300DC repair

Status
Not open for further replies.
By M.Chua - This amp can sound very nice with NEC 2SB600/2SD555 for outputs.

Seems the MJ 15023/24 have the same Ft and Cob and might
come close, but original japanese devices have some "magic',
(I like sanken , very good packages). ON-semi makes good
output devices but sometimes fall short of the originals.

By anatech - If you look at a Nakamichi 620 power amplifier

I did, and found it is a bit different... with current sources and
a EF'ed diamond. The only confusing thing is how it is thermally
compensated. My only guess is Q011/Q012
(see attached schema)are on the main heatsink,
giving Tcomp to the diamond as you mentioned. What trips
me out is there is no bias trim which leads me to believe
they used precision resistors for r018/019 :whazzat: .
OS
 

Attachments

  • naka620.gif
    naka620.gif
    81.3 KB · Views: 394
Hi Michael,
This amp can sound very nice with NEC 2SB600/2SD555 for outputs.
However, briefly.
Wrong replacement by a mile! They are much older designs for one, and only good for 10A. These cause this amplifier to take a step back, rather than forward. Given that this amplifier is rated in the very conservative manner that Marantz is famous for, and that it has 3 dB headroom, no! I'll go as far to say that if you have serviced any of these amplifiers with 2SB600 and 2SD555, you should recall them and install proper output transistors. MJ15024 and MJ15025 will work better as long as you match them (actually, any outputs), or better yet, MJ2119x. These parts will be reliable and are also closer in technology.

My burning question is now very important. How did you make your substitution? Did you use a cross reference manual of some type, or did you look up the original specifications? If that's the case, what source did you use for those ratings?

-Chris
 
Hi ostripper,
but original japanese devices have some "magic', (I like sanken , very good packages).
That magic is high manufacturing quality and low parameter spread. They designed a very linear transistor, for outputs. Their small signal devices are also good for the same reasons. they also seem to have fewer problems with noise. Perhaps because the silicon surface has fewer defects and undesired impurities. Basically a better manufacturing process. If these were easily available to North Americans, it's possible that our own suppliers would have lost the battle for market share.

ON-semi makes good output devices but sometimes fall short of the originals.
Well, not since On-Semi has been making Toshiba designs under some agreement. The MJW0281A and MJW0302A I tried earlier were by far and away the best outputs I have ever seen from anywhere. If I had the money, I would have bought a huge stock of these parts. Still, the newer On-Semi numbers are the best they have produced. Thank goodness! Now I just wish On-Semi would give us more drivers and signal devices. Diff pairs in J-Fet and Bipolar would be welcome.

I did, and found it is a bit different... with current sources and a EF'ed diamond.
Yes, I found this out after my prototypes were running, so it must be a good idea. The current sources were a carryover from the signal buffers. They really help to reduce supply noise and make the entire thing easier to drive at higher signal levels.

My only guess is Q011/Q012 (see attached schema)are on the main heatsink, giving Tcomp to the diamond as you mentioned.
I'd have to examine one one the bench again, but I seem to recall the entire diamond was mounted on the heat sink. That's the way I did it as well. It works, Marantz accomplished the same thing in a different way. The temp sensing is done by monitoring the driver transistors. The temperature compensation is accomplished by balancing current flow between the "input" transistors and output couple. The Marantz amplifiers are partially corrected as that design also warms up when idle. When you start driving a load, the output stage actually cools off to some degree. Same as my prototypes. Mine are far more thermally unstable over the course of 6 to 10 hours. I don't know if it would ever reach equilibrium.

What trips me out is there is no bias trim which leads me to believe they used precision resistors for r018/019
No, not so. It's the Vbe of the transistors that determines thermal tracking. R018 and R019 are matched for offset purposes. I have repaired some of these amps as well, they sound very good with that special property you like the sound of.

What I'm saying here is that the bias is controlled by selecting the transistor part numbers that satisfy the current, power and voltage requirements (and gain of course), but also balancing their Vbe drops.

