Around noon on Saturday, June 16th, 2012, my wife and I were out running errands. I believe it had rained pretty heavily in Richardson the night before – it was the day after a big hailstorm in Dallas. As we drove north on Coit, I spotted what looked like a promising yard sale near Campbell. For me, a “promising yard sale” in 2012 was one with audio equipment to pick through and this one had a huge pair of speakers on the lawn. Good times!
After we arrived, I took a good look at the speakers I saw from the road. I recall them being Pioneer or something similar and about 4′ tall. They were only mildly distressed – not an issue to me if they can sing and not what keeps me from buying them if they can’t. The seller wanted $80 for them. I looked them up online; most people were saying they’d paid around $100 for theirs. But owners of this model weren’t in love with them – they were only acceptable. This isn’t high praise for such a large pair of speakers, so I decided it was best to let their MDF keep soaking up the water from his lawn.
So, we perused the rest of their stuff for a bit. Little else seemed to of interest. But, as we were about ready to leave, I noticed a large black metal box on the wet grass near the speakers. Upon further investigation, I realized that box was a beat up and dirty, but possibly expensive, amplifier, marked just $15. I was very hesitant to rush to judgement, because $15 isn’t much for a fine piece of vintage electronics and the owner apparently judged it worthy of being left on his wet lawn. So I suspected it was probably some cheap junk – designed look good enough to trick people into forking over their money. Also, its cosmetic condition didn’t scream “working” to me. And who was this “Onkyo Integra”, anyway? Never heard of ’em.
I spoke to the owner about it for a little bit. I asked if it worked; he said it had a “weak channel”, might need new power filtering capacitors, and had a light out in the front, but he insisted that it “still worked”. He also said he’d had someone do some work to fix one channel. I could see into it well enough to tell that the tops of its power caps looked really puffy – a very bad sign. I asked him to let me think about it and told him I wanted to look up some online reviews before deciding whether or not to buy. I knew I’d be committing to do extensive repairs if I bought it; it was time to rush to judgement about whether or not that was worthwhile.
I’m a firm believer in the idea that if people aren’t willing to pay good money for an old piece of audio equipment, it probably isn’t very good. And if it isn’t very good and they’re still willing to pay good money for it anyway, I can at least still flip it. So I pulled up eBay. I was able to find a similar model, the M-508, which I noticed people were trying to sell for around $800. I turned to my wife, showed her this, and said, “as you can see, people are saying good things about it.” I happily handed him $15 in exchange for his (working?) Onkyo Integra M-504.
As I lifted it, I noticed it was REALLY heavy. Its documentation says it weighs 49.5 pounds, but this knowledge alone doesn’t actually prepare a person for lifting it. Reason: it’s pretty large and nearly all its weight is in the power transformers, which are on the left side of it; if you don’t take that into account when you lift it, you can end up being forced to carry most of its weight unnecessarily far from your center. Mine actually weighed in at 50.1 pounds the day I bought it. I still don’t know why – extra moisture and grime aren’t adequate explanations for an extra 10 ounces.
But if it’s heavy, then it’s expensive, right? So I was happy to lug my new 50-pound paperweight to my car and drive it home.
Here are some pictures I took of it that night:
Note the peeling vinyl faux ebony and how so many of its surfaces are dirty or greasy. Also note the quarter for scale; “Banana for scale” wasn’t really a thing yet. Some people might see trash – I saw a canvas and I was eager to sculpt it.
Then I had to archive my wonderful new project in a corner, because I was too busy to work on it. I swear, for 2 years, that thing made sad eyes at me every time I looked at it. Occasionally, I’d look it up online to read about it and everything I’d find would further convince me that I made a very fine purchase that day. Here’s a screenshot of the page from the service manual that shows its specifications:
After researching a lot more of what Onkyo sold in the past few decades, this amp is from what I consider to be Onkyo’s golden era of design, which I feel spans from approximately 1980-1995. It wasn’t even close to top-of-the line when it was new (M-506RS, M-508, M-509RS, M-510) and it had quality problems (cold solder joints, poorly matched power transistors), but its design was still incredible. It seems like Onkyo still makes really nice stuff too, but you need to buy something really expensive to get anything close to the stats of the M-504. (i.e., Onkyo Reference M-5000R) Their quality control is probably a lot better now too, but I can’t judge from experience with their current product line.
On 7/21/2014, after deciding I had too many unfinished projects in the house, I hauled my M-504 in to the new Dallas Makerspace location to tear it down and have a good look at its innards. By some stroke of luck, I saw Andrew Kwiecinski in the hallway mere minutes after I started ripping it apart. I’d only spoken to him on two occasions and hadn’t seen him in ~6 weeks, but I remembered he was really into audio equipment and seemed to know a fair bit about it. So I said “hey, you should see the amp I brought in to repair.”
Andrew was the first one to notice that one channel’s output stage transistors had been replaced with a different part. The problem with that being that they hadn’t all been replaced at once and we had no way to know if they were anywhere close to gain-matched. So he also recommended a bit of an upgrade to the output stage. He said something like, “I recommend not turning this on until it’s fixed.” If I recall correctly, he was able to rattle off a few part numbers before he had to look up the rest to make a complete recommendation. He recommended I buy 10 each of n-type and matching p-type from his list of suggestions so we could test them for gain and use only the best matching 2 pair of each.
