost the signal and voila, everything works. I've been running it for 2 days (about 60 records), Every one has selected perfectly.

A pleasant surprise is how good this jukebox sounds. It's a nice machine.

In saying much nicer things about the machine than I was saying about it a couple of days ago.

Mark.

Hi Mark,
Glad you fixed, and posted it ! Now where did you see that factory mod printed ?? Also what is the R number of the jumped resistor ??
Ron Rich

Ron,

It's W222 on the CCC. The W designation indicates it was meant to be a wire jumper but was a 220 ohm reisiistor on mine. Fixing it was mostly luck. Using a dual trace oscilloscope I was able to sync to one of the strobe lines and using the other probe, look at the home signal coming off transistor Q116 on the CCC. I had replaced that transistor. For some reason that home pulse appeared to be significantly lower in amplitude than the other pulses on the data bus. Then I noticed the resistor's "W" designation.

I remembered seeing a reference to that resistor when I read about a CCC debug tool at this site: http://www.cdadapter.com/repair.htm

If you download the PDF troubleshooting manual it has a hint about a factory mod on the CCC that replaces W222 with a 220 ohm resistor. My CCC is not the same model number as the one in the PDF. So i said, " what if someone did the mod on a version of the board that didn't need it?", and since my amplitude was low, shorting the resistor might just boost it. So I clipped a short around it and voila, the juke immediately started to pick the right record after a power up.

Now why this makes a difference is beyond me. But I've played about 25 records, turning the power off and on and switching in and out of scan mode and it has never chosen the wrong record.

I'm calling it fixed.

I learned a lot about how this CCC works. The micro has very few inputs, so many things in the jukebox share the same data line. For example the Home sensor shares the same data line as the Most/Least popular switch, the Clear Memorec and Advance buttons. Each one is strobed at different times and the data is fed to the data bus through a diode. Then an LM 339 is used to normalize that data bus to the right voltage to feed the micro.

I hope that makes some sense. The service manual provides very little insight as to what is going on. Glad I'm done.

Take care,
Mark.

Thanks, Mark,\Yep--a "W number" indicates a "wire"--not a resistor. Iffin you really ever want to go real deep into this CCC, you can always "bug" "Mr Grumpy" , at CD adapter--and tell him I said to do it, because he deserves it--but leave his wonderful wife, out of it !! Ron Rich

Ron,
Um, thanks for the suggestion. Being a newbie here, not sure what to make of the "Mr Grumpy" comment, but I can guess. I'm grateful to his site for the hint on W222 though.

The R87 is not mine. I fixed it as a favor for the Vintage Radio & Communications Museum (http://www.vrcmct.org). You should check them out. There is a close association to the topics on this board.

They have a basket case Rock-Ola Imperial (1937) that I may tackle next. Someone tried to "fix" it by taking it apart... And leaving it that way. I fear parts may have walked away.

Mark.

Mark,
He's a good friend of mine -- We "tease" each other often--laugh a lot !!
Ron Rich

Amazing! I fixed it. After much time studying the Home input into the CCC, I decided that the transistor that acts as the gate putting the home signal on the data bus didn't have enough gain.

Mark, were you seeing indication of the "Home" led on the mech board?

Rob

Ron,

Oh yes. The Home LED was working correctly from the start. The original problem that brought me to this site was that the record magazine kept spinning and spinning. The Index and Home LEDs flashed properly. I read a lot of the posts here and assumed incorrectly that I had a cold solder joint so I touched them all with the soldering iron and buzzed the cable harness out with a continuity checker but the problem remained. Then I followed the Home signal on the mechanical board with the scope and found that one section of the LM339 (CM339) was not functioning (I shorted the + input to ground and the output didn't change state). I replaced the LM339 and the magazine rotated 360 degrees and stopped. I thought I was done. At that point the jukebox seemed to work except when I turned the power off an on, then it would forget where home is.

That's where this whole discussion began and I started bugging you.

Somewhere along the way, I decided the transistor on the CCC that carried the Home signal (Q116) was bad, so I pulled it. I checked it with a transistor tester and the HFE was 39. The specsheet says minimum is 100. So I replaced it. FYI the HFE on the transistor that carries the Index signal (Q117) showed a healthy HFE of 150. That still did not solve the problem, but probably contributed to it.

Despite all the trouble fixing this, I am thoroughly impressed wiith this jukebox. It doesn't have the panache of the earlier ones, but it treats records well and sounds great.

Mark.

Mark, I asked because "we" had a spate of Rowe machines slightly later than your where it turned out that some idiot had set the LED exciter current too high in response to an original symptom similar to yours. The net result was failing LED's in the opto's several years later. When those were replaced we could not get indication of the index on the mech board. The dist (Betson) said that there 'were problems" on some boards and we ended up just replacing them...Your solution might actually have been the answer.

I no longer have any of those later Rowe machines nor any data so i didn't get involved here. I did operate several R90-92 combos and the sound was far better than the hard, PA-like sound quality of the CD-100's.

Rob