Which brings me to another point. I have asked mechanicalman not to attempt a repair on this amplifier as he intends. I am truly hoping he considers my advice because these circuit types depend on output type numbers for more than simple specs we normally consider. I'm worried because someone who doesn't know much assumes there isn't much to repairing some designs. Once you know more, then you get an idea how much more there is to know. Everything seems so easy when you don't consider the details.

Here is a shot of a working prototype PCB using TO-3 parts.

-Chris
 

Attachments

  • 2nd proto module (medium).jpg
    2nd proto module (medium).jpg
    68.9 KB · Views: 376
..... and this is showing the outputs as running.

-Chris

Edit: These were to investigate the effects of different outputs. The newer On-Semi parts perform better. These have a mis-match between NPN and PNP beta, another prototype was built up using a mixed batch of 2SB554 and 2SD424 from Adcom stock I had. I selected the parts that differed the most from each other in beta. This of course to answer the question "how bad can it get?". The performance of the On-Semi set is better, and if you look, you can make out the beta values marked on the cases.

Looking at the previous post, you can see that the driver transistors are mounted bent away from the way the board was designed. This was only to use an amplifier "donor" that had TO-3 outputs in original trim. I wasn't about to design a new PCB for this.
 

Attachments

  • 2nd proto xistor side (medium).jpg
    2nd proto xistor side (medium).jpg
    76 KB · Views: 357
By anatech - The MJW0281A and MJW0302A

I use the (N)jw0280/0302, only $1.25 each and they definitely
give the most "bang for the buck" ,15A devices as well with
good models.

By anatech - but I seem to recall the entire diamond was mounted on the heat sink

Now that is interesting.. I simulated the nakamichi and It
was biased to exactly 60mA per device.. even with the
newer KSA/fairchild devices. Makes sense that all 4 devices
are the Vbe (just like 4 diodes) on the main heatsink.

Wrong replacement by a mile!
Mr. Chua might just have been referring to the sound
of the devices. A slower device has a different sound, like
the difference between a 2n3055 and a newer MJ/MJL
device. Most of the bipolar kits on ampslab are MJ15024/25
based. Even on my first sub amp (C200 mod) the difference between the original 15003/4's and the replacement 15024/25's
was noticable.

I don't know if it would ever reach equilibrium
I can't digest that.. I run my amps hard so I guess the
standard Vbe is the ticket. I can't let any fellow hobbyists
build a "burning amp"😀 .
OS
 
Hi ostripper,
I simulated the nakamichi and It was biased to exactly 60mA per device.
Don't trust a sim in this situation. The transistors are not perfect and the thermal differences will throw things out. Any diamond buffer design needs some fiddling when there is no standard bias control circuit.

From a manufacturing standpoint, it's a wonderful circuit. You can build an amplifier that has no adjustments to be messed up in the field. Once you get things figured out, you only need to watch the trends in your supply of parts. Warranty shops can be instructed how to handle these designs. The average shop most certainly will make a mess out of these. No worries though, from what I've seen over the years, they destroy amps anyway.

Mr. Chua might just have been referring to the sound of the devices.
I'd like to see what is going on there precisely. Doing this work by ear when you don't understand fully what is going on is reckless. I stand by my objections. I'd rather see where the root of the problem is for him and correct that. Others who read this thread can also get the right information then.

A slower device has a different sound, like the difference between a 2n3055 and a newer MJ/MJL device.
Those are certainly different!
Even on my first sub amp (C200 mod) the difference between the original 15003/4's and the replacement 15024/25's
Again, they are completely different parts. No surprise. Mind you, MJ15003 and MJ15004 have a place, and they can work very well.
I can't digest that.. I run my amps hard so I guess the standard Vbe is the ticket.
They have never gone even close to that. I'm just saying that the bias current is always shifting around. Not dangerously yet. However, it is a prototype. What I don't like is that it's in the field, and has been for over a year. I had not intended this as you can imagine.

