Seeking advice on Crown DC300a speaker amp

[JohnRoberts]

I have already given my advice.

Those power devices look similar but not identical. One is a 2 MHz min part the other 4 MHz min, More is better unless it isn't.

In general Germanium diodes are specified for low forward voltage drop, so I would not expect a silicon diode to just drop in and work properly, but it depends on the actual circuit. 

JR

 

[JW]

Thanks John,

I have been looking around for other amps that deliver 150-200 wpc, but since I already have this Crown, and it works okay right now, I thought I'd just keep at it and see what I get.

But if you have other recommendations as far as amps you'd prefer for biggish passives like JBL 4412's, I'm all ears. I'm not hell bent on trying to use this amp above all others. Just looking for something that doesn't break the bank too much and delivers the above mentioned wattage, and has no fan . . . .

 

[abbey road d enfer]

For the output and driver transistors,

I really would not advise changing these; the stability of the amp would be very compromised. You may think that replacing transistors with faster ones would improve performance, but actually it can turn your amp in a high-power AM transmitter of very short life expectancy.

Also, it's been mentioned elsewhere D108 and D208 which are in the VI limiter circuit be replaced.

Why would they need to be replaced? If they're good, tehre's no reason to replace them, if they're duff, replace them with the same, they are currently available.

Also, I think before build a relay circuit for speaker protection, I'm gonna give the two electrolytics in series a shot and see how that sounds. Any recommendations here?

And this only has to be on one side of each speaker terminal right? For example:

amp output terminal (+)  >  +leg of 4700uF (- and - tied together)  +leg of other 4700uF  >  Speaker (+) terminal

Correct.

I'm assuming cap voltage rating should be at least 70+70, so 150V or more?

Wrong. The max voltage any of those cap is likely to be submitted to is 65V, so 80V rating is enough, but actually 100V types are so closely priced that it's a better option. What's important is getting the 105°C type.

 

[JW]

Abbey,

Sorry, I wasn't clear enough on what this amp already has in it. On the back it looks like this

                                L channel                                    R Channel

Output:    3990          3990                          3990                3990
Output:    3990        SJ6451                      3990                3990 
Driver:    SJ6387      SJ6387                    SJ6387          SJ6387
Output:  3990          3990                          3990            3990 
Output:  SJ6451      SJ6451                    3990              3990

Since this would ostensibly be a studio amp, I'd like both channels to match. If they can simply be matched by adjusting bias and level, I suppose that would be fine. I have doubts though, that these transistors sound the same. Could very well be somebody's friend repaired it, called it good.

I could probably track down more 3990's.

As for the diodes, the following is a conversation about D108 and D109 from the Crown Forums. Guy named Shawn under the name "Raindog Industrial Arts," someone named "DGlass" and Dave Engstrom, Crown Technical Support: http://www.crownaudio.com/forums/index.php?/topic/1482-dc300-recapping/

Shawn:
Replace D108 & D208 germanium diodes as they are the achilles heel, always leading to distortion and coloration when they go bad, which does not take much effort. I'm working on substituting some modern day equivalents as the original spec'd ones are hard to get and the circuit is still weak after you replace them with new ones. The old part number is 1N270. I already used 1N771 but has even less current capability.

1N34A is in my amp right now and it is a "modern" part but it has not been on the scope yet. It does work fine, for now. *beep* bad spot in a design for a light weight germanium diode.

DGlass:
D108 and 208 are on the negative half of the Current Limiter circuit and are basically only on if the amplifier is protecting. They're not even in the direct audio path so it may "brings the science of the 90's to vintage equipment" but it won't improve the signal.

Shawn:
I think there may be some confusion here; D108 and D208 are in the BIAS SERVO which in turn becomes part of the VI limiter when the VI limiter is in effect. Faulty D108/208's will affect the bias of the output stage of a DC300A. A poor bias setting will certainly generate distortion even though these diodes are not in the feed back loop or in the signal path.

