Any idea what this is?

[tomizett]

Not a very exciting mystery object, but can anyone help my identify this component?

I *believe* that it is an NPN transistor - it measures correctly for an NPN with a meter (ie, two diodes pins 1>2 and 1>3), although I've not been able to put one on a component analyser yet. I believe the package is a SOT89.

It is from a voltage-controlled-amplifier, which features a large number of these components (at least some of) which are connected as trans-diodes (that is, pins 1 and 2 shorted).
One of the devices has failed, and I would like to replace it - in reality, I suspect I'll need to replace all of them to ensue a good match.

This one (the borken one) is marked "DS RN", although some others from other modules are makred "DS RM", so I suspect the "RN" may be a batch or date code.

I've looked everywhere I can think of... any help really apreciated!

 

[Nigel Goodwin]

Not a very exciting mystery object, but can anyone help my identify this component?

I *believe* that it is an NPN transistor - it measures correctly for an NPN with a meter (ie, two diodes pins 1>2 and 1>3), although I've not been able to put one on a component analyser yet. I believe the package is a SOT89.

SM components are difficult (and often impossible) to identify, stick it on a component analyser to test it as it could just as easily be a double-diode, which are fairly common.

 

[tomizett]

Too true...
I've never see that pinning on a double diode, but it certainly could be one.

Unfortunately, just knowing that it's a transistor probably won't be enough - if I can't get the same type to replace it, I doubt that the circuit will work propperly.

I thought it was worth asking here - nothing ventured, nothing gained, etc.

 

[spec]

Hy tomizett

You can establish if the component is a double diode or a transistor by checking for transistor action- I presume you know how to do that.

Can you post the schematic of the amplifier. It is not uncommon for transistors, or diodes to be used to tailor the response of an amplifier, typically logarithmic by using the exponential vbe characteristic of a silicon junction. Sometimes many transistors are used to cover a wide dynamic range. Is the amp a compressor/decompressor of some kind, perhaps for music.

At a guess, I would say that component is a medium power transistor, possibly like a Zetex FCX1051A, which has a low bulk resistance and hence more closely approximate an exponential law.

If it is established that what I say is true, you could quite easily characterize the transistor by measuring the VBE/IC and to a lesser extent IC/IB and then finding a matching replacement.

spec

 

[MikeMl]

For simple DC logic and switching, as long as you match NPN vs PNP, you almost cant find a transistor which will not work...

 

[Nigel Goodwin]

Too true...
I've never see that pinning on a double diode, but it certainly could be one.

Unfortunately, just knowing that it's a transistor probably won't be enough - if I can't get the same type to replace it, I doubt that the circuit will work propperly.

Transistors are rarely critical at all, and if they are it's likely to be a very poorly designed circuit

My experience of replacing SM transistors has been to simply treat them as only having two types - NPN and PNP - simply locate one of the correct polarity on a scrap PCB, and remove and use that

I've never had any problems with that technique

 

[tomizett]

Thanks for the interest all.
Sadly I don't have a schematic for this circuit, although if I get very keen I may trace it out. I imagine that the circuit is the classic type of analogue multiplier involving logarithmic conversion and then back again - although I've not got my head around how/why you would use a large number of identical devices to acheive this.

Spec, I'd not though of measuring the curve, but yes, I suppose that's what I'd need to to to ensure a good match. In addition to your comments about bulk resistance (something else I'd not thought about), I suspect that they also opted for a medium power package with a thermal pad for good thermal coupling - the banks of devices are mounded 2-sided on a ceramic substrate.

 

[ronsimpson]

I doubt it is a double diode because of the thermal tab. (I can be wrong, check the gain on a transistor tester.)

Next time order the parts that say:
"DXT2222A-13"
"May 21, 2016"
"Diodes Incorporated"
And not the ones that say:
"K1P"

 

[spec]

Thanks for the interest all.
Sadly I don't have a schematic for this circuit, although if I get very keen I may trace it out. I imagine that the circuit is the classic type of analogue multiplier involving logarithmic conversion and then back again - although I've not got my head around how/why you would use a large number of identical devices to acheive this.

Spec, I'd not though of measuring the curve, but yes, I suppose that's what I'd need to to to ensure a good match. In addition to your comments about bulk resistance (something else I'd not thought about), I suspect that they also opted for a medium power package with a thermal pad for good thermal coupling - the banks of devices are mounded 2-sided on a ceramic substrate.

tomizett,

How come you don't know the unit function? Could you show some pictures of the unit, especially any text and controls. What is the power source. What is the input and output. Any information would be a help.