I have attached one of my low power prototype PCBs to show the mounting of transistors. I have not yet applied full power to this one. I do know it works though.

-Chris
 

Attachments

  • small prototype (medium).jpg
    small prototype (medium).jpg
    71.1 KB · Views: 372
By anatech - Don't trust a sim in this situation. The transistors are not perfect and the thermal differences will throw things out. Any diamond buffer design needs some fiddling when there is no standard bias control circuit.

No, I don't trust simulations entirely, especially in the thermal sense. What I do know now, (with all these helpful hints) , is
that with the diamond, running class A, it will have to reach
eqaulibrium before I can set my standard Vbe since it also
contributes to the bias of the OP stage.

As far as the reckless substitution as you stated in your post,
would a difference in OP devices throw the diamond's
stability off??
OS
 
Hi ostripper,
would a difference in OP devices throw the diamond's stability off??
No it won't. But it sure will throw the bias current off! It could end up being too high, or too low. Too low and the sound quality suffers. Too high and thermal runaway could happen I guess. Or second breakdown could be achieved just when the music is getting good.

At any rate, if someone doesn't understand a circuit well, there is no way they should be involved in servicing it. If the equipment in question is a simple, inexpensive old thing, learning on it is acceptable. although I would hope that some understanding of how it works is present. Otherwise the exercise is a complete waste of time. What we are talking about here is an amplifier that is different in concept than most, therefore what is important or isn't is not as clear. That and it has very good performance and a high used value. For these reasons, learning on this amplifier is ill advised. In fact I'd go so far as to say that learning on this amp is downright silly.

-Chris
 
Hi Chris, ostripper

Sorry for the confusion. I did not mean a direct replacement but rather something to explore (from a designers' perspective).

Anyway, unless one has the expertise and the necessary equipment, its best to have a very competent repair professional (Chris or similar caliber) look into it. When fully restored especially with all new caps, it can easily last another 10-15 years. Best of all, for the amount spent, it'll beat quite a lot of amps in the market, even those that cost 2 to 3 times more. It is a very good design. Very similar to my Bi240.

By the way Chris, I don't design amps strictly by ears. Only after a design meets my engineering targets then the ears take over. I can't afford an AP but what I have for the time being serve me well. My lab consist of:

1) TEK SG505
2) TEK AA501A (2 units)
3) TEK 2430A
4) TEK 465
5) TEK 575
6) HP 339A
7) Boonton 1130
8) assortment of Tek and Fluke Bench Meters and more scopes.

Regards
Mike

An externally hosted image should be here but it was not working when we last tested it.
 
By m. chua - It is a very good design. Very similar to my Bi240.


{BI-120 "clone" with balanced widlar buffer VAS/LED CCS}


Yes , the bi120/240 was the inspiration for my frugalamp 2,
I did use a slightly different circuit as to not jump on your IP,
and njl 0280/0302's for cost/space considerations. It plays away
in the living room and sounds wonderful. Thank you for the
ampslab site (projects and classic SS amps) ,they inspired me
and brought me to this forum after I built a C200.

The marantz definitely is not a beginners project, (once I
"hack" it into frugalamp 4 it might be). Seeing how nakamitchi
and marantz use their access to consistant batches of
close tolerance components to base their designs on, I can
see where TLC in the repair of them would be essential.

Below in attachment is what I prototyped and is playing as I speak.😎

OS
 

Attachments

  • fa4.gif
    fa4.gif
    23.9 KB · Views: 351
off topic

to anatech :

since i know you must be very busy and didnt want to trouble you with a PM i come across a wonderful amp form canada
(korean made though ) called sears AM4001....

amp is working ok .... im just going to update /recap/ restore / rewire it .... but no mater it really looks like a performer ....

i wonder since the amp is coming from your village may be you have some aditional info or a schematic

best regards ---- sorry for the offtopic
sakis
 
Hi OS

Great to hear you found the C200 useful and inspirational.

I will pop over to your thread on your frugalamp 2 when time permits. Looks like a very interesting design.