With Andrew’s help, I put together a nearly-complete repair parts list that day. He also noticed that there were a lot of questionable (mostly cold) solder joints on the main board and some that had already failed. His recommendation was to redo every solder joint in the amp. Additionally, he recommended replacing small electrolytic capacitors with poly caps. I think this is a wonderful suggestion for this type of repair – theyll probably outlast electrolytics by decades.
Based on the fact that the main power caps looked questionable at best, I decided to order replacements for every electrolytic capacitor in it. (I’ve read that it’s often the smaller caps that die first, so if the big ones look trashed it’s likely there’s an extensive problem.) I’d soon placed an order that would allow me to replace all its capacitors (most with 105°C parts) and the mismatched output stage transistors. I could see at this point why someone with less fortitude might do something like sell it in their wet yard sale for $15, but this stuff is really exciting to me.
After realizing that the P-304/M-504 set generally sells for way more than either of them alone, I also bought an Onkyo Integra P-304 (the matching pre-amp that would have been sold with the M-504). I figured I may as well fix them up together to make a matched set. The P-304 was in better condition and, importantly, still had all its knobs. Its capacitors may have been fine, but I decided to do a full recap on it as well. With no big caps, a piece like the P-304 can be recapped for $15-25 in parts, which I see as a bargain to stave off failure by another 20+ years. I’ll write more about the P-304 on a separate page.
At this point, I still had a power amplifier that needed a lot of cosmetic work. A younger version of me would have attempted some cosmetic repairs first. But to do that on a piece of equipment this trashed is just a foolish thing to do. Any effort put toward cosmetic repairs before it can be verified that the unit can be fixed on a reasonable budget is waste.
Here’s a list of all I did to it during the repair process:
- most capacitors replaced with 105°C Nichicon caps (higher temperature rating, but same brand and similar series to originals)
- main power caps redone with a close equivalent (high quality 85°C caps – this saved $30 over 105°C)
- main board modified to take modern 2-pin power caps – originals were 4-pin
- cracked output jack circuit board repaired
- cracked input connector repaired
- transformer housing tabs reseated (they were bent outward, either from high heat or an attempt to open the housings)
- new output stage power transistors (Fairchild 2SC5200/2SA1943)
- new TO-3P alumina power transistor insulators (Given that TO-264 parts are still manufactured, it’s surprisingly hard to find TO-264 insulators. Alumina TO-3P insulators work just as well because they’re so thick.)
- missing electrical components identified and replaced
- every solder joint redone
- circuit boards cleaned of flux and stray solder blobs
- calibrated idle setting on each channel to 14.9mV-15.0mV.
I borrowed a meter from Andrew to test ß (hFE) on all my transistors. Here’s data for the power transistors that were in the unit on the day I bought it:
|Right (Sanken – factory)||2SA1492||2SA1492||2SC3856||2SC3856|
|ß = 119||ß = 75||ß = 30||ß = 24|
|Left (Toshiba – replacements)||2SA1302||2SA1302||2SC3281||2SC3281|
|ß = 108||ß = 98||ß = 65||ß = 65|
The tech who replaced the left output stage did an acceptable job of choosing transistors, getting them within about 10% of each other at worst. However, the factory Sankens appear to have been chosen completely at random. The PNP and NPN sets were off by 59% and 25% respectively. Given a handful of modern parts, it’s not hard to get within 1%; all four pair I installed were within 1% of each other.
Here’s a terrible phone pic of it from the first few moments after I had it (electrically) reassembled:
Note the shorter blue capacitors. I opted not to glue them at the base this time. (They’re exceptionally hard to remove when they’re glued in like that.) Instead, I applied a small dab of high-temp hot glue anywhere they touch each other or the tray wall. It’s actually possible to break this bond with a knife before trying to desolder them, which is super-important if they ever need to be replaced again.
And here are some shots from calibrating the idle setting:
Note that one meter flips between 14.9mV and 15.0mV between shots. I see this as an indication that I’m pretty close to 15.0mV (the correct setting, according to the service manual). I suspect it’s probably closer than the factory adjustment was – those little trim pots are really hard to set with a high degree of precision.
Now that it’s electrically repaired, it does actually sound really nice. But the best speakers I have to connect to it are two pair of EPI Model 50 and/or a pair of AR-2ax. I’m sure there are better speakers to use with it, but I intend to explore that later, when I have more space at home.
As of right now (10/1/2014), one of the best things about this amp is that, since I haven’t finished its cosmetic restoration, many people assume it’s going to sound terrible. One guy went so far as to angrily grumble “this is NOT high fidelity” when I showed it to him. (I didn’t even tell him it was!) It’s so fun trolling people like this that I’m not even in that big of a hurry to make it look good.
Here’s what I have planned for this phase:
- Parts have been ordered to replace burnt-out front panel meter bulbs with LEDs. Original color and brightness will be matched as closely as possible. A set of original bulbs can be had for $20, but they’d just burn out again – I’d rather not mess with this again for the next few decades.
- I will soon be replacing the peeling vinyl faux ebony panels with curly maple, fumed to a medium to medium-dark hue. I bought the last of the wood for this the day before ankle surgery. It’s highly figured and I think it’ll look incredible.
Here’s a picture of some of the (unfinished) curly maple I bought:
I’ll do 1 or 2 more updates to this page as I have more pictures and I’m further along with the restoration. Given that I can’t presently carry the M-504 (or much of anything else), it may be a few months.