I take it that although the 'transistors' are in a medium power pak they are not actually dissipating any significant power. If so that would support the shaping/temperature tracking theory.

About the number of components involved. It's been a long time but I think that is to support a wide dynamic range. The exponential Vbe function of each transistor pair can only practically handle a limited range so the transistor pairs are arranged to only handle a limited part of the overall transfer function. A pair is needed so that the amplifier is bipolar, once again off the top of my head.

spec

 

[tomizett]

About the number of components involved. It's been a long time but I think that is to support a wide dynamic range. The exponential Vbe function of each transistor pair can only practically handle a limited range so the transistor pairs are arranged to only handle a limited part of the overall transfer function. A pair is needed so that the amplifier is bipolar, once again off the top of my head.

That sounds very likely - a colleague mentioned that he'd seen something similar, but I'd have to look it up to gain a fuller understanding. I understand you can get several decades of log conversion from a single device, but I guess in practice you end up loosing a good deal of that in noise. Dynamic range and linearity will certainly have been the designer's prime concerns.

I didn't post much detail at the top of the thread, because I was just hoping someone might recognise the device marking, but...
The full story is that the circuit in question is a VCA module from a Yamaha PM3500 mixing console - there is one per channel. It's a small circuit fabricated on a ceramic substrate (although I don't think it would be right to call it a "hybrid" as it's all packaged components, and no resistors on the substrate that I can see). It is soldered into the main channel board with a single row of pins. I have a service manual for the console, but the VCA module is just marked as an IC, with no reference to what it might contain. At least the manual does give me the input and ouput signals. It wouldn't be the end of the world to trace it out, but I'm not sure it would gain me anything.

While I normally just post here about hobby stuff, this is actually a "work" job. The viability of the repair is going to depend on how easy it will be to get a replacement module - if we can't get a replacment then I'll have to have a really good look at how to repair the faulty one.

 

[vtech]

I do remember the Yamaha PMxx series consoles to be some of the solid/well built units available as they came out in the 90's (I think?)
Rows of controls/pots/meter etc. I was involved with Yamaha's extensive Audio equipment for several years back.

You mentioned about having a service manual. I believe what you are pointing to is the VCA hybrid IC XK868COO which was manufactured by someone other than Yamaha(Murata.TDK, etc).
Surprisingly, Abtronics.Ru has it still listed on line.(have never dealt w/them)
A 33 pin inline IC. Yamaha was always good about showing at least a block diagram of the IC's on their equipment(A long faded practice)
Look at page 49 paper (42 of the PDF) and there is a block diagram of that hybrid.
BTW, instead of SMD resistors, most were etched on the substrate IC.

Hope that helps

 

[tomizett]

Thanks vtech, very interesting.
I must admit that I'd not found that block diagram but, yes, that's the part (those two op-amps are NE5532s). Those manuals are exquisite - as they are for most of the Japanese gear that I've seen, even the later stuff.

I'd not thought that they could have been made by a third party, given how vertically integrated Yamaha is (or at least was then). Thet's very interesting, as is your pointer to Abtronics.

Regarding the construction, there are SM resistors on the board, but I'd not be at all surprised if some are also etched as you suggest. I've cretainly seen this in things like differential input modules, which appeared to have been trimmed after construsion for maximum CMRR.

Nice to hear from someone else who has worked on these! It's a nice tidy console, if a little fiddly to take apart at times...

 

[pfofit]

Here are a couple possibilities
1. on chip resistor PNP
http://pdf.datasheetcatalog.com/datasheet2/8/0uwho09khdei9ec87lx2o1sy3eky.pdf

2. voltage detector
http://www.alldatasheet.net/datasheet-pdf/pdf/80682/SII/S-80729AN-DS-X.html

 

[ronsimpson]

pfofit,
Good find. You must have worked hard.
Some meters measure resistance at low voltage (<1/2 volt) and you can see the B-E 12K resistance.
Some meters may need (reverse the leads) to see 12k.
From my experiance the ratio of R1/R2 is very good (1%) but the resistance is +/-20%.

 

[tomizett]

Indeed, thanks for the suggestions pfofit. I appreciate you taking the time.