I had 40 inches of snow overnight. Took me 3 hours just to clear my driveway today. Apparently, 2 more storms are on the way in the next few days.

Mike
 
Hi Mike,
Okay, that's cool. I just get very worried because there are many service people around everywhere in the world that do not have the experience to do what they are trying to do. For over 20 years, there has been a semi constant stream of damaged, "hacked" equipment. Transistor substitutions are by far the most problematic thing, that and over tightening transistors. Proper cleaning and new grease / insulators seem to be another problem.

I guess I'm simply at a breaking point. Just looking at all the needless destruction that goes on "out there".

BTW, nice equipment. Most of mine is HP, but I do have some Tek scopes. The HP 339A is my favorite THD analyzer. IF you don't already, monitor the output of the 339A on a scope while taking measurements. More information about what you are measuring.

About your method of designing - or servicing. I full agree. Measure and balance that with what you are hearing.

-Chris
 
Hi Sakis,
Hey!, I'm just a guy who likes to help out, although sometimes the best help you can give someone is to tell them to stop. Odd but true.

BTW, you are welcome to PM me. There are times I get busy and forget to answer - just email me again then.

Now, regarding the Sears AM4001. Sears is what is called a "Department Store". They handle a number of items covering all your needs, including home entertainment. They do not design or manufacture anything. What you have there is called a "house brand". Something that they have made with their name on it from an OEM company. I'm going to bet it's a Far East firm that built these. There may even be other identical amplifiers with another name brand on it.

When I get one of these (almost never these days), I just know that parts and service information is unlikely to exist. At that point, all you do is figure out the general form of the design and go from there. There are only so many ways to design an amplifier. There are infinite stupid mistakes that they can make in those designs. 😉

-Chris
 
Hi ostripper,
Nice looking amp you have there. The front end is similar to early Cyrus (I and II) amplifiers. Should sound pretty good.

I'd be interested to see where you go with a "power diamond buffer". If I can ever pick up where I left off, I'll post my results.

I was lucky enough to pick up an HP 3585A spectrum analyzer. I have found this to be more than useful when examining a prototype. I wish I had one years ago! If there is anything you really want, it's one of these.

-Chris
 
Hi Chris

I fully agree with you on this. Though I don't repair or service amps, I do get requests and they often go like "I sent it to a shop and 2 weeks later, it smoked again". I can see why when a NTE transistor is simply substituted for a discontinued original part without any regard for stability.

Mike
 
Hi Mike,
Now there is the mark of someone who simply has no clue about what he is doing. I find that shops that also repair TV's and VCRs have an entirely different mind set, and it's not quality driven.

You need to find a good audio only service shop in your area to point people towards. Normally audio service shops in a store or working for a store are a poor bet as well. Shops that repair mostly band or other sound re-enforcement gear will not do a good job either. Their main concern is to "patch 'em up and out they go!". Said another way, "Good enough for rock 'n roll!".

Clearly not every one of the aforementioned shops will be bad, but the vast majority will be.

-Chris
 
By M. Chua -I can see why when a NTE transistor

NTE sucks..😡 I built your c200 with them , it blew:hot:
replaced with on semi stock and it still works. Down with
NTE (overpriced floorsweepings ,like radioshack semi's).

By anatech - I find that shops that also repair TV's and VCRs have an entirely different mind set, and it's not quality driven.

I now do all 4 of the major repair services (receivers , monitors,
tv, and PC's) but my real passion is audio. Even if it means
less profit and more time , I do not want to see the unit again.

90 days is too short for a good reputation, even if it would be
more profitable. So I will keep a classic amp with me for a while
(week) ,abuse it thoroughly in the real world.:drink:

I worked for tandy (radio shack) as a receiver tech and they
did not want to hear any music from my corner.. just run
a sine into a test load for 1 minute. They even told me to underbias the OP to make sure I got the 90 days or I
would have to fix it again for free.😡
OS
 
Status
Not open for further replies.