If the output of the DC300A is shorted or driven into a very difficult load for a sustained period, it is very possible the VI limiter circuitry will force too much current through these diodes. The peak operating current of a 1N270 is 325mA @ 25 degrees Celsius!

The Central Semiconductor part CDSH270 is indeed a better part for this application as it has a wider operation temperature and the forward surge current is higher. The Central part also has a continuous forward current capability much higher than the “old school” 1N270.

If you talk to any true professional service tech that worked on Crown equipment in the 70's and 80's, they will most certainly site these two diodes as one of the known problems with DC300’s.

Dave Engstrom:
D108/D208 are part of the current limiting protection circuit, not the bias network. These diodes generally became faulty as a result of the negative current limiter being faulty (shorted). Though a germanium diode would have slightly higher leakage than a silicon device, if Q103/203 was not biased on (excessive output stage current) it wouldn't matter what condition the diode is in.

The 1N34A may be a better part but unless the amplifier is working into a load impedance lower than 2.7 ohms this part of the circuit never turns on. If it is initiated the distortion caused by current limiting will more than exceed any component benefits.

Shawn:
The DC300A VI limiter circuit also once known as SPACE is not transparent to the function of this amplifier when it is "not in use". The VI limit circuitry is an integral portion of this amplifier's output bias, signal output and feedback path. If you disable the VI portion of the schematic, the DC300A can not be biased and it will not function properly. When D108/208 are at fault, the entire amplifier's performance suffers. The VI limit control circuit is directly tied to the bias of the output transistors.

Also, the negative feedback loop is feed back to the DC300 OpAmp right through the middle of the VI limit circuitry.

I think the VI limit portion of this design is being treated like a separate entity when in fact it is embedded deep in the heart of this amplifier's design and function. It is not a separate switch that turns on and off yet it is a feature embedded in the core of the DC300A.

I think I will have to provide some performance data based on a known working DC300A with D108/208 substitutions so you folks can really see the effect the VI limit design has on the performance of this amplifier as a whole.

Unfortunately, this is where the conversation ended.

My line of reasoning was that if there is indeed a more sturdy replacement, it's neither costly nor difficult to do.

 

[JohnRoberts]

My line of reasoning was that if there is indeed a more sturdy replacement, it's neither costly nor difficult to do.

Germanium diodes are old technology so unlikely to see much active new component development. (IIRC there was some activity a few decades ago for HF amplifier use.)

One of the comments you pasted suggest that the VI limiting is also interactive with the output stage class A bias. A better part even if it exists might be different enough to require revisiting the class A bias circuitry.

JR

PS: As I also posted earlier Crown pretty much repackaged these old designs into their CE1000/2000 models using SMD. Maybe look at those newer model schematics to see if they revisited this known(?) design weakness.

 

[abbey road d enfer]

                                L channel                                    R Channel

Output:    3990          3990                          3990                3990
Output:    3990        SJ6451                      3990                3990 
Driver:    SJ6387      SJ6387                    SJ6387          SJ6387
Output:  3990          3990                          3990            3990 
Output:  SJ6451      SJ6451                    3990              3990

Since this would ostensibly be a studio amp, I'd like both channels to match. If they can simply be matched by adjusting bias and level, I suppose that would be fine. I have doubts though, that these transistors sound the same. Could very well be somebody's friend repaired it, called it good.

I could probably track down more 3990's. 

I Would make sure that all output transistors on a side are the same; Seeing a mix of 3990's and SJ6451 gives me the shakes. Indeed SJ6451 (or modern replacement NTE388) has a much better SOA so the amp should be more reliable (although the original DC300 is actually quite reliable; the problem is they have been so abused in touring PA that it might have tainted its reputation). But really, I always changed all transistors in one leg when one was shot. It's not really a matter of sound ("sound" depends much more on the topology and overall design than on a particular component), but rather a matter of reliability and consistency of operation. The bias control on the DC300 relies on the output transistors in one leg being Vbe matched.

