R-500 initial checkout

[Hak Foo]

I picked up a R-500 at the local thrift shop today. $35-- expensive for thrift shop stuff when several heavy modern receivers were $15, and it had the input selection knob replaced with a clearly non-matching substitute, but otherwise not hideously dirty; all lights seem to work and it really hits the classic aesthetic in a way my beloved JVC gear doesn't .

When I powered it up in the shop, I saw the signal gauge flicker at times, but no relay click. It didn't dawn on me at the time, and then I spent a while reading "the protection circuit caps like to give out" and thinking that was in front of me.

I blew it out with canned air, and the first few times it went on there was intermittent popping and clicking, little sound, and no consistent relay. But now it seems to have the relay pop in more consistently on power up-- as if it's back to normal-- are there parts of the system that misbehave that way? (Reminds me of a Commodore 128DCR I bought-- it wouldn't power up the first few times, then suddenly started working)

The big concern I see right now is that the "Stereo/S-lock" function works pretty erratically. I can't get stereo on many strong stations, and only a handful of stations escape muting in S-lock/stereo mode. The signal strength meter is zero on plenty of very listenable signals, and full on dead or static spots.. Using some broken rabbit ears as antenna.

The SM shows DC offset adjustments, so I'll do that along with the obligatory Deoxit sprays, but is there an adjustable bias? Given it's a STK module unit, I'd prefer to keep it cool for longevity's sake-- new parts are a pain.

The caps don't look bad from an immediate inspection, but the glue holding the heatsink in the "12v supply" zone of the PCB does look horrible and dirty and gungy.

Also, the loudness control seems to just quiet the sound, not change its character much. Could be my headphones (those beige Koss ones Massdrop was selling a few months ago) though. Are my expectations wrong based on units with conventional loudness buttons?

 

[avionic]

The big concern I see right now is that the "Stereo/S-lock" function works pretty erratically. I can't get stereo on many strong stations, and only a handful of stations escape muting in S-lock/stereo mode. The signal strength meter is zero on plenty of very listenable signals, and full on dead or static spots.. Using some broken rabbit ears as antenna.

Try a better external antenna.

 

[avionic]

, but is there an adjustable bias? Given it's a STK module unit, I'd prefer to keep it cool for longevity's sake-- new parts are a pain.

You have the manual. If its not in the manual then your SOL. Usually STK's have a fixed bias.

 

[avionic]

but the glue holding the heatsink in the "12v supply" zone of the PCB does look horrible and dirty and gungy.

DSBG google it. Glue becomes corrosive by now.

 

[Hak Foo]

I pulled down the offsets to like .1mV today (they were already pretty low) and scraped off as much of that glue as I could. Had a few weird relay clicks during thw "wait 5 minutes before adjusting" warmup.... maybe it does need protection capacitors after all.

Then I got into a dilemma. I wanted to strip the unit down for a cleaning, so I had to take off the front panel.

The three screws that hold the front on are evidently made out of a space-age alloy of Chinesium and candyfloss, drilled in with Odin's Phillips-head (lesser known accessory to Thor's Hammer) and were gouging wildly as I tried to remove them. So now I've got one out and two ground-up messes that won't really take a bit Trying to glue in a screwdriver bit and work them out. Don't want to just Dremel in a new opening for a flathead to get them out, because that's kicking up a lot of metal dust in a situation where I can't really remove everything electrical near it.

 

[avionic]

he three screws that hold the front on are evidently made out of a space-age alloy of Chinesium and candyfloss, drilled in with Odin's Phillips-head (lesser known accessory to Thor's Hammer) and were gouging wildly as I tried to remove them. So now I've got one out and two ground-up messes that won't really take a bit Trying to glue in a screwdriver bit and work them out. Don't want to just Dremel in a new opening for a flathead to get them out, because that's kicking up a lot of metal dust in a situation where I can't really remove everything electrical near it.

I've had good luck with reverse flute (left hand) drill bits
https://www.mcmaster.com/#standard-drill-bits/=1dmr2z0

Inch Size Left-Hand Drill Bits

Also known as reverse-flute bits, the flutes spiral in the opposite direction of other bits. They are for use in reverse tool operation—most commonly to remove broken screws. Jobbers'length are the most common bits and have the rigidity and length required for most applications.

A split point keeps the bit centered without a pilot hole.

Bits with 135° point angle produce smaller chips than 118°point to prevent clogging.

Keep a shop vac nozzle close to the work to catch the swarffing.

 

[OMGCat!]

