Identification of diode in a form I've never seen before

• Identification of diode in a form I've never seen before
  Posted by moothefish on 10 Jan, 2020 07:50
• Can anyone tell me what the name of this style/package of diode is? I cannot for the life of me find it online, looks more like a little tantalum capacitor or something.
  
  For context, these components are from a 1970s stereo receiver, one on each channel of the amplifier board. I've been having issues with severe hissing noise coming and going and after much troubleshooting I found that when I wiggled these diodes, the noise came and went. I did the usual of resoldering them (several times in fact) but to no avail. One was always much worse than the other and I've now noticed that one has a slight crack in the casing.
  
  The issue is I don't have a service manual and can't find one for the life of me, so I have no idea what I would need to put in as a replacement. Anyone know what these are, or how I could find out the ratings and find a modern replacement for them? When it comes to this level of electronics, my knowledge is limited, so any help is appreciated

  ---

  
• #1 Reply
  Posted by wraper on 10 Jan, 2020 07:57
• I would not be so sure it's a diode. Tantalum capacitors look like diode with high forward voltage drop when measured in reverse direction. If they are leaky, voltage drop may appear lower.
• #2 Reply
  Posted by moothefish on 10 Jan, 2020 08:17
• I thought that at first, however the silkscreen on the PCB has a diode symbol and they are labelled D401 and D402, plus I couldn't seem to get any capacitance readings out of them.
• #3 Reply
  Posted by magic on 10 Jan, 2020 08:42
• Capacitance reading on diodes is only a matter of resolution, they ought to have a few pF
  
  Too young to have ever seen them , but some sort of zeners perhaps? What are the voltages across them, what are they connected to?
• #4 Reply
  Posted by moothefish on 10 Jan, 2020 10:19
• At a glance, I'm not entirely sure what its role in the circuit it (this is where my electronics skills are a bit crap). I've attached some photos of the board in question. Seems like they're connected between some of the smaller transistors and the main output transistors, through a resistor or two?
  
  Please excuse the truly awful soldering, last time I touched this project was about a year ago and I promise my skills have improved a little since (though I still don't claim  to be particularly good at soldering). I've noticed some joints that need touching up like the one that has an obvious hole in it lol. I'm also aware of the lifted traces but they're all making connection as far as I'm aware.

  ---

  
• #5 Reply
  Posted by moothefish on 10 Jan, 2020 10:20
• Sorry for the double post, images are too large and I can't easily resize them on mobile. Also a little bit potato quality

  ---

  
• #6 Reply
  Posted by T3sl4co1l on 10 Jan, 2020 13:08
• D404 looks like it would be for compensating output bias.  Though it's not greased to the heatsink.  Not sure what they'd need another one for, inland.
  
  Does it measure as a diode?  Or a few in series?
  
  Tim
• #7 Reply
  Posted by fcb on 10 Jan, 2020 14:55
• Never seen a diode in that format. Probably a capacitor, although be cool to proven wrong.
  
  Any chance you can reverse the schematic of one of the channels, that would be the 'gold standard' way of working out what it is/was.
• #8 Reply
  Posted by Bud on 10 Jan, 2020 14:56
• I've seen varicaps in similar packages, especially that you find it in a receiver. But if it is used in audio section than i am not sure. Do you have a diode tester to measure the forward voltage? This will give you an idea about a replacement type of diode, silicon vs germanium one, etc.
• #9 Reply
  Posted by free_electron on 10 Jan, 2020 15:15
• diode stack for cross over distortion . most likely this is a 1n4148 that someone custom globbed so it could be thermally coupled...
• #10 Reply
  Posted by RoGeorge on 10 Jan, 2020 15:16
• Other transistors on the board looks like Si, so I'll assume the diode is also Si based.
  
  Never seen a diode in such a package.
  Does it measures like a diode?
• #11 Reply
  Posted by Vovk_Z on 10 Jan, 2020 17:03
• What voltage drop it is? If it is a diode then voltage drop will be as in a diode.
• #12 Reply
  Posted by SeanB on 10 Jan, 2020 18:40
• Those are normally a double or triple diode in a common package, so that the voltage drop will compensate 2 or 3 base emitter junctions in the output stage, depending on topology you will have quasi complimentary output or complimentary, meaning you have to compensate for 2 junctions if it is NPN PNP output, or for 3 if it is a Darlington Slizaki stage.  Used to offset the voltage drop of the junctions between the output stage. You can measure the forward voltage of the good one at 1mA, 5mA and 10mA, and find it will be around 1V2 to 1V8, and then select an appropriate number of 1N4007 diodes to match it within 50mV at the same current. Don't use a 1N4148, they will not track the power devices over operating temperature range, and the new diode array will need to be thermally coupled to the heatsink, but insulated from it, easiest done with a strip of Kapton tape between them and the heatsink, and another bit of tape over the top.
• #13 Reply
  Posted by andy3055 on 10 Jan, 2020 19:16
• I have seen these in old audio gear. Mainly, Sony/Sanyo etc. What other members have stated about multiple diodes in series is correct. What is the make/model of the unit? I have some old schematics and might be able to find but just not in a hurry as I am away from home till the end of the month.
• #14 Reply
  Posted by magic on 10 Jan, 2020 19:28
• This is not the bias spreader.
  
  The track which goes to the output transistor goes to its collector and also to the collector of the driver, the two clearly arranged as a Darlington pair, so this track has to be a supply rail.
  