As for the diodes,

I don't know who this guy Shawn is, but I think he has a strange analysis of electronic circuits. Indeed, when one of these diodes in the protection circuit is leaky, that impairs the distortion performance of heh amp, and since Ge semiconductors are more prone to leakage, they must be checked thorouhly, but when the components are up to specs, these diodes are not a part of the bias control.  Dave Engstrom is the one you should listen to.
That's the problem with forums, signal-to-noise ratio is poor.

 

[gridcurrent]

quoting from page 17 of the DC300a service manual:
"the output stage is a quasi-complementary format employing the CROWN class AB+B techique which uses no bias current in the output transistors."
there is no class A bias, whatever that means.

 

[JohnRoberts]

quoting from page 17 of the DC300a service manual:
"the output stage is a quasi-complementary format employing the CROWN class AB+B techique which uses no bias current in the output transistors."
there is no class A bias, whatever that means.

Without looking at an actual schematic, I have seen similar approaches . Basically you use a beefy driver stage that is capable of putting out maybe a couple amps. The driver feeds the output through say a 1 ohm resistor (guess) with a power transistor connected base to the driver emitter, and power device emitter to the actual output (or through emitter resistors to share current).

At idle and for small signal levels there is less than 0.5V across the 1 ohm resistor so the power transistors stay cut off (class B).  You still need to implement a class AB biasing scheme for the driver stage, so no free lunch. Slightly cheaper because the strong driver takes some of the output stage load.
-------
When I became product manager for Peavey's amps back last century the (value) PV1.3K amp used a similar shared driver/output scheme to save one pair of output devices.  I had them add the two power transistors back in, use a proper driver, and renamed the amp PV2000, still cheap (I didn't raise the price) and heavy iron, but now lower distortion and more robust driving low Z loads.

JR

PS: For the OP I really like the Bruno Putzy class D amp modules. As the euro keeps getting weaker and dollar stronger these are getting cheaper here in the US.

 

[bjoneson]

Just a testament to the "bullet proof" nature of these things. I recently picked one up off of eBay for $75, that appears to have been dropped off of a bench or maybe down a flight of stairs. Likely while powered on (based on scorching inside where one of the main filter caps leads shorted against the chasis). The rack ears were severely bent, and one of the filter caps was completely caved in. Part of the transformer housing was dented.

Bent and pounded everything back into shape, replaced the main filter caps, and it powered right up. No DC at the outputs, distortion within spec under load. Amazing.

Have it powering a pair of JBL 4315s.

Post Apocalypse, there will be cockroaches, and DC300As.

 

[JW]

I saw that same DC300a with the caved in filter cap. I was going to buy it, but seems you made an offer or something. Nice turn around there. That was only a couple weeks ago seems like.

 

[bjoneson]

I saw that same DC300a with the caved in filter cap. I was going to buy it, but seems you made an offer or something. Nice turn around there. That was only a couple weeks ago seems like.

Yep! Took a little bit of a chance, but I'll be damned if it didn't come right back to life. Just an absolute beast of a machine.

 

[JW]

Abbey,

Sorry to ask terminology questions, but what do you mean by "leg" of transistors? Top to bottom on one side or left to right?

 

[JohnRoberts]

Abbey,

Sorry to ask terminology questions, but what do you mean by "leg" of transistors? Top to bottom on one side or left to right?

I'm no Abbey, but suspect he is talking about transistors in parallel powered from the same rail. If one transistor fails as a short (melted), it is not unusual for those in parallel with it to be degraded by the same stress.

JR

 

[abbey road d enfer]

Yes, a leg is the compound of 3 power devices driving the load from one rail. It is essential that they are Vbe matched because their b-e junctions are driven by the same signal, with only the emitter resistors trying to maintain proper balance between them all. But you may have different types in the positive leg than in the negative. I wouldn't recommend that as a "feature" though.