For Japanese gear you're much better off using a JIS screwdriver over any other kind.
I have a set of Vessell that I love which fits these screws perfectly.

 

[Hak Foo]

Well, time to come back with more details.

It was having a lot of issues with the relay suddenly clicking mid-operation or not coming up quickly at power on.

I replaced all electrolytic capacitors on the power-amplifier board (C3xx series) with the following exceptions:

* The big 8200uF ones (C340 and 341) -- that denomination seems to be unstocked these days
* The lone 33uF 6.3v cap (C346) - I ordered 22uF by mistake and I don't want to spend 9 cents for the part plus $5.95 for postage unless it seems like a worthwhile culprit. The pulled caps seem to be within a 20% band, so I'm hesitant to make another order unless it seems a strong culprit.
* C325 and 326 were listed in the SM at 100uF, but the caps in my unit were 47uF. I went with 100s here since that's what I ordered.

It seems like it's not dying mid-use under brief testing anymore, and still sounds good and tunes strong,, but when cold, it sometimes takes a long time to relay click, or it will do a soft click and no sound, followed by a later, full switch on, or a frantic series of clicks.

Working on the "when cold" aspect, I reflowed most of the solder joints on the power board. Perhaps marginal improvement.

It seems like a very sensitive part of the unit is the three transistors roughly in a line in the protection circuit. TR306, 308, 309. Tapping those with a probe-- once it's already deprotected-- will often cause a pop to be injected into the sound. Sometimes it pushes into a noisy mode or even causes the relay to click out momentarily. TR306 seems most likely to cause the relay click when tapped, but that could be because it's physically closest to the relay. 308 and 309 make different sounding clicks. When it's in "not yet warmed up mode", tapping the transistors doesn't seem to get any closer to launch.

Since they're closely wired, is it likely one is iffy and propagates trouble through the circuit, or should I just consider replacing all three? I' ordered some KSC1815 and KSA1015 replacements, but it will likely be a week til they arrive. The other transistor in the area (TR305) doesn't seem to cause issues when tapped.

Also, the STK3042II seems to get fairly warm for a non-heatsinked part. Typical?

 

[Hak Foo]

Okay, I'm still getting nowhere.

I've done the following:

Replaced all electrolytics on the power-amp board except for the two big filters (got that 33uF one after all)
Replaced the three transistors mentioned in the protection circuit.

So I checked the DC offset... and the WTF moment finally came. I tested from the back of the PCB, and using chassis ground.

Right channel seems okay-- a few dozen millivolts, but left channel from the L301 lead to ground) goes as high as -40 volts! It sometimes sits at near 40, and sometimes swings wildly. Changing voltage, input, or balance doesn't make a predictable difference, but at times it would tick down to below -10v if left on for a while. Adjusting VR303 doesn't make a noticiable difference.

Where would you look next? My first thought was TR301, but I don't have an equivalent spare (only KSC1815 and KSA1015s). but the sometimes variable level makes me think of the possibility it's a short-- bypassing the DC offset control or piping 40V right into the output.

 

[avionic]

Right channel seems okay-- a few dozen millivolts, but left channel from the L301 lead to ground) goes as high as -40 volts! It sometimes sits at near 40, and sometimes swings wildly. Changing voltage, input, or balance doesn't make a predictable difference, but at times it would tick down to below -10v if left on for a while. Adjusting VR303 doesn't make a noticiable difference.

STK2040 shorted internally or the driver STK3042 ll is holding it at rail.

 

[Hak Foo]

It seems like pressing in on, rather than tuning VR303 causes the DC offset to plunge close to zero; if you can hold it in, the relay usually comes back after a few seconds. Sometimes it stays when released, other times it drops and goes back to 40V offset.

Deoxiting the part and reflowing the joints don't help. I note it turns much more easily than VR304, its partner on the right channel.

Theory: the pot's actually falling apart internally (so it turns more easily) and usually fails in such a way that the circuit says "dump everything onto the left channel" and fires protection. Poking it with probes disturbs the balance. Is this feasible? I can try to rig up a chain of fixed resistors to see how it works.

 

[avionic]

seems like pressing in on, rather than tuning VR303 causes the DC offset to plunge close to zero;

Reflow the solder connections on that board.A cold solder connection on that VR or that circuit board would act that way.VR could be toast as well.

 

[Hak Foo]

I replaced the variable with a cobbled together resistor in the 2.5k range.

It comes up with that fairly consistently, but it takes much longer than other units I've played with-- 30 seconds or maybe a minute instead of 5-10 seconds or less on other units.