  The other end of the diode goes through R405 (left channel) to the collector of one of the input transistors and presumably the base of the gain stage. I will have a look what's going there, it's probably some kind of clamping, protection, anti-saturation or whatnot.
  
  edit
  Looking at the markings, these green transistors appear to be PNP (NPN is in the center of the board), so it has to be the negative rail. The diode is mounted in forward direction from the VAS (TR405) input to the rail, in parallel with the VAS and its 2.7Ω emitter resistor R413. What's going on?
  
  edit
  Maybe they wanted to make the VAS a "precision" current mirror, with current gain exactly defined by the ratio of R413 to R405, so D401 ought to be a single diode. Too bad that the ratio of those resistors appear to be 1800
  
  D403, the similar diode used in the bias generator, doesn't generate the whole bias alone. It's in series with TR407, presumably a VBE mulitplier. Quite possibly a single diode too.
• #15 Reply
  Posted by floobydust on 10 Jan, 2020 21:03
• These were a unique Japanese part, as a bias diode string mostly for audio power amplifiers.
  They were called "Silicon Varistor" or "Silicon Varistor Temperature Compensating Diode" or "Varistor Diode", with the SV part number prefix i.e. SV-3A, SV03YS, SV-04F, Ohimuzi MV11Y etc.
  They ended up in weird locations in the circuit, wherever the engineer could stuff in temp comp.
  
  Long obsolete, people are just making diode arrays from 1N4148's as a sub. I made SMT version with BAV99S and bolt to the heatsink. I recall seeing some that appear to be several series diodes plus thermistor all in one package, so language translation of varistor vs thermistor is my guess what happened. Those measure like a mushy pair of diodes.
  
  I don't think they ever worked super great for bias temp. compensation because they weren't always mounted on the heatsink (extra manufacturing labour) and accurate tracking of bias with temperature is extremely difficult to do. A VBE multiplier is what the industry moved to using. Here Sony subs out an SV-04S with a VBE multiplier but no trimpot.
  

  ---

  
• #16 Reply
  Posted by bob91343 on 10 Jan, 2020 22:59
• Sometimes they used thermistors, so don't rule out that possibility.
• #17 Reply
  Posted by moothefish on 10 Jan, 2020 23:51
• Well this has certainly been a more interesting thread than I thought it would be! The receiver in question is branded as a Rank-Arena ra-402, which is either a rebrand or bespoke design using NEC boards. I wasn't able to find an NEC receiver that looked the same, I did find one which looked similar but used a different circuit in the amplifier board.
  
  So am i correct in thinking that I will need to somehow create a new diode stack to replace each of these? How many diodes need to be in the stack? I'm definitely in over my head at this point...
• #18 Reply
  Posted by floobydust on 11 Jan, 2020 06:20
• OP, what is the amplifier not doing, that you suspect D401, D402 in each channel?
  Zener D405 is common to both channels, looks odd with heatshrink so it might be two diodes in there?
  
  I did a rough schematic but don't have much for part numbers and values. If you can check, it would help us figure out the mystery diodes.
  
  edit: corrected errors in schematic rev0 -> rev1

  ---

  
• #19 Reply
  Posted by Jwillis on 11 Jan, 2020 09:08
• Doing a little searching and putting together bits pieces here's a short version of the history. Rank merged with Arena to create Rank Arena in 1971.  Arena audio components were manufactured by Hede Nielsen in Denmark until it's merger into Rank-Arena in the UK in 1971.Rank also had holdings in NEC Japan. So that would explain the PCB branded with NEC. So maybe you could cross reference the stereo with a similar one Branded NEC from the same time period. Might be able to find a service manual .Some basic performance specs might help to narrow the search.
• #20 Reply
  Posted by magic on 11 Jan, 2020 09:09
• Seems about right. You forgot R415 which charges C409 but who cares
  
  And yes, R413 is 270Ω not 2.7
  
  Maybe it's time to just put this diode back in and see how many volts it drops?
• #21 Reply
  Posted by moothefish on 11 Jan, 2020 09:33
• Wow, you guys are great! Was not expecting anyone to try and actually create a schematic. What values did you need specifically? The issue I was having was that after about 5-10 minutes of running, the audio would start cutting out and this god awful hissing\rumbling noise would come and go in one of the channels and to a lesser extent both channels. After recapping and swapping all the transistors between channels to see if one was leaky, I found out that moving the diode pack on the channel with the issues caused the noise to change dramatically, and i could get it to appear and disappear by moving it. I thought it was the solder joint or the trace but i verified the connection and desoldered it several times with no change.
  
  Happy to put the diodes back in and do some more checks. Dumb noob question though, If i put the diodes back to measure voltage drop, what points on the board should I measure between to get the best reading? Just across each leg of the mystery diode?
  
  As for the NEC history of Rank Arena, I did already try looking at old NEC gear but couldn't find anything that used the same circuit as this. I'm sure its out there somewhere, just haven't been able to find it in my searchings as of yet.
• #22 Reply
  Posted by RoGeorge on 11 Jan, 2020 09:43
• Why would a diode have a package that does not distinguish between anode and cathode?
• #23 Reply
  Posted by PA0PBZ on 11 Jan, 2020 10:28
• Quote from: RoGeorge on 11 Jan, 2020 09:43

  Why would a diode have a package that does not distinguish between anode and cathode?
  

  
  You missed the white dot?
• #24 Reply
  Posted by RoGeorge on 11 Jan, 2020 10:42
• That's weird, I missed it indeed!