It did go into a cycle of repeatedly throwing the relay once or twice-- it's possible the random value I dropped in is on the edge of tolerances for protection to engage, but once it goes on, the reported offset drops to under 10mV immediately.

How does protection actually work-- my first thought was "if DC offset exceeds a critical value, throw the relay" but wouldn't the characteristics of the circuit, and likely the offset, change wildly once the relay closes and a speaker load appears?

I'll try reflowing the STK3042 and its neighbourhood too-- I thought I touched up most of the board, but I'm not sure about that part-- it sits freely on long legs so the joints might be easily stressed.

The offset measurement still swings wildly, up to 40v, and with not much consistency, in the "until the relay comes on" phase. Also, when you shut it down, the measurement jumps from near-zero when operating, to ~20v, and then it trickles down over quite a while. I'm not sure if that means anything.

 

[Hak Foo]

Reflowed the 3042. It seems like I can easily get pops in the sound and occasional cutouts by tapping the PCB near the 3042, especially the left channel side. But

you can also sometimes-- not as easily-- get them poking the PCB near the 2040. I also found that if you put a little pressure towards the centre of the 3042 itself, on the top side-- you can make the relay die or come back. This screams "loose connection" but it's been reflowed at the PCB. Maybe a loose internal connection-- which would still fit with my experience of failures tending to go down as it warms up-- if you power cycle it after a few seconds, it often comes right back up, but from a cold start, you're looking at a long wait.

Since it seems to be more reliably on when the STK3042 is bent towards the front of the unit, I'm trying to push it over a little, and will see if that helps when it cools down again.

I understand it's possible to test the 2040 and 3042 independently, but that there's a fairly close coupling between them and if one's bad it often takes the other out. I feel like if one or the other is outright shot, it may not be worth the cost-- $20 for the 3042, and who knows for a *real* 2040-- to proceed further.

 

[Hak Foo]

Okay,I think I've reached as far as I can go here. Replaced the DC offset pots. The problem seems to be very tightly coupled with the STK3042II. If you have it sitting roughly perpendicular to the PCB, it throws the sky-high DC offset and doesn't work. But if you bend the part on its leads-- like 20 or 30 degrees forward or back-- it seems to work. In my case, it works best "back" because if you pull i forwards, it blocks the pots. I can't imagine this is great for thermals though.

The connections at the PCB have been reflowed and seem to test fine on a multimeter. I sort of don't want to throw another $20 at a new 3042 without much confidence it will really fix the unit

I wonder if there's a loose connection INSIDE the module-- is that a possible failure mode? Pictures of disassembled modules show the pins soldered to the actual module, so I could see a loosened joint happening.

 

[avionic]

Okay,I think I've reached as far as I can go here. Replaced the DC offset pots. The problem seems to be very tightly coupled with the STK3042II. If you have it sitting roughly perpendicular to the PCB, it throws the sky-high DC offset and doesn't work. But if you bend the part on its leads-- like 20 or 30 degrees forward or back-- it seems to work. In my case, it works best "back" because if you pull i forwards, it blocks the pots. I can't imagine this is great for thermals though.

The connections at the PCB have been reflowed and seem to test fine on a multimeter. I sort of don't want to throw another $20 at a new 3042 without much confidence it will really fix the unit

I wonder if there's a loose connection INSIDE the module-- is that a possible failure mode? Pictures of disassembled modules show the pins soldered to the actual module, so I could see a loosened joint happening.

Sounds like the 3042 has and internal intermittent connection to the solder pin..A legit 3042 will probably fix you up.

 

[Hak Foo]

Broke down and dropped the $20 for a replacement STK3042II from B&D, and installed it today. It can at least stand straight up without the unit dropping into protection.

It seems to run noticably warmer than the old 3042, but it may need an offset adjustment.

I took a small heatsink, probably from a 486/66 or the like, and tacked it to the module with some Arctic Silver Ceramique (one of their nonconductive formulations). I've got two zip-ties tied around the assembly, but I'm worried this will loosen or break due to thermal stress, and then the heatsink drops between the two most expensive and unobtanium parts in the unit, probably shorting every part at once. I should probably go straight to thermal epoxy.

 

[Hak Foo]

Potato pics of the heatsink attachment.

 

[avionic]

I should probably go straight to thermal epoxy.

Probably be the best bet. Hate the smoke your 2040's..That would really sux.

 

[avionic]

BTW dc offset won't do squat about heat issues.